From 145041cead30b4483ab53d3e1b9aa966d9cef5c2 Mon Sep 17 00:00:00 2001 From: dincerti Date: Mon, 25 May 2020 10:05:40 -0700 Subject: [PATCH] Update website --- docs/404.html | 182 ++++ docs/articles/expected-values.html | 183 ++++ docs/articles/getting-started.html | 365 ++++++++ docs/articles/getting-started.png | Bin 0 -> 45105 bytes .../figure-html/ceac-1.png | Bin 0 -> 69213 bytes .../figure-html/simStateprobs-1.png | Bin 0 -> 179611 bytes docs/articles/icea.html | 413 +++++----- .../icea_files/figure-html/ceac_plot-1.png | Bin 107551 -> 90308 bytes .../icea_files/figure-html/ceaf_plot-1.png | Bin 95474 -> 80391 bytes .../icea_files/figure-html/ceplane_plot-1.png | Bin 141689 -> 103687 bytes .../icea_files/figure-html/evpi_plot-1.png | Bin 92620 -> 78357 bytes .../icea_files/figure-html/mce_plot-1.png | Bin 114932 -> 96632 bytes .../icea_files/figure-html/totevpi-1.png | Bin 72041 -> 58210 bytes docs/articles/illness-death.png | Bin 0 -> 23316 bytes docs/articles/index.html | 90 +- docs/articles/intro.html | 778 +++++++++++------- .../intro_files/figure-html/ceac_plot-1.png | Bin 78196 -> 65110 bytes docs/articles/markov-inhomogeneous.html | 630 ++++++++++++++ docs/articles/markov-inhomogeneous.png | Bin 0 -> 30987 bytes .../figure-html/simStateprobsPlot-1.png | Bin 0 -> 144778 bytes docs/articles/mlogit.html | 359 ++++++++ .../figure-html/icerHeterogeneity-1.png | Bin 0 -> 82637 bytes .../mlogit_files/figure-html/stateprobs-1.png | Bin 0 -> 109994 bytes docs/articles/mstate.html | 463 +++++++++++ .../mstate_files/figure-html/costs_plot-1.png | Bin 0 -> 47545 bytes .../figure-html/predicted_hazards_plot-1.png | Bin 0 -> 118022 bytes .../mstate_files/figure-html/qalys_plot-1.png | Bin 0 -> 36453 bytes .../stprobs_by_strategy_plot-1.png | Bin 0 -> 109666 bytes .../stprobs_by_timescale_plot-1.png | Bin 0 -> 103955 bytes docs/articles/psm.html | 349 ++++---- .../psm_files/figure-html/stateprobs-1.png | Bin 110268 -> 92367 bytes .../psm_files/figure-html/survcurves-1.png | Bin 108083 -> 88776 bytes docs/authors.html | 66 +- docs/bootstrap-toc.css | 60 ++ docs/bootstrap-toc.js | 159 ++++ docs/index.html | 100 ++- docs/news/index.html | 167 ++-- docs/pkgdown.css | 161 +++- docs/pkgdown.js | 15 +- docs/pkgdown.yml | 9 +- docs/reference/CohortDtstm.html | 533 ++++++++++++ docs/reference/CohortDtstmTrans.html | 286 +++++++ docs/reference/CtstmTrans.html | 251 +++--- docs/reference/IndivCtstm.html | 484 +++++++---- docs/reference/IndivCtstmTrans.html | 367 ++++++--- docs/reference/Psm.html | 346 +++++--- docs/reference/PsmCurves.html | 310 ++++--- docs/reference/StateVals.html | 231 ++++-- docs/reference/absorbing.html | 86 +- docs/reference/bootstrap.html | 87 +- docs/reference/ce.html | 102 ++- docs/reference/check.html | 116 ++- docs/reference/check_edata.html | 91 +- docs/reference/costs.html | 206 +++++ docs/reference/create_CohortDtstm.html | 224 +++++ docs/reference/create_CohortDtstmTrans.html | 229 ++++++ docs/reference/create_IndivCtstmTrans.html | 104 ++- docs/reference/create_PsmCurves.html | 94 ++- docs/reference/create_StateVals.html | 126 +-- docs/reference/create_input_mats.html | 140 ++-- docs/reference/create_lines_dt.html | 106 ++- docs/reference/create_object_list.html | 87 +- docs/reference/create_params.html | 163 ++-- docs/reference/create_trans_dt.html | 106 ++- docs/reference/define_model.html | 376 +++++++++ docs/reference/define_rng.html | 292 +++++++ docs/reference/define_tparams.html | 267 ++++++ docs/reference/expand.hesim_data.html | 168 ++-- docs/reference/expand.html | 95 ++- docs/reference/fast_rgengamma.html | 119 +-- docs/reference/flexsurvreg_list.html | 105 ++- docs/reference/formula_list.html | 105 ++- docs/reference/hesim.html | 90 +- docs/reference/hesim_data.html | 159 ++-- docs/reference/hesim_survdists.html | 129 +-- docs/reference/icea.html | 219 ++--- docs/reference/icer_tbl.html | 112 +-- docs/reference/id_attributes.html | 293 +++++++ docs/reference/incr_effect.html | 127 +-- docs/reference/index.html | 337 ++++++-- docs/reference/input_mats.html | 261 ++---- docs/reference/joined.html | 130 +-- docs/reference/lm_list.html | 103 ++- docs/reference/mom_beta.html | 100 ++- docs/reference/mom_gamma.html | 98 ++- docs/reference/mstate3_exdata.html | 251 ++++++ docs/reference/multinom3_exdata.html | 255 ++++++ docs/reference/multinom_list.html | 220 +++++ docs/reference/params.html | 194 +++++ docs/reference/params_joined_surv.html | 112 ++- docs/reference/params_joined_surv_list.html | 134 +-- docs/reference/params_lm.html | 123 +-- docs/reference/params_lm_list.html | 121 +-- docs/reference/params_mlogit.html | 220 +++++ docs/reference/params_mlogit_list.html | 200 +++++ docs/reference/params_surv.html | 203 +++-- docs/reference/params_surv_list.html | 111 ++- docs/reference/partsurvfit.html | 105 ++- docs/reference/psm4_exdata.html | 120 +-- docs/reference/qalys.html | 208 +++++ docs/reference/rcat.html | 123 +-- docs/reference/rdirichlet_mat.html | 127 +-- docs/reference/rng_distributions.html | 322 ++++++++ docs/reference/rpwexp.html | 111 ++- docs/reference/sim_ev.html | 233 ++++++ docs/reference/stateprobs.html | 207 +++++ docs/reference/stateval_tbl.html | 226 ++--- docs/reference/summarize_ce.html | 125 +-- docs/reference/surv_quantile.html | 111 ++- docs/reference/time_intervals.html | 217 +++++ docs/reference/tparams.html | 198 +++++ docs/reference/tparams_mean.html | 251 ++++++ docs/reference/tparams_transprobs.html | 279 +++++++ docs/reference/tpmatrix.html | 224 +++++ docs/reference/tpmatrix_names.html | 226 +++++ docs/reference/uv_rng.html | 188 +++++ docs/reference/weibullNMA.html | 92 ++- 117 files changed, 15013 insertions(+), 3933 deletions(-) create mode 100644 docs/404.html create mode 100644 docs/articles/expected-values.html create mode 100644 docs/articles/getting-started.html create mode 100644 docs/articles/getting-started.png create mode 100644 docs/articles/getting-started_files/figure-html/ceac-1.png create mode 100644 docs/articles/getting-started_files/figure-html/simStateprobs-1.png create mode 100644 docs/articles/illness-death.png create mode 100644 docs/articles/markov-inhomogeneous.html create mode 100644 docs/articles/markov-inhomogeneous.png create mode 100644 docs/articles/markov-inhomogeneous_files/figure-html/simStateprobsPlot-1.png create mode 100644 docs/articles/mlogit.html create mode 100644 docs/articles/mlogit_files/figure-html/icerHeterogeneity-1.png create mode 100644 docs/articles/mlogit_files/figure-html/stateprobs-1.png create mode 100644 docs/articles/mstate.html create mode 100644 docs/articles/mstate_files/figure-html/costs_plot-1.png create mode 100644 docs/articles/mstate_files/figure-html/predicted_hazards_plot-1.png create mode 100644 docs/articles/mstate_files/figure-html/qalys_plot-1.png create mode 100644 docs/articles/mstate_files/figure-html/stprobs_by_strategy_plot-1.png create mode 100644 docs/articles/mstate_files/figure-html/stprobs_by_timescale_plot-1.png create mode 100644 docs/bootstrap-toc.css create mode 100644 docs/bootstrap-toc.js create mode 100644 docs/reference/CohortDtstm.html create mode 100644 docs/reference/CohortDtstmTrans.html create mode 100644 docs/reference/costs.html create mode 100644 docs/reference/create_CohortDtstm.html create mode 100644 docs/reference/create_CohortDtstmTrans.html create mode 100644 docs/reference/define_model.html create mode 100644 docs/reference/define_rng.html create mode 100644 docs/reference/define_tparams.html create mode 100644 docs/reference/id_attributes.html create mode 100644 docs/reference/mstate3_exdata.html create mode 100644 docs/reference/multinom3_exdata.html create mode 100644 docs/reference/multinom_list.html create mode 100644 docs/reference/params.html create mode 100644 docs/reference/params_mlogit.html create mode 100644 docs/reference/params_mlogit_list.html create mode 100644 docs/reference/qalys.html create mode 100644 docs/reference/rng_distributions.html create mode 100644 docs/reference/sim_ev.html create mode 100644 docs/reference/stateprobs.html create mode 100644 docs/reference/time_intervals.html create mode 100644 docs/reference/tparams.html create mode 100644 docs/reference/tparams_mean.html create mode 100644 docs/reference/tparams_transprobs.html create mode 100644 docs/reference/tpmatrix.html create mode 100644 docs/reference/tpmatrix_names.html create mode 100644 docs/reference/uv_rng.html diff --git a/docs/404.html b/docs/404.html new file mode 100644 index 00000000..39a3b119 --- /dev/null +++ b/docs/404.html @@ -0,0 +1,182 @@ + + + + + + + + +Page not found (404) • hesim + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
+
+ + + + +
+ +
+
+
+

Page not found (404)

+
+ +Content not found. Please use links in the navbar. + +
+ + + +
+ + + +
+
+

Developed by Devin Incerti, Jeroen P. Jansen.

+
+ +
+

Site built with pkgdown 1.5.1.

+
+ +
+
+ + + + + + + + diff --git a/docs/articles/expected-values.html b/docs/articles/expected-values.html new file mode 100644 index 00000000..55523635 --- /dev/null +++ b/docs/articles/expected-values.html @@ -0,0 +1,183 @@ + + + + + + + +Simulating costs and quality-adjusted life-years • hesim + + + + + + + + + + +
+
+ + + + +
+
+
+

Simulating costs and quality-adjusted life-years

+ +

2020-05-25

+ + Source: vignettes/expected-values.Rmd + + +
+ + + +
+

+Overview

+

Before performing a decision analysis, (discounted) costs and quality-adjusted life-years (QALYs) must be simulated. In cohort models, they are simulated as a function of previously simulated state occupancy probabilities. In individual-level models, they are simulated as a function of simulated trajectories characterizing disease progression. In hesim, the discrete time state transition and partitioned survival models are cohort models and the continuous time state transition model is an individual-level model.

+
+
+

+Cohort Models

+

Costs and QALYs in cohort models are computed by integrating the “weighted” probability of being in each state. Weights are a function of the discount factor and state values (e.g., annualized costs and utility) predicted using either the cost or utility model. Mathematically, for a time horizon \(T\), discounted costs and QALYs in health state \(h\) are computed as,

+

\[ +\int_0^{T} z_h(t) e^{-rt} P_h(t) dt, +\]

+

where \(z_h(t)\) is the predicted cost or utility value at time \(t\), \(r\) is the discount rate, and \(P_h(t)\) is the probability of being in a given health state. Note that the state values, \(z_h(t)\), can depend on time since the start of the model but not on time since entering a new health state.

+
+

+Integration methods

+

Three types of approaches are currently available for numerical integration given values of state probabilities at distinct discrete times.

+
    +
  • +Left Riemann sum: The function is approximated by its value at the left most point (i.e., start) of each time interval.
    • +
  • +Right Riemann sum: The function is approximated by its value at the right most point (i.e., end) of each time interval.
    • +
  • +Trapezoid rule: The function is approximated by the average of its values at the left and right endpoints. Specifically, for an interval [\(t_1\), \(t_2\)] with value \(y_1\) at the left endpoint and \(y_2\) at the right endpoint, the function is approximated as \(\frac{t_2 - t_1}{2}(y_1 + y_2)\).
    • +
+

The Riemann sum rules approximate the area under the curve using rectangles in each time interval whiles the trapezoid rule approximates the area under the curve using a trapezoid. In general, the left Riemann sum will underestimate costs and QALYs whereas the right Riemann sum will overestimate them.

+
+
+
+

+Individual-level models

+

In individual-level models, costs and QALYs are computed using the continuous time present value given a flow of state values, which change as patients transition between health states or as costs vary as a function of time. The state values can be partitioned into \(M\) time intervals indexed by \(m = 1,\ldots, M\) where interval \(m\) contains times \(t\) such that \(t_m\leq t \leq t_{m+1}\) and values for state \(h\) are equal to \(z_{hm}\) during interval \(m\). \(z_{hm}\) will equal zero during time intervals in which a patient is not in state \(h\). Discounted costs and QALYs for health state \(h\) are then given by,

+

\[ +\sum_{m = 1}^M \int_{t_m}^{t_m+1} z_{hm}e^{-rt}dt = \sum_{m = 1}^M z_{hm} \left(\frac{e^{-r{t_{m}}} - e^{-r{t_{m+1}}}}{r}\right), +\]

+

where \(r > 0\) is the discount rate. If \(r = 0\), then the present value simplifies to \(\sum_{m = 1}^M z_{hm}(t_{m+1} - t_{m})\).

+

Note that while state values in cohort models can depend on time since the start of the model, state values in individual-level models can depend on either time since the start of the model or time since entering the most recent health state. Individual-level models consequently not only afford more flexibility than cohort models when simulating disease progression, but when simulating costs and/or QALYs as well.

+
+
+ + + +
+ + + +
+

Developed by Devin Incerti, Jeroen P. Jansen.

+
+ +
+

Site built with pkgdown 1.5.1.

+
+ +
+
+ + + + + + diff --git a/docs/articles/getting-started.html b/docs/articles/getting-started.html new file mode 100644 index 00000000..5d6268b0 --- /dev/null +++ b/docs/articles/getting-started.html @@ -0,0 +1,365 @@ + + + + + + + +Getting started with a Markov cohort model • hesim + + + + + + + + + + +
+
+ + + + +
+
+
+

Getting started with a Markov cohort model

+ +

2020-05-25

+ + Source: vignettes/getting-started.Rmd + + +
+ + + +
+

+Overview

+

The simplest models in hesim are cohort discrete time state transitions models (cDTSTMs), which encompass both time-homogeneous and time-inhomogeneous Markov models. cDTSTMs (that can easily incorporate patient heterogeneity and parameter uncertainty) are most conveniently simulated with a four step process specifying:

+
    +
  1. +Data consisting of the treatment strategies and patient population
    2. +
  2. +Parameter objects storing the values of all parameters used in the model
    3. +
  3. +Random number generation expressions that randomly generate probability distributions for the parameters for probabilistic sensitivity analysis (PSA)
    4. +
  4. +Transformed parameter expressions that transform the parameter estimates into values used for simulation
    5. +
+

We illustrate by replicating the HIV model developed by Chancellor et al. and used for didactic purposes in the Decision Modeling for Health Economic Evaluation textbook. The analysis compares two treatment strategies, zidovudine monotherapy and zidovudine + lamivudine combination therapy. The model is time-inhomogeneous because patients are assumed to only use lamivudine for 2 years before switching back to monotherapy.

+

The model consists 4 health states. Ordered from worst to best to worst, they are: State A (CD4 cells > 200 and < 500 cells/mm3), State B (CD4 < 200 cells/mm3), State C (AIDS), and State D (death). The model assumes that patients can transition to any state that is more severe (i.e., progress) but cannot transition back to a more severe state.

+

+
+
+

+Define the model

+
+

+Data

+

We set up the model for two treatment strategies and one representative patient.

+
library("hesim")
+library("data.table")
+strategies <- data.table(strategy_id = 1:2,
+                         strategy_name = c("Monotherapy", "Combination therapy"))
+patients <- data.table(patient_id = 1)
+hesim_dat <- hesim_data(strategies = strategies,
+                        patients = patients)
+print(hesim_dat)
+
## $strategies
+##    strategy_id       strategy_name
+## 1:           1         Monotherapy
+## 2:           2 Combination therapy
+## 
+## $patients
+##    patient_id
+## 1:          1
+## 
+## attr(,"class")
+## [1] "hesim_data"
+

The model will be simulated for each treatment strategy and patient combination. Such data containing treatment strategy and patient characteristics can be generated using expand().

+
data <- expand(hesim_dat, by = c("strategies", "patients"))
+head(data)
+
##    strategy_id patient_id       strategy_name
+## 1:           1          1         Monotherapy
+## 2:           2          1 Combination therapy
+
+
+

+Parameters

+

Counts of transitions at yearly intervals between each of the 4 health states were available from a longitudinal dataset from Chelsea and Westminister hospital.

+
trans_mono <- matrix(c(1251, 350, 116, 17,
+                       0, 731, 512, 15,
+                       0, 0, 1312, 437,
+                       0, 0, 0, 469),
+                      ncol = 4, nrow = 4, byrow = TRUE)
+colnames(trans_mono) <- rownames(trans_mono) <-  c("A", "B", "C", "D")
+print(trans_mono)
+
##      A   B    C   D
+## A 1251 350  116  17
+## B    0 731  512  15
+## C    0   0 1312 437
+## D    0   0    0 469
+

The transition probabilities from state \(r\) to each of the 4 states (State A, State B, State C, and State D) can be modeled using Dirichlet distributions, which are parameterized by a vector of concentration parameters \(\alpha = (\alpha_{r1}, \alpha_{r2}, \alpha_{r3}, \alpha_{r4})\). The Dirichlet distribution is particularly useful for modeling uncertainty in the transition probabilities because it is a conjugate prior for the multinomial distribution. The posterior distribution of \(\alpha\) consequently follows a Dirichlet distribution which can be thought of as the “effective sample size” for each transition (see Appendix).

+

Treatment effects are defined in terms of the relative risk, which is assumed to reduce the probability of all transitions to a more severe health state by an equal amount. We work with the log relative risk since it is approximately normally distributed. The estimated relative risk of disease progression is 0.509 (95% CI 0.365 to 0.710). We can then calculate the mean and standard error on the log scale.

+

The cost-effectiveness analysis is based on life-years, meaning that each health state is assigned a utility value of 1. The costs of zidovudine and lamivudine are $2,278 and $2,086.50, respectively. Medical costs are split into direct medical and community medical costs.

+

The parameter estimates are stored in a list.

+
params <- list(
+  alpha_mono = trans_mono,
+  lrr_mean = log(.509),
+  lrr_lower <- log(.365),
+  lrr_upper = log(.710),
+  c_dmed_mean = c(A = 1701, B = 1774, C = 6948),
+  c_cmed_mean = c(A = 1055, B = 1278, C = 2059),
+  c_zido = 2278,
+  c_lam = 2086.50,
+  u = 1
+)
+
+
+

+Random number generation

+

To perform a PSA, it is necessary to draw samples of the parameters from suitable probability distributions. While the probability distributions could have been drawn manually and stored in the params object above, it can be more easily done by using random number generation functions with define_rng().

+

As noted above, we draw the transition probabilities with monotherapy from a dirichlet distribution and the relative risk (of a transition to a worse state with combination therapy relative to monotherapy) from a lognormal distribution. Drug costs and utility are assumed to be fixed while medical costs are drawn from gamma distributions.

+

Sampled parameters values are stored in a list, meaning that samples for each parameter can be stored as either a vector or matrix. For instance, we sample medical costs for each (non-death) health state from gamma distributions and they a stored in a matrix (one row for each sample and one column for each health state). Likewise, the parameters of each cell in the transition probability matrix are store in a matrix, with parameters of the matrix ordered rowwise across the columns. On the other hand, samples of the relative risk are stored in a vector.

+
rng_def <- define_rng({
+  lrr_se <- (lrr_upper - lrr_lower)/(2 * qnorm(.975)) # Local object 
+                                                      # not returned
+  list( # Parameters to return
+    p_mono = dirichlet_rng(alpha_mono),
+    rr_comb = lognormal_rng(lrr_mean, lrr_se),
+    c_zido = c_zido,
+    c_lam = c_lam,
+    c_dmed = gamma_rng(mean = c_dmed_mean, sd = c_dmed_mean),
+    c_cmed = gamma_rng(mean = c_cmed_mean, sd = c_cmed_mean),
+    u = u
+  )
+}, n = 1000)
+
+
+

+Transformed parameters

+

The next step is to “transform” the parameters for use in the economic model using define_tparams(). This typically involves modeling them as a function of the treatment strategies and patient (i.e., the data). To maximize computational efficiency, all operations are vectorized across the rows in the data. Parameters not include in a transformed parameter function are assumed constant across patients and treatment strategies.

+

An important helper function is tpmatrix(), which allows users to define transition matrices as a function of other parameters. In this example “vectorized” transition probabilities are defined in terms of the transition matrix for monotherapy (p_mono) and the relative risk (rr). The relative risk is simply 1 for monotherapy (since the corresponding transition matrix is just p_mono) and equal to rr_comb for combination therapy.

+

The only costs in the model that vary across treatment strategies (rather than just states) are drug costs. The monotherapy treatment strategy include the cost of zidovudine only whereas the combination therapy includes the costs of both zidovudine and lamivudine.

+

Parameters may depend on time (since the start of the model) by using the package defined time variable. The time variable corresponds to the stopping time of time intervals, which are specified using the times argument in define_tparams(). Since patients are only assumed to use lamivudine for two years, rr_comb and the costs of lamivudine are not applied beyond year 2.

+
tparams_def <- define_tparams({
+  ## The treatment effect (relative risk) is transformed so that it varies by 
+  ## strategies and only applies for the first 2 years (Monotherapy is 
+  ## the reference strategy). 
+  rr <- ifelse(strategy_name == "Monotherapy" | time > 2, 1, rr_comb)
+
+  list(
+    tpmatrix = tpmatrix(
+      C, p_mono$A_B * rr, p_mono$A_C * rr, p_mono$A_D * rr,
+      0, C, p_mono$B_C * rr, p_mono$B_D * rr,
+      0, 0, C, p_mono$C_D * rr,
+      0, 0, 0, 1
+    ),
+    utility = u,
+    costs = list(
+        drug = ifelse(strategy_name == "Monotherapy" | time > 2,
+                    c_zido, c_zido + c_lam),
+        community_medical = c_cmed,
+        direct_medical = c_dmed
+    )
+  )
+}, times = c(2, Inf))
+
+
+

+The model

+

The model is defined using define_model() function. The names of transition probability matrix, utility, and cost objects are specified, as are the names of the health states.

+
mod_def <- define_model(tparams_def = tparams_def,
+                        rng_def = rng_def,
+                        params = params)
+
+
+
+

+Simulation

+

A economic model (of class CohortDtstm) can be created from a defined model (of class model_def) and data using the generic function create_CohortDtstm(). The economic model is an R6 object consisting of a transition model (trans_model) for simulating transition probabilities with sim_stateprobs(), a utility model for simulating quality-adjusted life-years with sim_qalys(), and a set of cost models (for each cost category) for simulating costs with sim_costs().

+
econmod <- create_CohortDtstm(mod_def, data)
+
+

+Health state probabilities

+

At the start of the model, all patients are assumed to be in State A (the default setting of the transition model of class CohortDtstmTrans). State occupancy probabilities are generated by simulating the discrete time Markov chain; as expected, patients survive longer with combination therapy.

+
econmod$sim_stateprobs(n_cycles = 20)
+
+# Plot
+library("ggplot2")
+theme_set(theme_bw())
+stateprob_summary <- econmod$stateprobs_[, .(prob_mean = mean(prob),
+                                              prob_lower = quantile(prob, .025),
+                                              prob_upper = quantile(prob, .975)),
+                                          by = c("strategy_id", "state_id", "t")]
+stateprob_summary[, strategy_name := factor(strategy_id,
+                                          labels = strategies$strategy_name)]
+ggplot(stateprob_summary, aes(x = t, y = prob_mean)) +
+  geom_line(aes(col = strategy_name)) +
+  geom_ribbon(aes(x = t, ymin = prob_lower, ymax = prob_upper,
+                  fill = strategy_name), alpha = .3) +
+  facet_wrap(~factor(state_id, labels = LETTERS[1:4])) +
+  xlab("Year") + ylab("Probability") +
+  scale_fill_discrete("Strategy") + scale_color_discrete("Strategy")
+

+
+
+

+Costs and QALYs

+

Costs and quality-adjusted life-years (QALYs) are computed by integrating the previously simulated state probabilities. To maintain consistency with Chancellor et al., we approximate the integral with a right Riemann sum, which is simply a sum of costs and QALYs measured at the right endpoint of each interval (e.g., costs during model cycle 1 are measured at time 1 (rather than at say, time 0)).

+
econmod$sim_qalys(dr = 0, integrate_method = "riemann_right")
+econmod$sim_costs(dr = 0.06, integrate_method = "riemann_right")
+
+
+
+

+Decision analysis

+

A decision analysis is performed by comparing combination therapy to monotherapy.

+
ce_sim <- econmod$summarize()
+wtp <- seq(0, 25000, 500)
+icea_pw_out <- icea_pw(ce_sim, comparator = 1, dr_qalys = 0, dr_costs = .06,
+                       k = wtp)
+

The incremental cost-effectiveness ratio (ICER) estimated using means from the PSA is $6,300, which is very close to the estimate from the deterministic analysis from Chancellor et al. 

+
icer_tbl(icea_pw_out)
+
##                   1   2                        
+## Incremental QALYs "-" "0.92 (0.56, 1.22)"      
+## Incremental costs "-" "5,821 (3,027, 9,548)"   
+## Incremental NMB   "-" "40,376 (22,948, 54,439)"
+## ICER              "-" "6,300"                  
+## Conclusion        "-" "Cost-effective"
+

A cost-effectiveness acceptability curve shows that the probability that combination therapy is more cost-effective than monotherapy increases quickly with the willingness to pay (WTP) for a QALY. At $10,000, the probability is 0.944 and approaches 1 at slightly higher WTP values.

+
ggplot(icea_pw_out$ceac,
+       aes(x = k, y = prob,
+           col = factor(strategy_id, labels = strategies$strategy_name[-1]))) +
+  geom_line()  + xlab("Willingness to pay") +
+  ylab("Probability cost-effective") +
+  scale_x_continuous(breaks = seq(0, max(wtp), 5000), label = scales::dollar) +
+  theme(legend.position = "bottom") + scale_colour_discrete(name = "Strategy")
+

+
+
+

+Appendix: Dirichlet distribution

+

The multinomial distribution is a discrete probability distribution for the number of successes for each of k mutually exclusive categories in n trials. The probabilities of the categories are given by \(\pi_1,\ldots, \pi_k\) with \(\sum_{j=1}^k \pi_j=1\) and each \(\pi_j\) defined on \([0,1]\). The Dirichlet distribution is parameterized by the concentration parameters \(\alpha_1,\ldots, \alpha_k\) with \(\alpha_j > 0\). Letting \(x_1,\ldots, x_k\) denote the number of successes in each category, the prior distribution and likelihood are,

+

\[ +\begin{aligned} +p(\pi_1,\ldots,\pi_k |\alpha_1,\ldots, \alpha_k) = \text{Dirichlet}(\alpha_1,\ldots,\alpha_k) \\ +p(x_1,\ldots,x_k | \pi_1,\ldots,\pi_k) = \text{Multin}(n, \pi_1,\ldots,\pi_k). +\end{aligned} +\]

+

The posterior distribution for \(\pi_1,\ldots,\pi_k\) is then given by,

+

\[ +p\left(\pi_1,\ldots,\pi_k| x_1,\ldots,x_k, \alpha_1,\ldots,\alpha_k \right) = \text{Dirichlet}\left(\alpha_1 + x_1, \ldots, \alpha_k + x_k\right). +\]

+
+
+ + + +
+ + + +
+

Developed by Devin Incerti, Jeroen P. Jansen.

+
+ +
+

Site built with pkgdown 1.5.1.

+
+ +
+
+ + + + + + diff --git a/docs/articles/getting-started.png b/docs/articles/getting-started.png new file mode 100644 index 0000000000000000000000000000000000000000..415d6d48fbc7692be485bbe122e27101d2591d5f GIT binary patch literal 45105 zcmeFZc{tQ<_&$uLj3QfEvWzW~eF-rb*=3Ea*%Oj|8#~$egt88)Xd#L0hD<2?zVEW{ z>x}*V4C(24zQ6Z-y#Kz(@#~+)xaa=d_qCkYd7bw?!qrvfi7sBdh=YSeq^KZs4+rOb zKMoGA8~6h7jt|9NFz_F)^F8@HIG^tQx+=<|=JvMSrVs7SEV$ve4!|fJ z94K5Acxh|lV#)xwePZV<3KwTO86gV1#=gzN#Begi#YUV-S4Evc+TO{6;TE?bH!qXK zMFs{2sMEtoqW5Ivet#VJNu0^b#l=CChsWLBo!ec2+uq5Nhwt|7+dRDdJpBAzzz8m9 z4?7oAIG3F>^XVjK^T=2@n>$%MxLDiUF<|F4HM4hh5ocn;KIqTC({;L7Kljlm<3q%4#xyn=l^94E9Dm%S>4-Juz19g*<*_1mE8v9pQ z^@q_Pew3KeZWQaj8TcFT+I>J*`YTpbJmvl%p$R{@IV@eZ#Vsl?{+CG_XY7Zy$IdJ zVH|4K`+J2^Y5sWoKL*O){=ph6cf7CxfiHneO*6*yo9xdkv_H3Q7Wk0WZF*lEpjNgV zdR`MDa#IHVFl;?+LpM)kETa8*X;7tU?C1&7ZRkm2;h$y0@J3W&0aZi!pJ&;A<2fvc zglfV3i< znEt5-?VkYa3~e{UADlhL+KD6&s{-VWr8;oP8-!tEJ(I>3FDU`ZHkLdrhdxA z&9#4drI%FQ#JHjM_L=ECV_{^Z@d`&@V6ZJXcMSUe=kADKzs$YzBm(pudBTI)-SfuN zf~0$AJ1RK=a`02R?s1dS{B}0#r-8j!SBP5Xz_Hv>b)r&hCC zkzDoIeg2r~*!b>oVD1#EtPGyfdq2YQFB3ci-}uZ%rh5H%jhuUk7d9pnAfQ6}OA)G8 zI~=8iKqmE)&oY&9a1lJ7|Fs8r-`O$xm$uBh+?0CXi^S!6J%Xb10b2MDyfiv#yIPjK zvD=B&Hh$M^rF~H44}Gk&BXj2cpy8X{T;0@Rs~&?8n@%&0_&-22au;x_SAL!)>#1Aj zWbnYhz{X*dsUgf}+rey7(C`^W+NQevAH7#!$Z)E%hEFM0J}!!}>hezS$TnQh8CP_< z5rAxT%-M^mYeBtl*eKHSn>O?DWJp?o7B9PeWc#iupLf|(=}?aDMNVDT@5cbNKs3-F z&&%=wTkO44me3v!damA0iJuW-Bn6W%UEw%^9W2(WT^Q zVimH|-#Rt-xf&-9ghmP4O|P-PO|d~>}(LM2`V^084kJgKUEBXtC?}X2XOcco$Nr`M)8>V#1eNE+66N~pEn^u$# zTg}hSS?)Z0+#DnDG;x;#@%txRp0mOO6NQn4k$ZE-?*4i+!C8{Z5P*q>#o!c(R!UY1 z*;>DlGwQWD^-%y4_ACfKC^17iX1)4cq>X>GC~(=N$h61sqHHbZmsE?B81@ld6I6bVoLidAPh?*8zW-+)6u>%CRjOOK7 z)F$C%&0R#iwH{Riw z%U0RbZz{ssh*m$oQ3;=yp*nnVbWJ`;4G(&@k4#Gdh|$e&cVw_o>_Qaa!ki4+LW&eV zkjZ#=Bi}o%>&ie6q@mW^vn7Q_9h>~85ULEq-BZ{Ke-0kg(PQP>PBe2HwpCerRUN}o z6D#KFehTaHQm_CY{&y8>C%IrH=MUbrttD+w-npMWU+?gkGq!~2I)XV%AUE}^qa&kBURXk9lKHm#?(RDLE7g2K{*-Hmiqc2`<>)YFdOcOShfcMQ;V%-dHlfsm_m|7wn)+%F2{zKzWdAE2UWn@sEX&A zvLk=3Lqa)vhXDL1l>vhXktR`lEnLdsX%-A$Rc`Q{OYWwW^9Fu=CV&nuY<3Vl9lQkw z97rbFw7Y_`>!=yiR&|qENz>uaYsu;PU~9c5MXlmp{v{8fKBWGo8ZLF$xJw$^KZMDK`QY=&Qfv& z=Z%$%9Vp3I{Ds&T&QnL7CSnnFK(ZsWY%OL^wK{cPSAW5oxi|+Ytg+fu@5J`v6tu|! zvg2dsQ6Y@|U^JNFO*J2svSYqIUP8M*R5l)t5Zxl2y>o3*@b8HgOy~mG`R-<4T*C07`4suUBy8Vs&jrFO~XRuuM0M7#sWKJJx5{0Eq4itcz9P&7c7kBnH zZp6lTkaZ-JQA?Nq6N6YAMImqw_gIi!89A{TAPoC!9wevf{R7sC!`8+FI=<^ z5Ms=%8D*ucMFB1?kHJZQEIJ@lv80`xGc`gfp)F^bI`^V`UR8seqv0|B3+7X1^BbcG zBVoP}|8c~3_Z6b-%OBs=CzJM*4ikL%wt8_Efg|<%0R*g>H4@)yM$!R~CW7`Rb2}U4 z;*(S#?&->W%8hY){tO}7PPq6#tdl7K-l{Xr-oF5HU!hL2qm`hLTT>~wg!!vMbm$_# za7mv)NNKw$uPm>S6BrpQtF;%X zaanbiLTcp4bN|e1u(;x&Ac3za_8ea zgI2Z{SN!{1#R*m4h@p8-^F|QsMNrH~Y|gX5X5;&K+%M%wN2n1)qVuAn!xeG_8m};> zUuK{GD<@AdbDfp{HSq5A&DoXt`lF#vifW zz28H^mPF99l9y&j6*|Nj0A>JN1^TAdDw`GRyO{))_LdxiTts7NLSwt{<;2A> zA((Dy#}vUz{`MlU`t}Q;?C*Rm9X{I4&a2;D^f3xtBa)=^F>{EO422Vd>T(N!lTXM> zL>j34Ma`B-p>Svy!SL(JH__cBEvrO1%y3KiHJkQUlbZIwU8SV*!=#>uJ|~Z4jGV_viPxcE~4;ebZ~4pRco-S z(ZU>`QF(f0T&IiJzwDBkO3)k@KV|epdj`DD{57$F(M$|&E>j^;*&N}3bv>*B;6AJU zcV`3cX4(*tuuHiB7WwY-LjJK!_gvA0%lJ~|dC-4jD7!u+92k0u-lk-4xAKJ&&TKbP)*gA}$h z7~>5$INHlUyV6=Uvb7dRHMy5QAyYmTw$gMOAGV4Bh2`Rj#J?W!>3Tb*Ww8#Ntjj%O zmwf{D^?m^^9XH_22Qkx>B>$b-R>tHJ?IlzbLvh>1SqTTgflnNte6NDw-UqJ?6Q2~c z!KVOoFnDl*_zs?-qRHd5tLZmDuA*8l*^h#V`?MP_phLRiYP;|i%(ed$91H=MnPAWE zU(*lql6g_)PWve@il|hwS=Jq|;A9!(fa>5=8&@Gs;KBuH1I;@G8RV++1dcu|bZ?LP zV`9jBoiXVXHnqR2_GvKyQh;oZ8*f;E=}XQlX@b^;OOQ{FZ3-s6wGJ!S!<|-%PogeL z42U{M`K8M&?O#DoQa-9_*Pre=M9;Z^Tts0q^z8)t6rcW$9e~U2$&yHOqQ@oJ(#G>f zYIA_y?U@G#Up=-42>r<#%&`k@j?>RP4P}84uVl^jIPgN)Y*|uFY0QuRF{M9ffs~<< z?v=P~SzGvSG%D2Qu;5#51V<@87>FGz8X&37Q62YK+M_{X*A%3Y&EdPtAs){wvqJq< zABx(!c*6dj6xar_8&oQJ9fxot-P9t)tNL1M0_pN4^9W-aB=iB+{=84H_$D`Vfox76 z@AcL0-OT2^iLiN@bG+4}Cj(BQ@{Ht0vEece5|Btcs7bgciFxyG@Z^f1i!QF5;B^SL zZU{0vD*&L-K9G6tiytOn!dv%|NHIVE*fk_?T5H4dgiPL_ehurJlfo46lcna6`;M0< zN2*6>h`FLeHLFm$L|P_5dH^EA0IV?ZSrUMgth^kpEB+2Lu12qERPoPA`Iy#`cCQlt z)7k#<`Le1>(NnQ+4@*LfDO(-*FC?90pDs}KN2);tF2Vpq@Z1lEP{q7!4OKGuIvwX| zT4x^v@V@_`EnvaSE`lg+2R5AOA8sJ`tfAk&#V)+uzotk8pQ*cz$dBP5*c8(hhwfaP zLX}+lKP;#UhV#uw73gY&gCxl%()O)as;g;z)kV)cvm0%G;^s7wd!{-P=E1E4v9SQpIwy=0wWW9n9Pb0sk`z#GfMfvgucRB3yg zN}wDYKQ+Jm?W+BDV&Tz`wo->ns9ln#M(jSQ4Dh}(-C}a#ma)IqROytGDsu}wEYH~HXqa$Wk0+$u%0d{XpY&fN>z$cbr}XY1S3P$N zb6b_p0tnzvi~%EAoGTq@;tV>Y{8;**fl8Y#M+XD2yC{x;2cK8-m#k%#72AT3 zST^F1Qk9TLjauK3c(vhY&Gm^$@pR&0Y5tgGnpd)Rt|Y(VRbPuz>GBIeSxUbHtVi=x zLc>5Fh~yk;PwYiWPH#_)da)=!Ct=QntQHm~d6PlUOVfkaiKa{xn(FLi>|gBCcDO^0<~~;|hr`hS=`u9?-q5?vQJ!9kb9+r2+weXa0MCXgk$;B9G^&CdEBk2~z8 z4CvX$jfkYqZ*@suauqX2!0j`?vV8WcXGOZs6}dF~ShHroUttLMWXgv)@kzGHkJ=S# zsj=`$s3@u(qq`y1m(*PE3}z4M!X^ITI|O%62R{fxf%FqO%zqVy{E;~EGL)m3Shi`K zED4Lm0PAegIvAF zx8qoJuS?meykT}R-4$h0-*nyeP9A^&wP_gqraVlDr$?IoUPd{4_*O2nS|g-zI#43J z+qK(v#H3rJXjWM#C_gvZQXv&8PVrA=S0C--v)ebFzz@msy)PgbvD`9Os)|Rh02Ew2 z(~6G>27fs0HC!3EZI zM;0Pze3dck(Z-K4B)wOA&fUSM>ANOHTVVOIF~OoU9^NYYg7oR&2#F_VUuI5^d=!Ed z_4WWHX|S0ts8jdlMtv?e*fMRwZ#+wQVyprkUXYHyPt6I|6xKA(5ls{9jmt^r>Oo!H zSoTgl)0%7g$&%>mh;*J*LyOBqyiKu6W0U}v_*{6hZm8m)x`_^gkRuyNEF(lbEJUKy z)Z#<^Z~gojkJCljAob?<(@QZT@iSe5&j(w3zmqan{(6dO)NMV#-c!)bdtO=5Qv8(9 z)QIvSJpE7(3*I2iqb#ULPEbi|2fcb*4*TX|Vi6mMX6GMLDH>R_B+@()wtN`pCnaLG zKY3Cj@PR0s!s%hENWl?q!w+h^Bq`$k7|tv`exP0FcdhGIG=GW&BdAmx%pd|eei)(q z)ZsZ`F)~G4Q$-y40|h1ziE5_KV=C2a9%r>W35w@aU^7TFNKv0PR?6SzRJE&McWE*9 zbtc!hc&fH5YImTvF=yEe+b2n)3+1ORIL!VbDS1#aaX}YL9?SvE_Q_SyS1lp$XIJGV z!bPb#F&mI`e7bn7ErP0M@O(U_hGR+_M>~cC=HTwp7)NrJKb*QYZ zlHc1GvprF?4xk_JMZi9<{b#cc)lBr$CcIE;;EY?}2S%?gQU>{n(SFmS5yJx$gJ+@~ zTZ{7fTV``)1|L%6oioGhOW$QlLN`F?!86tf3oNz5KI9%=-o*{579tnFX}Acjljg?P z!dxq$mU3o}8VsNs)vb_A@EnA#y`N}_3Z1%C%X7K9L{F{{C}q(Anh8IeN*TTW*)VNy zePXkg)SX7pFIR!6V^S_z*q|YhED+Rs+2wdPaW-<8W*_ zmpN2jpp(#6Cu0DUcM3#a^+LRw(uH1H4C-bVq5f2c*UU9lE_-&vX%rFVw>Huss^jAe z6dS3scx3NLbk)d!giWlG;&Z=;3Ft}qCLqILOz-Q6Q3=#Zy%!2IepAJoW}W98FWJ!x z5tCtkyk|v$!yR{0mH#lhu4Fl{?m^MU?$>x3YiTAB(Hp*h}3!-!nbc>EVk*A2MHnHo>EhffNHE*{hYk~F;NXef5nJ}%n;~dok9P5XjDgb# zA04_L#3>liBLAV|nGyj5aZ&xgS{QZWcKZmZF~KbA-AUh*c^OD&QFY}70=&XW&RF6M z%F#Kw_MBo|wkJw(V-dVPp{tg+TZAVm_3E2R>i1DE?`K;dt?NFb`$uJnSMwhaMy?$U zv2#v=*T5T7w(qOgOqRgk=#2G~JFEZby}BQRX|FM;f-?j_Br6p>glr#1_E#BA%F5B; z+&-_IC2F(?Zb-ncE+&nY)-Kjnkd`tOtpw^$5b3k6Ey+_(UTvB&cN>19k}wQWWp25m zwmD2FFviixsCxf*Tj+8t;Ju@p&jff)*-|^Zd5SYOCoim)(lQ3D_Df7N6mh0jZm|R` zPv&}5_!(ig*5v9B_Y)^(gL$w`K>)*N#1)oyW)S?QmnSCo!fd4fexAcM0?>}#S87Ky zu@HN5)pL!4Ee%VhRB>iawfh{1nU~s~dcA*?1sGF1r2w3n)K@8&hS-2Uc$6jvCw9Db z>s6y5>+hJ%Tm;T_1fJabaM=38OH~Bbkon$qX_KVuy4xnEjj9a#bfiA+`&7j#FM6KG6_lKm3dEWkY#>kbpPgoxX zdfI^C_9x|RrTl#vn8YTQuhIXO2j)catca~R*@9QQe=TV@ zly9Pr*}e7Ry+1t%Qyq)2*n4hnaj2orT{(8ODQqo_ z_{%tKBFGUYOGU=3im`5;Yp-f!=oC5+N-7bo|X9*^d zL3a5``>W9xzAjGGs7=2r?_JHD#_pr{8_L^TK~~JOA$~W#20eIo$3#TfM47Zb z{ORLR=NQyKf&xs3Jn5QciJ=If+(Ty^zhgRfHs{(p8!Eo?t7M6t0SK0at`OaNXiqjs zZJ1d-qFOVBJ(_R73(BtliJ(0SU87vUyp5|@X?aKusD^9A!Nz31YVfKlTGc^KjryJm znFR<0DKDZH{tww>2-sov{{eHTo0zZ+GA2c`c#9lXdwOI<@>2x+iKq2#5DaG8=l<4b z2E%J!Ir0|iY!C8xA_#JPwEDm&#~K0QtZRt=iRWvciCaW*tt$00G{S1msA`JQT-^8W z&0npdf~9Zo^@f^>vUz*{%lNw>5;AZCjP}?`D_M0zCRb9zyU5<>E2_+s5L;sh5in!giRg)KJ0bcQmz&Gsymk?n!IE4D3u`&3 zMLJCmYg)9$1Y0QN$|&VoO~hir7Cr7L@{QzO$Y#V^L#@HQno#4W*4s3-dY5X4(a+t# zVIp4EU&u}a{@N|lHSdBlQ~m3v)1zBnLc9jKQO`JW=se4wB{aijM|Rg4+PVI@X$Me4sG12h@4ADe1C{pflAOXozTy?|SPQ>Eu zq;8!BW)O&YKKN>IzO3MK^m!j}N0r;5bzJsI@WJTt=HEZGt^3KZ#v#vPqA4NFAhQUi zHL{nC?nJ~K=wQ{H>b3Q-%viI(f*T;=5jgf5WHKAu4tRI2ZQG~5CGUS)8t}Pe$@NfI zoJowi_?qpQuZ)-*^NA*H#ep}+5xqt2k|bx_6K(5CK_7)ZI22<1Y)Feww>#MAtoWvd z0P>!6-PZ}&N?UN>i!rNY^RgQU{SGmIZV6zO6BrvA1Mjc)wL+9V0Yo(;3CX`}55#2} za@#5zd3*XE;ytZ3liIee;;4M)W{fM!TvA1Y6(Q&FQ*DGYD!++t)^(*ksc6=o0wm#tfUDU{IPv9 zJ5VFwx?WBghwrX~&-2~Isnyi6*0}_dxZ?3xZ0#;nUZMg8lRXu3n$5NGhuE_9(Zmn= z*<x4Ns@m^BAJK<>O=Y#;tCYpbhAYZ3`c(9rK?up^sNKEDm zSkKZ-Zwk*4>@0aI5g%Ezmvwx)xabKKEad zssIqdc4NyM4f&zOJ;FH|s`O8-HYB{Nb|aZ$PrY3wotaqt%E(J2wf@H`?jo2SSeI-V{(0!*=Y8Y9jTFy-|84t+9TYE&x7qM`w{lUq>{|RdVgku&m*p1tK z-tWef2E78jcLVy79W;6I@W(6qC%sR79FAy5*Mx@`Z`!V9!IP#no{HcQ~Y2=Nvs?o2I>N>5V*(NrbpUqJ0ghR}&1T;B+B zVCq~XSaiv4BPwyG`y0~%0hA;leoA)}dF(K2KbPset$uPJ!{iCUmaHS{5NAQ$+zb2y z>%`1MaC5YDqgW3#`e1Bg8==I%(rY~-@E>Po7J_L>LJfHQS*syd7QS|t>BaIKCzdx? z@Q1a;h7jk7nwYBzJ04N>zsQ{r&dArBvHsPET894Vn15SNdED7? ziEuj!eJn5Fr-5I%*8Y2|V3N|GREyHGg~p;n6@^!@IkznD@A&ze>j!F0o01)|hp`HO z`+$s!F~GWD#0!-VhFmJ}QB8RLn`G>OL3e9utd%Qw-(cSX2ykdU^{o#WOewn>;2nLL;dhjN@iI29}IU?NYbaAyk1#K~ww+3^LD)Ip_K9nYB64~1I~e;4=D zcnEj}j*@#KY5T6$kmtn0;6~tGd@aBPpV#u2~ZqM*6^u}*j z{(A+aq*nKzRGZ4Yt%y5`Zx^uu)D$S=vb?)lH)!j|sh3GaSX!Spa_wfABU%gp27hc& z(FeTxzEid>baJKxKo6p(nPCm(Ab11he)eg}R;KVp<(IB?dLBS0{~Hvro2Zc?dB;{P zc~$I@0?zNSQID+&kAvG)UCmZBsU*l19u=cn-6(<1)Z%9t=l8o+qRn^n$aS}CTSH_S`Ss{l`l2_BI_Ryo~7h5Tr0^>AAaqyN@hijXMjW-`|cj>E$$z;)4xQ$ zo&)18ikB`uH!oQYlip^y&Z~rHK)1<5eC$!RT|#ZaXD6YA$CK+(44)~GtrNJrdZX9x zr(7MYy~Lx0>*c#}k~MmB=sW#KV&we`NtY;qPZ1e)wKnueHprzGj3y9yt#=RYrS9qiOTF1V3B`q`nV zuJlj@yC&L9{f7qRum>g4R*QFkefvrpV8vW6@y_;8Xma`038wvJcHlUYENO6k)woZb zi~KC2<{W_5md+Pf-1{Vy|8~}>R#eC|>x zg?L?U-Nf^!#b)Gw_1>;o3qVxFHpQ!56bpJ7-348FX9d^mR)IW26YGZH5JWOcA4z5|LfvERSjy@zX8+=koJtj1ZHUAo$ok;i@f z%Fl6K8N6g)=8~sy$3>WQ5#^JZG3S!CXN1{InM^k3$&c9Sp(An**&ul=a@nW>63(aX zK;}iIG+xK|h_@I0$vmQ>?esjx8pP@7C z=Il7v2H^-ZYcTpcCRHarhPiipY=PKy^KRB5T*I3=Xve&CS#8<(#;S$ER$_x9jc{lA z(fs}@?G@LyGDqZ35xbrL!Au-eGF7mw_e5>w{Kh++yLhDMyCn^P8!sjZ{=xvL&jDKt|ll1}r%(}C(U z?zg9x%pwD0sM_iIv_Jagbb@VZ3;U^91lzjwd$P~+~CqGi>$pR2a^$8Lr_@>|$RZHg`8DqEVwwLEwp z#-L`NpK}%5)t?Z?|K06a(wM_Yllo~e_?sk`v33MnCZlk9B}1&5lg6~Nm{TS-;YAQS zU2?l@gE`k?Vt{ye=V11{#SwH*qKV(?h57Ktvb!byS0aXzV)L0;J zB{*1l<<$gUV=02bTe9x4wqDoKd8sMUYwv#0aFJ>K_|Se- zbkfSLG4R`EKlD#6m)(w~u~8o{*MXBI{HtidRBQP9ov+bJ9QXH(aUsf=vlsLjlo^bg z^&C_jzgMYI+8-Wq)x|)dWeQKvQ`=wG2;_6yv=V1(QDpW^f2H%hR24K9JGQsyG;+1V zPMcTN_6rvWMB75FSWT|%P2xV-6rmtjzgia-qr7hxH6|zrja!zvQEVoo&N>n7%ytnm zr7E=PkGeqU^{g}fgs{|z;~&29Ucy6Xx}f~pd;)gSt^j1_Exj~kQ=CD|VUGZu_-In1F6>slt(lM`ebl&8`Id}Hp@JG2M9joGK0SgKc zg%R5F$!IkkU=lAv>3g1Zl`n>IEAI??oRi(6tO9VlZEvu^vrA|0zP9nToJ-_i}wHw zkC`$*tP4EePzx;1S5a8ThsaN-QV~kY8Ekwwy!@#}^q13z)t^7kO$qx=WZ@3+YVRko zcQyL(b`wr|anui~MV-$WVijw?*%rS!lgIJ>Si4o22j%fUtU&}&ZbQ%Xa3&cC)hvOm z3mR}j>OjV8Pn?_A-%?3ld5UyL*7i<4I{RAyCH=cymWf%i2w~aMS?iV=>J;E>t~T^O zMiqv8&x~w6k;;PRUuKyn9d~wLzO9>-goKNZDIfAvC=OqyiQK4T9 z>O$XJ_dc>%JlmzkY1~q6%wXcJdk#->&XZ^@l%7V{nj*(tJS`| zDJ}IbjC35~%bL8}znCTT`}=RFJv<;Vxv0PUp3ZH&H@4ZPbU=Ml zoERLYtQO*6%rXAPRhe{(`UrM%(ird!bhJ=;x9OcjFVx}ab{&c`L#IW?NBQBtGvCPK zN1M>e4~JCNx15CndE4+6<8K`65>gg!;g8+a=JI*-BxAG+lgx>3FO+%5Hbuk0VQ!Tq z_oBovdV??#@v==?|KBPHmY1u?0^N%EO}Z4hZnucFp}nxOIzvmJ^7-Tgb_r3@gtxJ6 zajQr_2?2?x=-mZYLPp*P=H6Rdag}5BRFzTbsRm>}ZR#*k;!W#!xHF`NzMGK-$#f=R zqBOP@n#mzl`85(av6ry$dVxeC>o%m==%D4&2lEu-A6j}=*YVOlMZ*`Xkrt+oT?F#l zkxI7rZ@mgm#sYI3Qy=x0w?YePdx%5Q07)&H&f1lWS*!71osjJLn*Sr(xUkXYQuWf? z($^JCRUdWT)4|#sv;t5W9@)^mN%IVL&-5o(tcj6J9CnU=_#9Cu(>jEFc-nlPQX9=3;l5eYu#c|q8DtL_)K$JeF zs9MTYwZK=4o;O?kyw_&|miwLG4ChVx_Uk~cSv40lQ!_V#uIk%ur_J(Q7Xg^qt- zcpO6e&N7W0HSmTwd&5C0po*RyxB(Dg#J2WbEuV|Nd@Zskw5U=gVf4la3uzblkz6~3 zLQ=Qo%kB|AB_@B;^d@A8H6_i>-`cJE)8woUk6V~7V)XbC$ya&zv9Vi}8-9-X z9g}L|(Knul@~S+LYml|qLrPMXRYchdO-~{}OLh+#{A!BEzc}CW0V>|&42^6(fa|+p zF({LbfFyULUH}QrBu2p0{UrCMnwspz7-4L$+mMdk1yp7SR%8HF5Qi@OFaP&QvpA8>c z)LuZ*8SdJSn~L(YKkL#N>ASt&`dWM6xcUktjUy!@^-d02eVFjIVSW4t%nF|Lc}<)! z>a}9oSjUL1@n&!vUK6&b6Zm%ia@6)wyfm8D0T~{AlEhVa{L5d2rg-?@^<5ce zUZdzf8mlT0eQe@Ab;ueHwREXubcX2Csio$P{ZKxPr38S?p zG+-hkE%)1rLsJre?_cB@n{JtY6voOid0$;79{QL$7AE5`A3I@vgIFVVO3554aJ;E@ zkk5~|adka{eet-`({;D}I^Iw99jmY->uOqQ0Y1c(uG@0C!r>#2(82sgEctd3!(C7w zW$^SM-2T$B=*DS8MjF{Y03MAF!GSFyflh-`Mp*TuH=$toY(m}8y(0PdTo`OS0~blD zF8iLU_7ff%E{uTFL_)iVisXM{a2E5k9yF%yHBwvfLq_RUAO31#qc`+XPi<_;AykZ) zH=S8bJ89l+Kt>Rhl?KFoSaaBMFf?+Eb5C!5GFpgR$ZzP?dHXUa2&}4VWF11c_Ck~G zOUm~9wnY2kZ1Y)~dxXx4UJYw6hv?B0AJ&jf?6X&jY<~T@2m&U4 z^ARiJ#v)Y={pa`1wRXWxpxHO3?8aZmzM&}kNoPqvc`)xJTgfBBcLMgq;;v1rL<&D7 z;Mc$UO>90dN4t?(#CphF@{Jc!S+zdt=a{pxWQO%sfDM^^CAuRhS=tV9Nk zUI#?8ioQ{S|A=Po2{n6+Ol)_@eke34`|U>a+z)*GoY(gU*WC-1>_Psj*Xj+Fs5I8? zF&OIUHU6(C5$@{s@{QDMpv5&h;*caMUbW)sq1YSM1a+u!7CX%g9Xms`&b9ugjO>f8 zZN8OnZ*2CpiWfxHTS`Kss+NxrD{v=0m8ym51($Z)E;XWlN=6v6Vy{}CXa=x0Qr$j- zg)&Wa(0@8Tg-mMPtnT>FNK|;nP06Q+aPRm);c(pslGv{(X!gPT2TWC)f)f5ISv5bU z6Ly^^Xzci_t~6N6bVFrK{ zwh-v=^jcK%m&%TnH_c6PUIlCGM)jn>t|~B$%c~xW{a5=7CuJ&Yj&;?IQW|M&Rc%)F{II{ssB}0(OnaXzmhrp7-Hye4iG_Zru&eweroPjkQWF(*fn9j^)IdkBB%&?cXA8z#~9MP3>?5zGZmE`A(EIdRAD%UNv30 z<@%xl#pNs*%U<)j%N(J1cNa;9HQ(GGgXvdI{(O8yc~qHwzdzVdMr%CU%}%7M7(mBjEIYe9WS z?oW$K)HdRONu}PuyW>s!*TIE8NHHVQ=oiwtYozTu%kJv8 zm76)xX_^e5*@mo!(i^RIY=}pF+DoU|)l#_kOI9f{7x)e?4Uys=ceXsD=yK06e1bgw z#%^%8OJ}LL&;Q4k@ZnS+{uBAKpJ7gZ8`+CplbP(0y;j*BDQe|^-02EPj(q`U&;MD+ z>5>RpU_lt^bq!5g>CZFLZ^M@RS;bm!b}h(OyrjMiG-7lF4LhxU(>&t+id@;tS>PdCr#b}Vk++NeJ@%v9^ z;M8-U%Jcen`iZY%)nd;5{Cq>Ls+@^)xy0Vvn>`1uK+@4k)$Eu(<<{&}c=vJ9!{}J8 zosq~!c$bgZMp^*P65uMHihv$Yq;G8*qHnnEg?3?k?|iqxxgc@)x7)ISbV#d3mEN`- zbFAx13uA)OAFfZ=iCbpqAIyhaoZHYWSXy(>;51HDW)h_9zCM<0JnYl@Ubaey9REst z{Ush^JD0}=lH7|tzMp4%iDKOK!%h4HHpJwlO2qMiplo?~RZ7ifPKjo43?ay_43W;l6uSEnv?w4Y+?oHLe# zj_ic+*;6+O*=N92jsOkPdXKL})xV_AR(07md$}5YDTA8fo+IJ*r_EJLN!{V&p27u$|aUTtuNm9Tdj9m@8s&`pwee0 znBVt9*PHFC^d)AAS-&4)cIN&*J58#1ytBB%9_$*DS9aJs6cwF zCBw(F&L2M=?K9u}BE)I8&hqFLToF1v!W0tr?VXqK!2iSETmQASZ2!Xt3I%FNp*X=Q z4#hRNv^Xtoad&s8v=o=(6bMi(P>Q=#v{-O=O>qgq0^}R$IrpCX`4^t&m%NglWMVrnrZ?+B%R^ql<{eA0u- zqgNhK4G;`}8iA}!`%^mi$8Bd=HuB=|(ivj-@CeT|e0311%7V#*IRpCSLTrWt^woB~ zXY1eZXA=tl_T2=`d^343CwcX|=6hQl|LhKH@-x@_@pUS~st8Y6ZntNei}s*V^$hq{ z<#$zn141Zpq^9?Y$lJ3_VDSJH_KO&!SODa=@?mE{`KMM0#TyE|sjqv^b0}pv-V7;C z^Uye&>{FJ^SV9GH4Drxtr=g#pj@(F`%XfrHS?T1cOrMQrCV={zqQS}%p?M_=_)B z>8XPaJkA#4x7+Y+I(M0uQa$@=dG zGp5pF^sX0PgyuE8NL!)d?MVqP{T8Ks?RYxmG=H`+WUj-TA$wSVk((1Srz&LiVK~m= zXE_qF)Ku-c+p{zn4?@{{!pR@@y6AO9E{A;C(tOgJIh%~z?oEaXQ+&+&-Ld9tARXVP zfJ5?g5nU+5y!m4O;*AkVg%Qvu*`yFljJ+bGe>DXMvan)jGja)A;>rYx#Ll6ln3NYx z)hiO}vF3EZ20Lsp3jac#9gE$?7ohzne8=?>{u_NjEG&YSwm~N<1P2x8{?X@I0pW=n z`sJqM-1V^`Ozu@0WUD)U4&OZKnV!0hy|LYFVs3ETpEOYRUu*O@w~P;>JY&>t*|wbBE2A2jNX8-&oB|n5 z#oL(u%j@+d+-6P&p+pIJ{FG63c@agd#tE8(!f4h|&sYr2C(uEIx&jvCljfZi4MPjd z6jKlXea*XA?8a&i@Z8<$FbX#r#KVG%@l-fDl(Es zk(9J(4#Gy`Kv!{rfPTph_goq_wyS`f2H&z*Gfi5mjX`RXt}aEfJ_&qDNwYx~{>{F_ zc+Ay^Z=kQwD7dzbFrvRba&JukXwF~Xou<%hFMJca{e`_XJwV?{EiW1zwPM5ql?DE_ zCcNZ$qEm!wLpBCE;wL_xt!GN2g|MDYgUPP-eL&0+HTT?6g3Wm5?Q?-sd2|9x8AIzI z1$I0w_*VU7c7~V2=-Aumj*;#^GN_fW6T4N0`so6!5yvOIL)TrJt%Yk2rzj{q znVH0Cv$aUOWYO$Jb?iJ}l&u6&jF+HY#^Vl?iMI+@pzB}7gyR0X_lnFR;WqV^{G)t;t59!VvvupIJ}llX6ik}fvI*jy z;^Tz(baP03_L8-PP)X)sRYP-IXK%rKn|hQt8IHm|@y6rKuI*g*r}1@=7p2Ki`~cUR z^LM%xc_S_B==ZX{DqbUPwei;!4V-7L&_Hs+;ND|*=Df!q;amfChuE{lBn}owUj(ah zB8?OpB}kbGL=rnbVFu3|iyG5^xPFcbaxnS&LSMcVIi|C3BmI;+?mLgHJFrO}&70&2N$??-o!0Wj3;h zmdUPrWhH*n{E=6c`JG1-*h|Y*(CoP~G^9LP?mFl{(sZ+SeLyly><=I5c_IkBoNii4 z+tHZk;mTENrdD=k7Pt0W{}k~2o-XO-X9QjnlBH!&#(+XPgx0UW7B=8lj9(>9cE(!W z?S31rYBDhLTyl-+@cY1shOLC(F^rKB^Y&s)S)`O>Gs(U{QCd(oVKW%8W~F6^aW;0g z^uA=xBA(Lx^RlynTfKg6SKqJ9AVPhQ0f*Tl@+aIKK+%y1E5VQSoBVT;;o(Qdnj7&@ zSzFInz6#OU&GdvWwvmr@SW_J858dy$< z4sesv*0{*Gt4|6g+hRwrqvzLMn`KI7t(-D7ihtV>%2_Yi&P`LnV3a$MhI2R=?M1(n z;de~KFJebO*ARul_@ynKdwMER{eBCrucY7u_;gT}tYq{`w3=yajm>H`AUTTXX0umc zhp&3N8a4Ug%&LB-D&x77f6&5}d6PP+k4%+50R~kJtN~T^lAQe)ewc4t)PpFet;}h+ z%8K?`_U;7LRo>DM$1AiDBvIqRo5S)c?<9FDj>glz)o%{IR=+H}yf@aZ;4=7=nuR6d z;GDt8a-#U7??7nEqi3N~qIk=nI4iw*#@fE-Wc*1OPv?{Wu;8!F-!Rp zObG?giTg>(>@Vpaas}eXYVhM)hmjdsvL7=!f!1_fH+xp;+?P1#2!k%=Op3s@nbVv3 zYcUjfdzsN@bz`$wl~xFCb&))0dOMO(GIw~lfZe^z(8oG-Wy$_>)fInHNv@|;g+F(TTVIe3C}qmoZb2q z6+9K~O`XrJ-|wFk zp<1d-GP|$H5}K1x$v)UrfxErxm`$BugNk=`&QyRj=!-#p^TNzhjvo{0ADbus@T1?n{jcVw@gB~Qaj52n4&GVk* zP;dK*>kjt3j9(xH2}NGg&Qz)*83(n%Lo$A8v!JGECTWg>w*tS+^@ZuW0Y5a6xoo>W z{pXw5pt&aY9Lmem&r=D~xku05EOO-8!M1jc6-rFcu3=v>m-5=C1y(P_VWlj|D19Bk zTn1a_pW%cCwyPGTrF4LkrzW+EQd(kH)Bbx!Hd&B zFOJ63NvC?<&ARZS)AFNbu7qwao<#W3w-=_qC<#R08=QRVof!yx1h&3iLtG zUpLGB>o~7g8g+L$e`G6UuAMwcAtUnUMaGWfNg??H=Ymt8;sibgcI51F7TtFT0J6)b zhH~OsL6{K?QbtuyrcA|XQYYVh*rQ6KI8HYE~22TF6oY z<_(T5B$^KUy~a$_qp-fB(DfIKhJLS6s%>sIc1J_5kts=|pcwt~PV}6n^2@f^M?WGLICDM*ZttNt?Z*lE)zq5i+_iD(!`oI*4x;Y&C05=mL zZq(iJKD+L#@%Y=@nwSZQwm&s_lV^u-dHRqnTsMd^p}da0EeXl}%j4f9Os@bJD*aWDRlA3UnZF0Q&T0i8X8S2Q0Q0Uo zKm;_cdrknVL&Hy&a4o<9|17|c0CxqpqXfPh0r)r0=hj@8vNJg#-kG+;~aWJN7e;O!Grc?gg;vT=esj6gZ zwci;Usorl=ba$c;DH*6kG<6~DRH2KlXjT6^dOB7n}D`T(KwRp751t;lCT*USxo#wqXgb038(G+MsdWrB6on#2~({9>c zQYtV`yV52oSE2WP(QR5%edWLa(9h>zddKEaw^c*va``U7jFf;x9^EC|)|c&(TU^=J z!;aF&24Jo3b{SIKXC7kL5pTGY3v9X{4~6Vyw|JUud^k3e8Rg!D+gVF{KE*bvOVRZl z6!8u%84P;6Jwo_KnY1>9Wd3E9cl}ywW|KI%JGln*agpfvAU*uTxa@PQvvWN47uhHA zc_(txCtqUr#&qt<6mnh;03)Aki|xfTI|wx&4V1p_-?+Tl%{c%#fHb3A zAN6g(9#ZA(a5~kdFQ*IQ`<~P+UUu1QW^$fJg_JpjCaX!COvx!wBgID! z0Pl+=H?DPpcInNh=4409Quc4^ONr7BJPtpM(2h2%ZJ$lCniN#cit}qHF$Si$3B&nS z-Td*;Iy=U_69s+^&_{h9u}F;|yT=lZSBm`dm`}q2n z=jAaYde0I^$(8Zr-LHn3^M1)-c<)tPRZ(tXKVv4ygcU|*|5!-EB3?HUiEcd=W;zPl z(_(+GE3T|;Mm;>G?X4hj=@{KuldhG-3R)VfiklK8NHIlsk*g2zwt#hTJ~5yYw>s@U zSHWCzZ}TMiHTU8yU=KDCQn}Suz!|pz$F?eLG0M0M@eVDY8f`l;KhVGS}b2Hou z2%nsIC2hO3QRUMA&5`Q*XDwFM0{nRD)f`_9bcL(itp!SR;XDvV^s4bcaF+=_`aDks5&1qHzV>9yD@O~{RO~!3U zLmP~y9?Pis`q__d>FQLnGy1mf=Q9?5>yt6j70(5rR3S9c@!TxvNjn<8#E0bogN%wt zsU@gH#S3;M9&<+*+Xdum)uXkkl!AYAZ)#yo&4=uD77hUk8k$cOyTn>Bka+5%{m3ok zK<_Q~2IHjwK4u8b`E8l-z0)#OX;e`^t9@%dA+bh+5UO`MC|&6_?^bZCbAaEDYSG+e z+RCe65hhJeLB#gMVFUsFc~glaeyGBf_=%8v@S4G~W6M5LZaMqny`#KWeJ{eE(8qDT zF<+nK?S8JR{>4f}f2y#W$yjvBK#xD*izW`^n&7XFVfLuV6`f~PE`}EqRwU&PGHfiW z!2?FNOXjOeEXgZ(!NO4s14l1D;TOeAS9T%!Guu5HFcZHCf2=oCJh1Pc#;K7z!4jYG zg7Fug)S6NlKrRIMN~!XEQ4Q<^ldyk8>%uzD9xsVs!B*l~N?_$6`Vvhb?A!X)AXu>F zyBN`)R7DdN&E6Y__iu-CFD+#A?)6=M#iHoWwdSexj%cX(Dt4Qm^N-|>GK;qC*Acn( zON~*bXpgP?Qxg+wX{zA?;ez1nX`WdEwy5}DR|EGZC#u=TqVut^*-Xa4@LAl^T*1)M$TN!52&M;A;k*$Eh3qjq>78!TYhc(Dj(u^&xa6wv+y>m2 zOy@T zHFRw+)I{25O`i@o_5GEy+;QZMbJJehqro8DxJ?!6EkH86-cAf0clbT_Vg1bl;#o}R z@B1e)TWQc7hFoXROG7rCAI+c0`w1G$*z8lsfTFYOZm$@H+Ckm%U`&|K1-4&m)Okj``Jo0aoZ| z&zYycPwnhDhWqq1!g4?w4H9UCOOpQ?rVLFwfZmX_RWy^{z!as&n_x1c`>=|wYMATCOSGNyz zn~Se{`GIQs^UBgA18_x616*(;M}KuN)pR^D6jI&JvqKvGfF|9WPRwzm=(H&oHu_bE z$RYBSo)lM;d$Hy#k(3}VF)7#=6>KGhU!CQM`Cdy3u{?d1Vcy|*%`(fotshwv5>l4_bdMB&wU;U?WP$|lLdI`<8@LVc8{Y~Y zXiWD#{XkRuOO@9n2iFAlCD|R@CIPJb{^A>Dj#xW>jTa&Ui)hP;9aqk>C^^V;$V;GN zcIWEk+T$4~?tRZ_Xp?H4eM|TFv`mcti7IT7{L) zBJE~IE>X_nMw%uns1s21yR_%tepRVfAJ^?bJhw5{-x``wg1Xemnm|OfC^dl{8|DsP zLCSu_uvHyUlB;J}=X??0&uV-Bpw%jZ7N8$r!R#@ZT3!1YgNnh9eXn&fp{<9gc0sqL z_dADX%onRcR5ut~Go~MD7cCDB&U?tC$)}|KEMzLwNgtVo;zq*zfH0jKIMbWhW_JR9`<$iojwaFN1{s@tJmUAu4;ZFi z=lrfKht|PfZrv6zTpN;S3nnr^qmQ&jh$u&9H^$n-mt#rzW(CkiMm~{9Kd+)KOa5x( z@BEFPa!hSQz24=CVtOHSg{};?v1OL&x+<=whIVRu1h!ldZKN}9-FD4($`P|cYV@Ws z_=WaYf8(Wx5oXC0Q%6Wdae9%LaB3Us2?JnUr~9nN*R}2H;s=zp4hs`&3&Bn9`(c>4 z%i85Sz!#;FeOWiN7B_DuF00&isO~S~*(to(K3@RVsw60HS-q4`49V!=6VB72I?k|^ zRw;O}nIMUXoh|h?(||%QkR>JJW%Rgiz}1)EW3~igKX-Y9X47BYT?Muapg)A0F9y@g ztruFzbc4ScSp)O6IRlz`XAxIMc70<%n!c4tvV5`<4w-f5AL{V6(-~CmN`q{YYWR_F zxs@1buf3l8(rv5DQ;4B$#e+9W9OUBRZ^0@rvS=s6+B5vv3wwBGr-*mTX72_DVc8en z8MAU~ONg0nzWn87M&>}97m{}&*8bsf1tyg-jFKf-popmkc9<%W65D96wg$z(rAsY@)%rhNYglnvxDJ{CU()vJRo4*@-t{1WL|AL~|r z*J{zt!t6YaL%zhtOr@3rtTT?94sLJYN<`gCus9AbxcFAH=0;A1UKa4kNjz*PU7Dk$ zGQ+YdEgG0jb`i)WohOg!2@8{tHIsU*j>vx+%~lf0^tI_G6OuC{srK{5iK>{KYr{bV z?(E_vverXj*g0|ITaN%)GIK7&c)>ReQtD=zP%>Uiw)!bl^VkOuo)JDT`|yyx>Rn+r zEfJ@y5%KG*x?+v0VOia2^5WML5)y4eA2I1k#4#VHd*R4fT1SD!clgEas9=)$Hw3oVaP}fJ{5C4gsa7ZK(jh&=(JdH z6L;<=<9&;4APqqs_i5t2eh7X5zEgmV>ND#%Roa5~j`Qp1@AaqF-IVtq*t<_6A$7U0=`P-j2p>5N0lOO^L`sOiN zcH*J*DGg`Pg^)h+(ARVi#S)JGrHbLA*^e=bgb$w^Ccg;~^$L(Un7eiN8<5Y8^C zD3HjIpjRaM8uw=JNB36fs_^bxLkUc+PJbuJH1|grIn7>p3S<=3{d3V{K7~R3hzazx zJ2GV%i=|&pJtSOZg2-G@Me|#smNRI=`Ji31>}q)~$}VlWM5qz{fl_o^_`4`*L~l7{ zs@;vPo}OTQ9yiw~qzWgJE+p<)Xgv?LTtwV~-+HHLJ>+D-XGig%cs;icxs0WwO?mEA zwWMN$R%lDC8kK*%oGh-XSOYe4C$f~!InEc}S;hG5Mu;#JW0iY2Gi!mp7(2SKi!O zd8muIn$w`>dO-A>Qg&#k^NSy5uXUdgcWfQ0rp{Y2e$`BGRR;p`coDyu6wAMs`bl>` z)Xqe71v>%A_qq=1GrG90oZmStQ7*nkpQwH{**VD>m9@EIDI@=~c8G?;V>D-f&CU^3 zH$b6k^)k4JQV4h0-lL)JM>Sh?X62b)93UWbsfzj4!9@X|U1P^IWA=;*4`N8X%+7Y~ z60vVxhfCojWT9cH9h|zat1#tL?hf~N{MoGBr1?pzI8i0>2m<(YDkk6zqZn1_angLQ zwv#St=a%6}X6Mn^;1_M_(RgVk9^HR%b2g#FP<(+_`3|39PqSRk@Vn%%x^1%EE~Y}k zp4~F&NA4Z!SgDP~*Ws!!)t#&+n1B=qW8T!I5}g%ejT$Nb8($BGsG=G9*o$#I{VfOk z3&BcRCeVjN_-g?nT|4L6nzf@%#K3I()*k%}W3r9O?+0OZA*(G}b?nv7%IiOB=IRqT zGhrA)l?gq@hVW+Dqs!JyaSnq}WF$UGuuErB<1lyQUWj$zHd0|?jogosbo4omh8QOw zB+NcZD{jXYi?76HqM@WZ>dI3>r+gysy>#4>jKIsvUbo3XwCY6geeWZm{orr?uMOic z`Fk!)qU!`#FR{hX-mKJ(^_7F`Tqsm+7y#j8ga{$`N#UDk;IEdzx$)oftJ@ou2USFi zwuNrH3}2UQrCN3D7^2T8pUj@k@1AYAJ~&B{Sjyz_9;l>CtK2c_(IQx)=vQ=9bC1&A zD^6egE|A|)|0<5!;TrLZ^}sp1o7QeWA7)&>DXi+3=jIaAuK8mibCE|@rpXnUv&uCjjFT1SDvl5b-)z7K75 zRV7rD2$UpPPHq6vS>kS_hPJ#a>EYKwT}8uWly7tCUKgxwdPcBSIDb(duj#Pyc{CC2 z=`Ib9R21A;HhZyjmrlrI(Da4?qnEA4sK*N-Q(N4(8?n1m$H(T0fc!QC2o9-Ksf zz&TyK;Fu@Cyq%61(~{0(r@PRA!XXp=!Sk#S=!$JFa&f=Eaa4DYiLiLK|L%F~$L??w zr%b;x7*QqizM%R%$5Y*krAvFK3+&`|aXyy=a@Uv1SWJ|PH!WIgrjz-;kdb8a77g?M z%Qje-gdb=>61|zu@ME(_%fnh3)YCC<$@%j_f{ixIuWAOLO5SX4>J&=UTib8}add6o zvqT|K(PhZXBGRjrXt3>zCD8HW(~M6G()84an5{o8Ve90yPV}Eqo}jqUaU|B5e7F98 zUfOc8598{mnqXyp4Jg4ob?o2JRG0Q*Cn|!4?-vY>@*Fw?2a;DNoW$Achrevx8-HEO zF=iknGPkiHZ;FIvmi#(JbhDFYvi~!n1S>k@w9bgL_PcP^4Q5|nKTQe8+e!g;N>Ww(i4O%?ohP4ao5rg`(`8apEkq-6|_OqV7 zPiMc}W$2PTWkvD-Zo0JxO5`37nEuPEKku>>0~caM%U^6bG+DdXtQ#b~6?dq0DI7Jp z%*m_XPq$Y8XIoP7%kKN|yi^%GL-)D4*qQ+>z+StruW_#HiMvSJ{rOo| z>|dO1Yi6G5KgzwALPjrJR*STk)$hF|1dc#_ycJ- z+4Co4Y=K?@-rD-C8?8ti>F&>`4z2!1H(#4`H^XJ`@PvvfH&Mg{3FA;(i?jr#MjshB zEg85=yAS@d;eoI|yB#mE9KDg!Ny_Xm<@T#7I@~wn;0)5(An5eV8Mt51hT=YN35cU- zM<8H3J96Eh1$ngb@K`^3g1Jz&A`^bI-OFcH(bPEjs%P;T5PIYn83Xtm%30-Y)2w0L zW5VvgFbXg$_IBBcTXwy@m5w)B#aPp7_LBIC>d$v&J_bHqaOJ-F^|mF!iF>i|cs3rG z&S`xSd+^KO|DbPoA|hGFdA++SI0ENXg}IU9mYqK&id|%zVSUp=I>^`FL+YfxWO0JP z7VG|q6K#XONtdS0g%D_13X^@4Frs#CDN(iU;bFno3z1h(2IM%A&|(YcP!yiHI?*Yh zR;LV!2yq|x?rJ}`mAdQ6R^&@@PsmM(Dg7^9 z*i*FEjS8$}HUpVkZ+-|71DOA*5}q;?GAhZn6@K0*^b zam(NGA5_0qDACM#kC0i+W6ZwW*entBY6uY=!uoUuoqz?Aj_49CuQ(n!Q#DoB8B9z4 z=tWDwSD33ELm$g_cXUiJEr&!Wl)x8>J+1Ik6Q<_@+OR#=N@(Un0#lQsH%wz(zV%X- z4_wUmWz&3{DC1r8-na=%j^cd40UeV$LjQ20DNMn-lJjzRwC~&)ch2{1$Ph!<#$4&& zNAnZ2A^Gm7i{{e>{UgigW+&tZThp(i-%;WC_>&#kO_qU2XsD@9MM`**a2oaIu@S!kHc5W$!M-h|0kpK)7Voc2*{N@9Orq z)HsfQena&Z>y5nnBAXt*zt49Ot?-}RjYC&w0q*@{Dpp@h1d)ujW5^0$HM`h+RM z84}{~@lBi8mg1i~`Kz0(Y|1o0_lSD&^70MF&DxjNUg<6Q*oud1)5$bTHHPWpIr#dz z7V*!{;HKi7n?)8)_|xsY^Mnu|_ox9bN5PdoL&(lL;uUyJ$I2}wSvZdKiyRO@lYxmz zI<7UPd;~(NT?n|(FXZR6cyu}T@V7#O=+i+TzS(CGz9!d%ic5o}%>{3y0^t=1Y?b$)@JWibD#?w(;o!ik`KmED)V^h{2&>Y>Oq-NYoKyWO~j!UP_?_d+hPf5 zzBEqYe|Q${KJG=T#-jV{X89wmBIM?X$4aeVZoHkjYu2++P8Rva3HE%baG} zQmrsa_5j3|+Mx8U*5t*lBDG27G3??FNEt*fzn`&L^m@10`LH9S0XDfi8NZmubWH4b z8pl>0pC*8RSlnN6;CG_Rf@ntIGnqPwBUeS4hbmI0vTu*YmknT0%SxmRY>lgD?X{rF zu{3@4b@*q~lbuyh(X-$2Jxxl{jI~-f6Pe1tduEr!P7|&eTnhIVtMqG(-i%xRfWHGC z9Ar6Wc&&;|tzV1gbqWO?pYtV~0mCk=a>Uq^T~@!V`U^VjcpSuJW)6yNu8ST)PUwU8 zMc3AMfey13Z2r?5vfk+iLNOzzO58$2qIcr{A2S2R_LXN`KNq1(NK~DxO3f4&WZs%n z;lY5&R>w>c@j>!}=WirMye;F1n$iazp6Z#in@(`$9&~WXQ@3(LYSgUNVO~SgFt>-d z@S>;UOmxGDD9 zjYH(|>qe~_?KexRLN5Sk-IYiOadN&T50T>_(NX8wnJL~O6m5He!cn~lM@``Kr z1`2lH+#nli8rO1|%!^uYHhg(|wt;oWHGR0LU0wyU{%vkW6%K%_$r`HirL`cbZLcimnSRoMsf2~fT|%az9*V#&0S|6P4t_upfQy4Pkx;Wu5k9G!=`Z|pVC>s z8^ne5A{)-bcW3XNLeAC#I^AP+Ez$&A*+kFU`|N>E3q+)xTa~wF;Xelx_3lya2Tq~A zwS!Cyg_k8Ruz)$NXh_S&q+TUC?ax>$A9`-(#&SX`Xuud#>lY>*1^pj_mmiArm)K`3 zu0ae~H9soW=1xoLr@V|tHXP=;4fr{Um1Dy9{nx{E7@Rfc>XD?dac_!*$15Y(i54xu zMh*h1Gr!Z_z8dv=nb$Y_nSj_BpY>5rt#0)>5dDL(;XJjX*lSifv->``DxT))GKi4A zjG3bL&WB!e7ly>CF>ZmPzd4BRxFoI%_srQW-pq23@ldJ~R%;3ZbB{ruKPjLuAarN` z*8RA@*hE16qppKZt)IYYlf?~UKu&+?uWJ^lT#}?Ss8bZ0q@yC3J9-sRQrvE>x@DI=09C6-#YL00=sF%-tJ>#9~p~%-}1-pWHzRn~INLZRV z$l47YLMrngsHzr;-dAnE>`f~j>fsUN>n{vc`Br1#Eh-Ke?gTQUb>#BU#;QgEj1$(s=Y!O1`#ZtD_|i60_jl>m`DOpInFTZTWc4+ zc2F(0qR9Urm_6q@N3T90!wP7{GR8Hy*z8KY_Po+TYHxu<_-&H+ng&yhy5v{9cr!YE ziZyiYi;t>A#u#Ta$gj2l>)yn8}{r;@yvf%O=y0)#otY5-7$nHP)7v{miTU zumm^yfH2k2>(v`&x72UTG|PYe#bK`~Bl#Nr)kb3ZuGZ6CRlkGDzH$T{aBaHygM^bH zZvkFiyO?r)a+=oD)bRv_5p!gy3}f`COLEkC~g3(-uO!=n}O#*Gt45@~rk) z{`bH=|ANv6Okv`?&z|VY<4xs^1N9-7t?3X0FIV7FN$`V zO{L{~3-Yact%flr6b!1U(Pb=CQrWJxQX)%UKIMcjE~z)`0{`Qo{4EcL?NEAp;o1J` zabaT58iH_kq`1T{IgWx;wAHIE>1&?KB6;(Dd*v#s}8J^LH=ro%Pz# zXncw_N8hAyRA{jq_oY{iv#uPY5+3}eHu~_yntjaZZVnLfbAAVnAAT8=o%I`L8?aaK z`j2bpIsF_e@%9r7ho5|E$oLok%HY4eF$my7;-3;Z`KMH$=^Q7!g3WCewp(u%@enr~ z>$OBbLZl$i{y%=s-`2a~cNB9~6MeBJ``7jgXNU=AyD&sJ-F=`@XB##v+&=ogmoOhK zEvld3ZbYhsfsi??f;a3pFf4l21ZKkc9Tqqn^dnxXJ^geUXI9n#<rc;P{w49h4cH4 z0))j?!(nu+J$qs*SN*$FL-u6x%lSg(P!p1k=&|UbQ^b_c@tP=bbe~AkDSBrsZ%oPM z01O$J(*U>5w7;-)Hs9xFg6@brc6}n{)W3G)b?ClePaJ0(EIAP40GzC``US{Wlo$OU zvFC<9X_zUDi2iZ0>A5u*z;-Q_mn67X0$#sdo6=44gZ}b(9Y^=$@KoW;70T{o=}?Hw zT^HsfYlXGcLUH1?1fo;F^drwHXD%&kf~6k!EYRXe4Z1(*NbadoBX_PnqfExt@!5&e zseYUpbNI(QDS7BP(8?TrQOGC)tjV8UqbPDfw2hefV{8aQPPPPx>ggD!th6IducVRZ zUu$a8Han@piH+h1EJQr<1%T9JLtGG@Gylt{7eWktvXz$=oVC>)78?J%i)vZ%=yR^l zMw;b6gnD$Z+QCVgf<#68g>Iw4dqvNh*;SK(DCe@~rlpZZFZldLiPoc2pNJD4&1v{Zy4*?YJjY)@xjx?b)pHlF^XBdd#8btBN|?q;9}Qq>BJv}AiPuC=!|Q?LC!6=1W26)`Q>tG0uA zOsCgKzt&fn!Zgc5rV+Hj;G4~TpLKv|oYewoiPFWkxMtC)6mw=k5Ta?>@Y&onyvHhd zO8D)#NGX@tXP2kY7n^ynInz#8NmZ47W-UwKoW}}XtEpYb9jppMDAT1ij21YTjQYI= zkk%&sd`Dx>_nUWwOov;I&Ptyz3uPA=wW3F1iUN60opzUTJflgP4#vf{GupV)>1=89 zeFA{-4W@B;yse|QH250Vabw$FiqYU)ZzqZ!+HE2N_(bm|FMGf5GHgviQd7x0*Gsc@ zq)yAABss|n6^tV!F-P;?{c-lmf}Pm99Gtpa-Ocl3vOas!i-G-Og67vJo#9FtuY-c0iD??UH;~AF4x=j@uGxYI2XJE+d2tj$it~A-R0qmw)B*>t57hld97$I z&z*LE)X-K~{)zbUrAkA>f%vsM<#u1o&}et2A{0t12-GDo z`eTw%;~XlCZQ!TVMJwO?w*|=c1081EQ!mKTebW)X;(u#_NwY}s$@s1(l-=>8v{<*uW#tQ~FRbu&DkYSr@i z+4i8$n6 zerTB*QLcO{d7iX@3R{!h4>X^9Tk6)N+VO7OMOZ@T^i zt~=zI6EHpssAD>VMBUB$JO3|LF+GBj-&blB^6Vb} zvpwT@S^O`LR_hy(M2W)IpXP|uJc#4mceaztmAS+r=R8ww{#t_H={UMF#P%_ zqP~JB$mb6G8ma{$B7TAuHen2Rfmi;B~%Dk0>>H2gl^koRUv}#x3>-@mzf9 zyhmM6<*K8{1|O}x+jsJ{ZV{RKeoJ|N+v+>gL2-T;`+|&M*rUDbKW5$EY9|XPP+Rd7 zYSLmL_u9Y~_Vpp$gw9r<8t*p#)<%d;LAJBLm95?L`zRR1hdH;ya)(Ps)eknV(eS4i z-0m0ab|p6dh5{GH%X0FnUL8=%h0&uT4=2b8Ack7f;f|0LT{LUPnp9H@?=(c}{6TN^ z?cT(*J6ioHyG(7F7}8z)-?ho2>6?37$iOiL6;HK0#$4%b>%JHe{t*(tTiV*+$%@!* z`cM;xv$sapGE)C-vJUL++dBVG zzU@|DqcCzZ{W6Ywf_Ndhv+PWEUd=?;NtWE1lfMFJRz9*)aTI?irb;NWY9&` zz5Qj5ux3`hzUd>&rA^mM{^ef}y9H|eF!Dm3;=>|i!D?em=MOS!JTX~|1Y;nwpT2(< z%#Kdqsv1r$slO94xNXAKOA!5Y!frh-z4dvvw|A1w9}rWa`Fm0x^yY87$X?pI7XkWoix9ZlU6TFV0#g741~wcdWYyoX@6CZ1Bs6uWpYf+ zoaWcK_k{5CLghYNg~N_IWAD--hKHD*Kes;PcezNEBOAH*eI9q0hCAIY^H$K@+G@Y! zB$$&LMYBC)GcFbJD9K_U^iohMGLRL1vTMs8VJmdgfDx(Tc1 zY5&`l%j!p`w=1aE-&wt`ArU7~XVV=pAW(P6*xqowCo!0IO;kk?AB`YbYAn)1%8)QMY`$88+`aCE5|2)YkPfiNsZP&XUJRcgG`SV zem|YI&}1xykl0724}YJTNT#-GxDdWB=UW45*u~G#*wCZC74Dyxphv>y^@A`5djgKIf+>5`=8ckYFQq!HkEq!gl|KY^DRjQa=Z%=? ziwWYrc2>twwor!+|51UcG`6lPXH0>Iov(Cc+FTzxNbBEJdCPw=9kIgSq1)CwagpXC z7%5xD@k ze7HT!_TvDq&3dlY0&E)?#w}UhM9U=Zh8R8N8~DXkJ}Eb|*wK1<|8QvUftpFPx-ke` z0EsEv>U94MQvEMcfmF}39Fh`dm6bc{6qqx#S=XNi_YonLo? z@S`ANDCy*d|39cCaYVBx=zdkDxdw|1_1>$RqRrkr4P%_Iw_1sv?Wa>X;S4LguSWfj zkVGx;qTGO_G9*BjxzhN>7F5aw^+EI|MHNBbI~L71R2qNM5vR~?QJ1G@h+ z*NeaeOF(JxbAiY7Z7D5@qlG~t0vzLV&AcB0_vL7n8c@Ip6n~oazTgv`g4R4h6#Y{G%zA7&ABY<1c-N}5QJL0Fl$?k z%N1ae^&_(51Qy__(kb%*X>br-$G2L7eZ%Px{zsWSA67(zm>}dSh@I4Aa%`3!bNgF$ zV&dz+z5mh76H}=S_rT*ISVd0kTw6^W{c2|3d7$oxBCm!*mBVk|js4dlosM1>M~b;u;2zF1Jm zihV?XAs?k}AE)r>xSnvNuziqdPMhfC{l0*vOgYNv$$yP9 zOFcuy0cYb*8hrXaG9v3D{bwV^ran#!JZtw9rL7ys^zbBCK`S7R9pk{9jb%nQI8c4x zqDeNemykMEEquCg9;CVdBJySiAH;FB2Mb8g?d-!|^pXH7y5DNfI}>i$TEkD;cKIz` z3xyNH)ZxdDW)=X2m$S9nP5yuI!ujw^ojA)!Q^ZF`{sb8c0q%G`p8wb0b^bM#tzBh; zLqHuvFOE_|8xVsCks_c{5(oqd0!k4@0d-^mK@`vkB0~`Z34}-&36VkcBE3q-(3_$Z zQIw|iA|2j+LecmB1NY11!}*loK09l#XFc`g1XhV-=L-fz7mdT+N6p_0aw$lCAZWoC zat1!8HYIB-u_;|l{tEMw4WJ{{Y*I(`#Lli^R51zd>y;d9TFdDDtl&-CSC-F78T1Xt zY69s93%x%fHfvMBnZ)|L9oY;8Op^Bag8zEDZgl{0LfpI9%^T<90Yj_IlhCfwHF(m% z%Z#GvqS>s2?$|odBA5GlQhxr+g>HEZi_#CKKCH0U+~ZqYTTqYXnF{9MhRyQKGMaIf z*}VXeDdJSb$&~1wp5DC*=UGzgYYyHg{A4luzBs==E=2GJ?u;&JDz|=$b2bXhgp@87 z_E>JkC-%O&nPbTA-$F#s98FXy=2Z=D;V4H-gRmJiMYFxxMH7Eh zD3p(z#Ov1u{)p(W*PHI>Fefe~^c=6N9pHp@v@&5lEJz`Q_Ar$*_Wc3I>7)th=vFO; z_p-b#!15qxi|sN3(X%_^-u@TQLf~N7P`p2tgS)20(jfD+j)!+A5-+65$85J8Gk=hx z5kw$Q#!Beb_s%|2c;I^UJf+U*Ymjs1E7UDfuf@{*D~yIPTEhJ0_vu`#S8m1R@_wfM z-HdBqu$5UkFBJMwoo?8&Zdp+8xXG?n@vIp;v%Djh?mo7GilqBpXVAWKis{5nm%a2+=2u3um7$UW>c z?tfi=GlwpVgg20sZkNQ)JgN1HA`)$Z!tG4evxoff{oXjsdz$%CNm$5ZUdxGriKXN>* zc^-~BZ?RXCaW`fK3-`ENwkg#pCpPQ8$!rO@hu^h$hq^(UlWr9sqEqLg9`Imx+)yir z?PWSE42CXU`z!@#CX@d;B|eB&)nd-^AuRM;A{$O2Doqw#lWF(*BsD2XL%X74P07>l ziw%>e{`6(GmE&ZYKEV@yKBrWN9W90Rl{99R;yz?sX0iuz=DYjGB z`u7v2@iXbb<4Bz?gDoTP7peVZe+7!oj}gD>$~#|JYo$?R75uu!kjD;L!7zn9$ici2 z+V6^T;(J0q7+er0FQmChG|X046e|V)vXoCW`g3`3`a5OegW33ylxA-}W9?|d?F5ow zsGY6mR?e7|TjnR&KnYpaeW;V|wr9!&N7RME=ei}@RMow}>>`Q}C<5bTug7*OGR77D z+Gxz>z`j)<=Tw)OoFf-?RWbRDsw=bS3Pnx+IVXi~e$!_$Rs9&JdFbNap&c+^!`)r|OEl|Gfc_0W9qTk>&yNSn# zD~-03d2klIGvO{;G8Zp!%}}fvZ`*@8$YXhW;rDVjEe=ZDaiWqP`WF?eMgz?`qr@>! z+rK1sI%T|8&De?rSn)7w=T7GjVWF1%nVS6%`1?m_-@5|aB`_TXO4o+5LwiJ(?tsrx z#PfG~A4!h7hDyDo(N7MnlyJ}mR7oPio|S*lKTnN?$_1?ZXNzk4`JApIgSfvMk%4Ty zz5kIgp~H<4pxz6)sZFPAM`KT|ZDOc&6ucD3f< z@47nYZtb`wdB+%WLPU|WRz9%`qdKUY_okvzjQO1p(HCmT zvl&PL`Y)YL;uXQ4jAyI|vLiC6ktXL6N@NoTnJtPB1Yf9oV98%64Lzc_msEinC^mPC z>(!kOc20i)Q%8p%J#<6AfJkuP-l+Cr&Po5i3J&hC9XRR7nsUi+UWz$jqW0pR=KV7? z%Q;&B)LwI)t=DG@u_Qwf78Immg}_tzd6@^UUhO>?(d8yRvW5|jDdTwVne{}hmZ{qo4 zMy?jKgm3lR3;>a3%rLUOz~NKZgO6UR@h90zXeI}*JBl@d7C4yW^Sgt3WifBCl_?D8 z?jV^FD&u3{L~uCU3cWRcZ0=p?O25`t^1JTP2951*u@6=~eb#0z}2urh7fM5v>Q6wZ$~5x^i7U z!ojj=V2w64&OsiZk6cFUY)oVUE4eR=(z+RPNv#vXRi_wn4gOrj^vH3^juI*dFSv8| zVK9Tp?GPZFUd~U`;%h>W z=JWuiy;)xaHtptOZ^WbIUeKDmDFJ02oGCZ7wDOxnhQAaf>Zp}doRRHleJDNgea;Ca z<&PI!7C*L35S8o|s98yq>TXbpj&hJ-HZ`iCKe!Vu$MtHrXkfEX)VEY0X45p}u3C%# zGLo}LQEOO}8KEs4;K&HCO3yk0^TQD)fKnXljQkw_oQ}fPiaJ zhm3cq$i;92YREO66QJ>Awb;s`XSnLi|HH}{xW{`MDEFlE)6R=*+AtWXPOUYj9K|cQ zsXxNh-Y}}Fb93378iWow3n7vKjkUF(Cy8umGGrYs0cFot|PrecBm#!v$c7!#4c$95P81W zVsO`wk&kRKl*FJIR=YnQS-#@tdDC6}3?`RHb|;>oVzFbyBYV$iN^~p zo`|^+|Hn1^6Uer9Rq_qswVOkI?SMs>Lt%no-xKf0(4JY6-U8lcNDI|DX;9pbj}%us z^S0X0^rG(<^%EJMg(4DLxuT(+mpmO$;%vnU#&`Vc(W2wKqyoI z?Fp}QfPZ-6ap$-I|q-wfuKhXRZxtmql%}ODm#bO>k$NY4AreTF;eQlX?b zLZ2F!%`8d}e9Hah;6C6<1UY?8M>=zEV9!JCj_6>u=DnUBek382T=`Wb+#WdR+ztUyjQH(O1B7F?7i~U21)eB~Z30I&H(3%Zd zj}7P9Uo<@ZJlM1#mOSL{sDrzSt(6NSRAVM`?P%~>=9;-R^aKzJ4x~zY=le51=qS_j zHybhwQ`Kou5ncB=RoC;6x*FihZgHE>Qr?LVI|_cjE2JoVB8yWcV;@LOy>R!C$Cxf* zN*zb=Ot*m6W|t*+m_FewE&)$3{bmjj_p%1IL?WJWPWXvG$K}JxlP|)qBZa8RX2XdG zR!dCu>?Kt6%%leiY?_h-ghK?VgQR_fjXr%Tt;q0`l2+9j>3Y?hvvrx}Uv4kfzl7%6 zWLlsn5We@s8E&~Xh*2>-N#A9e%2z8G{yqk|bm4UOAX~Xu`}?McZgWpoU(#SMMLr}x zEB=K1s=}8l_gn50?PjjXIxAXXHB9wM+oTsr!gG!Yi$t#VO%tQm6ITwV_BmK>mgZW7 zKsFZd>Mr}kcK{R>eXl5b8x;R_ei+|)Dm+dao=Bd>^?j3kL&FDo%XlY4*j@pDs29uK3j5agkmapHsNaG%(_eoL(HgC zx@f>*gLFmUx87af)qD3R(ghIg^hQ;fp4>sz-(v}FnPvx?9aDw;$yZ-z9hNsQ)^6n99jHcp`p{o)jeS+f-6&_vYTi$^y*(As#&MKfPiZ*(jn=poYXrSAtiVu||#D1vyRNKWAs#k;3;GO3F-3x94_sQx=|XL%HD0d`Y2WskfN5Bepa zCkiAfpLq^4tPS(ZCGU;ejZYyjr zb9t&x;j_`c&*?;%(?x0kg_qJfCx!g9s#6m2;8S%>|ATUSFa27d>H9y~;m0X|q=W1Z(G?fm`^Z@z_) literal 0 HcmV?d00001 diff --git a/docs/articles/getting-started_files/figure-html/ceac-1.png b/docs/articles/getting-started_files/figure-html/ceac-1.png new file mode 100644 index 0000000000000000000000000000000000000000..f11ebdcd36951176b3ed38ed3610d82b6facbba5 GIT binary patch literal 69213 zcmd@6WmJ^k_Xdm)DIp=LNQ_8`loBF6C`xy?N`ruOm#8R+ARq`xsDwyKcc>uUFm!i! z*K=kD6+gf4|Hc2^v)1!w))MF3_ldo)eeJ!kGp`ioB?$1S@E{Nffs~}UG6aI@1A$;b zaInE&bStUPgFjp{zk65F@~*^Po5wbGDo-978cP^k8{3&1DofmfKyC*Gsp*+ftC5M6 z+M8ZwXi<60^W5Q!C;=lum^#nLyb|xBp|x_&kB!)sIQQjZ(-o}Q z;;)(!Z^_UoqBGekxyJ4NgI@}>hCK66pv3;Dp+v)~mnw^XO99_GUuWd*4|jvU?xExO zjKcdgN4e8q@LvX_t){3oI-$AkzoiqM*QVWmG+THRR?X>FG-lZI(_X|Y$lPR- zZM^d(mYRyi%ShtjrkdUR=MQdvgBv8sSuA{=o%-}B#&IYi`^n0~<9Ms@X3saQex|fW z#fv;!d8m89%F8%2rc^ESH8lSIAfx?+SukT&!#FZ}^wv$!TI-FuDfA9--(pOZ6kf<|4}=yA)vX(S-8yOqI~O zKweby_lh6k$8~iEetShNdDo5nMV?aRP#wNWr|2|ZOM9+sNGj^Z#)mG-bO}n*=hv#w zCAIz5mJ5yEH9$yiJB&Jq)M#bDSNicieIY*hQQ2#5<>Z)qL;QKA2@g)VugmJH^rW*9 z7QDFH$Z~+CSAO`*AX|5ijZdjy{8DU*M$4YT-ohc*HtV59;;8qe##*6Wt#(ssj_l9H z$C~!eYE)L(6BMiMQ8lk>Xf|xuCN+redqvbo&17xr32E7I5Z+IZ#M-ymQ@K5M|*XUjlL;& ztfr=^_d2kP_hM@T-ptJVM))r)`sjQrLEOdN>qmH&H92N(zh@r#U@7c6tsQzq`!hdI$a!Fp<85VXf%tg**2dKyxAEAY zGB_tqE%~$@VZFu!G#k|d{xjAw3nj7+D?M594)a|sWm1m%PFmi*PaTJ2XjHC$bKkMP zB;@i$dI2`}e@~stcjR!I&u%U$@|WsuVV_5jWR}E;M0OKi(+UgZzxyIF=f1soVW~VZ z)o}qnzV@WA+mCZvkBCl$?HnO>NP*Qr$HH}!nthrt!}FP`t{cKWAwTff53o$FQ@t$TYl$NR*~Uh(t&phVk=Ut5hn<%T3}25ioYy;tO7ZoL~H zAAi3-ny`&c2%_Al6%da1>joVw{S*q05I=>$k z8%~XOl$n@%FOZ_1B}930A*+%izSWe0Ypzt;+&uOC#TyA?rXe=t1miNz-}kXn5)z=i z&ubMeT=X2HAd2T--lYCyV=3GxF{tHN#pDCFB2a3KnuGBwX1GPvRdBLOK))CvdT~U*Z38SHwRP zRed0e(t9k_|Ht=;CCeaa3DuWX(Eo>{E2E*qSf)~*{MTWLVbC>nOdk>(S=|3)Cq6J~ zEF4Huc>IqG|9Ux?qi;1nF+{rT=Ks4F7FSLAwuIyjH9vjfkW);a)quL!(Skv(>f3zF zo`%i2PUFT1c4cK{`f|6e(6?758dy?3C>wfZQ>z%7JpMmRM#H4K7Lt#a@ZiR^oniN^ zbj|ELLUFy({D!~IM{yYkT(R!@@-{p&vauscX6iYme#pql@s1r_%T#S5cis?LYgVF#w^dAim*{va;Tooh}uHSKIU@`}1kW zz26>=+Z6*BZP;C#nCg(>d>0Ua=Q{OHgI->WtybKq3jcpt2^xB~gd~Y#)C-fn7QvOe zjKX0Hx9RiLw-vp%bK`5mlitv5woA}O8A=4x>ai2K4*`$l5Q%Snj9B}n-2bt5NgtRx zme2NL-_@mo!bpST^~m*6KO*Uk>1IQEom`WkxuedHch7|xbjy}FE-(dM=8HHw*lC1k z=rm7jzDbSd%=#ZjDUMTpOTdoag2Z1)B8ZA{jy{M^sTrQ z--hxdqlbS;0u->yWSq;l-XWIts4v!gdQ|bLRd=DjnBqs4qu#W(LZ;w!{Tf_(afeb5 z{t>MRu10^8v@%{N-Ff+uUGmNZ6xZpqOq`` zi!BHyC9mgnB!){|iKwdZ9M2gxcnW75iQ3ewKfavk9sEd6iyUIn{Z%cU@Ioz(h(~Xk zh(6?tR2Y*9ox8z7u}G~rnfBhDb1@YQnWdks?A&;0Za;xl9PclDIY>N^UP`(rommB7 zHdF_ibP4UqTUt63Cii_~5>CG(-KmqXB_)ma|Bg+ zh{gA`*6X6>6MT<%d(jE6-c?2zaxUbmC097jdA-ZhdpFSRmOF$Ow_v|G!b~6YaVd@M zXZS+b71emfy^s6v?Up>#JU+6dHsuR(IqNejCManR7x_PAaLG>K@@^tMPZQz!HeXld zK}dK|sVV^S&|{hoaY)^b>J_>Jzx3g0HI)wYG2V9(FX;1uhpG&x{{hY%X&446+>|MD z3%@WeuSrN|H>b}eqo-93&2q#+1})85+!%T{{Xs*vF+rv(QJ(l?q6OSo>*h})NcSmb z3{eyI8F!qQPS1Q0-OyWP*dRVthcHrjNi2V?f8dbm9&)!*~2Rn%`8CsqskNe3uw{$~*`*TZIZ%7}V z{rhjja)l!Xo3QehEjMOHeb`CF{76?(S>>}WGTbdV&X61FuweCa%|o|L)$NWa@V@d} zi(C6qQhv9SB%1B76&YgL^CWy;a1-LE;mb? zGB=K?-9B7xQzkPk0lA#}gmw1D-HO@fo=zC=%F%Bs44W%&zy=DEt8@=CtCuIz78Tqw2~MOS^;M;S^7HWaQbNDqw{PEW48(k(SnA}g-ar1b!bycm57z2lJ}-_F z9S{)ERE_hBlQ&leCpud$9UAC}uc%lHZ`1KjpGm;Jf4x>1cZt-^3a+tixA|Tgyjbqx!E4DTB#9X|>6>Z$MQ@WqWNwd>ErnY>=&U5d7lTI3w>-0mhSu>p0?S23Om^)?~>#O@(N>cowTj$3+H<^ zy;a*G_l==R+M|s6V&Qog66ldH>SF+8p`lM`>=|j3cI;SVMF!M^?+BXkRy(LH)@?r3 zM1`&P*-SmxDJw{f^l6bM%R8Q?<16~i?L5DvIv0K2-B5h#=wY*uU3$v>YhT7T@v`0g z-;ch)7R>9Bp0eLMx+^T7?%pZ)pcZ%8EFlHD;Vx!jbKQH&2v00cBc;@J!$G*S3m+NZ ze6;~*xOc^3q@vtm0WJs6)ME{#U=O`=_l5W(uW1W4cMFIR!il~6ypmvE;zkX>gN|*Vf4bq$PNZd}qrJL(+c1pW$ClTUi6d5szJG~d-X?ScOY6?}sDy5A zlaseUl?M}3$u#%6yGUzr$2DDJJe8BBzyCf0d31g{h}+50PV;8IUWvoomRLc-ndt!5 z%n*Q>^oz*gI?tl8u}6^7sS6FzRe z6+oSjeIrw$L(C$Emom~(M=@@gs%y|K?=g$(etOo{$AJl+=|Do4<=kKQZ0B6u-igAx zQ%V*CPxGu|_K*#<1QwO#It@6dM~{KGr#QrHcXh0z+}&9V#6H;%x0UX{zW5M3V1K9b z#0vYe@WEoP@nqfG3mOXlOzdT;&u|0zu16Xs&uH!yvQ1p;Lbsi+P+fNIfHbu>1fm+2Aa8jvBYL96q!7nuQ+ZS%1m0w#q2~>f zg9bnzG_6X;K)6lyfL}$CQP^X2UyOqcec`ckcM(z+PC+jd_{CexEh&=vkt#JuK^*Yh zt#9Z;48JJ^R0lIh0K)UQ_ePYc_pxnjH$G>X;>0%k!p~7o?>(8~rGf#YiC;nR*QsK2 zZ6E!5RVpO}Zvfn#`YCYwh%cqE2LqJKH&=w$*tOWcXF0>&64I;$NoQ>>6Ke7HCu6&O zhlTRLW@DThxKR#F{T#saM?!FD008d{4JQyUve0`kp$wFo5=Yx`dzZrV0lKK0yGBrb z^xKBb`E0lC+1vrV-6Cmei@{ieKihN*j~LSMD_1%OmFd0VR#z?XbCiuF6qK$_Y> zcg{N(e?rpWtvbu@l71)kKUqRCBFFB@g7gO?Ed~~ihk4|+#%W;kRODD8B9Ql@fstH> zPM)ym!7I~U9NV4SvHBu!1TlYm%uc+!Ecp+4X`0d4xI3Ujr!=oHczAxQ zp*5=@mi+HXM^Y$6m5+J*bP2w<@i09_Qe8F(Sa&yaB=6l5Z^@)!PyhbGU&l{?D6$NG zewNzz=WpzQLGOe7=H?IEi>EeF%`~c0e5c{q#$xiZzO}O%%rvbq_+T?`yrD6W-Tuv= zeIkB&9W3`cKI`x04x{@DyC-_|bB`amD2c=4zTagyS!ZR;GVgr3RrWU}co`2;tzE59 z($n{eUuc^fWF7XBX)y3o`cb=AFSy3QQnc+pz6=*`P)a}Zz7#Of#UhmQ_duN8p3H3P z9XDn4fdU#G14Zawq=!b?Owff$#|28kzEGIL$<2t)u;se4M*wV-3 z8ZvRbZQN%Dxq{R#)KpFiwarH5up7i70evfydx4$z-q1SsFIWt;3@%S)W4kb2`IBoh z7=R3Oy_w|#%3adQR&6vnjTN*C-K~b~(mnU~LcPD&j2~@l+Pd!c+Ma*G_{Y_Y(LG3H zGWn8F&kAuw2Sht*@ws<^&G6vMg5rbCcTdx@-j=CnT!nA^HcgO#?jdGrBme4jmi|7D zrIBp!IsFaX?@g{gOp(ikGfRHw?Rk~56V)O>@!P-9FqK)s+2tRHzd`xPeY{D}wPLc0 z_0XbpQ)aC~7uQ;S=-x=NqAonc;13=_d^axuzZ98y8jbR{kZts1_87+1j`J+~6|%Yy z^Bv>c2TFz(1nMn2nI5wT{x#7#dhkBI7R)V_9YBMy4rBPF_R=^5z2`_&WCV9f9}nvW zq*n!}7@Y8A3akCCk3k{TyubIC7jKu(lW7W`q3@MetKfo1K_R`@cb9RbWU8?<$9R`1 z;%{Elp$JZbDvsOt5Xk7(CDAv3vJaDeJk=8 z0Xh(Mg6O!5U!S9P=Q`%$Ww^iFVfrpl&Tc44J)LKTOY3EJ2|sx9(ZSWi4ZOdmmVwCn zx8^?Dp_a}dQFRjU`0=Kfdge>EZ;5XwgZBztT7?yK@7?=oCGkdl`6ugd3xN3MQG(6g zPI-xkLV%FlNVudGP1CqX(NUC5zI%yv*d~^h!?KKy6E$@UQXE#_6%#){w?^xtZ~3j25rcq z0L@Mk>N1-Ue!SH>P88EWG?bD>HoS{rOHMu$f+=_Jcd#J-1`fQBMCde5VIlnC?!?jJ z@KhetN$5(z&L^4bDLeiJ&A<3p#}4+P*s;Q#9kCBE3Vc;I4^f_Bg?snDyt{W#dDo9F zV5TWIr>L8U;E!D;aDY9Z|5T02x^NP+VQnjs>)!HUMI&|=b2-2ESW0TBR#L14O=%ua#8|`5naCNUnfU7rsuuLb=RStm_m_kvXXf zFMB7B$S8Ftjt1G{hE17@jB@@AA+8e#I}zmq|H6&ji59%~YhH2g>eq*(rY=s#x_lx# z?gDF))A?t5IM0B=7p$%uL~a_id>EZ+Gc@^~TWER;7Uj%3~MUq3!7P>C!H z6ObHJu+IN=U_OyT^uFGiFevX}8RCABKqgxzri%ss zknqx+?2~zZr(-rXM)h@zyg(DDzu}l>luD{4Kwvp?Re<(!b9Cb=2)_d%Z+~x5T{EAKxujXZIr+G?&4u)^p!Uxi(IML+xYN$ zXsASHEZF;c=KY{pN47r<4hKq%@OsUu+l0vX``*UEBwUQ_@TUu{+CRBn?0Wxo=u^u~ z(#V0OfA{bO@L-FVw6!PzZG)hlL^Lb!x5YCQnL1yKEEB4)oTsWkfx8may}SE2R1kjw zyzNT~t|clwN*4LptVt;E??~KhXuM-y21Sxz?O7BiXu-!pargg79D$rTmkpHTZg(25EdK6YO7%&Y5b^t^Aw zayS-Tt){y8@3LHg#Rn9cg|6eP%U}4AI8LQ13ySglhmL++)1v~oqk7EAMJv_$w9Xw= z+<94R4tzwK6Nlk)QEX_IgLHzfwH`WUk$N1Df{qLee-1}TwMOag$U6QK;y2CwYTw;c zpL)+a@~Nv`i+52jhkvEKJb_&)S(A$MG54Hy;ev%|?Wg;YYZUopG=}lBo(J3Y+e-r~ zs(~-aOD%gcUl>=lZ&CDS8%mT1kNLd7pFL3l#V@CoOlnE~;u?5rURj1t;_OMX^zcVN zlamDAQjNJHnW+aO$Mk~EaF$?)A&|G+{Y0d^j=8-u!nqG+)JVw7<1rNVX4JUul>{nu znxOlW=`*r2qpiSE%)+ixIg;daJ@cMPhjS8F8gns|AO|r%CNPyQWrx>|Q(TOcl$0`^ z$#P(V`;=+kMV=z@{NF(51Wq*My4CvZ>O6ab4e|cNe2~CCG7#Kqd4@61_ym-YSvWcQ zYQf~QdUUcT`uK05d3v|+MBNXfdL>AZB1h!4>x&tGB$_>pbC;lN=R4gX8)J&c)dK94 z?@Dg53tO3~2?;PfeEiM?^*#fvqbldoDm0es*Fz5v4;O8oP~|CB;tZh4dn)woI4Rp< zZfa}Tmk&$5qWm^ukjHS8o*CaKRe#}|-v>sx5iw}pPz$*zqWk)=u!U8y_Ms)b@F=_r z3IAZpmYFo`Um?2>`6PxT2uU;`5u)WNEz&3EHA?ob>(qkqF-On^SY3mDFaq=Kl7>%#&g`7 zLjH15V#xyF?M8trv`9~aEEy7BnDB$WUnZ&BuG< zK8yF>J?8iyxO|fOxh5@^XQ(={WIos))p)c9R6sWKePEs%F0!w2VS(a?T-@6c5{39& zt`a#_+P`3Z0D_6}2!GN&j% zjshk{xvHIZ`nBeGO6?UVnK3!OKvvPpey{Yu;0J?tBX-FX`x-JrK(AtMngp5Inqu9wFHV|fi6wD*2B7+zRGbEr)fY<`4 zJLX}`EvY>j?hn}H@S+&lq)zZ>}nCFLsg)W)CfnSq#YIWtQ#VcF+Aw9sZjs zC1!(|QakQ$WBujDcK8YwmbtFs*E7n2s+%B~)Us)!l3nOm*n<^)mZRE-w9CuF)$Yf=;qX^ydyikTfD!EnWQ!t4 z$m6OudE7lr(|F#mtNciA&nv8GeRm^Dg%U03kK5$Hd=$QEJs~kio+IoLpzF6}_Fhwd zXu_`)Ra)21d-Ct5CL`!`IQKnaq+fzv$%?pl(I8FQ8ht9Of^B>w>QKm(hsUr=|`Hxb*1kWS=r zbNc0IZNAzM<{Zd@(RksNhaK6&qMuvdl`+Y6<|V71W_5^@@a@GSp@Qu1FJw&>3|NPa z{q&aseT5{Y&-b~^I~Y_3fZ(AXum$5^uP>3{U_P=zlR)`D@n_g*BWF$FtY2IE9#-hd z(n|GvQvWlq8#0U@C$7qA7WDasaRXWLOq{#XNQI|TU~idgE}Q1JDxu7>b?GPp$FD9K zzI>3;(OYk8-tr-K z=5(i7$X~ic%xVJ2?Pg+~qwfdN2&&(*9=s+K$w971r~u}j*_T}CYu-T%TYzWak32sG zBN8UXW9i;r(U~9$jfr2ltbwo3a9hcI*dPWhj9DQfCx@$Z!E}Fft|Qy9mTI6EbcmyP zEhF;G+V32CRSKJdJA~SmUY;2Txu@l10`hC#eN+c2T&}H`uX0$1qW{f7An1K;#66IZ z?QyMqi+5}D1BEuRL*Qf$w}(m`En|ZpN}wk4=)$Uwx8-4=C?SourqR~%o)cT&lk;cx z>=Gm5F#}7tRLaw;j*ku&&VTHzvFu14|9z0kRZWL#T0YKs-3KX&4FB12^DBta1z((3 zhLa!Wnzm-zV>nVn<%He0(>HD|zC@l2%o7VBU!ZsWc-uobd)?mW)+`%}MO1Ma^@aO* z)xr>-C!;V7+m-U2h8F~vE$+Fl*E1VWH$|CL?hoElxBRt;I%F8R?`pR9CY;mlL=|oP zZ`KHdBc9<SpIdRkXTN6MrQ8zWxiT<5qza~O3-uLjG+0T1p<*Wz}aPfcOqHHh5_g^I zTFL4Ud{Q;6byTpb!U?bB`T4evj$2(94r%*auSCPNlY%}(gWhu}*5&d#blIq5RP7Hy z99oD7`fXSL80@A(UeTJZ5Z z);sxDgPLVCaWPkTQ=FK&G8eaJtn;v|$#hmnN5?Mw!ywB2 zzT8XDUWrF{JL$Uf%jC~=ec6!E@#7h;-FBbpC0BMi(ekIx_VH+gZj$fi1Bm zC2ApGIL9!0(kif)rEUIo(qd4&?DOnVnEZ@?$sZ)bVL(hK0c1|~PqFZ2ITL^&x;ARq z6X53h?C0QyDqoDL%}#~Lb95ex1vVpbgK%Xk#7@9kz)>82>QrBG<{;J(^Zhr-+Q}hS zczHEkMgFb9$$pQ?ymKs}zWWJ)pdXL*%iZz^y$+T-04^1eKEF44MxDua)9oquhco~Qpi3S5q{I~*UN#W zOqDoRjm+@E5wBAD143FMO>6?H$wfiXjNJq4kL^*T%!}2~b5sN6OC(bTswAV8c>J1o z@pQF=^=Bz|6RD&$q|T(iP{hZ;uAXt)cDC(G)Ge#%!}Z8Y9^YutYRPelUh7-;g^ z{YfW~8*Co@8AO*xp*+ftM5!?G1x1~H5V6V**BkSfObQAg_i@ofHs{l_8bTRmEQHkE z?_Yp?ed~{Czkvp&hoySpHl$V4kZ3S4K|dnZ;X6-1ehprdSeJaD=ot1s6y zaJx5KMM>#x`Nhr+&z&LDr=kKaAuV^{dFCp81y(VK^BR?*K(JW{SVk&3w z8`ebb*XG1?#|JBsKYp~SRx-!Ad^(&{@Ft=8;0Ey4ZX$?J+Dhy_5Ge$Rd`cQBb#5e8 z5PEpaJY;S%Ou@-%05q$qfvP7@qYWsyVz*FLCs3z14LTszEQFIyTbO{TZ)#+RmrQoX zwmNhHkRbGH8#AqZpr98oGPHS^&92K4ImiE5jE3Zw3X+#0M~*=d`4afbphX5uHNopw;eSt{y$-SdSI=kucDmmw0SB8>p6c zC%1fd$)G)y2G~;JBLYbIMvI_{2OYt+p(j5Hx&ujgam_VOc%S8O9?^3Ao8D=yn%JjWxb#*U51hHCnHc^i}F{{@<*_(-%7|>lP`79CK0Nh+7H=dAZ=!?>4i`rf^D8b>N+d@VTFy6I; zqv?oT0Q7-x@DmIDX*jL=a%!k4t>S^AF1MttOc!VmNvtCCbW^aMI-*9_*Vh*#g_sp# zS)_7qOQE4xADqy2cWkQnz@Q-?p4yj@9SA1_%2u<7&%N)#|8*tm-AW9@;cE17ew1*nWrr zz$OU1-N|gYtXM09at>_L=cUl?1pq+rw;h`b@J#0O zL+56EMaS!f{gzMCNqdszlG}W??pBS(l~=ocmbGTI%i}q^@yHumuk+MCw?^kyuErT)0=ojO`VT2_TNt;ed^T>EE< z4(csAY4-@TujW3crL}K@)JGbC4D^J{U)Q1i-uRkoaE$qbxkk=Po&%tO*3M%N7fy@H ztOQ(D{M^^$B^z(nVBu}eE}?lH=51v80tc6@jJxsUwI1`4&t(dNf@iAPJ}^Q&)nFqY z;xV8xF?&F`C8}$~((xehm`Yv6r&d3n?x#bQn`IA>G|Fbh0V%`{3#|M@ps1<`lDtU( z*{^5s!z(=9i-E#)>T6miNwbV8>t!J1?(%TJb1U~`)9!#pZbLgRwOHUe+!N&JUdUs9 zHk6;8nD|VmD%9c5CM9?IGVg8_%MewVc-Xf`T?5BS$|IzoumVZdT?kDVPmGF-M6;;v zTc|!~wBbR>ONRK_LU)rFtS)U6aqc_k1D*8RbLFRdYWm(!lvgxGT49>rzN-Tjo5DlK zwpqvgw&I>xgfs$1$8Y+%58yj4+gb5G+t(EB<@Sf9MRUV|LGMrn*H%l`r`k?Oe)5)C z8mK~QrV1;sj4VAOgv_za0aa{eb6*cFr@%o36AeV$k1?B}JpXI*K$=BkmRg$nr-_3u zhrqYxA`yJKJ^Jp=WLZ|}=I7}=KQ97#h^u&-Wy~$BM2#%{>039-8Pk&?A;Y(lgUzID z#BwoUPQl4$2T210M|U2d^Vtjv;;gCXzq|hn)PDKlPAvQW)nEqa)J5|8xn6mSJl;CS z4SG$l{G|Q%8i--N==DRd2`?%TJ8%o;%A!pRF=&V}~=7Dq(q+ zJ@Rfl%X$-D{R>j=qH^&f<-a_gmIuuXVjVi9c$a@7lO*2~{N3rwnZ-p~r#D3x4{F!W z)DR)QPki6opp^pPoiY_Yh^l8j96q_kZSp=U>j!@7f}M+qqwud`1$V;cXx_%b2q&N) z?oVsZbz14rf%DAm3T;UN(!o<24-{y%!jD&m%hz`&f@8AP9kBC&?27-<2}n@tzN)1) zm2@hI3O~BciR6a!oWyx1TIrg*75q15?Oq+kwAZSqdACp_5&ks&*v74i7mkEKGPC= zwO|fEwpd0b&&uTPlqi*E2g~fv(f~7f#5;qdd5soLKDWz#bjPdbxFxs1bMz)IijM7} zy2npE&`#Eq9tCNdJFPqOtNoX%6TjpjwufAEb9$6~yCYi zjf7~}^Q~cvlmIV`#e1g&6lFiq0N7wH#BHlvqghrB_)~zg?xvdkObf3VRcWEzX1Rrd zAVRPuARt08rTf(;i%?@b6zK_hp%-$m;D{EQ14B=8rZ+G3ESX%uQ=coD;o}mL+ z>t6?(T0bOob`jp&OsLnPeU>;@JdiqH<_AA9H+SMu0QNv(z5iqOGpdVaRh^ zEk1GLKxn(yFlS`n?$Z>X^N*eap8R^T=`I8TOx7 zg0k$Jt8EF8HqiVr21@h`eHyB&?|UM3?WBUlC=Vb*8t#3)QtoZltoe`olsDgg?E*xy z{T#AH3OzEJrRLPeJjm2vvbv-e32xP} zP#SpH{^r>*=$)!cTR+Uo^Fgf<&f~8_d+MpGM&`Gq{>p_w&J~SCMFyFNXT)6En4Sn0 zb&_uT{sD+oElkGdd$T|1iu6AvY|>R$e)OxScrm9j5zyOqCqo)kO7{llvO?GT&HlqF zMLf-oV9n?Rco6dfQYP6OM5Q?8?6Mdzz7dbySuRxd*VhFXv#}9TjZ$=i-?oagiDR#D zF_u4VAtN7RlnwL4fT>w!7Ee=wi$>oZmiqJOneU3hYufC3wUOmwbRdar$ts9PG?y}p)L(TS+P7e zM(q2iXaokeO#Zew5NF3p>}kC@L8p?E0SIj*K1UI(MPMQQ%K^yUyF0$Bu@=vNO)~;a zRcYtpX_cY?Rz)@2SLvoUYdrDmldum8X;05ik^hUP?Vq_oQEYCYFe_bKZSP}+^_B@IGB>6sU1k54>_vnvmsalQ|FV&E zM?RgON!lKj_#u`J@_3MsZY}gQ0%7H~ZJt#?E;A`0%AD#O*zo$V%QlkWt`)t- zCo#I%D~l`MC#8o6BPSHWTdd0<$(jU!7StPjdUNCpA{|9ERc@>cI8r#*zNtDj5mZq~ z7!$leyXeVBmSgaUir3WqIq%x|+kK_n(9a$21{FjOntTm=K76beC3JW0 z$WUkXuh_;1riV2S2s_})P8{rI^N%0YrnP}8X!A6$C7jxZuK-!+oJ@dyOIP)vW5>%M zoWD6w!*lx_xf$1)(gXqNOG`-IlwH(K&pc_j%bI#&^5j`8{SIr6&3;EnAL<=`v*0P?{6r3R#O z!Izr!%X{y|72HDasowgO;=itKirDOX&%BTi#F|A$}%|J)ak$5&$B_(>uLBQqo z$O>Q&X5?|xbu9HuqIpT3IlZoZvvNV^)~NGK5J8y#(daqR8-;c>pjQeqas~m`Uik$K z4;MUaKa=zX9qZzI9A&F_xR2Zcyw3AlE4p~GA>GR;1y*xt7>41}LVKMfErsPw#|8}( z#Lpdt4usOV@@UB>naO=;`CaL9Nuw-omdbZfA=Il_3$l>}DYC=zpjNVM6g^O5^Jc|+ z?VEGccpSx4T|h1tRDsT8Ru?{~=;Yzd_2#EID{D|RK$K1M&;_s#S+mOM5vi+6Jjfav zK@K>T+bR;4)9_u$al*M`gGQ^uqZo+5oBJY9+Z3Q#BlV=!8slxxlq!%e(Vk6CiC-#J zkxtm=(P%X*PO0Nug#c<3kU z?oZe%4ADMoaMV~N6p#*q`Ml-+pAvMZ#kBeB#6ouB9x@_?OL5?Wu`kAg+qHu;HFh5u zneVefbc!VETxlB&L5EpIxfm@}jnwDDs^_k+6Vk-mI?6gW_sv$)tn z{U=5DRmEYLq_`0~t!KJh^1H@Akng%-qV?^8uPdSy#^pG#y5?4{v}25fG}tl@X2l>v zGarpzs$D?hyhB^u>7tFS51!Jcr6= zut+I_L+^SCbzF<^j&81;DSBt3un)MaMOlH5Qqu4PmdNZSaGDc0FW5tKOW ze^>Zf;s3uY{Jx8ntKKELPY-rtUd(40xMnES9pCS~^Jg>BFhik;sjg2o%H`)%Dk>={ zWyrO8E)I(=^kii@dOIO~u}+6PTBO8LJn~#S;9J|sXldsNJt;Y=i!_??i8DrHUTl8 z2i30M9R@Iq)#pRUQ&Zbnivq=YfB6k`3IQ_&)VJGs=7-%^-Vt2B)zx3c9PgoQ((=K4 zJFV5qFyEw!5b5%lu8xl6Xrb#AzTYqD&m-S4*W*rJRwet(m3(E15!I77g3H9QLkpwd z{W`J?L|wt9KjYdrG~H|kkNfCJndQTANtxbT-F3$1y7Ug~Un3svt?%x&hLjTz7Jx()pY z8xOHS{*Hx?rV;~g&DK<^Pa-=@A*6r?58L<8(fnF6AjCfF)YV$hX`L!;B3d@Y1t3J( z6(oA>7})1;5iP!cH70UhYwR3*;z~q;(nfr7}g6nYpjRhXm;^p!BOp_n`A4 z{wl5?=Yky{+mELtaE6j-fozWm>bO_K>2-w)M;qMSDIStJ_sr{Xt=1S2Vy=fjFz1HL z-OG133%HQf0Je?48{xsG?ArLP)S0f{vV-T~vK;;n`IU|*K{awQzvObQ2RlyXHp_ut$4 zY8P75WXJs`eXs`yPHS0uhBtKgovY{3#Q(`Rs`09Yu-gy}HHeBg^knDc*g%{B+zfI; z6#C7eFHg1TYQ6!*D*|akKU*$nyrU!e5cKan1`+eH8YqyHmXWzvdxWc?p=p?oLye9yJO+GQ#fu@b#}Q2O83cmuaL(U_51J;jG;wG4H6`|$o3~!oTbMt7cY1}*CjD!Z1lc92_ zWLQ(7j~=K!u2*3Ysm5TjabNm^l=Z?sVO{^>cH;7afHIPKdIU;yE)sE}9W+)G0pHs~?;L1AB{qG`34WpnAbX&irn&Xw zKj{+*CrEum)K~I4<++JUzphjqWa2{fKwF2)X;IUi7X=IaY@<3knorLWT|{g=?)l>n zKO-^Z{?19#7nuhx45ld1adU*oV=17F35MliZz4!MY=7~v@B2ShB?eqXlNgj^dNj8H zZIrojF&tHkQy=gF=BH)aE1j9efuB~+ynMxI+Bxy zNOa`nWv$2_P>1j&V3Lt}Wu%GrxS@vxhcT{j)_#Ar!F2S-ggMhuhn~N6+g)&l6LE?A zN7fEP6KHUCY%42E8)y*5$afLi4kGVeKM&yKIRNnYK$&I|Q6M`6iDNhrc)V(%-R_J( zRs}s!W>?pn!4fq3M(Ugc%=;ka$uVt>d;NF;)WW%QhNE=nKvXxe+u|-*-aiQ^BdPN~ zU}yPKeQjYtfshtw2#>-h9s{?5Z+NSz;P4Ht@O7XqRe*UbcAjG?MyE1r>1s8oe-~h2 zo|KdXG1RIkvmao8taTY5K7}b~e;KvCD?OTO&Y(^@+nA(9_QO<+! z=o>}r4XnoVU8&<+Zl9MM8$x`B0zU&&4wWkGo#TCQa%bezJ$#clG96eM26a3*?I4|Y z`@~SMTfW5kWB(*Cp^DI*Ar;+R^EKhIZo%VMD!IQ5TO!MDkI<(ZH1HIGEY^PQM(%sV z*gCAEJ5E3gpdt?9cX)U>L(c8>OtLmk<5K692(U#m46S-oRya@Q%k-n z^zt6aw~SsGh9 z*^W2ozXqDIQA7ctTk&N6i?<^0=x1MFP-vF|xlfc>Q=jj1Dy+OhRVL4R+?AmuZ$+-u zW!|`F=X)QNra9joX5dOlHCMSepvwv*$pY*ZN$RU9T#BP2kP;G1>F zU*W_xeRBS~&Z=Gq8s_E!aILCS)l9~%tGz((V_84| zE@dyKAz8JC|Mgtt1r%5^j_QqF`{WPT=TUbSkg@ywKU>N_8jlIsgfzj^||qpv;$ zzj9y%PuEHg{CGb%&Sm@+)s8VzsAGINmHwlmC(scf zi8Mj*HhRZEiG@r_@HSaTBE~(`b;q2$36GXb>5%a^BEzK;YSK&avlC7 zkO`?kA1m~Bm3S>X5}?d|CWWvQN7QV5G0styfO3b%j}*_W`i*+#_;65Zj&9y6wn|Sv z)I{cpA=(Ip{Q54L1b}3}z1Z9t#P3CL-67jrnoqgzh;$#+H<^P)_P;FlX8+`9bsf$z z4n$n}Ws-gW08Vg;OC0#c%wtnjj?0bp!yQBfLc`UWCZV9NGX>K|M0Lgjo*pH%)uj>| z8Oba_*e1s}!~nuBuP-zo$Q@=9!7uo9q^dA;FS!Gp9Up(x9tU$fk1Al=V>OMo;!pDv ztj7HuZT*+KB`QL%T!_4I^?sc{o!2!#Yy!Jzb#--By(WZO9nHqW$ot3 z{Edm8~MXV7h=Z^is`%F{|Z8gV4Zkg|Gij?y!wS( zjKzkpYlhN<3S(UdT8Ht>m)pCDV2|iyKw$^@xebm#%Lf@l@22n<3$iyS#lhYqx@Zs zjPq)hN>{kC`%;qb?wHB(akLpAm0(4;ek zzkhDWLDkSrR%cu{K8%;oy#B9a{&Sb|67#4*K^N7bYkX?q!1?1rj){KQ^E2GG4hwPH zEkPDPKXQ7Bs?cwZ z;003wiZlSRynv^5kls1_3s=+<2!E}mb(gg+o}$nycXJfa-rL*j0P$$!N$Jpqfo0H( zK@_YHmz^sEHCvH4h?3 z--Jdw9}?a#v4Ijld4u>Fu-qn;Y3}2Jkgy+j`Rp2CJigdiKF7PYq79c!KU@J#i83Fv zg&&0)ddq#eDY_nOHEJ3gNN9mJV**%U6(Q*U`s{mr=i;U!%^5iSgvWCY(v_jZ=L#A= zlZbW%S3JLc^5dZrwHDD6kGa#1M9;#z&`^@O!Qxz9r#@p7pINhmY@~`9>WO)MIOg)L z-3>)st(B>>DmP+D8^j#*%u!v5C^wz=*gn=3dnWD1r>PSTX2yV%=f<=E0l3mR$97f? z55)PbZKDvy?qogtW?mEN9nel!AF8%QhL-b`tDFY&z*c(lFZ+p_?(QY-|HbCwtBBAx z6=|TKHVl3x2oWGj&SOtK#pehimB0NFZg^U1+(IX}A9lPhwnFFsue0lq?shuVXnMxC zGG~>o2OuN`>&vCJ6>Mfs&Lf_MA3(cxx*!_}AEiOnVY1&ci}!oN|HIZ*KvlJMT|kgh zK?Hds2uet!gh*epkd_pXP&x#qr4fB1C>_$MbeD9QbeA;J-JSnBmlv1!{o@*Q;9Jl5a=&I7CG#nwn|lQe8Gp`s+IV%2jk=94H489 zKp>GV$W8TFhnDPt>GUhP9?)+KhwZ|C^P;s*4gHz`#^q16d(~!Lm44-ibf>sPV5&b( zd<C{BP;|J5O=32ea$}*D5MP7nSTrPB}lMcf9>SHMmak~lAB;Qj-j0a@V{Ca zn39-!(TX@&-{M1nFs&IlVWRA`zF(suP;Y~l^rgS#>2xBvfEL@b4vVJ#y}SNCg7p4G zk{ZESD?nt7pTn=D!PevlM%_xW>%l$}Z44tgXZwoUO3h(G{Y{OE)1tFq_1x@w>0l@M z>`h^bjhu6Q5;g{&K+rBlmw`<;*KY;P+X$jxBJQiNPMvTDmR)YSNfVuB^2mV^()D{j zPK#dIs{l(czwv}}XH+29w=iwP;jae7;riAKL1$j8H$0A?b1zva*FbX0yFS#Xqs|V< zqRCv$P&*i0nVg7m9Y{ORt>>Bh?(isIC0*nqjM(#Da%~-nlmEwN0iM@13+aYhrM1I` zVf6E#^_v~wSi-{0{1m8+=XmZotWG{k*#ugf_i>N%-DRnj;+wBeSwfq+D|dSqEz|dt zaF1=eTrQV>0To`V^=w0sKCiyw-`xo3!hIZg+;MHia#4aE+tEXaH`!%~XH$uFVh3At z_=W(u$u1^BUSpu);!>U)b$G$;#BczTvxu0$!zxPz7*$c5#pDo!{p*pfo%M7ip-Xp{ z#UAV`mO&-D4CWaJMp`urd@OSXkbYVz3+QOA2{Pw_Xo|ochO=~r_=kC{0_TuM@%Lnp zL{s}ebOx;V-!_^4iIyD-4vE4E( zWSnOec(cD>(}utrZ807gv>$m#f<>~NuKp=fJdeUFH zcX1*_dFGydHe;L*6w;UJH6p+W07a~D8u<-$#se~S&t|#-z3vgh{8g**`m0U5XsF9i z5c?oU*JH(Tn|nufx)GNVqg_BMhO`Sn3z4Q(*3)iY>4t5S5nq<+bMg@6lSl^*S9Whj zpW)RVRgmJbo-F zIJgD^hBVFkd_P-av_ned^YJilgZe+7mGrOGYFpgPwJ`R!)d45bQDDKPi!5mq>OFv^P_>{`(@n4>Jmi4wc(vr3NxMXCU7bFZLd#!?xvHV#?kF&mmYZVm z4aq;~p5R_GIC#7plPAEXgo~NyB5H}%RP8-Uqyy4j;aYJ5@N6jBE)>ld&!nuC3)nBl zN=FN;Av8w9;~oj2^^m3v#IE;2Uct2a=|&RL*zHq*tmRsZU_%7Iob{hMz2+t9jrwWY z6$7(7fib(ch=_<%Xu3dY;c3IBl6FPk$nPHhmkWZ{Nt;fhh~;BsP9k*wpkjP?-3Nh( zQG$?Voul51aUo0bS`!_ge||?$$y*R3lW;>(qwCJDj3^|B4x~NUtXlp-sr~+3Iq&4h zGJQdtX+4MUW&@>a$Z50wOH3=X3-{f0n^2-jug&5Vl)Vz5wHZC+K8xqpy}2CC?7H0+ zqt{pXjN58lx(Mj*@g_qFCR*DL5Kk?KDzS~bbK>99ii!4cc61sKNqMM259jlV^VjOl zy7H%%v|YC1i}++bZisDuAxWXBLj)xt?c7&pyOq{!#8>3FPNpj2Ae~S=AA})z0k0xT ziRz(KV%sggqjLp=A%U`jH9;A=OOxqJNC-nLd_Ril&kp(eG=0*YKbr1?U?mitbJ0OPmOx4XNh)}t;^1>+yG~pSfebdI4)N?8gL2nhKGO=SZTwxOd=hIa{UI=n2nQJ-KL=9|pMd#$?f5eugSZ`g(Sf&FR0 z^c{|ZzTwsCmfla+30Gv0)NW+I-OeCmaraxOO&YBXd?T^;9DDD~9QMbklYHz)Skk7C zsVfMFR~H7gtsl~riPXz}4$j;$rm>!-ProVSp=&C~LgOT;o|B%p5p+C&n^F0}h}W-s zy)jj3;@9qq$2_qsQ<*)mBjoU_m*Kh@@jHtzs<=_|{Z-?UahjG7-uE~b)PDV%XPK^+1;hV@Y=4ZX$5?m^HacL|!^5dYwUWcu_X9W#B`U7yO;2P>i{Ni}3keTJCeKypIhbDYG!@x0{0B3x zIfcb{i}hlG+ujK6fgcDL-}uCdw|f}b!O5YIHxE7Ij2$h*!b!GJ6iX}-lcU`y5PF|3?jQ>9MDXl3>@5FFKDD?+z~UA88g+g zrI{_B7DO^uzf(BT%4|_4n+}ol|EV#$FOqItKsn3=HRkExo#Au)SfQjG^?L>)VTdb` zxS5hZ=DwIxS`zfMWG}#@<3*{Qz^@eYir)o)5ilLmT zNbv>G`8hf_taqwP`%X(}O0B7P%{;6kWn|Q4P-Om|P8|{AE7j{z5CaQ2gENmWK-CJ26e1w`RxyIlEFkaVzde9M%{-q2mP9QS_)jKLj_L!?5 zvyYt>rI0_(eqTvY;*7k!Lnd*F&{)ZW+PsZzSbtO)^1mIirrgH#iRhJ*u1DihiUtA^ zWg?FMZSbxWFz{lO``Z1dacY*dwwy2To+Ff&ZhBo4bDpH61l{)3fBHw;;KhI2Udu9pZDHUHQ z)UW0~bHF!hp9m7m(ynkpm?Jz_avOUnqOHv?&ZWSC$p)NFi%rfJQ&(`;UIy4hsFzP7 z?pTbm_AjLTJ=d&G(yY^gsK$;*ih`G_baR#HGf09Bo&icvHb5bxL0ksz*#?8)?QKNc z$Od6cegwxoYhJrttIV`inEdYxL|fDf=`M8Tf|kcHPzJoP2Nt$a$Y40eUh}@dk=R6|cEplwVw&G#?mmEg^Ru20#I)H`)9FsSFamA8tNJGa&gc4#n`RCQ@*RuCGlpMrU=Kn|u_0 zc;?^K8kk>~{T9!n-~`GO-WmEnPxShrJXYkWRkgc zK;}=yYectADFxy;(3rk{Ap83PlI`l_Tx9!$_CaCns>AcG;W;Pi+Kh%*|85z0#Tr_e z$Y;$?H^&-3GFS?6?_*n*MPFe{y<2?x%T>8V*RE%O2~U!tp`eTBnnvfxl*xxxahZ$j zGv1I1X3FZP;2sB1?ODOIe))?n;75&A4%-Bq>0jpNMrUjy%c_y>@ell>myG&~@7}0z z8ErVtpN54pNgS%hyc#QT?dO26=zsh3y5sSj^N30A^Y?iELDZu7e(UnY`|_{h6BI5ym&ArnNscBmp0F3Fm%g)9)%B zvVMn27n`EzBFE-zBLC_n{f%`O-lZGEj_vNGJWWGh3kKnfm zsn}o3wCsCURgz#%d~<=bEv%Vyqw?kU}o(C`W+xs9Es z{#jDXWlR4#xK8Ti>oxG(o1dyBa0p_2CC(NDtGu;4e=ycQ+pHcPUbE3oRr27=u93vbmGVcqQv&bu{Yfn zW`tUpD?A2S;fQ_=Hp*ArC$%KweI<4lXI9UNIu-Jov?nP%i4xWFZ8`VNp#J?0tRw{v zTja3y$Jn?nIsui-1h{o##xL_Azd#d_o3ctT>dYsn_q^ zM!`B<=T+q{T^Z0EXxjDb8xXdAiT(R0Zhk9vZP~tjO~?n6aM_Fd*tcHkBB-zV5h~(s zjeVud&(ph2GRn87r}e%dA_=!tyvXhfc06t0=#2c>W%M>Z{HIWp?sr>drdF9c0&lg1 zm3)-xX`KK)U?`9(_MLSMk8{YacYDOQYc?3<{JH=jVnCd(Rc3R1>dSyL`sKNn}6M4 zGxwKv*ttAD&sBxuUW# z9eD&_nPQet$7=Q<5oHF)m@`6?XSBXZ^nTq4)$U{&ZW24#U5-#gQ|&zO(`A_@OFa4@ zP}f`JZNF^#6ytqB|h_J$vrFh{WTMbVr-YiVT_ zFB>I@G&zbUKYGVO_&M$K80}$=V#jfRZWZ;-xj$^L?g(KIo01qlI?f!~;terRlKQ{$ zV(}u9`M=t-CV0D66W1wMZ5Mw3r*xmENV3IB67Gju-8{j10qAdAnrY~(W>BLmG>7&2 zgjco3xJAJd!t63d&A;Anuo1X#AV=JzZ-45Gm$9G;1Hm}^S=|+-B9r^U!Te!Ie9c~? z#-Io5e=!lAbF@dI>^HdJkS*G{*2_}vRNuz;Z_lOP*qGL6R8!33mX8FbimFbT*z zH5!;MWP9{zR>pTu(w^MAFX}4zE;~$jg+}H-(QSIkgK<8hK$3KsGr8VtfgBrQ;zo5eLNf2K+yJb` zv3a4G*#dI+x0RIjWgw_>ZNuKJs}}`A&`$XSlP2NjcG^=2YZcV6)$gzB zlGP?{oyL-6W(m6p83yi)nNCTgwmXV#ks;{}iTBP>wuiGGDOcEWtg*dsd+Wm0y7Q99 zuE|jikb7Ztt5D&D$lU{RTo5v9h%j9ia)vYEwD9p*IOe2_p<(+`ORjxSo05^W2LBEH zPd3qa8Z49MSocZV!#zEYEAF^1@pP*o0qeqyh(MxDZ#dftRg5?Z4%ig2#1DD~sE7~> zi4w<9jF=n)IXC|FP}2q9#Oc86{~Dje{djm8g;i}+5J?1C&ox5a@rGw9Bf9_P6Ol5r zk=jvaAua*!$%-=OQ1ceUG z)l$u36=$-$|NJIoi5y+Hyfnnsr?6nck}Q)h}3fWiukoGm$q@;q?7v?s@a`*b5^%||?j!c7^ z?=LwMX}X+4OhN#Es#;na*@pF2`B(kd>3uHfPY#rXgp+J@=|V z@%c_~!#1Ws-O1{+Is9D?VvVsIuWl*e;$R0xl~>r84cs!2(V`%DDb~B)XV_YCk3xxs z{@UwP^_z(?$7Ux)>4~DhygFgm*~B9}lUx$ET?(EX>J~dD+kGo(y;L=>pDWlNMLU`9 z5Pi{U*5Yk)NN`eRWo6|a%NHZL9_;z(;ylcs>w0>sA7zj4x4+0$OzKGz zxuS@Pi3#t-P5$hj-%iu3jZpV83a~xb`;aU4Acm<;#BQBEyH_j3dXo9axz%XD$@4lT zrp+8-<8L_am?jJ(Olek$)$LKtfiuDL`UU-K$8_u~1^VlvYQ}OKKA+lV}5sIkk3gWBIk%OmTf~bg_^@uWq!1RoP>prAqE#H3Jc1)E)nVAEK#;mMbFj z2{3I?Y6ECRU^?m3DD=+#5tj|8@}Qp|6gKo2rzAmO6adQrZ6biY>%@ulKgc{;}^7Hx>wO zX~@=6v|obMnH54d-@#7?R_*2F2Pls%A-j2Ad>|pIPIsiLE`6E3_ug&)anZApJ1sT* zCeW73a;iysjC~RXJL3m+?84OcfUM*@H_@4%T$HOY6QefDgA`c7S14N-43J5JROm^9 zAXiE*qEq#Uv|Q+m{+?D8%V6;ALVL&5{lK8|>)VfvFUu%7udePmAa)G!EuYguA;k zVcct<+htl-VNjcSmO|-u!{0j+iClg-NEI{Sv?B9-_|0YM7CFHwh9Zy&j}?yi2qm!! z7XFgrDq}zd>iUp(a3fNGkV+VTb#4A)hE77C;^J7jhVRvNMNJLT{DDd4Ma7CExf4`W zUG+a4)}|Q&m+<;yx8<01LI_(%p*SMPv6*g157iubO?ZMDW+S@`r75wYOks*(vJuWZ(w*lRQRaBW}1|8UD@V zGIuT|Dx20A_x(YE34oyJSdm8vj$ZVFSgSh=6EBe)b z7rC7_7m0lJ@!_ct*4;rT(cuoh&atl(LYjZ8Zfw4NlrDu_H#sZMg;0txKhfm-` ze+gsZ=3o zGDGWIHDXIPedL3m5kWIY2?mgY4L#p&V`#yIU9y>M{*uP^ASAdwHvNoimjgD-NFolEui}r@zHl zYF&|aQ9%51*M*Q>nDk1)fJRxA?=vC7wSOgDumnxe0dy?OaG7;5Div^zg!h8Kte!o4 z7Iq1)u{VuNJuj)=Rt5c8M7SZGr@P^KYkP=AmkJjs$Xj5915yb$R4WcjPHfLQY}lBX z?g_30XZNvoFucUa14CsPr_Af<~%dUIxTAQu_UsFJ_2g(8L7^Y!=H@7F9^#|Gatz6-nN8nF)nMV+Hdv`-CJsjaJ$lj zo40KYbQ}FS7=WCRgP(OIvmy9?(JzQohay-PA_Z6+v-UI~o*I>FIKTTVO3RG)Mz-;B zurL+~d|f*_*Q|dp1{=!%PA6RgGVB!)SYwzWcmk+;(N4?vcsQJ1Op9MbV%1M#C_8^_ zik)*6Oa7})LqEu=OK{=xpfV%eHHy9KA$69fRJV z93@FDl9#l71{10l5ed@%xiM;gBy>pex;>%ylhVJ{{YN@sSKN z2($sftRh04RvZH6(Y<*X{VOnE=#h98r07}>Q;b_czhDN4FAyM*uyfylp&Fi78-UE| z@N0x7hx9>oYYSQ+WbQ9rzBQX`_C!gw zWmjF$sL`*RZ&mxs>kRm}z8!io72b-K-r!1bwX4xQVlEqp?)=Vm>f^hFu&a^TaD;HM z&8^3J9E#8i2w$$x1y&BcmizPUH)FPb^p!@lORo|`pz<0=Rx<>TU%lk)#j+SwUrPYt#&l=ZM0tGRa z4|H9K0qkfU^37jQW}VAPxaI%xP`n@&Lf{}+AF+vqM4>;75cjMz{bQR~9iwx$ zY4*!2xycUgM>Kb>(4o?-Y>m?*M<6Vim4E~@fvNp5c}yV_77pQA1R4Nx*(G47NEbK_I>J#uL2}9_m$d|j$N48!XAN9<1^tI;X)c<#=8r-)NXVJ+f70$JDP7^Gk%!fQn zXNt_f%u!kZk~Yso*(fTmGkEBl=C6O;XUX9`6{6OLqz95&@ z{Zts)VthF6-2D{}hB9SInDktNxEPc|6&93g{*kWxTdi;=B&%jdTZ69-&>oRd_5M)r z-pVMRX64a%u$?V`33={CI3MICU}p$XF0Idw{Rbo^U{I%H0l&cMJw#Lt+3`&vrZ7YK zCy>-9xVHm|5gS-X2SQSD5W9!2&T`b~Hv1(Bp9~tLytlFzsxkh>9FZ#3|JMdXhw<1y z@d8@$W|33jOnZyD*5xSc0PD%-8%U`Q%QFYW6iCGf#__0Go%^5&9s^DJ?UPndpCi}H z2k|90l6cF3cH)07_9hYwfC%XYG-?_o`D}9CZ_C5=Qrd*7?XJs4$zToa1`7jnA@J{r zhuGzRN*@glDdRxMQZ5Lr#n?qKdJhR%zjS@?0g{ds!5~m6Bk~Yq1zP?DP^hi1U8P~z z2K-po@d^h+%?Eq;2!W~3yWSW(iwf{Sv41haw2_woZaGmz1`i9E2)2Bze$80&va&v#PVs7hXz@>bO;t9B zL(`vhO~C(DU$*XRbTbo#SKCVw#21Z8h!QS9On0p2>5mCMEj+o#~_iVSA%$%7;fi{FKDXi+BLljWd;6gKRh$Q}t zxo^i>ov97>Ar}76Gpjj^)W$&W1`&#}^9qc(S9RF%nGb#PONMbuu7Lb0@3o*#M zgN5b`zu^WU6MzwZ?OTzs;^M3Vxuvm*wt6%+&z#WLyE{}IN z{60>wjW!M2MeO-IEf8lT;!F1+=i% zA@2mx8CL^aT2=q|pzDD%ki9R6I4C`Lx{Yttl?x}Ddfx3upo7olF$oTP(vL#13%eeH zFhHE1!Axvek}uJER!2VEH=dhw2E(-gp#6eoUSKar?)uYMWbp7u-DAOWRGuwE zwUF8}k`sk>iUC5+=K7$02F`(O0hLTO;8<}_7^_~PF-!uJYF+{*WwbS$6{H~bW(H%v zUFRB435=2|&M(r1_HSPn)=}Dj*m)XNSq4Mpov#*t%nOc1>TH2&qm9S%RC!YTmeB;tIuk)$ z?z5@NYX9SR=K(s~1#)W8wDm!&>GT7!@UO?X4ZgV_{OSc7^0$p(?M^OzfEZaTy44Pb zJV>PJr0~-_ISp)cpxWO=23yqsA>OlnuWAocuaR3-c&+n~0OEO?=;qqlCHv!lPb(Hu z@IRNROK*#FU1{K728*AtF2IQ1V+Vcy>?_-l;fmxIh|9Q_sl)p6we>Vz+t|ci7ss<) zJB70rk z_3=>atZ9j4=-L_N=Ck&gH82FRu#^%HA?5LVeMuVb>`1_w$sM7674~5FExXlbZbrXO z-CKctitMWwv#XrKc!Z+0AZHmWBI7sNVPP zZV&n?qM99J<7|;UKXGm&NjWnizi9j;&v%2?iOlJa(($IwjdRYf_w9go{58gAZ~tp< zw3F2i(fY*Q`X~OQ71KQCmu;GYnta!H9&JusDY?*6&IBE2TAa^pp{HWgIP?&zR3iKK z&wvvT@v_96RCwz`e5B%P?pP5~eESOAJ$;are{5?o<-bZZ)GfPxL)znJxGdC-ThRBn zI3Cw&)4va7fZNuw)4)2r8~uFz+|t&iHq!xHsY@dK!rkcsdF8)Jo@`#7DmSb667MLG zo9q+@82YwI^MTyA!AR#pc5>g(GJzF}luM?~mu6{Wq%$o9gJL#9!;6T=V%EZ0UX^5c zJ#Xl<*rTqiNKXD-CH#5dx9W%*NtZ+e_jvL%PS;@R?cMjDh77+wP%y(>9XU-AQ+=MF zzV9O64*8C!e_@K+lTy5meZkSObmYVy7*8$t4(@ka^UZ)1_FQGp!^KM1)ony3YFzf| z*rVRDmVW5a77A!3zUmtCX;irbg{mE)5mvSAK4;Z0?^4Vb7ffM1?b&xCo+{J@SQ66S zp$OPuAh>R#N70rH2ss7KV4`%WcNX>RoGpJ!mZmZ%C(MP((`yw5AK)j`%A&L<#i!C9 zZBw#9xl&(meGLw?HpW?K!uwjk4-?%7#j9tbT#4M;!rX_w_&PQiqQ%Xn4d6K4 zh17Tkma911`yUs01b>Fx@&NR%6VnlYzeK>K^=!T^TkoiLcfV?Lh#0JTvbj#^zdy0l zjZ0c*WMurvQOP&u=y3by$+)EJ8O8O03N?_VtY0~mk>YexNku~=xriyV%hW(P!~JpL za=`YvJ7><|$Iq|Nke;(O4M#j%_)S`V-7IDK07O+${S8fIG|U<9e7QV|R939Gk#VD( zlKZzKGJX230EF;%!D`OX?1|?X!1@VifC;CmtDmRR@x6~)_xHCA%o3%y#c8!vG*4+4 z%zQH)#xpalKVxxT#^?7u_LDuuhk}2)7ijR;WUC(wAiG?d8-2ab-YRF|M(Ko=mNibk z)#i?rwV|a_DNcTov&)u^qfjPH&5N;ukI+!|$#+(+$`-qdv(95)!NjC$`9_5FaO8em#6u>&h{Snt=g;oFq!<%^iU|dJFmoi+Ax; zL49!W@X=EPIPncwcK{{1TQ9{wooO5HBfKnuhJGq(AhqL(^5WhJy@_GWkHlGdgjLH? zkRQ^L5Dxx$i1&kM|1wjz4J03aMc<+ENKzbh{9_)YFH@AQX!t55Wwpmhgmu(YjZ$f& zTFZ0da+{FZ@rcWF9h}l6f=UtV9O`-1W(%$HbvJgpp`h$$AeoH8YEMH@Gr$d ziW9bL<2Q!zy#OkUGDB!YZLS*liSuC5nyX2P=(%_ve3oq(0MPemz=bF@_&Ijo}gouK%Z<>RA6o5D?)rRhAXe}DBEaOuP%kDJby zuE_Fq6l*DZ$P!a>@^*E)-{50bE+6dsv5c55V_B>AngWP`R^{$L^?}9Z*WaI7m^) zi2#393OwHPLPBF6F#|%xbSsHtK(eK>pE5XEP9+mA=2~v`{*;-f8WHV;gdu?=B(S$U zB&9;w{kybX`y*2nzs_K_Qc*wd^r_|TevPsvG-CVeMsea=4UvMQ<7#aNk-~Y5dqG$r%ak zJR%ud72w}vV6y3zLHnef8j$F)%cKvB#Knc|-oKd0GvyLENu{U7$c!Fb3hNi0q3XRt za;t3F+-UY6(8~7Q?@%()ac*ft2&#pum11mW>%c+kC%(~e1U3DMMxzQ39zcZSo!Z8k-zwu{QM4?K2_9KZ+XF{x_-C=)T_zb6^N~F>)Tg~x}QhBBE|Jd z!k1u3k}*3YGei5I#5UPgv9nBlpmJ7OpBE!=A-eX1KN>aW5`-9Tjup{I0&S{+{`>dxBwqDkoCypA()dl zUPW^=x;clsi*M`AQ7OvE@tB_dO7By@&Z;Pe`C$@PNb>|L+-R5-q$rZv#fI2GB9wKv zX2dZA=zNF)Tdr$Qp7AoUR}uyAejr6W8U2vP>bt6|*W;#Oh>P_h&aRh>n&&5un%;W9SC@UXnP(zPH;>f}zQ>UY z+Tg0(Xu0_nfcmjK95}p{|I$0ag!|bZMgg(3Yc?yqChP*8v76{d-(P?zfK~~}6+!o{ zfo11nu(U;Aj=q#t-*iX3wr6A@x_Of$Usk;e4 zg~eyCmu`S`Gq)%%j-k~Y?VQ|AFJGuyI+)I$4wCudxnC@f(?szO@ps|e; z?szLd&YTF36sfYMlsjdSU8BP!hxtT3HQg*OXGc6kR{v8!RP_R76TV{PTy%8Ve@yy7 zrE!)HOR3GyZ3envDg;5&F-*}75i*)MoxcpB^3MIJrF4Yu0zOgJZqX*4@mX0}U#Vxa zUBy56DcMFE_(epR?z7gvr9sv|717Ie3dyDK{80fEQO|uwp+kFQE&(>KiB@b6HeE%> z+y@`NU&hMDUv0yj>?D2MqDJ67T_~R41QZGL@j5=eX5_zPTN5G&7MEBM=MQ}UxB2dg= z4b_$2DTc4G?-%2p7+a||M)A+$M~XxVA#5C!v7Hn%DP#&IZU+@H_ZF*YWv+@egnKW< zDYlkX2a#aD;UNqftsW!6>VT{Hg)9E!hr(_G-%6Kzkn7UZCHt~<)=Z2`hexiC%rdR( z75@OYN)t~qK{}v$eJ;UXV6MkIQ+2_bgLMo|9{ZqqiB(R5^q?@*3xCTwE%FN7ur7M3HK72WSWXUBUBhK@L?mRz@ z0w>?wnPtNdc@`p9zAF|)ANGj;dHQsIgdj8E{>^#+4lZfZRqTTFQ-&T?YdW5ZMb&IT8*CC~`hokmHpwWq*4K(#Z0sy*-EZx2I?lrh=O5S!kfhEb$f z3%p)jazmxyNs5NE!dX|ZZv~FigG-m=iU}SbdH$plp71+gUI`#Xudk#md9TamV@hlm zoAJX-_K`M6;ypN6k-?`QvI^+169k7egC1^-)jS2`Z7g?k0$i0$vIEMPP_3HqCANhZA6N=mLyvqoZUAy4 zi^@lVL3xfzes{V~_Kg*O45Prw^8zzR=#B8vtngCB1?sUoK;|R{Dnnppzz-&ExxMWG zyYX(qisR>wE-t={Fw&i=y$BG(xTwQTs%=vn%s)CYoUB}@W_V;60%^4#tlTiC4Q3%2 zB`jSDQvt~ymi!!@5SYEPbW&L0(euF5m;r-Mz73g&x_#=YGc1{B?(+Bj$twiDu-x6h z{q4udn&AIjJTg1DZGmnKwzLUxkaQTshbSRwQ2`o3uS~;ClZkuBk3J(oSiF*dh1W5h zFEqXJ$R3>;Cn@>a%d=-IkY=fyMHjQ76Hpny!6az@?7tsA zg=mAK7EVDY07jT@QgV?Ugz`vc`iF8or)=B4oIIq5CbRy%H~5&~CjOi(x&p~i9GqpL z9}k2V$S*GJUJPi9d-Fg02pJNaGi9zo3B57(SiM?3!gOszpABl;!qI5a>(f>O7_0^e z5!JlNO}DLL1`k;!4Z1>`RB}0|x|+)elojN@O%rKHJ*ptq!nQCg`P2?`SX0QE9ky=! z4<9=wwjreD1Qe6!r||y`=*VAT$WIMixS$XB{5=HqaVeek`qmYbj)2nNCw8fME?vfa zJ?bZ6^Tfpw^RQ~9wUOHKicwJkV0FcIXo7lHjXdjAlZZyqkzNU#$$hbfZRq-Dv}`f* zmt5?@TONH!I_Ay;dBUR8E~YxhLURLU8eq&}nzFqc7=(kOa1;4&%EE9voV0aX9J8e|;&s^oU0AwWi>GYlIc6Or{ae2ucqtc$vh+n)PeT~!Qn^J@yg4CUh-5b?AQlCT1Pjf_rx>D8ryRc zYtns-|H$VV7xN#J-jNUnE!N2!rb81t2kwN<7oJ${c#U<<&6P5#RN9WK$)%a<7j=)X zUp@S9cijv{1Ty?BIS$N1wSBEm_5M|yKd}5NayEUMHg$l?ZDcQgJ7j6p?_QJRrRdW* zs3l=WJ||h=#Kl8d!uQZ?lB)t?F@6d@NySS_>oIr7SM?qJ=QU(lSAA}PC%?A|IU|Sc zN%H}QSG3NPLGg{j^xu2)l5?Xy@z(M+_S#cbm}gEfY5s*w6CvkRNIeD;?biI zh&hx5IXStf=tA5m2u@sp1P}yGBnfaj33R=Yp`#vSCb0PJhL%<|2uKKm$q6TOGY)(} zcqMjQ)p|pp7c2YGx8oi|YYsfqC3g}A1wanY5?DT{U%UkjOc^QuUJ=D|w(h~?)VATM z{N=FyRRtGV({Kd-p!~kgVBxd2MoTxaPV-UF*8?+C#d0?zg=l=z&L2!aQ1PH?%VoIY zsx@`>^Na+91~!psqyzX*8ih1V;grf3l5~bV`7J2>@^U8`O?x(uu^fE${(jTSGQtB> zy-WJ^4HhOf_!E@^k%>`TZO3?{1ggWb^>>zpv;PTs=AVD;dbkXmDfiao^4HjyzkabFF6{Pu zC`PFHlhoXs_xBWbr=?^sE_0RN5>_Go%NhIc3n^5FOV?NsUb{o*jpsG!a8tauAcAU~ zx?k354`1)ulVtR;(Z64fnHg}yr0?;&EL%E~#&71_W71X8GN*rD3G59VlzI71I}OzO z%-c;Xm%hX%tIa1TzA4a8Wlv*NKAI~%J_9!-f)W%_C@SJ0r~NS8P!2xERnU<>r#vAu z(yG9Vnfs&YY#;(E7&bupLiYye!&h+@mpO_FqUYbZDCBNDF+8-}`X|?`Y=BqyIc+2h z;ixgl@BX|_IM<_ZGCKF`o~5g;_}LDh6wI{(ygIlTB@wYkI01P{3DUF}8EIuWINrc~ z?}|-*svxgRM_)cg4*}459*^+`miu*FC>F@euBd!>=m>J$)~OK~uD)(Ld}M?Oh#(O~ z=0aJ#0YRa7WR?p@5ByxX zdXvtWR9Fdbaeic0u(}snNjB9s1gm#xaYdJw{$cg|>5;Wisfj#Yn~G!(oO2xV%S+ZQ zVca#zH}{n*PA}ZQ`#<>&!7Zo;$-7^r!`?H%+5h1kHz}qYr7)&>Magn!-Kz7*@FMuh zNwB%)e_We|hs%2!Nws%qU6mL}uP`ptxV{*T;%AVNynFR%Y+9v+d%C~jNp@+-1fG{j z&cX-$TSOaFuOn^e7!PVwA3hNu!^C#5+#kSw`oa1W9C!S<-Tel+b}l$R5I@f!ZaV!| z8;`$_;ezR+;nDr&hFXQ6{jp$J0#-#fs%53=-qY^Z*3`4)5?_)cekfMBcu9R3J9=I9 z$aQUP_)*dM1>^i!D5u|V5@kt`sG|RTXMj%NE_GP2O}xUw-(8i`RlX`F*}(Cd{Zi1d9!<^yX(?-J&bOgB?11FnF1<&>=WEDN)Ng zIiAObqP=n6OHl~(n^1JO;SMX{Ht55d@z{NMDO0Kv6yb4w50h<1m}`t3Bz|1Z!WdPL zTqYe8?2oI@o(oc? z7mAJw-qOPS3`V&#DADEFWzITI(q8IJO^E#D6S8}{i(!>Y?%i;z)4; zB+TxEO?tAPtbvCOZE?xxR=Wd3Ltms?5gxf&l~Kqpd_I`W2PmTaYRm*!Shth+i;4mqm+5MRCWM#TGY{HqLyl}VaCbmzk*V(wp|&GW9sx9X-eh+TX%k!=Fv05a^|$n8^bw>oemg8vX>nx6Qu=?tEbcXY*=iw4 z5Duu(2I{8mf zwh1x7Ufv>p-_=3ZQ#uvP{ruDugOl^SR!eJpVfJZi)nQPy%dR!XHE-PhV7c-9rE=_JP-cbON4U6@57ZQ70MZm(11;5h%KW8ZQ_7>Vp+?QC;?O!w$k{eNX176vnZfL z3u>r7-4i@BM3|w@n4s_L^E=l!d^r~MQ#4^6%%~kFPqlKlkPH1cK`K`jF%kWaLJY<8a zh#P!h&+(f_?n2`_6l$>x);uEu6S3025RZrizNQRKvYvbyNCze5wbv}#?j%2HXjO)Y zBXRFuhrqVeN8RnwJQIB!Q2~|a%T?h&`vYUN*NUjoGMaz4tY0UN8U+@ZIh05*Ccc{t<0=eCME%c!pE!!RDivV>dOXqNWm|(J7-N)v%nIv8nLG&1(h^pdAFT7*x7l%F!Vo% zejBazTl2OIEyTDDnS|)1lKRHJQGi+Z8iQsfu6CYKP)9hMl7iXi$d(~V=8Ih?)pVzW zj$5VCHD+YF$=drFcFCSLRkvX)e(z`M#dj<;@e%#fgC={QwL}lzP)>gTz0am~(S5?k zo#i55qQ4sj5+BLo{IYS>&Y<~+MT9+jYzjHYP>7#$3JwbTbXU_|fth#_7zLX*x@~5E zwd7suF}7*g0`^zETk}$+mHbb&(u;=Q9BTGFybs{K|Uty~M<#0N>NczS;eq@>u+9W0#Wj5Ss9qMuM|V z0{*q~kMUTnnQ|rWX`C}|%`pBlZAC?e%HXiyrZyANt4v|i27HNQVcZmXv4r%HCTe6HPzu{O9@Tyncyp)wf>DV#VYE;R}{=tZ$j5SF?%Iu^Vs5MzkHl3wC+`1WP!e-SGFp)AoIP0X}BXg zW94Xc#!X^Ob=JoHQ6j%XX&KQ<*?miW>;0cQaecbOHJXL(>Z0Q(Du8cZM2QSkMMKck zjO2Zl0eX+%ea}VN(pwmU&zHgFN1zK7!jfblIQzn&SKn(fuN+L!x7~hO(OkmSlGY)& z9;>+Pp*g~)$@`aGn2gQQh2GOad3A3!4e|(iThEm(dgb7$r>Hc=vM{M-eX)sKR>LqeQ{c z7W<$kG&BnL1ytUrP zV$qTP(O)MoSDLGFm+I2D=#Rl9-~9>3{Q0}Wq6tI)9lWr4S+E#-26h}oJSv^+8m`+$ zJpZXd{JpSzv1CO(jb9_@l6*tM0OcnVT7f!`oc>qoPdof0oQSO6L;>LPV1%wTPs|hJ zK*3b@gCrr8SvAbyoNieQxw(79wL1c;u|p;Swi4(*16xksSm{m~38&_WIFR!KMr1o?CDo)hOBru96)gFzt(q9!42RJm1N=A27x=ON8N?b`MM9j=&9?+y6hfI@@05j`>j*)AJF*8=6dh4P~R^PVA)(r zSXzd&I-o+t<9$L@$4`=Pf^EVNytn$K@*aGHIXyG(Mu*@8ihs9;PBKPOb{yDDj9ubl z0U#68A2D`==r@IYVEpokf!7z<-&g_Ti~vLN=({d0tDr<(d2L@bq%h8~(LrI?6vEFQ&OmF=#^?6VtFj6qN z4sYY4xZwZgbBRu=#GO8Kwh7wgkFi8YA3SgU+K^MF}cdg;tJ|6#56ckZqxNtuHsiVL~VWsLKRMcVu4yo=x4 zAY?@7wI00(R&s~F;GBBKe)V$MC?@$pMlxvzD1;1Q$nuXFcj2&{`+P!V|BtMmkp8XH zY0A-bAk?OK`%4KYDPjgrRW;V)iF$E(|BaTM1RAK`Se^9^E~@?aaP{vDx-IQTV4p%J1d#(Hc|lIX7mV_|pCT z*iY(-|LN(9sV|N=k};)10Pq z+VgprcTS-o`iv+BT|k)jO1xsy@lt%mmnU(nZ@e9bdcd#j9m)_anb}u19!0V0 zO@3Y1qKMC-co49^9zS>17vq@(!oH^2dqCoCh7s`jmzJuMS8)Y@emt`d17fkUBA`eJ zBm&B-j7lav=P|xdHZ@xss{I(tPp|#aiOP2)ecgUom=b5BL!h(IrFHv6-vI0 zRbrxbgkXr=tyYSmBDk&_^Qgt~SE_x_E$*=QYQhYr{PI|_xO#7bAWC>s?Z)qb|3_So z(Lfjr?Zn3Z&oGvSiX651Rn__*KmL1({x>r8uV}=CBr6GoanB9K(tn|Dn2-G^gtc;Q zWkhI)bJ;Rm729QHY&e36jg4)&Mbei*Dnb-WyLhu2*n$09kRHG=@CAKj{?0wBS2x3) z(eFji)5;`6xzSPlEkz>mzKN_b^gw-}i9frm9}NuE4(>Ao(# zY?cLi7sCC|9^!avNC|J+|bK)2H^qY7#N^R6%r!ZFA4 zaNR-M+T91WP(;xh_$^&|t>80zm;ZKf%gxh?^KB^2h!Qs}!NvZKsin3n^vui=K-UBc zd}qA4stWcb!FNQ+85Zdb=#UcT#uZ({#wxN$%V3lx&=?BhAw~QPQTYo$1qUJo2j!8c zIvtCQn|a2upAPgc2P-8S<3jI9HC-`@)qA(qz!=zG&DF%NG*GS!E^@VacQwS|EKw42 z*HgokGez{rPM}sN^n!xD32;dQhZtg(*TGZpNGb((r(~t`G9!k*j?qaw zxP!d%TADI(R{0DPB>U^STe8*Lt%P7j!7<_XQWO`of&AhMPaefaNUyO|AL6+X{z0Hx@S% zvK^M#*JH0?tqJ%3DAj>{U<~Bd(%nBnW@z*5&X4_N^wUJK{jv3eJ=>3^gZfM&N59?= z_k&*OaG7-+Vu0ct%s5#e<((H{Q-^XR9KCr(3SI0q6~u#aiek#GJ|ftH!4x$}xG+yK zk>cbj66tFod+&)>xy_s5Td%3+Xs~QN&pIozC8=~t;iWj$HSh-bJ>>o!%3>{yUfVEi zlNVBWV8i63OwF6w+}NNv`g!ger@2y$!=$Hht1)32VJAm)w!`OuW~12isq2W@VvjdG z<8JS%s0fNh@BDQ!0p_FLfrQCjc&ThL8+n(vs1~mVy(Kx1d$?U*$F-kpnrE_Jc69C% zU=t$7hxe}T0hducHiYC9E~3sndI|&!rn_Juxi$U^E|&TQc&e7Q1FwiSmJk$W;y^ci z^Q+u>c_cf)aJzBCBV!>xlkuaN1&vaI*+_nUJfDO3Y`jCTYwcEz+sYVfJDroP=8&ou{syq5tZ0?Nt=6*7094Na4N31{gDK`CwkGez-+_`*!0l z&XBOEw|X(TGHR))wXJ}hJVX+N-{c)Mg(P*Z)WC+d{iB+iAX2ot^Sx=U$F9A@rNgad z-~a!7SOpEhWZ?nzxtXj=> zhgMwCNw|+%1K;b>8iwrlet3853?0lav0ck~>5f~~g$=lU@J&Caj#(HTkMu;)D`qx*CN*~`@Z9fsNp-D+; zso%J$5fNoGXoDMD(wcfCN|0E>qy=S(T$I5jV8N@;4h>Fm)nKb^s@J^95gk+30i45+o~Tn}vwJL=U8#`jkisC+*!RYVt2Si!^cKL(cZeR`-)OT? zk`l?dZ)sQ1Su~!~bM!fnD`x@4a2&6muB@)_X0v;s;-p}LHZ^F+OF##!eG32Q_#e1Ij%?y_SxAuM z#ZiLSv;>PAnXW?yXGM&6W5bb`tR2S}y( zy#ly?6(3Aqt$PxjB}03yGf%(x;Fv~T?{*KLdSt>O*X?An>j@W>b-EWXBKT$B-y&mw z2=bPu8U`!o<|w?R@GaVvsEr@COc(0oJe)%SkZvTX#956E|JmbmsL$0*%hE!8@Hz9((1IW2ak&fm zPiNwO3JOkQu}?}wsV-P5bZXM4#DZQD-cRMBicm_$J}>WfPK;lQAm|h0$2^*{tJgtKvDsyJ zrtvzC(nee6l`GNKWy9icE--j*-ZHQ*?H|*&$|I>2Y)%np3kMqtruaSU36`wZZc-dq z;VwPga2WIVHJy2%y=R@3Fd@#>rrNGY#NY+h0;%rji1Y?AAqMukQS}V{iDou7U(;E~ zVnr2We2uDCq*V5^p+jb_Ww~jKDB9)h@a1!ZXd2ArIH=~=24vXxDNCj~sMAjCJAA%y z`-NWQrd5%(mU56vsB2DHK=njGU)F$aPwPEHm_%@a$tJ}oXH*L|zhp7LCed_&J%0rZ zqOyk(o-QcpU3@Xi)b-Ltm_(L%bnnJ+zvW|KKR2p=yH}l6Rx0`SY$bWr>U#TSe%$1l z80Gi4r&(g7bZk%``d2}x>RuLTC{=>A15G?e^Yuq2WmSKD+1WDhoBb`J_`Yt~3++2pwb#fof$(x_6|5?Z#JOhhl4l+W{h;<MA0W_L&`oK#jq!)m5eKTkh&cRU{~0Oj0vFO;|`m?c#--E1qn zVlZgMw#ref=QQbMuP=-xBs6%#brR>Q?j2$)+B>$2Z93zE`{Y?JN!9#6sD7zxWA)Ee z*7E4wOaeiTBH81u!(Au@F5FO~p9%ha2h8a*W%=m(;&_H#k)us!{#zp|g^%5rvsH5@ zrniKBo!hqDFrl0OLg|!~iMY%eY~ndrfb{W?He4G>#uRRO6Yggi8swIIiSqi?6j0_p zsDa#zv>7&uYBzPb{e7H6EiZ6Fa-PLHoYA;2dCEdTD3G?O}_nHbF&J!o} zwhKQD<U zOhas>e+akj5(iI5FDk7usbQGi+1i?m140#ZHM-a%e1p5J69+@@Pgy2j;}Se=vArcR z4dH?s_5o(RJj z(I4oWss_PN<7(Jcp87@$LtfK?jb?(S=dGG*?0}AW)1=I%o4HOQ^4!C;#)~zPBsjPF z6=cJBXEdr7jQ77d@trTR%3pN%pCn}bURQRhBf!hY4Mx($?P-#BC1$KccXUGehs~xc zZ(B1AQ*G!p}17pYDeW;l7=giH?KUqXEn&r;62@;B^%JzX}L@i^Bmr8 zUT5kKtFLp1Q*A4c$N;@htl5K0ilV-WX!qO%+>#&@<`JQLftH@bpz-7V8qaul)Vf@c zmrTg5Bv9{Ds+>wSWi$8aW1`>#U(xq)Cz3dhA4P5d;`fr#!8GNa3!>; z1*|D3WKTlbPK1w5#5cvEvMC%cV~`}~}C>eX^; zr`+S4trCXMFL9qIQz0;8pnKHq5&%jJG_>G)$lHD$`qJ(o3i&yt01P}$&lkQm)!<`T zM`fq~@>0d2ZQ%4Yp3yYf7C zER~X@uw~T_bbnvz)X-Z_Huhz}(#t?U5BH9ul$&i$2_VE=51bB=14?s9xZ zzlKhlNUUY8Mz{$Qk#4jzPMwtFQJ3`~YSZ)rCW<#gVfPQhhu#0FGXFi%fHYn+l zMOwzO{7{3IueUHYHviB)ZYz0G0l-6#mgwHAbnhpaGuEN)u4EKHt!VNUEQy)-Ub~Sd z&&tjrEz107J`fA;*&`z@4c@BdpDe8mTRA{em>OFz?y02JBSdy1MkIo%6 zveU|BT+WOlZd9Oq$#$vu`(^|!>x!Wd`W+LJr2NQ!qk>r0JUO0apMvC{Ki z7l5!|c6pddh8JZQ+xT8x*J~#;NI{l4uhq!FHR#nTSlKXstgC(A7aJ$UH5lhwH=InI z^~i}28R)}(KeRuj%t-`bf@dbF3|`vfy{5vNyI#GPdnJSUsQgaJ+Sb7&ThE27@365K zh`O%2dwxm+`0u*PQDy^h1$c3a?@@MHVoA(S83nS;Y-Zoq=4{eS)=2K6-l)=ZS`ebx zMbkh8Uv&*Iok`ZMw5&Pj=zXZzm2=HPGr@H4{_yq{SapNQQ8zB?-`9o^3G6GZgoP~n zkf>zM-x#2~{EP?@%ewAP7!`g^+GB|c<~@(;Dv!hNd#BZsuFzely#vCGW98S4B?=R7 zcZF{#uLNQC7ac=i9JeE}iZj^4G+p5im53(Xc-BGe2G^5IZrjozN~#_B5M;?05IbgKWf;sjvc~(xPeYlf)?R z&=48scQ(CmEjGFE^YE7Ji^Xe^AKL^6l~1*27RSy$JB3Yo>hyWhQ&eh1dHCs^ylfBwUe7OoS}UJ z?iVP8X$)PpO~sFwu>@hXc7X0Ewt8>GP2?iY1 zQ5&jNS%W5S|6LYrNjs{l;7NRVGe-}(u^Rs}W()nf^%E`KgO<_%nU z%;S%)x+fPN&3FEBQS+_&=igv%ectt3GaN!QfDtE}j>D&|fW?4R!G;2$d5Lew%54qG z2ehrwB+c^SMOn%g>5t!AtK`pK(1ih68$S|_C*@=IvYiue1pa3T_p%(6L(LH~$INW24+vs}jKse^xM4lbJ?_~1#75>_wT%504u29+m9|NG zwPq=jfB*fbSm~z-FQR-g-oqIr7`9dgMurtz2UK{25+WoXeSKqx5!G>rYGWKwQaxq; ztE9(fC^XUmFo37AY&pr$81-8O_pu?Cq9$q!{m^i*K?PC^H;bBjJQXVXgz>clcll}q zK+NuDC;86gb_a%L%$ddlVy!zU-YjW=EmTJJLPqU zEb^oRs)ClxZR&SezAJLph6t{|b_X}0F);2TPSoin@64(2pqb_7E|Ma`YY0=>ce|Xa zI7uxrt694qS=rv4vJCO;&?&s1yARSAH-7MXKn|HLvK_JN4R-(lc;?X8by6YaM&S>>p3ThB!ulwt)4i~4yK z+j4XD(ly2Xddx0lrByKv_d+m-@Tr=iHyH*9Mz z)t~c>{*o`~f>M{V|6GwEIdlWyyH{{^AaW+?CLeohC*N$mzb0quJU`z|^ds@aT%QB+ zTIvEGW}kzoclB%)L

yrfgGCE1WNOjY_Jo3PL#Z2H)UeZKvRUVmLYel5xGpKIL;t zL)C%WbX=tWO2Hr;VM%kl^2QHoau{3dAmL?U5+VlyL5m}>G!Ad7WJCfAa;qp><;FJ? zh7K7xTVhM{HI3D#U2*j6;p#GW5wa00eR&ni9Sxx5*Sil!f?0fYVf9QZJ?xXn($AqL zZOYVEIx}wmaah;l-4VwsN^!(Jl2}{z1L(+cZF4AAfZRnRm~H~ zy|oshs*siG4F`DImZwKvfMC_nd;^OZSixgqeB!?o7q^I0q@qV6JXpU|nP}w;##5&n zIxR?r1K!cdo+TOmA=txBI4^ zk&;s;8~h)sUb-=)ERq0XqQsq(?y z$42uC;}r?ivu^kW5f{}>tm$lq$-d$9em~^`t>#zkaQ>w3H!FInOl!OgW`6G|Qt&}; zR7o(y0MJ;Ac&dVr(i@IOq>lm~rfpQ@uXC-kReXQ;5iwYV9U$S*IjP>v&aCL@?Y~Fe zdi&4**@X{T(TG@#`=7+Y&EMzQd<&OM3S~uMG?vt-q$!E~5S^*&Ek*jhS0V>gdDJ{~ zI!j2C9X6Dv^y%>1g#tYw0619kdd!vJhlThN@*8K1l8B%(8HBHx- zm{yx`!e9`3&YcOwj|o*D_yx8lkvR`hRlHU{j_lzI?KD5XM*^LBDejfkaQp4lce z-{g|TmRh6)cI=H@%EgqaVrGRzZ8-daUJzx_&!m(j|}ku`^k z2!YEG{Kf@^D`4q5y7rc&alEjI-#No(Od-@^`72%bP_z+y0hwJWwMxotnOO^y4vV~i zBjRFiHG2_0dyJrKQNqh#_eUQW%fDucyyfMbLN;P7<^-F1xU_Qq9$fi$yp;-1e# z5yCV0>68^Clhqehtl}I|gVFV;Ka?(4NaxZPmHf(N`LZZYt62w8X~)vgwY4P?&>zR@HlTDa#K$k!u?#uUqVn63G zPt{Y+dM&v$a;Iti!g85fD2s?KG>lBdb7@JI)Jd5I?!oP%ZzcSsG zP%I0G}Fdc2tsPMkr?>MKL(_okjZ)V+U0~ z#g7Udn}r`{vbq3e>MeEb5W;W zP9i$y^Vj_1#)`{@;tN1^y13$4B-rCFIiLo_j)zYU!(x{;g_k~J^#KPsgGkH$syakjTZ`}Ebh(pSS+a+BoClN z1U!wl-ee1ol~_`wH*c}kY=$MwkT0TVH7qH+_mY2c+pG527{l$jWG%56H>a-g12`}S z5~gHWqL=@=LrEOR@01-LsKo`Z%FFDph%Q9>mmZiJ7bgOj4&4gqlRQXYo#SKng8yRu}^H zTLSJ?9wyb*OfxRc#_FP?c=p1NL)RS^cYgl(lGEYTz|LNtuPt+{cpU}2gL_@(qw1YK z%SSi=Bq7U`aARZo!a+zM|f&Tk4G|RMfvKK}z$0Ef07aYYa zC--i;&Aoe)X5iqb<{37gvDcR_cdv zb2(!&8Y@?F+zi}%6qBZ2(*$Q)zH{7GzGSDZ8y(e=8YY9)hj zUQ3jIQ@J5G!2LTUeMF1lFyNAXv;HR!!>3J_$S|svN>zFr9jIB+wELbcqbxTZlzv<}k}|L>X6N)K%8b z)~0ek{lhYfL8D}=ieaFy5j+ogmuyUz8IYlSF5MDW{b&(Z^Sc27UzMuH5I-$SK{jsZ z#m=+6J}}g90oTlob|(F>pwb9iMk`!e#G~eCHeg6Y;xGk9UkZz6iX}ub4Dyno!dwJ+ zLMyIj(POgwKQCTLB3V(9ORnngS|B3OXYy2xPLP|1;%Q?OX0J93s+ddHvZdDGMkC>; zJ6iHrRM>Yjk$uF|hCGIqz!vjv3Z1fglS&6+qx#nb1(LY88#ZkV&>{4j4ObrqZT0vF@r8^;4iF*aU~mqe(Njh^v~wkl6{NoI z_Ds#a+e*ujZG2d|ke!p>&Q{u|@`-;5SUnd9_@AG<>YoLl0uTNk0`W7m zPGx~!tCaa`YHBMHYH=Lut}F>8!}s-;u(Q82%sFW+ty5dHdLu9YTH zG}wna1mMw_yo%sdQHoxdY>5E>mu~q)LDc7GjVlF}BPqJ4-&0QXcLi46IXv;wIezwOeIMVe89qyt^=(2^2MpH%5p6Z`}cEcE=B|!^G1C zsPPhkC?Z7vQ~&7?KegTW6*5JZfqJMVVGO-c001r(zhL4A05)m9-Yteo;G9VZ41!I8 z9@;A<3FLj!>EV&rA(+~ANM6n&)?sb`4r?RdpDiNeLfd5B9a9}FZO2~wg} z8q#)@CM^)cL+?Y9tsB$mLvccKJyQ?5It;h&=SvrfE9c%qq)5dW2Sg_3>J(TrJV@;e z#naBS9y3wjs3%h7B?DH1UB;OWDxM`Ysy|(VzNY?h>R$ZqNC5{^Ss%$%E(T%GnTJ zQqn2oNhNeHiS*nd#G<(Ct)GQLl=ynriA8IMdqH%B4{^FVb_rk&ynOSTVTzSYC(*#c zc3StN>9@k!=06LFgYGm-J%#(yT^GpHUXDi#gVBjDJ}$nsfS@(3YLto83sDxVSP|SA zmhODFfO{|gPR{~G?e<5$E!^GtlbXWU$7z!j-c37iLU*`w)pONd>HXmHC#;E0jee!CMt&fu;ugw!4ilSXQSvJN`Zz zZu#J8Rrx!J{7$+~VY?;xl${#XK#v^h>3A$^2T;_cp6@}TacP+sFD+ku>4<-NnL6F4 z-EHe-Q+sP!zUBvvKHwNeqQ0h^WaYx|MU_b@N-&3eQ}@r)puEFD*rlL7@<0!ll7V27 zrOoUCxuC<&dGpxieb@I1FuU+_#rSVny zo#MZ~{3k+_{ov7*65Ns(`xgrS=Rja56g{o>zrT83^nV}I9|YO=e?#JbSI+MP@9z!c ze?#KG`ZWK4+f3(FzbbU!{sS(M`+`+k-y!rXoS}VEt%IRqfoQ%pY4=g^ia14g{9m)3Hp*qA7<}bK*vu=abqqRE(>nrIr68JTKj^a-n|03=VcSM z?-&8#s+G!|D!}L}H&v}4Lkn82_w3HjJzjXI4-pS*ux&M7KkWGPv`xTjUbQ~BR zk9zGdkAu_rIDC0!K>gq|8M%Y?JoKW}j6}!a&)zNj;#tJZk6p_Js>Nyep}%J<*Yx*K5NglCyhX zx0-$l;#O8dzSvu;0AePi<(6|DBR`Q|-=ZzY(tA_Blf2J2CIZtvcKq)881oKHZvVu< zv5t0ty4%CG*tf5&fU&{Kjfr0m=r`ahPyn8knYyx5DV@c#KX(&GME0SNu~IfMQ4kApr^S1EV2eiDBzqVr!rY!;xeT+N5DE*nsZ!VDtLD|7d8InXiO=8zZ=6;d>#Dl8X^hEO^+v~yg`AKU)eqJ}_gK1%>tJy22dbYfl@BZS6|O$( zH^`ptF$i1~U-8l2VioYhUfJ~u?yrS-FL|v68gRP24~radXIMS#1A?$som(8<&PjRW zdo2*dPp}eny^eWN92b2tU zenT1vzjvkScL(CfOyfWu!EQ=CVW|Sr_Qp@yXyCZv<0>&sBMrQEpV%x9t-&k3Aft$i zu}I7cCiYhO6(io)a)f5{w!_R-?|;5pOEA+4k{y-^lq8DAqN5$+@RPbjs08V}ib9JP z11*>mE|Nl7`qtJ0CJ%CxQd9d6x0v0s+AUMQ|5TBZxxyqd56yQn;*S{eMh<;&@VBc) zWnoQdS+N;tA9rB8-6{IC z(lKL_pI%6WY2gy9q~|t+&k?lVSCQJ5S;<-hx@|;pnz0=AZ=bB7w^ZMccWM4Crk*Pm;BZ5D%*IrG00UJ(VUs`j{9)e9O+s?=`Fy(dF3$bQ zGlMUqG}m8jSJORc+K~yWXCv<|a`v{BCThXZC!}i+mu)k4y*|ZWyur=9`{S~A^u=rJ zv}<^uIXjsY3G0UB_D7{e)P7;uCEUPcC?Tukm9KVEE$USepz{h%Om}BWu4gOj*D8Wavhzr=<|#zKG4v5FDlh+xOE*)jrlw|@Wl<{{)*Ylu zcAbu-@%?elEM7UoG3=?IP72?gOU z9i#}Y$wbual+#TduDCF*SNma@eNk$hF3>}K#U6W%!3rf@x4tlu5*5(qWloXAy&HoD z2QuB=&sa)zzh$T8LqJ(IJJXuR6N}wn^-S7C6r5El)hl`2(CfeQO_6bxL|CsSGk%Yy zc`>SIi+r0k-M^eTH?v?7ru=T{@1`1MInE+_9vn#59F2Hlyv?Qo2$HKtKax1Ew!gL3 z5?j&yBp)?RGNRYA=!zjN)-w@R^{wH?-%sNCCX@H<`DI=579g|b_+K6}u=9?4wI{eG zQy;*cw+c&K-_Xir`B%cKN5-=pm*;raRtlQoP6LSB5o`Hh`UBqLpPWC5m{w)orQc8l zn&xMot!FESDX{{HgbT(sW<11elV_IOUHmv} zk(?u!>!KY1H!;fuqW3EdHR9ax5nE5wRAa9uMc)Wx7OyU1>{!^|y&a<9Ofli0Sw3oE za;6x6q-hJ0SB9g_vD3@jv>+B%>fB@{X%oPGZt4IIqS4(3`@djJVK$PxgDGkZV}6F3 zINC1%U0q4gbzW2IkBDn&%(hXONTDpsySQM(ZE%KX- zBn!igFVU0UDry7n?bI{6;Ia8lxxL(D$D2^pVyEdp%KOieBqPJ8=^Sx3oUI z9!-z4xsD~S<39ZU+YMaKf}Qud;n1!1F+|V>{#A;Oborzry*oT>wT&p+u_>v8H0uJA z!;RdNwS};yVv(bV>BV!2q*5Sad5gTsU^b~*pHsddz~EGmNE41D!*{>5w!1FgJ9K z`3v!?yl?l`E+dnS|5wqZM@6hXo+eqOfr*Huf+GBuj;)HDvnM;x_Y28q|8VyrVtfn; zXua%6yvF-{NlF6PezcNe1~4#RDw9 zwZPVrcmvU|!_63$@VfDp?<80xu9ZLWYVQwF8lfZ4>8)4A{p_Z&k^eU8y022xm!T1G zc_n^*$41pJ7ReQ_S^vVhjgZEssSyv?_1mS4&#~gG*Mbr>-_XY5-71>=u!D3d&HBq7 z;TKIZ4%{aTb;9CPDnf|4hLp0JXLm~GpX+5HT#?+SS5lm-sTo;vh!Bj!9$PKBoQfJU zMHkf)DmXscDF&6}^h1&9i}yE1n0t+w?j!iK4#dupsR6GH1JyQd1RS!u(*8!T z=zLqwM;?Wf`2H@tBm8}si8pgLfjM#0$hO`yJS`yC-uMbl&Wog@&5Tc2)U3TrC@

e@FOj=vicLnwRQ{PMs-15M{964MZ$Z(?%WO2B_5hNR(TC|ZlzQis$%3^;SV z9ocUA(dD~_#N)q*Y{kxU4Yus!4yJS6mw)+R?FC2v!ZPS-U z*E7w3zPm4p3=fF86-xK^{(xxFNGJ~CaYCBl_`SkTTGuH`St`$OUx6r59IHm@MYOLQ zhLb8BCwEYE>(ZN*$U$V=Z{k2UL8dj2K^kH6pA+1f@NL{PQiPkumCW(lX`|$*<48V7 ziBAUT6so9mEDr(#u%a}e8X3}1K zO8`zgGDkgedR*4V)Gg-q^O?6}+tbLj&MVU2c+FR4P(NED6)}oFPx>fu5(cXKel2WF z3^AG4e7h%nR_RHnEYp+Ury~l0S+o(=@5kRFBdUQLAn`8B>#*NOY!62_5X*n^f%y-s zI#ZtgnN&1wXqq4+qI>{^Kz&CE{K`Vew|B z#)CQ#CFCg0Xs9B7LfBxqwj%qZI>XC4sX>Qro{J`srDW@kc7%)MX*H`{>v;LlD;(`P zW-V8X*I&2BsmDk9Z=~k75sh-m$!H^%>I-w%TN-sIJw_#DHUvo-ZV>m*VtQdpwXy74 zChNp_cH?fdZUzkbp_8;RpILPz2Y6IUr0+(mPjl@v50~AjLE1AbK8}AAW3)xj9eqD3 zp|rm0%**j}P}FS)81>I&CF=jY$j60mI_s}RmiSoh*=+LNzbCdbhe)sFetJ;0@M zcH$bdlSs>e_4G(|v%SO)yPu4%rmN|&5t^kg#;lNEBDk<$Id5;RqB*iZ!m@P1c%ulH z3e7}cJn`@jy=7lI7H6xTl7m{nQ`E}WeBEL0jr`?ebqCHwp6j*@`y|u4xDt0YK3YcZ zf4c1Tc97WOPlE1)MStdn> zIX=TcY8yMw+Xykj<)KHN+Ka-|XreAHYd{pJxPzSq`2j%L!1F@#Lr3`oM;L3VKtjDP z$%fJT3#_9MW)9jQ?Qp-(I7e#5Ud8($?Plrs+L^C+tWPlF*{8Ag-2D-hoQgc32|AlB z?Qyt@6ZH8YVJ^F%!>gOpp7J z4c09;&aJf)$!dDM)B2Qld}g6)(qKMKaM?UVZ_4BQdUd+gN#+0(pPeT8s6WlKh-W@y z?}&Tvv62V9CXEJ!>AVE`x)EOSg9UDt2Q~&?gYT8paIZJ)E8m#|x)IM~5i+Lx%-^&2 zy59_?=(7rFdoWU3(y$dPeAp|S%XYR!Y4j1{RWjl{_d-WijY8Q)ki8YLuURA9F0~t= znRYi35E~92Wc6$~Nl=Rn@$7w4R>B6KFsa>#NMWiU#UcUHX{YAHa_y&!saXWf?BIVZ@Bv#(~7N^1*z4#vM@~nT_<(s^BE(OGE<_lN^ zZI_3-55;F~$1vgqC!&;H7(F}Y9mG<6#t`U`T*tONE&=@PDVH0vESZkhMx$M3hLT^wt{j_(`(lh+Exk zb7j9ebG*T8a)wSb)#57MzPcW%eBJ)|s(5Xp&L^GFl`CCZIzHAcmZiNBIBliKw9N3* z{(66+FO%AK8gj;-w;BS*jxdu90STegq_x-F)4PwZxFArTzzwG26Tp{=NEyk+ z%gdh(ZuZmcOvo7Ou8!ggXPzaVbN}IY)fvX}v(#tLpOjBdFmRB4=Jg3OMqn;-+WH~0 z)j5AsY(N=pEIq>pZ=lQ{mybJJ2ohUAB9pzjTJJm;uyKR0{OK*mIK50APX;gIjL)<; zlG5}=N^r_p9+&>r%Qojq-Ac1O`CdZzX6rpXWjrk=m(pr8G7eL@At9E6abA|-94ef* zha(rr(oUmys5NwqHANI%FeS99 z+9_&D#a6W&gOOHf6m1nvqSjcZRaI+h3vELYMmx1k7yDpB!%b-|)rin4Gc_j?&z;_< z&vXBPJNfnVOVV>Z!VbQT{tut(8%!!{y-v!~te(j*-Vr^_3JEXI;Dvzd@QU>#9O=`XF z4%L#e$hUQg`V&ijylpsl{a0{pw*vW)lMLyO)B%r()~Gs-!L+l!hS;fY{K=$?aTb(+ z5y%-J5h`h!%Cf!UTI%UbJ9qPSu^FfKo(d$3{DadZMR9F>_-h}mX11$mAta;IcSJ=|TtNDHHJ{1rG_1ZD>&M;FVIw7Inf1=@4?m zoue_&l9w@0x;C)4U(Y|$wHjQDPXT3*Zi$GRjF_rxPrA~rxv5p@B~GV~we_{u6ezSN zp**j0WyBYCbP3UZbC2D*{aOZ+Z2ajJ9j3W12dM%_g!)bSS;M_g<=MG;D<2dR+SO}q zcHHHb5zW>_Q5Qkx)|WR=eOyH-vjL;Oko6{!oor0~#)_*a8`TRkZzR1stZdd~HU}Mo zL4r!&Ia39yU47|dAzFbdWHrc=WM@gu1&4RBv{4eRw-=fg+IY3C7zcckOZQ%Ig_`>d zEvzJU$ePa5o0>Jz0)-COrCd8gIwzZH8P!^>!f8pzWG+*BD^z;3fYkl&@C5z?2N{xZp&HU-vCRU$Ill#Htd09K)(e!hjbig*nqGBH_604 zWvu*sEd`QUOZ#ytG8M7rT&0MZsY#abpkr<_r+hwA7dm>v5#kr5J# z$13{~+PdmVl(k8(5qBnzwj;_x)1oZ6+d0OyN`0D24k{O^MeEp~US}%ka?s|HSdSjk za-(e)L-Yc0%UpEGuUNHUNvuDQ^rbeaCoH{W2MTgI2Lc)GHEpQbHYY) zL{;y$g=Oi4Z-=U?-%msLjg;YRGd~T_myDaQ$f=zEWz>Y`-AcdMA)4Ki*Lmh#y<9&G zQJ90%3^k)qyK3Y{t%Q~81n}JD`CE~|%);Qnly)xPk%GkgWY|!2WlgFy41Xs{220Gm zgDSdCBp+h>Drpeqy?*o`L!+E&ca+K*z8*w*woI-zn%mqTe-ZQZd1f2fOQy>wKS;_D z4r{_%&Cr=p*7c=|7m4i)N8JG^c`#}8 z?`S>IcT$XIq)b!La}c)o>`ac@JHwdEzrHztFiEQui5x94@NIt>2ly3Y3>{ ze9MQ$`G*r&Lv5`*!m9i5@%WYt6+Zo#%-Dj7Ig^v<$uXeD?ECogmjOL{!EXs~D5G?t zmP=90{C+>dNUp*sl6UVKMl$McbB!gOKRpLvru{5WnN_=F89=-FlbMamf`xwg_!084 z%U^Bbx?szLB-@24_q)=Wo4zI_^1TGFYlK3QCo1-Qvc!`v04eNfHk-f+#b}Wf74X{=SiI30gSf}+Am88cp3gOcui~YG^F+csQ7>pLLl7v+x`NV@8J>0hdT-0 zU$j03o+NEyY0lq%!9)e0+b73!v(6*^{M=3OVS->p$uS>%l^)Mf_z~oPt;i@57<*vq qcmaj>`Fo+fApW0C{lD7O)r7NYR+lq=m&z9cFKcsqvszQ1D}Mp`$`bkj literal 0 HcmV?d00001 diff --git a/docs/articles/getting-started_files/figure-html/simStateprobs-1.png b/docs/articles/getting-started_files/figure-html/simStateprobs-1.png new file mode 100644 index 0000000000000000000000000000000000000000..6d71e94d75147c31b79c75260cd608873dc59ac8 GIT binary patch literal 179611 zcmeFZWmuG57dAYMqzEV=NJ}HBr1TKdDV?KqqjZO0ASEG458WUw4Jy)I5(B7ocMSDi zqu%x0bASKdU+?>84rVxJ&)#dVJlDCd5H%H9ysP9_K_C#Gyqxqy5C{VT0-*;Hn_gMx4~O$R#L4qd(k+`EZuY(F?7PoXf1a-TK+WggY7+71=9+-p zEC&QrVwGHPuQE=K+-4E0JPuM@4$WCh5avm+t}B}{-}Ko_*-G`^thL<2SgtM5jf)~i zBS|dbAm#q<{4(;DDBGLhmoQSybRBgnHlutcFuy9UdzHbg)JJdV#3=kCt++;p>gMgI zG$s{`dAP44(YAB7y4}#c5K%PZtNPT3Cd)PVAfLIsK72DB|Ku#@7jA7a&;EV*6^53E z&8zt9kv(n4GS^P;y&r?Zlx@}umKQQil3d}ivL{=QFVgH=t)B1Mf6DDoND~X*dTe;g z#>cqyP5rZCL3EnT8TNNud*5oxxv99Ms2k##Z>DDwOLJ~)NgEbTVHGr`WffPwruAD2 z=HFxg&he^WV5v)^^=n~4Oj~lR+vgfS{oP$*^p=o1gT$S7r&D{4)ZL|pRldY-Z%^qn zM1!Z?p|4Y#;Kqv2@?LNp2i943d~DFo&q1&J(`6(Mm%d+=O)Ff#LS{q0An(r^or!P2 z^~1J8c6IGUy{(VbC(Y+Me_^11 z6J7-l2Ftst4BG#sa~dCHZQVAB2y|O)hvxp*+Ve_5B=Z2<4;(qJ`U~wk0+k_KA{l?t#hfz8TTREnN8N5|GdsljZ?dRPkp)4r*%7BXLl`fd(0ms z+k**hA#GH9Pj}{Z2xR=?Iun*k_l-pK969l2isCU4HYeI-D&f*a`$79JE7=>FeOb=G zH1_&-`S$wu`L-H?qlCLtel&gKd_v<4{c>otN3K8XP2zd0#&RXjwn%}N-tHzfiPJ~^ zZ|aY^cbTtYn@XyG4U8*ZJ0(fuN)@HTH>Fw3JvD6N)J6+AEd%c>UubD*t~`moL#6Gg ze7{M8D2zeMEG(8^^XP-lUOL@d)rE)F*ENJSi1N!uBnVzs6^CL>$V>TI;}TbwU*+9;(V-M!rSWpoUY`PEYi2)kKZ(1ZfXIMT zHu>~gOOyD{`YGr4h)Pf#Wk6s%y~C~kF02`h`+LEY8}N9&3q?iL6H0m(~CXnLTn!*SlsoBJML zs?s>mwi{Q}v)8|WpZqK!lrmJ3xS+^i=@moS`-l15CMKT5FS6Lf6L)VXemWb}En_E= z%FhC^?LZ(Bb1z}yE~K#<)6dN#6Z9mj_NfQVwJm|;k9y|!X8$M~p zaJo2Is;AWR9VhVHzLJoVVstLu_e&};8H{QNIP2-*M&aP3ItNmsY2d)%I~R6|m-?=& zBX??dyV%ah@9GC5uUt(BTr=$ydUgQ4JgA%W8P(G1?2KU^cjX|Z{izaaxZrrGU+uK# zQc9KD%~j-QBjF%=i0Ic*8pOa$v#ps%>LNic8KhkE+yk2th0Uyqb5vU%FB?wbTbu~7 z>0c5LRKPv^A%0<+P1CveQ^sL{eh@U_ef;*Gm!IG3{h}t>PXPyh$B<0Fo)*-`mRP z-MQB<&eeQzP`Z#2%E`K9S}X8t7h0rN4i0g`n;+=4B0WSnadC0EMMT(VFAi(WpFB|$ z+l`9sfV+71DIph~yN_H*+Efe@3 zO94k1T>NEvRSeg!VJ?L6^mP;roh+qs37wxWCOO8ftXO+4CRid(L2s$-${&7W`x!|7 zO~ZiV$PF~~F4pb2AOiiP=JoP!v)Lx!jOJ@LYA!xL4MqJ6zr=$*QVuaNNHX;J_FVg> zYp>d$K6y?HDc%ji)o>7)@y7k9--AWY_lJiD25vig4JQ>@Z<9iO%08Ri^PCIBQrNxb z@a)(1>0syUelTt@yH&p> zHdr;l^2^PPBn4bZ!Aed~RtrnJa+H&#FQ2~owKnpOv;z1bjg!tJ6lmzLNe1R}s6bUg zsb0f{C3Oleld~NiuS5<;G`~(XDT69b2&i(*`xSyV05XBT z&L`%-b>ycinXWeJ>>{0q3GqKDgoc#(S9|U)R{m)Cr6?@~V4?E8!zo8q+pdw3*aL!K zEPTrM7LwCW-5k5E1b%8WI}z?#Debs_1f3Q(Ab>RvX^aq|p)0W76&2-B6+6jt?LXaW zhWa6f({kX@T#(J+7CE->miKIuTQcb{i62YczI82!yV0_G=?jHptfSvv(r(I@k8BfS zwDfgJe6QSq5~DHwuRt9@iZ1tT{<(2R1aO;9Xm}p0J<9i76e6gc%;hN_<(jf!Ur4&p>MKvZ)vP%%? zxAA!L`A-Fe_yS8-iXjrT=>h_zkgg^J8VG@u)Ncy4N9`*sM^w6R8QbPENoD-Gy*m6n zsxy(JY3H@#lSQ$*yyjF!JeQApA5EoQ}fBL?kdv0I{gxlRH z3%~Kt&wmG4;M-^e4I;k<`&(81SQ>=Jg$ra`PFw2ke|_}#zdC@FikFK33-#J9-AE&u z_8y@^y|rB+mFw;PB=PqZ|JvksIgmYY9Z|c={%cprF9bFKi~pZHc5tKNiC6LIQS+s* z=h;sCTGdJAPc3Nq1a@|IelUKKpPL%p+S;12^$G8nDDx^?3QL<C#R%8D`l^5-aA%bqozdeGg90;rHSJowvf@CQKF1&mj4ft$m-E&>LKd1uS^9Kyt7@o9i_b`vl%{XtX+sFTMCd z+u2#e{=vab`-jC8v2|P*$B~yGmUE_EhLv|`j+$!nPM4Qt83@ZZ2PY=(C@U*JfS2l% za=Cz)9|2Qc4(~J+BlB#Oxg=*(wZ&K)P-eTMS1HOadG-YU( zx8~m4G=$_8;Us?h0wzm!8&rp{QP=)dUW}P7@M=RvsrR&jEc|>cXpok z`^OEFj(TBbi^n&Uu27HqJb-Fyh@VenZ-$PIjqPJ?UqzY#j1FuTSc=U<@bf1AQuW_| z1=?U=2Nk~%>i^~C{#We-UnBqhf91D>*>+=DMHvV~I_zuUOi2QcajL$n8fyh@j{c_x zwMR4F3msAPLt|r1q=rud4ZSz2%5BF>l8+4f`2-!mASM65vIIWUtqE|v0t=Cr6x&Pa z?{MxHW&XB}NIZdW`RQza_<@gG)7WT)%uH8u)8YVtO}1(KIade8?C}@&;{(GOS0cbq z|BQnF^8%4a_YuVKb8t{;I&S0YicuA-bzs0t_B-i@#(nuBb8@yM&UG~F-z;)G7qt70 zz+ZR&La2QIa|n_35wxGrT3`<2kEHvb%g1WAxT6i6pB{Cmolm57A`>GV%<8pkwrFq*ZM)te-TI{BG{5_frC^0!++3Sym7mTEiX(DG*j^G|`Wuvp;QvVB_gluB+gczOJ3A#zOw4&xn$ve@!{X;LfGa7&UCWL)E|>9lm~DdB zK^K(|*C*(K6oXq(@Q!RGbrpW;Qew~PE;;JEQ($}Fv<}HGmg?MegS(D*(wq;{6br9J z8AY|WO4wG^*H6v7Jur$?R#ojTGi@&l+g__>TFbF$@97D%x;Ne_-wOY(~SBFWfhHdvBZ0tY3fLG&+XWE6V`f#~ImuGpF zW~&Y=3y9N#5;k^MmtLJ8eY$9(?ze$TK{e~Kv{Sd%Lm!@r8q_+Q*Q_@a93FLvHS(%v zaw+`KeO(W@yM4bl_EGKE{7(sx4uKlO=O$=lYl};srmwG0|EQA`w6VTk*;m$I)m*wj zUeje$o~`P)8{_cuA-z7|b+6rd`Elp5IvyLb0M&>Hg^r@9s|ycVBW)<=ulniK8{2w{ z`IeCs@1`n3W5?Y&+;h3KD|wIgl++~spN^OXXY-Ew4~%eek&65H<%#>M3vWrD(#+WA z%Gj2MKrf63NNu-Hos869H|ZK0Ryrxh7Di0}%+XqmT6ECVj+jb(OG}pWALQ09v2JW^ zjHREPc&?p#m2|ppDXjY+j`w%KO=EUF4C>E@Ux?NEjQ1&s#95-W5;b+eKXXpo+uM&B zpO4oXnm>NrDl{@OV&)6jfWy%kpAz3%gTbJsbw=)!`|O9#@IE%P*{DuyPgsQkp~d`? zsRhTdZQPW4M^%+ego@rslysYyW`**gmF9XrA5Z}jZZBII0-<3mW8d;Looj!oR8%<; zaLnfhD>@6yp(g%)&n43{{5PlkRYx?$bq-RE*%4%U_B)?ui4YlxEy#tjW;4M$`2cCP z^g(lc9hVEjyDTkeR3{x1$khUl0DP)$KSA`Z*eo%YcR@1Mv0*y*_>6ecl{ERty4FRR) zE%C4W@hw&_J?2^6o{aSIHmBM;!RrZ6y#*Yj6jj4tHKC~_LYWAZlr+DWiTqjX{@sTL zz6Me^*KadGid?0`)3CNb-@|R46SkHoO7*pr9Wmx76H%o#=lKhSVr>kZ36jKm(Y(vS(>V`lo4lr1B2opNGtyGcKUIS zu_0+fTmU9Sx%TpNh=N#(Syg6FMyKO zy13PphLVgd>hfaivf6kCNIni7kh&~iFYz2|x3M0($MmH*w~1evnM-*gbr<@rVCO!{ zeJjwZ9m&5zMhr4)>bRrEE#GqIr9R?SAd{j85Qd0`{UOC&QvV&&-Laac(XMjD_9uc+ zCi*%c#afuOEmck_r2x*gfqB#K##E(z355`=>d`?GGgc4p@%Tb=U&Ph%XYHtuCl**o zgO(Vo%`@V+zy^(SZ}>sw1o(%OKKCYi_gUW>rFUi$TVCV!^IO0ni}@faR$0O#JD|r`@LP zfislKeJqNv^*wd7wzXwj0m7~J;*}lcB=e(@;PrP2(eZU&3kwN)f+Pts8^s*ve%v)E zvw0XdxCuWX?)6g`qRI3)l^yZ|{0DQ8Zve4%#lS;djcqrKvxw zcSaq`3bgANO@hVAL zZYeTSzW-x`>&qqP-1!d)n5w8{LjV*X=L4Y+DkX?$SCZ!JEyx>4QV9Kzc4%#r#8IXj z+jT%WQ+o6$>L4-!loxlowft=)h1$Sfi8>!0o+KD~Tcb#jugPL=?bYqUV=2++8YaJ)(B+Yf!MfyHA~vwrKZc+tUQ zho#KS3vVanLef5OpcF@55m1J7d5zBxgRflCZWM~k*nRQhO_eB3Xg84s@v^2%KGT{~ zHp;KENC$wdXcff7ncs_mHIq)pp4V#H5`?LC4A#CK&%g2uC;H#8Gl#4Qs8{lww0S-- z3St;DH#ggRs71%jJxJw88~9f1srQPx;!}JD{`8;4{y!X^6Oemzyew$nB>P(4R`B%W zS?&2(^@-uZ$Bn7_`qA;Zmh-y7>M-iHhXtW1Qla-i5guFf41IEvVE}F)er9+79)siN zntc8`4MqG93t|>ihz0}7jeE!e+WAtbFj)%^-%fExjQp)F#m9VEs(0_+~A9Vrzjt1Vq(n3KFv@!mcHV~TK8lt4N8Of>nD}NL3{)& zB2OhbH&#?c9h&7|5QntQk9?H0p`k$iRS96Xq+a=8*J}>S7SG0I`3R_D08lSKcaw&u zraDsQ9cG&3V^o^dLIhZp=%DgKso9fm$SUnltAt>4p%lM>fZi!D&QQl$Ll`G#S6Jt@ z|1ikkc$|5OB`i}4oRTu8@tyfac6~Z6fbvj_kSDM?D}NA0nOSEO-3I9kDfZ9a`QR!@ z)Ihc}!6!%?Z&q}r0FN>i`sP8*&D1|3%fD+C0MB6uc%{d)YPQcooTvr@JdKmcFub(A z>-h*ZY5bp9XftK1hcFXnXnaTQcsza!OVYKHh?a5RQ%Y)R^qm;qML;KJ>B74N%+kQW z5aNHhtltP{-m?Qm#myQ(JIh+GUYbNAZL>E;bnV! z793rIOSv0UJfUDBN{*qb7xUkH0EAG*0)V?D-r28Rhpe~Vyy(5j&rVWOhOSqB%-Br| zkq`OuT^mo8gI$3kE}0P8#=2*PQWj1~;7ML14QpL#n|w2jfQg8f@1>;EW1U=#@<)WX zMFsfYeiXFf{;2UkM)m9MUnhu<0FBP&R^czdkq;jE`e<~1C^c0x9AAlH9Bq^;hzUf_ zC!zi)$p6nP{OcSyVB}{?5g@8q5*|cFCxc{NUZLW%z}~Vz!#w$k)I4Zv-td7D z6@JvmLIW;_ZvdxC`}y3MeJ5*gue^`l$UVF!VXe_~3*tPZMP<6I@g>srcpua;+ybzN zF(sW5l?N_VYV}r3$Egdh5Uv$toB--4=#YbH`lwgFII}3B!aOqMZx7$B(zs!SrKxjT zIflj;@z&P$v&Jeyx@Km{^`0nen0osk?J{N6lw;sM*=P3etG%&%h+lh zZcHBV&Fb9NBbOD$LdtiyY6Ob6a!E+cy{zxp({Q1dWe#vFHml>y6+w-@biKXNccmb% zcxoJn*aLxaBxF|B)rWYA>s!n{@f8i+72F&%wQ2S!HMM{i(0B4xYL6V1Ac1bb*F6sy z9Av!Q_<}X6nnt3obtd2$tKQGgg!Fp-M;x#LoQxI6z9ill?FZK+KCotCP9}s8=i!pm z_5GpLw|3j7?ZdgPYbT}UBzfFws19QT0an0sk>t>B)yEnM16A>~+EQ&;ZwpUNW!01o z796rMeHC=z{wUnt@$p#?XIpap*Por@zlTp`dIMtxuBGs!@hL%7PD_#v6HjYy>6L;i z_Xx1)U9qT9n2v=WE-&tLDpY~xCIF=BKXq~xQIid@+160>u@VH;CF2+$84Wc?@I!(s0OC`aVH(+6?usF?LArs^iLNDfMT6x?N~ z4AuNn`zCjQ^f`;xBIW6pfOzIm(``K zaCPDe)<>FNz4Mp;ZaH$bIl>Fayyup#U>zdPcy@A9zcT+YJR)%^AKC^Sh??V} z!2=tr={Tj}`5yFH3z^sa>{y}{z1XR^^}@w<>|cm>8@lVVg!s0$m9pp#7HCmf4z?LG z))9YmRcN6`D{!9s5!9}6T2DE=zy+jq8}8khrC2X}_eGSGrBN0-up!`@*f=)RmoR!J zVHQu-3cvJ$@|?-f9$3ID=bXUVmP|~Hi_sH0vTvF!#_)glA^#ETfaaw&;N9Pn91ZM& zEegDW^sf91lb*DRpq_M%Db;{B(QQ#j!Op?;hf03+$+z!RBm^;mgtiDNdqpD0Q64Gw zAz%e}_OP8u=~JkNPk& z)&^7{KT9sKlG7?;ZiXDhSJpp|Jz(1YkYm-_O3r@miU)*Hh^nC}JK+g9h zR!q31r~+yiz2Ya0c^C^NcDOd?`?+?NlP?Qxzf zac9lz!v;Mh{RHDIiSrULn>$zrVyRG82@6DkGZ!U)TNFVt@OLTJzezS-EjpTUJlDnRVM)gz`fa*TW#t}DR3&9{ z!rSqsrDFRUaTksOs|dRthouTs2LlBj0FmA(Qa{Cri1I6I$O3nS>dIj!Q>IjzjQ zVY?z0W8iQgyehY&>Y@g6BP+`*1+xzM33FFk%|2!xZ>#C=ap$%YR91G7*%yxq=8}*u zWaqC`?xk?K5HH^#LjMml;{?oXA^Kp@_sX~T9bdw=LkWq9qMsCPOvKy+(uBH#X0B00 zTKJ;Mi!5fUq@+syc*|-ungG^unIx3V1riqjYqaN1LAW#s65a3NDX%l9t#JfUNZSCJu z6?2SvmiBXXY~@%!9RFZ1-AjD@!+Fo)(^0 zZR`w5Ov@_QxS{NcRS>fOwC@8TVi}bhQW@uJ%xy@^m9_q`2E#cM<_>~6iEjA@bz+yA z_o0ge@U;Ha$;H6Wl~iAu{#)@t;OD?f9Gb89qyh29$vQJp+6J&H$8;iQ2%Y?eHw=F8 z9%IM^?mD3K=}`gRp9`b~miR9Sv3dDl?=y|>D}{wCm^{5#99Uc4+MHgmIQfZoZ5qG{ z%|w3GRVJ}F-zQ8kCC;O`c8W@E#sHV>qHLH$J+`(T`^IJ_Aq|vk<80g zeVD$>)$^E4(=oRu@x@-3lc?wWAls%rK_rpy@0axRZ9o!`9|+t$Tp-4FfX4}u0POLq zVv2L+O*A!8H`iGO{$pm9@YNEnZZqDxRl(m(l>b=>A_n-26PNpOjMqgrc-e@2BzanC zL=)W_2E+E(?XYtdsIfA(OQ>`lb#1MSWG7U;9z8i}4dlp& znVB3*$H_EVZ|fW$y)L0DSh!4dg~TgWANepkyhu>FhbKmJqaBsBwQt-8)m^>*au@@w z+)KEur((tHr9fyAn^WP!$NNg!qJ z7}NAxwhcJBd*8P1T*3ABMGk&rj!%G=J4j*+glB|9g$6(m+og!eQS;(3QqJJLr~ zfW?jnYH!7(*Cc?Jo%l14=XTg5Fe@;vYJf6K=c!82r5@Vw9J73IY-sd3s@pC^MhJ{o zpC-Vqfq^4^#~}ci;`KyI!*n&dC%i&v(%e)drs`DAOeNw0T|#bFW9sW1W_<TO{jUwJ4c|nJ6Xl+RNAd;oMagYY0FEMK^Yt#^&-!t z^aX?U`q$(hbgrEb42ImTH7`F6QFfgl)|I<6Qk=zDPz~s&x{BqRhJhB0{^XWbp-N@e_H8CF7@U zH=Rxv;&c0a_XlMcfQE{4W8d`h)>eFyW0RnB&64p;w`5MV7#w5El6F8uwn1JL6h&q` zi^StjdX6>Io(wFw7Ubu{L`2;6l??V4t3WGOX*;PW??1pS+Z#_J`HL6u`CK-GaWZ`8 zuF$+OZI95aTHV-qi$&uzfIt8oV9vCQ{oL%bd+u&d^><1W?!Yfkn=fsfdrlimOSw2e zi1RIfC_d%=H+zXj^=>D0AaWofb9=-SMNj0@AuZ<}ezg2OqfoX2N6z;8eu#!?Ind6y zq@svf<9Bcob?-u#MUAl*X#t(APSs(W)`E36u*0W#V!*s9h{NUN9|r3Vqyb7r!X0tgh}dnAwf1+{339v1+{7eksnbtUN(a zA{;l~s8rt(#09puC$!pn^LfAWGarB-w~=96l;**P=bSMhRTDZ1pmIz|)!F*I)g~xU zk2w}4mfWI`;X?%S-Dx9Th;Kg2t^_i_C9xwxYX^sFb|ldsiWrKYM-oP{s68MvmlsFf zHWD4fD#FV9-Nj*vRjii)?f#Qcmc$s9uv?Z21<=>b&%ZOMo-P1)(XeTT3Oq~VSL{KD znBTE5e@5_59q@UxN;2qV&vZ%j@T;kIcuDPIqUQE+NruXE|KX$=hENOEYD|gHJ3?J9JxZe&pTy6k zYF7&ihaAi!FGNn)D~1nG4<0;vbbGY&>$gw&_DeE_IJ(t~dft4Gayp^b;4f}vvO$xj zDYMKZnSJqgzuB<=bg1tO0vy3Gu7>Srqq->^SLX;NbTz{TI)Z5 z>5Wc{z+Q-Ju__0^zAM42KAOUbHdO7ZaPCv;y#QG5HYw>Lag$S16@ArLVC2UQ2V({c z36`l*y6hIpwMZ_w^f<=o%FfB_F)p1q+C0;?gHc8LX(q47x;gju_6WOp6o$pP_PkBH z<3>uB%Q6YO)pH}H4cF^(IiJ|=zf*@hMXlOT)BP_3X6KwtX)ue3 zyQzlk((q0N+4QPPI0B`H(Yoaqo!~LrcX^Hct==CjkjWXunZzO z{emNYh<>m3txw}-Q=S>4jwgw4*;O?Rf??8|pN$vP_uHLrz7B8kaA%NiSQYSW^0WS zHQ||+`|RV209fIVsQi|MXYV|Ymlh${r6MOU@2kB)tQA+uG#^%#b_z7@;GG=SG;;uq z(2}>eR{(Z-&9jV?)P)37CMzWF( zzK6I}e3`@Be4iSr2C5>959;+NG)fJWZ#%<>@ABIb3pz=IEm&nwd%|AP-;1tLnfips zEgVCcy*BOWD3BB$sw8uK9-~1R&ODF7L$#w2zi`b&S$SwC!;(4lu<(M-ri}XVykMu% z%vU|UOE^^?djD=ggFIBT_6lX>c+#U_c1)Oj7T3|bDdWz6F=H$ zW}c)_UaV*<%zV0ko_&dJIAZDfb~bwd!CT>$9S=+B@}*kQ2v!vt;?DbhFPU6=G((|Q!fX9@icvtHREWv*Kva@$O zeoL^*tw_^X9d{#x%O;&P+7-rSlgqVl$swMpjmRWMQ#)_(dH8O%AIgDi7hL};^#Ya9X3-%V za~TmOE^LPv&V7RI8&CRtg^Y~GkETtyQ=7lxId*uaszf#K^$9P`cyIN3_5rgcu69ZE zR`2%AN-4kk=@;ber2ojbuL-}%IqN0}?y3C3cIvGWFLqBTZn9dy#eU#H{DH*A57d3B zmOFDvpuCiX^KUO{+^YL0#~9=jpBsgi^w-Aip6&G;Tu2QQelk<@+ zR^;}o0uaD|SnG{#Wzary@L*U~9In8<-*SZ@tLj{M_9TLD!v{&xcHf$*G(`@PhyViw zPRHBdD$fr?fw`Z-$?4i*bi^((UXfc;mMHoA)6z>sX{FMZQ20eC^ep3T)(XJFi52xe zoUe_k!N;hpbJ^i79D2%Q!t+?Bdyy56_f0eB$Z*Nd);=l)K$E`3;D%;seBYsdytr-th; zE%P3WG6^?SC_NxUTit(zil}iLAoa=w*bp6UOfdsux@1Lig=J}cHJgBeGJ{rCDl$a> zPc$mtQHyLjlJ^EfSj~`St!#BjvOU(hr7n}A@&h#rSYY{ z8@3q}bLZONB>vn)V(~M5n<@LuH`d!o7Kahy%#7Pp%jV_8qpJ-XL8v-;fj;0SGBg{3 z-S44X)r2>!a}G?I*lL+%hXSky%xK-&fQgvMDF@p$iH;THh!xZVV?o`aMpn{b^ful$~is|!?7KY?G z1;aozPAtoEo^GHP9{Yx}nfn{%(0&hjG*vxcYhmMc zHmb9W1eaV{eVWQUmD0Dcna7VsxJhpq9S!T{#K$pwn{NH-H~*Ol0$mMG@JXf`C+T}! zG=ac@2%~&>S!SvHM+UAgovQ%#V{w=FINL)Kzh7p^A-7dEr!_Zf8;k-^Z$UE7aIe;n zabBPFwOGwK9(dESuC!V;CFB2Nb$G7dspyJV2ZLl6dm{E<({g_+Q3L)!Zh$jA87HFy z;8;J&D=^MAHzCI{%P}-O^ynqwBjK8ULvyMDAnQ@#tVY!&Fa(5K2YYY30S@fr#Ic(; zGkSM7lh3q0a;Z`NE4p|+#YfM6BxA^=p!4<$ZG6D$JYmAz9@q)hpb-QZNM&LPfxzE_1>ji?7KYrKYTbAh=}c~J@ie!jj)zKdgjtdbQr}Rq-62R$ z124XDkwJ@@D}lpoRGo+zFH1&>BH+p9S2a3ta+r)i+NFU?hq!&TUDs z0cTXlBqm}8hE1S#kKkBjic-c|5Z&+9aVssyU*0DSd9!=+6vY%eFz0+WnCms3er`*M zZTFQWFam!q@pESkAq|bu0b*1pFb4yYTXChp>>2d!JEDI>P=R$ww0Rs3o&18OuRNq~ zdffBJvj8%HCXoQ2qVpFh-&(+zAD%Sj>9u#ZcEl>Gt=vt6T>Hvu1rz#pbzENGfO($Z zfAudV#2%>!%cCUm|( zV*g!As*p)?DCSfr0dva`YyXi9NV7C8Oy+P7#>5bc z3zo0id-r0i^Kr4_YEy4Ie@LXBhv;QeT_K3ZA)#XbVXte4-__+6hs~6(?${7wQBYKU!BAm^F zkT#^dE+wBRPC#2Zcw{E7?z27M*Z>>!`D@VoZzd=QsNq72nipC{oaK{8kgTeqt=eRs1JMHllyVnJa{u? zASzZC(4>8NBXBRfE`_rR8yorXgyFfRA5Zmo{}>tln^2H+96Mky)#$T|tk z=q!v>IYPcYmhb5!-VYOVwl68KH)Xsrv}H)h24L&)l&AlW8-WO7z_F#B8%Jql3Dq;a zx2`xSlKxyYae`Qj2F6AXCQfo%(n`5df?V#nCgY*#;i3WZ-KAzpgV+`ZuS?4< z-r1pz!M1jD8(li!WMaiJ12+ z-c$oCT4`Uc#y}g~fqjPN?x$V6$2v;<9+r>)D>nm#(ty>)JTMeZ1>$&wo~O-n$$R`( zTCUUraPA5W4AJ*kD}C_h`~IdMe+J{)akgEwz5T~^1(Y3bg61CSG8^1-|sY8+0= zK?LYvlhP~XKv{k)>1E+56F~C0TmIk2f*1jAt~k`8#F+j{u|46FFi@bWN2icG+YT0}Xd|o20E3|Mz~km5#a()cln{oWXP=kP zR|d(~?*hzC4m2Os0)-5ZTTLmD4L)J=mA3vIb~T*jeV^6M`yLGcK3V_-18|($*Xtc< zve!J^T;v{ok_o+iPn2UP1ShJtR!crHs5N;lKu9)#x(hpCH9S}8I!Y{BZ5I3u;m{;t zAa451H#0NyftJ{6N3_`EXHkIxDG&4VrWPCSTaC?MvtN#%1Il!j`B5E)YU>IH#Hq8f zDGj#tqM6O(R=)(qj9vAl@)K7kl9=%xK~o?$xmEtiQwef(P)LCmQoP*^brJ;T2NNV3 zLo#NQ77C!WN7-Cg)eCwA{Z>OikcAts_pf*5)r)73_@-0e+jI7Ad{COka~Ysrlz-7v z-d%a}vR014qNdK-=>j3FOw9%<=3NY-{+&LRsMG?1p6_KGC-8P#$;#+3r zL>m0*jNI6#{|WM$Gr$uN&gPAhbb6DG>{=d{jz4-!0_QyREByE__(6!WfwZy5y_24C zRk?-5ZX}+{1PH;$*@gT0#%$p5d&gd*D;* z0VP%+_lchzJ3R7P8e6@G$l&*uIc+Q4VymAyJenltY|yS6BF)ZwS)r=%wE5%BZs(l% zDF`%nNwh4rUg%V^wBM&O)mwXyy!osVV40n8c3pVS9W;}Uj?S+DmJ09~flKc9gR43S z*g}B;vyxM<#wWmlB}tsD7Qm!?dk&=~;@~K+w2fyM+0`hVN!Q5D>J6K@c5@(NOxG|1 zYg)EpbJMB@DCCAEc&;we@r;UT_3WRRq&z%g{UMWmnqjCsctf+Iq9XNfzc~atADw^Z zCw}le0=E0HKf|!maJ30LF|C)@h}=BoL8+sI1Lvfszr_l+00*LW(nKt`L3eR!2|0zv zx}#^zNrT8REVpgu@-1_#a59pB#wxmP9C5fpc^Bn(>kcQM{&X9VbZ`-xejf4CO~&hY zn{_eky&2voe_(m)Bu%&&UmZ+_QKxT>zhI2De=X)wCuc&)8xq$hRdo-5E& z514r4lzVESC#*{O*%e@LD9%H&$Zng5iNXhsQHTGSildj%(70(I*4>!XGnt7Bun9x} z4^q)@_J!`mVF5E6ci;0JOe3o_J14|r8_S6B8A^=@&M<(xP4}9au_eeZPQj0{w2WnVdm2L*d=#+dG4U&twA+`FTHR!^uFsO2NbeI!YDz zd4&|hD6#;r>wtg=8V#lFD`f&FL*6Jg5`GGbewV64^in0#ooA3pmf?3m-;P z@-yO)H|!J_|H~;x@-2ZStV9pJ_@<^!2x;w&Q^Ag_>1zS{wgDt}6)NuBiq!^KmIh^4 zU|#ELG7#5jz#W^RbjQFWiL8uVoHuy6A91&?u?#? zR02?Hy}JP`3y#1;NyHifxETXpYdbbI0@trwJ4C?%lH;-3&RW2xm3`C;SZ&RnL>#fG z8{5ey#8y?_sY@uXTMKCN>)$*?_~ildJ1&q|k{cwHnOin^@Z_sf(1>npGnW6D~VrPQS{Ye4uH29Q`Q+*ktzl)j}_-%ZF|up1lx0_Bb%7gV7||P)d*Q> z&@phg3SXZrY^L_D!^Ay}WZoF3Xh zn^iAw+Uq~_qQ4n|rV6meLXqcCIiy9&lz^94Zh#uYCa0(!>JRTRO1smMSJUZB?gQjW zdpjT4-*KU;m>|B$T%z=%BxP)sWgyHI@?&@yW3*IwJr$>^8(m9)pHJ`TR$`KXZ6Ieh zd~4*R$)v3@s=XXT0e!nX{lH}K4w#r>l8hFtz_&un}D6K5oqBioX0snU3jQKI1(*LRv<`=vP@-9L`!IOF}TJPJjquv?v4v<`tql zT%`_CTrl=Az+uhfs;@evxM^iy0bn+EU#+zG?@>e@AAgYhlGkgRGb=H?=YA zRtTieN);ft7Hze=8mY?)R^4)KFae%~^grbW@Z6v+z!{&E#MiU28(E}+iuSz^^CXWgWOIYEsjfBa5ECb0R z<2?gix~$YMGUIW=0COq1uZeVc&S>H*@Qkv*sozgE_$}>G`fwW?ZoN4vw&A&6r<9#< zkE15`=O5JS$JkO0@Wcb$LDz(YIaeAqHLbs%srA=de|T{V1M6@w{JUJRSxze{a>NA9 z-Vxifs+fZ^4j@a}UIKFvN;1Ra=Oaqni8oLVkInf58UMxBk4cAgfMC;vpC=z|HntLb za`J2yZu%_6!Mr9xA8_M5EA$F{`+_$AG358hO}XF}-+-N;(}m+{i~eAK!I&QMhWlAu zE%{&y&y9!NL=DF6yvK%K>nNsn&bWXn!AoDmVUBB+fJ?Jr1=!h6{Ksv1pH)+Mhv$Z6 z?H6sNYZAaTPXyn^VOzhDmgAIO|JfYpvFN-%cFlNNpBvC${!$Srz zI?KAQ_M5IVclJ<8$0z8-r zz@=*4MThv6gR_R`uAEa8ku&GyFgQftX8ix?dh4*Nx8{AAl9WcerBk{?x*HLY5RmTf zRJuXBq#I!q(jY0_NP~2D*ZbM$obUTQ=l8&M?LWNuW4ZUNwPx;_x#wO~)JW!UH6rDX zOj2%VyBN{OjM(YI@Kjg#Pp_Y%gDs%&*jj(0*n}8WwV?* zaMcTx@a__g2&GJV``8eZqTx_FUlX4wH9{(aZ>;0xb?yn3FE9%f09=0LF)^uFDJf5T z@*3tP2}MKEsTJcm!tLrQY>USv%|nNQPkns5zQ&dRk+S>Gh8}$dxP^L71wjNkK)%4* z-jgs5`4tM7f8yY5;9La!<{Zt4GA5RMk}&;KwJAV}61ej-6&V&^IiRXb=I{}Uzbd8? z#(dVy8Bc-LYvWB{K>j$*IR_9rm*0ZQO^r2K+lP<0R5q;tPx$O#_xB2z_AON89orp* zucpGpezB*VRr#Y+F%6BST~R)?TFVR`Hm`#czG1Eu+B{K|323l(v%6{sAO7ds{@=fR z9|7-uQ=!;AlMM?BEV%9i#+jF#*m^H$bO4rF;^s3IgC`Cob|#lnsiJk=)PsGi+^s3ML7Db?vHZoH5IYn7L1R zL`L_Xldwx|4bTX_$*LF_>dXDbl&RT`os+G)_GD510jSCER6LwHTgXwlAp67a%N^x& zI76r(QWEKxB*e0-vj05!qkme_{%ljY0{y1-<+2>#6s< zzUC0@(UvlIHPf-XzA1^v2$$QtLlT#3QlA~A6*YF)>mUk=`McP!;?bWNM)1{xslt&v zbM42*#$Ib&x-SPaRJwb5K3yIx*pG_i{6glMCnW>5tZDV-LwqTY z5bwj>(!;xh=1W~3`?cWenwnoPa6XVu>w1*!&DPNGesDH|uTF9jKJN!;r7u$`q>TIA zJzamxg^@&e<|6o7FJ;)hz3Kmsw?Ff73Nn2PJtwlNrxbeD;Kpaj$irupLM?@BDV1#z z6{t{rT^AwJ06h`cNCD+Yy%)mN1Muyrp3Bhod%syqgxozm*4yDos$0AtPBZ%fVNEl$ z0Tz*wGU(RF`@?=eXt;S4V5T}N84nm<-6e9F|43wGW0NS(R($k7Qb%c*tG;+w03}OH z#N;4>5))z)L+FeT0O#Y&n8x^?Z9g{X8&FJ*A=oYRr}ZWlC|hxf;(!jAw*U)m;^le% zf}_$cMJ>CvQKST*O#6V0$fE=WCbcfs44VXgMzW{=B0Fokh@XD9Y8rr7QOOVBG@ z78Gj(Ogp-PNnxwGcM8rkRctLL z#0}pYRYimX2o5$;W0u z4jN=A0~%6>5&Y@wuTi6j1ybzd6V0H?exWC|vklAbM>fZPh-{l>;Ygm%v=E(!bfV`) zUS#3{wKFr!792fUNLYnxNrtMh@G{?ahWU}^Wq4SJNx>6UZ#4(0gzUNKzFT%Oz&0#3 z5RN~weVVA_oV?YL#VfGLR3e}^JkBeZsvM*h^Z{?h=QMTcz3Tb?EpIE_K(7QNFnl_7 z=DAr}V{Os6$7+ay^nxwlf3BzaM5mxN6}@K69D^L7*XY~#r67{<*oowVR0rIlV&4-} z7E9-27xEA!n^0dk=E z#K4;)gUZ{24Z;3tUunHk$dZAAF`Q4Zx*hx5VI2BjuP0P4zdX#B7cDjS?e**f#Q{Gp zFues}1WvKmK&L0SrQ>t6bpfnK*Nt7#e%12U^D>*{G=8#C^t79g=U}bb*HQfBx0i=? zRn3|8{4%30HszbU_eez@{miWlOLrSwY$b-x2ix0~EzC(ybo-n289%zy0qLW%uhG{p zxb$adzIz3Irp^Qv;|N^wg`~T*0P+x>1E%wzU$dy^OaA`Qsz219Hejv7@M~y;iif|I z6f<8)k8@_(C24wy)zX_*P$20gC*P%pK_m99LD~M$URaH;9{nc31{}lu<#Se)K3iHr z1qckf&~pNX={}W|oy}rzxB9CUlZ5}F4EhvI@T32gy7=`WUKt=ZJ&<$Q`0HrdQ*xWz zC`NA4ui+5h=gGujn~m1+IW3HUnItC>HlCs__xm>efTWH^1b_Yg#hmJ{+YWw~mc4*p z)Qr>j-Qrss9vH3I@#RDXK1L^zt*w{|XHvsXlbyDcz;ftWVw6w7$dulQu|xt>;${N$ z=7q&QF~?!Sue>A&oF}6?@qhS&=X%%TkvmDNo7k2^4)>zd7g`3?A%-yCE<3CW@{us(FZ#i*r6$6epplA+Ws?`1oNp<*OGb`L~$qyumZ4Hm%na z;bcrzIM{!oCythRbReGujIOzt4-XIaXEMR{r0+-7)JiFSscN70)Z_km zV!62+#-BJjK?aOUH*_d^eG|Gp9R!4Boi07rE@vDwXlg#ON)u*tnL~T4Gq8jJjAY%sy&>ZcH!*TQy0QdOBybQ9~EsK8BKKKK{ z)m)R`p{{dawvabdz;Q45GqY{>#vegRDmW znH>P{C#imV+t3-5|GMO(V6VIA3JGKJwpW-*tZx z^lPC}E56(O+Q1xoE!lQbmp~29h#r9Bgyp(V%m-lxhsV}9fFg&@XVBSl*d0MoCeKGx zKYMz@viaGDbMW`Um6Mvt~mfrBPuzWmpzF%C3_NcC4Be;72o?RG7 zp*?b7YFC|xt9+`IkYx;}$k#x}@{*62{k|RM*(9v%FuP?(A5C{zqy;1(ZDzeU!+_|z zyIyNGS9*AH%F)qys)w*|K}AjdX@KHjbTkjpGlV(YTlE~qhG$@=N1Hd6_`^g^7nXqZy*c@S-siA`&&?8Df2`m9$QkP*Tu^02}bC_E{D)TRp6g5awQpbYFgK$bUc z5;BECnyt=altVy+{y&65`?X$~KSZ~@$VM$kKar;T0ixsZO}RyYdA_So&@Lc!R;4KK z$ZJn+-cZjRP&{8>{{tTi;YuK+G_sHLh@iEV(JK;FNPFS0$O|U z%WlEtwji&U%o|6uxOguK1Fi>Rt}R_F^)O3)A&}5NqhZ9N!RJM2Sw({QK#g$@drC`> z?_4I;Tr|!H4h9rv9|cTmGu?NJ3W~is<~W>o7jK6-Ii_PE5Tc*@3_wDic0;=F*G%`F ztoJj1Bv-d`nY+JSVooV;v7BL@skKV=_zKp8KIz0%>;+yWNcVP%E!O{nK$Zd4t5+zK zBg*pU14~ab9;XAdF75WFvjkwIRG9)+e(y>Ttr*P)YgyME95f!*2Aw((LP<#fge>lu zH#<5uluT!gSOhZ0SAr-?Wn^`Eo|X?f#L8JZwxQj}e8R6GF-V{KBsIbqC!P#aa2Jd2 zf_<9j1t{|?v)8ulUwqM`4dB@Y&@7$pQIVn-ov7;yz*+qYNWN*#O-)a0OJ@B<1e5tZ z|6pdKMbIqNI~2@&-|i=v?}LT~=>eU9Y(t@kAS%g6wX&4#Q5kw_5Z)FM*YFq?D2R{=(gh-Nnc;g9cV8t* zxa}MU@`A?Y$CrH;X04qL#RNvR&5Kn73r|4~?Bwe84>e+{oXb#8Z%}kw>FYN(0beDL-U_F`nywQ;NaC^Z)z@cCyFXS zfy#g6yWTtB;??L0039s#0GCv{8?)IvAg>?;Slrn$1r0h}X%>gBJ$!2<4=W_4_*n;5 zzlRh^l~M{P&Dvt{-q-T8;3`O$y)aMCZ|}g{F?q`^Bs-zJ^U)$KjFLI&irE$*xC9D_ zMa}NH=Mc|;q(9DG>45cmoBW6eL}_;yA9Vsr|*D9W!$PO3=mmf>_4 z-epc*s!5v^`NIb>yf{sXueGq7H|C{EYKe z4rD(+-7IxVfhB=ct@%BggJJpp<1xq0O{rk3Pc=%RM#S(NZ)(o=_8Wfl28loUd;14p zt-h|)j#o#0M=?c2V|wytM7I&Xesf5e#rfFRCo`H^^j#(Rcx(8}dwAs_i>cE3XnF3b zFo>}^T~crN80) zZlBjd>djvSi&{y}CW~d7jY0zx&(A>Oq|FtNrEL+Y1byZ!$?HWhzAVsN`OL;{z(Vp0 zZ+1I1G(}Yv!XqAqm?O8pGDI2qU1^AdgDIjqQ~Kf6x|iJeven$=3-xRE-AFN|*X%LY znYE!|PycvfC{P*%Q5NV9>i7r)COj6S1f7Nn9Xk7+`MU`dZSCv@RI*$Kg}zV*l##(t z!2gsk7&(DLGd%wEx$}kDAO?TKMUF;>sNYgkkdtv59IibI8B**kckId?+!d=e3 z26DCRu4T|EjEIgU^vJrC%i|H~S=(YiI|sKS&=9bug|9iI68>`@W1d_XZfm|lH}QNlKexRC zUp4@hGegF9l>Lgq{D`Bz-k>dJ=Tq3p=_h6HjGAIM)Aw3|L2Q9Fs>p%?2+#an{aVZ{ zXvbPTGU80pHX5R^v6rQ`e8L3jEXJ%rdS%P;6H=5;`EY!=_D5&5vukd9-6-TvP~`p7 z8~nxsVFZ%bf*Bc=Awc*Ijf$F)6Wli#8!YNs*!bnu`}=pTl%I`K@KR@qYJ#E8LGkRa zsBoq}>~tN$ta;A{G9oBP-=TDiuGcU8eCZL{f7sV_iN(#X_C!)$bs894TvE;wKTc(# z6uMwEy_Y%4TM_%brA*U)={j*|gx>_G=KS=+Z46X^Eg4$4`PosSD9hPh6Y)KPB94hI zC&v|Kr9S0b0f(!<0k-Ms zw6CP?Hy9b&QpRuyZ?D8@>e|Ru3_d&lQDSQ37G7!%C_(e1e;l`s9Yp9!M^e9)7G!n* zYmP;Mz-PP$so#vKFUrV_1EVViCWU@{q_~-kR{%(T?>BluIwK-G4968+Ci~T!v|@s| zQ^{qug}H^@0W+@Iz+_k(&@%PYr*yy>yk8geovZz%`fHnE6kVZ4(&_~yB}Mw@E`MhlT6-PS0HI%C+|rDBFpVU6;oAR?U||) ze{T6ycwqU|a@eR(KVibFSU9eL3}VQ+$oxo0!N*;cC~ZSJMtpv&a9yhhfI{V84<*V843h|?u7Hoc%a zO^!G*T$>QZ^WMe1u)QwgJ4zu5-$mS0pf0-266^si8}RuPLNJA43Y23VRW{b7vaHPv zwBfP3Qd2OjBE|^uPtsLi3-7C4p1n=&>;9bJ>~x-6k-#(oNu(efF^t>!>^H`v238N( zgCcGd$E2aP<3uk1uwK1jespa82sl2-okUnYl!BempHLP<3fwA+7pUB)a}gWgJ~sYR ze9j1(GJp{@-9Ng%U&Va048Z{vqPNl(`juS5Aoc3l&>$thMn5tRww085*hNyiuRq<1 zk3U;D6)q_4o*R1RNstKqF-aX9?$i?l;2hqeLl>JJNrw@+@CXZJbM4#RBnu!f=D{va?k^6QyO2UTzA56(tc)dk z&f~Bt0EWeDh2i&=Mu-2>$siT{>P6PPwVH5L_vLHa^^C3E}zsD-%L)NXzZ7iz+eic*wf`c!HF_q0=pZhHw1Gl9$yE<$cl$Jyj2v2CI1 z3Va*I(D)>h$Ss`1bCBkwtd6oX@A2=fQ8})ZgUM=M-byC%!gazvw`4VV$=Mi(AXNnj zatTf zgGoFpj`lr0qruEJj~?;tip7Kz)I4Zbz{NfW;w-w#XtCCju3US8`KW$zdC~VhXiIt8 zR17wI^i00~86ZIMrUPu!6{NNWMP$evM+ORIx6gqeCPqjbo8i0#Y7=TdSrO7`EPB@E z&X=W9#qF<`**0VXqUvgl7wR%`$0mazkjdqBtx$JDyGcFct(CcjhNF~T(xxLNN3}7b45h?rqP;WafWr{6TxJoL+@j|1Y0n%7rVPLd;0#A<4{<_CDRaqSD9AuZj_7-d z{((2N3rXkQCM-jsbcujDXKSSaqg4@EzDD;u&cyt0WJ^vTo(*Ho;DeyMI6p^8rBAWz zyJHPx%XTjK{8L@p+MZFJ!;o4Taj)rP|H!20d;QbiBsM!HuORaE^UCBshYv zK|_dxPhz~U@(1!*fCn%!9hT)>ptZLq&0@%6vSh50SV-5D_>~MbT*W7(q zH1?VspD@n)X1{+Ezr|4gw~uCc7w^0@$a0B8$$0jj2?u`3;Fu^Ev+2!@2V5?>GJ)UU zMLMPS=zvMbq~JTpm!bRlQ7Im@W6P&;p(jIEMjd6akExuGG%COxL)_{!XG;#81QcZQ zT)Gw^d=@dl!q)O&IYFZMo;zfo8W4aXG?!N80xIM9G}EjUm%6I!ILoE3&a+=IH*3_X zu05gY*X;%%yB|d52VX6p4&@LCa!7t7)q7v~l2xDU8Vp=SSO&y5m(-P+&sdnCWD=_eg)b^6<^pjnVi*W8lH- zPfgOyZkESpP13c6K1sWp9=QC6K&Brc@mEBB(+qPp$k*IdU}WP~_*OF>>!$uU0vtNB znUF6wxfW}*nnLUEYh!c1duRhJC`mm2<%#&kwT``6CvrYre=N@r{?w7OM}gBJ6uYiF zb_9j1X7AKZ()=5w31V1;wvMyBwVgN!vkJ% zg}ZBxDA@S;fUi63v{IAQIxV92Lw+MK)jBmUj&1u8|MuZq?s*IT)NMCmwJAHDsHi9_ z$-#PP=Cpya55o1Mf$??Y<&ecnJ_~5Ss0I6N{7g#8ur8Yiu|r{}j)$_3yqo#Ch9kymdr|py#phyCwjW z17gypV|B%a2QfH<5SNf3)zYfH<0L!b9S;&&E=bBr$S-&t8K{+M*|GcA_X}*)Qm0=| z{}deMIr?S6;=wfL*?_&1mD0})n}4HK2<^a%>fO-HOHbXec1YrGZSJl&*wbvlH2dbj zVFx08_tIWBZ02~b=Me|A%Dbg}q&2W)(Y`f2<6z}R0U(P+9LOn%=BDz?PfoS822!b< ze4E5`jgoRMYShI3_gE_yD@|vtUJ$L#&J5;fcHE-(rDh}Aw2L?cUzmJAP&rGr zCzfrH6-vS^*+Hy5MGIkqvMXANk13EwsV9=>fes}ow@niHA+S$|7XC${#(9E_TJn-!;%k8bojpH{ovxA$}Pm^x-ReAg? zYD9s3cQ&)pGKU|FDV89=Q9N4}IKu$ec6W2I8JBDG#c%py?i$GXfh;yvDlCdbi0U}w zApSM$T+{e(9AUEG!?n{>+$9;NRyAsT=_ndG2z1Tt&sln#60r2y$;M)SD#XMrvIc(B z6b!PCo%$adr8KsLcSsX^+vpBP6st5iq+{d5Qwdo8RsNn9PW?y0dw!&`i8T@7eehkt zBc0!$0mMvrz&&Ed(^BBqhW`P`?IrmWaE0jnU;}zvcqob$T3N*~6`@Do&K_1}pq=A{ zcLn>_N%_A3F;LQHTxuIT8ej>0QU@G1*|ty!@q5KNsJuLk_cgqg+`M$7_qTiMQw1|2 zycxtXDK(m%(d8vHiLpvQ-!`9dgBcASynk; zN`TkaZc#7c6Nb?p{J}VO2F)$)d>>Xk{(5Mg4kj%vc z2G>(oqXC!s@HoaCOf+2VFD6i(eBB!*RD0|~p0SI*Js6G=3Rw#T!Q*yy8LVv<5ZZ0^ zm5BXBIKTKE(}8h#sHkaETbIYM`_rGnK&rv})}KXkcU90pDhEx}oJivNY+^YUxRqzI zf7aJP5yBE!G=$mhP6QZ1cww@ZhX?VZoQ(})V>?b$0B#DG*87p)^zvL7(?DAC=4O4O zP5%bXqMnwqVZs58?>Rq9B@UQIdS$iSA;MSr9TkR*nCEN{@oe>sHJ~WfY#>2CR7I*X zE{Dn1hXk9;AdBz$1}&OvLXFRXe{t+j^+LqZXz)u3+Q2pApPbL&G!8HTN}<_IU4U6m zdzm{4g{<~YCaXozZoY63kI#{Go$yZzY$la4dj{?p9#{ud+HNlw{$YO17}ezI_FT-|5(%!Yj67aeU98&B?qI zx^xnsx6W-#i^W&-`6p8}W!0wBMTtgSQykQ9pY`ObqNI~Z>>LjE*ceF*Fj;HX@fx{= zhPH0axY9v+Goxgcs?@4e1z}|Cr4exPzmR6Va;E?(roaM+V2bD=E#2n_H(ieNYZ!F%hgwmF zFeb}7tS&zk?-gmR<40X*@vD8L?rdQa8#5=x&1#QqWjra>6qk2Z&!!->;N=lXxH9@- z!mDYNpq=YSR=$* zdEV&Y@L6Qihqnn2nN#>ayqY(oy@V?ys|83zmc3aP==Ra*L+fe^(sVx3(9B@1KnIf5 zO-JU<-HQ8V1VKA~HjtvVtp8vC^4DKf--6f9>gCyB@H+!0M3iG-G%BRiP<+Ri<-6WBfqD69qMFW;$_910O|NT)~{8$fFIauwPFnK`H$(s(bA^a)L(1 z!;;$Euht{CcW(Du`f0~ZVfzvAoXu}^2?bCQaHT$?s!^#-?2jXrzNeKi`}xr^DP!j6q_)UJnz~k40 zLkJQsc@lmp%>oAbc_lz|Ov3+W=4)PF6?{yVxx8pcRcpbt!xqF&_>TLt5NQoudTKa}EZE62mf_9K5%r+I zz!#N&z9j=Fg+uc4c+a~Nqrhrr#E&G@1r356uQHwm2FKtd;_5taPmJ+~RE=WAwlDb$rKwmga0MIvblE5p!M0Ri6BL##-;RL@5hi#epG|= zFp6LQ3!Lb)!rZMFQ?t?6QTo-(LBym=eEOQ?+FQ-|!?;Y$=e9Hn8X7~$LBgP{3`0)h z%dFv_L3xY`EO)V(Vnz|HErjrCb{FQx=Gu1;t826tlQz+?tz=;xXHt5tA2c)%IFSFT z`fZRwo)U^(?48irI;KDcnT*Lb`YFEpgM=IvJTpyyeCNsXlZTw=22-KBrc`!TR$zML z-)T+Tr?;dW1N&9npE4*A!WBuCbkbRSeO1|A&m{_&H_%_u#RPW_XF7Bj5;$t66gPn| zB6jKe@+fE%9C2#|hx2F$8hgx%MnD}M2TZkp%8w|x9xqnAp+j=0;IkT@H;~z$%(#OIIReqmT zrDHpugsWnCs9SQO)d#_lBp~pPE_;RSb)7kXb#;vQwQ}qtlCNBI8l7dXQmih_H`x{)CHyU>HT|XVSL(4s4|{IeR5zRg z8jDMexWpnSAa4vz>gH$1>33yi)9U9VaEZB{GMtoR3jVh300|VGt5F6)$!jm#p;8j& z5pn}5tjkzjXSALNgM|eHCS7utiH68;{H6n9fdS=;79}0yvqM^mg|D?f^G*S+&|E@) z*?;G@NrN!BK$ouqCCaKB0XOfH0f4>$W|0yNr93I|B z9IrhSlzD%}^EuHPc>Xd0(`w%ypZOw%tj~q`GMRI}hqtnY8spWgKjXKN#-xP~6P)`? zqmzsk7iYiVWjg+GAODzKzZN6O1wHHaRlvQZ&`r$V?@G~zh6ba_V)ZJ!RSCCMf+d=_ zVq6~z1@6zrPrzdlDJdyOA21U0WE1-U`AO=!1?SXFHQtA>8oI|H3ewesNj;Zt7dIO- z=#5Drp+k?T%3H148}fn=K7`?$2QiR^^s*^RAL!VCH25-$;YrkbsK|}jU$meoXBBvw zyDolsN6vh9Ax-d^_H5-8$jPq*z90dchm47u8RT*MN!6{WG4=0mvmcKPsdufWkt_H5 z)+$W_2H$r*80HbEu3khMrUYT-N>hOyq#8jZ+&zMbzNB4ND_HrE9)bQg0eU6fm|FY0 zE(h)lPMXmQ;Tc@76%z^z?Wk5t$$#;A+Ro6DG*CvYbM3rFv^^F7R$9@vpaoT_H)lDvyA+u!QsBq7^bTt#mGhMLa-z<0lik-&O>k@VpLP+uSUW zyNt`c6LUV8PafOUH8nNOf}TJ&LaB}jl|h%6brjbLv~JQgD9p?cJv=t-p>#g^46{q1 zGAuztbF4)0zF}ZT4;F6Aj9a+bz(~aPeN$tD%ESGho5fzo;O1IGA7bt+_s5o6!&V*J zgoM}b|D@AzTo4LzWB3e(hdT*Hnm2=pu`vI*;$b6(LYfqQ8&dM@R<7X19g1<4uaPn{ z`zf80xM|7hpJ?oX-c_`{yP6z=(_`248EIk)aS=a?i~o}smcmpA&F0u=HE;Q_ilcg; zJj>31gRLYgsM9{ko2bXDsDx98xmv3euvn=p3e8vJ;6Ct6?&)Bi;1jUYAy~25Uc~fq zf9YY+-kV9U<-=d^b*FGxlhpb#>O# z6~+g+C6L-ns=jQtL?+Cwc&oPAuOqQcBgi8wZJ=kmOedgRE4qjpq)Nft| z{rmLAqbG5pj6lPo^vW_f8;KivuOR6~HyNFE{OJl?*LzskTidtsY2*iVnomr8E=*Q= z$>3J9#^8q`9!xb%OuW3Bp*0f74N8)v+U0{u2$W@b^?rSoqqbYl#(hA zEnoICqiin7V)xT)$s(hS;P5uC8BZ9`Ex!()2n!}I8_!2g{@XSB)k~6kH=t+qV@>CY z>3K{4sl6g5O#Wsy9s2U6R8xhb+?|w=VCA02Yvq@P3qV7$w?v9(3y+9?CA;M1T?EE# zlX|$wqnbNgprf0Speb)tewSW2|Ee#@q5HY5LU?T?n=nuS zBH3}Pchz`iba&uKdxwP{o7Q2x0A*jZij%wfIvbLzbi=Bb=1bO_1p@DHihICG*^qSX-ZJ88x$vrjP`w_B)zG0OTp?t{~uDd!& z+{yLNTQw5i(`c0+xaXrO(FR+ezDnVF?q+{EQNrA5Z&yRx4kMV7?xptMzy-?Q_6`)$ zlqo$HLjPpxNt#f`+9gS@W;?a-9pc1%9{CFBJ?ESe$~1Qw=8!Rca!lQu5KK&A%yx6q ziDJX5XtKR_Xv7~!^&_j-^Afg4_s_kt+%~2$8q9NGc6Rn?1%y_0(hb5%DO(yg{8j%r ziM)~OzrEM-gLdQ*BUk&SGfvpN7lCBHHgkcZtQ-Q8ew)kZ2@?}|gysor zSl#kZ7NH-BHp1>u#<zuI3jA9p>25XuWw=Z6i^JpOa!>le!O~-m?j)%iEO%7D~jLFAPsNZE$2ez2y)P zOf3q%^SSHsaYdbEn(?vk#U18Z#|VuQ&=$hWx_ znbLO(Li3*Q%7yWb@Fv!pEm6Hm>}A3z`8$-e#=RJu-pkcSJu*|t+V0e5Do%?$Mk!9a zOibwbRVD-d9vam9mLKoO_k&2Q*);PH3`cHvgl}AKyNH)v=Ta zN?YnPhC*tcHDM*UFz$Yv9kV$zQbAonb^m7ORF8Kk$FJzxtG92$KE2f1axm)MJiM84 znBvFytB>Xb(t?2T9(+9<%}buX4UoFPEJbk)sOy5e(F;FI!)o8Nga zY4UOD@qXzYR^SxUB%GvtoDV>SHv$BF?GKwT`})4pd$|;toqN-r)T+LFBn>q!p2JGM zU2zW0PrES7kCm(q;F3Z67?)5L+^R2D+azJ1ahqOacGQJK;oqzDcD8BmWiFot`n7=w zp#@pC9=iy&0PMtrp$p~MEEo2x-VARV$kp~!AFbIn(cqWotP2}3Dk}uW(u^b94cNL2 z`BlJ{<;QNUZ2+B7!q%dB_u}#|`!o%|4OOYP)V!y+E(`DPJt}+HNB5yNhpEk78#v*c zt0bWcamtKE27U@k5oK`{1UBumbr;FtK=k#ae8SU1o7IN zX)t_eZ_#vnk-rU{yK-_;|Al? z!A@4?DDK@7FFLtsHSWeZf1BQSq4r>{ux_Wb^`QwHM`mv@ujl04850r`7|Jd<4##{-T)UW=SD&|1z2xIsfk~z` zvx&;$A?2F=xjJn?iv4=frqAbL)2Gkta)Ekxx}t7urS&Y2SyMsTM3i#_3lej*2@1}* zq3d0FEH{?JkqY9Jh#GH&lbsFj-Jc8&XAOe)3*RDrrAdr+9 z;m=3tze_!BdzAX5KiPa6QmjE^BRZGi@SSIJG(YqTw=I-~HcjYN7Bj^VZ)myS4wIb} zs-`E}R%;CnBh4%S+6pc`*p6jQbhw~V*?Zg`(7bk&EC~Oj%Y{xL2D2?VKZo+Njh3(7 zii(xF=U;K(7YFil`Jk6-jk7%JbUe2ZT){#no8J};GkJkxW`d&kNt`_jN8?0t(aw?e z_`T3k8}vJwMPN2Zm0T2u3}30I_0a1OwDY-G3mD|oJ6xPMYh7owiQ#HmK_04V{1!RQ z)x0gAJzC1au{i2`a6B{1o0zXxXr|So!^JSi39?J)%{Vuy;%8c$m%D4}N*yw?G!Q8z zKLPZS?`ucC1;nP3k;9zAc0CQd#gIE3dZnHw&IZ$-l}>{(-IPi8$4d{Lq?YUW@~%)!#(ocgmu_ z8cHqUoSYVsEtlKK$XSYc@#00?!{T(jT{9=R@7&mXlYrS=PL)l}xbnJe!HnR+8t;Aq z6q)S-;UpARO||(VNrK>Fm0tlk?>05f)eAnE3`Ku3-W^ib>nDdvjeLmz^vh3_4OOV# zH8~+@cHQl2PFQumjHQr=cS_-G+sTs$E%vazayH6Cols$by20HDhcM-^LYq9bk7}c{ zhOo|+M<*aHIIq_zDbiwP&V&f)delKk7pF0~Vv;jX`$`c}GOau4ll0^lmb8yENiB)Gyq<;BGqWrsyhO-%y_%%33!zoZ~id)QN3{s_Z)QfX-ov2^nS|~~y-gPa=Q5+Wa zq3U1tk9;+>`Q8FaENV;zQ(1d({)c9EqqCO6$ksTJbTux<{hXae9U}?&w5}2e*Q^zc zlP)LdyW4PPaZJU{?-LY^|4LLvIalwS^K81Hzm!#uE4;*xp_!}kgZ6N5&5tZ3t_S^x zR@}@{$B5;cB&M;|LzCi)FZVhu4;lB8TZD1jk-la!7oR%fD#Z0oa`>+$8qM3W|FdR5kaV>?NolO~C$lB2`9F}WSTyqCT z@7}HcsOxk-tcHmy-5=rz^D7M;$uBclUQfx?G3;I|&XZk(^MXFX+EChU>!0+9%JhY< zRulYQtx!y@lp*fmz|LEPvuU3$JMb&T7UkaXG&6FkK=wkx@+TfSK4%x$}}03L9~n4pWY|txBZ{LGyUEbF>{8 zBnskCwkZ2HLjTUh6531no|Kr>$|LtW&9T_OB8uq34RV<+ZQ<{d)3|V3=$%{%36a;f zxGeYZtN<&t{ei1P@hRqIim+Dz(4D9hgm#jp{I37a>2Z_+N>p#_&zEhxrYrMPX)+QK zqYa1hRvzo3qIiScmPP za;g0T3-D4*(SG?B-mN||e|VsELzixR?_tI(Oxnn8Y)P0R1W8j`Vln!oxIxbNgx=av zdy8PY(?y^BUoNhw#KhWY;-Y`|D*$pSz;_7A@r;}^Fj0^_yWUJIPn%n!@vCAr(IoKe z3R7l!v3(+SIjsmyyWQ>o!d7LfolJ|_PrW3s#`_C9Q+XWs&C5F&9L5BQDs(7UX&jo0 z=pH4$21?unNHK4ww<(W(mNnwaaX`adrpPV{qETB z-$^XBAdu?C;gZd{&F}AIh5~#W*pSeuL6H2W?wY29S_ll<3Gcs5I63(n<9bsMZiuNx zOz{&8FNV3BYX!;6le`28ND+(ju6I(G-igY9j4LZM-C%r^(l$2!R<$Kw&fu=h2c$93VE$pJqug|`kl;jivb)d|*a`VdVD z6Kt;wIwaSZqB7&H2`^naq5N_a?J6wRk(!T{>;(P`rMJN#3=qI=)d2Ec{fW$VVu8dS zFF9&HI6vNBhUKR-2dJ6@m56c<@?n9?H2&jGO#72H=>oklC!c)_JlIX=_u%Ji zOw#p4iAp}XwvAY|oTeXgXLIuEzS?QMTW8iLm*<+d4Jg%VoU*~c7$$?7>i9MsV085k zUK3#IHXGbtNWa`Z+BUv$3=0WPW}Vn$K&CYRnHN==NFPmnUE77UFMe&kinkR1^X-HZ zpFlL)g)OIkz+aVl{aJ+f+pAXuZJcGdcy4>jEOASt_@j943-T8`ohHPc;R#fF*?Hx5 z31=SO%$?46y3R$Uqb>^R&fQBP_Y6QiQ>j-b2enW*QxZF%*)K+q^kd;D-1!%)w(kw{ z6D5&?w*(f)*CaMI@aVKBo)BDb|feO32p<4X;#C~|HG9NABtp5r3J@p0Ri_{rvK(=wKu z{RUJPa=QYG{Xag}3X>KU(m0Iw7OHcxf47Y(l|)Enwti3D+IGJ~ebp$V`X^|)>Hcus zr1W+sC}H}SOa;#e5OI66GEoaqZf1Q5bU-%zHv^{%BZ&(=%RPZ>(KY8)eLX}BilWec zwdbz(QdL?ea1#A_wo1F&x<`dj}7$4!7!#JO_ZJ?t-7WiJpaozMeeD3KKJw z5r;zZmLlT8Ew9dI9zNCbc9r+(W$~eP7+D1W&CgSXMDHsvZ6AmDgCfuQ<{7_o%iDw^YAyRR+%&K?*?6(bF7{A<;n{|~4 z_+d)u`!U;__)^cckDB&>l3eS}H1G|=$}eBX?j~z2tTS+7ZD(vEib(0aoGD*tT59OQ~OoBzRu4DqOB)u zE0fjZCXP>0)fXzJgvGl!wQub3x^kV5v&Q3xG8TUAt?&*xVT2B8=^=?rDdFlH-PNV{ zy7;%WVCbv?vHSuLWk^}Yw$HDXW4He_a1|Vwuuk$TU!HcyLszOT zdQ~Cp`B5v+t}&3)fo4`(ysvdUL<(tMN^acRwc5^^bK{1^7-t40{;YRh!FJ?hEGg*} zlMwmvnoDO$`|_fFT`BXijpP@Nc|lJUQO&zo{~uTH;8@r4eh&vHww=bd%@Z|d;|7fz zG`1SsXpBY;8rw$W#{r0|pKu-3|o_S_HYt~x(Xqh~%sd;ND?IZz1QlxLG zO*8Uy08cB**!u@h(UQBTPs-i?!O=yKGP9 zVH0DxEO~@8TXoHqVpI|rHkHzA+HvNty&7~OP~_LM)o&27RJ)_GaQD#)dX$w=(1cfV zrtG~>B{U|%=V)-w_)Z4OntCaQkZE*jl@P06TS-DhR{#D8^?Dhj8~%``zCV$A0FWpU z*JY=SOA2RonPp63)~i&Qp(e^}I^a=R7WUa{ARMzV&udsc-aCk$sVPP9u<`zn+!fRt z^NF+;f%KkC{=SKAu)Cj{ph&pT#168pL5w5j6nPVs-AnqwcuXlBU4s7=jI_t*Rv|nY zqXOMwW1;g`wwE81jl~WrR&Aae=p}T$PI38G?k-o%)QpJyl6>-3cYyi+hrh5 zLcD-R&@r@Ny+Dun2*tAfdpb+LXfs6;e--kV-yHrbiL1*qN@{ERSGL{C_~s%v5Nrf$ z34>!@pieT6>e==pwDsU(6d(DWt}=gcav8$fb5Plh$4CM3vd;NoVsGBA*H)WaPu$UQ zF<^elb++d)4hm7yF3H2Lnfu4ZMWn#Ht^0>aflqZ7Nx|lo0>W6W2j~&YRGpq`TsH^p0Ju@1bg)m5Ie=B*fo@8)`=PpNc6%612Zg zYlQ~U?O6)`t2^oVPHsRZzs&XHWY@U7DN9KnC6^*>v2;N3LqdT^MqO>vE-}x;^m_6y z_jQgZ))UA&FR(7D^?Y&VM~>sx$xWyX!IeYRJW)uinXom=%q*x6r3vZHtBjI{U5%Ex zGy8ie#0lRDrtI}>G_k+)I&Tg*jwiUv3~Re09q9Pj4;pDD2uP|oyzYdm=3dKHXFQ_i zkG$)N=0$$jzG?RN@H@8O?4-lH+Lv;Le!lAcWu*efiqOOm-Susq`_Z5~*1fL=K{*OV ze4@2=T3#1^Y{qg`#;uegYKDv!g5!_p*(tw{+_UHV?1=tg$BteM78r~T5$#)tD;!~* zsq)80yPN?1$EJi(q!12#lNSgHzB~1nhRKA}^P|`(1w4hQMHPMcoa|}*Eu_sQMbF*- zRvM5O1foRY6axt#m~S!Kyt>=6g%N@kDMq^DdW|(7gArF|7l7Sh8Tm~qHabf&5=u-QD+RF34Px+2Fmp8G}oCHKXDJn?6 z#h$^c`GLCA=@3Iz{EYPCszT7kSWKWh!2Se2zYB+26+sXt_p((MN4%07bi!|s+RX26 zt2tSJ{tYRSkNS^H#~=)D&xL;S_wU>>VQN#=i|+~Qx0ebNIEHn2620h}0Ui?AR$_u9 zCtIFcL4$_V!%OqYb}c)aT~fyL;7xj@2VOI4xKzfHZe;jsA1~Ze3$AzD4h7LWyhho8 zc9J9snD@)oD$u*TPHw!DzK>cYS{&1a#Bq1Lz|w-{XnQ;1N1EIZt4-!P)oHEEf)|Zn zgMq`XcUmy|#3eM@k|Y1EdmV3HNqsetfH$5r`+CZg__H`DT_esDH8P8bv@$=;=Ha9i zQ?%$nAN>1LejgAdY@J63iuFWTuUd^Tg?#!A`CA%KApPIpEQsr;n^ttNA$q02= z32DlrWa`gZAwM|GxYkFCd>~Z3lCBRoT=J4P%b;C>=r>gV=I<_Um{aYof^^t`Q7hU@ zjz29z=QFyv%6gNF!XMst6$eCxv+9^=gov864uNKrZL49IlA>^~(;DHRt^g0-f5)Un z@lI!g@h(bKBXEfW`GNdAhQooeY))`kPiCOXta90mZy(849jNP)Po#q-Jpf za6IcBdPlun;Ed+ylodVBuYv0O?pu9tIyaoIjgTX`eg<8!^QJ{alm>2I$|G22cv_qhR6C%o5z_q2~$B~aZ_(0cv}E+4!UJzmbPxusnnofE#wEJ6O-F~z%o zfkZz&sSBDW#lLE5`8jQ_1-5h;F@w!2pt6*l4OxtvqFnfG&gDUyOO0lHc6*4U5He+D}Knql0}lWKDSNHL&Z8alNaI52JaF4OI2^d z0evnbit@(TJlV-H`*Jh zg}JkA%v@)!rY>X93xEDT6veth#eU@>n5>a^=F0Q~PI~q&`LSSZ6-a)u@EcX830|0& z=TyWa4Hq(0yWnI~S-C^8LXVMK{R8iEyTxIQ51%82Rnw#Lw4~s8__u{ojn99x+q$D~ zlD+f_p|R&U4A|-Eu-Gg-zr2!0P?D@^hap~&K~M$hdyaP~LC%1P)rQ+?#-x@{HNT?D z&!J*9QKq9Q{&TGYZP`COX%3Sv@~LD+s%eGxf6-#_Aqs8@vqA+*$R-v1ci3>POrieHv0&%ybuviWB0mcb`o z&oS`l$TnR6hu7As-pm#z9{X+^Q(|CfOhE;(?wYv8vCed#&Ys-hGI>r|=_as*eH4;_ zQs2p_kNEr0f!ZwtTCYBenWOkbd5~Ht)1AG4e95{T;22T-)p^ScD>RI;243BBx<%a< z3?ENY>4dmwNH^GHYmRAhh1!ET>=sx2?^-ZLy!Zgcj; zl1HST48O)8`=gQTIkj}8VmizS?%5O!=%PjTM(a?R%HmPmZb55v$_Z|A)okeM2~yfi5eJ;3i#P5vss8=h0q`D2 zux&3>knLeg3aA@3C`S0oIpADGjF;3fgK$j)_Y;g6^;D}~gDWYKudrD6M9^6`NU-lP z8#X-;WH6?pVA+-%`r8rT(Lx&Y@ZqwBmoXC-ZiEfORfO}PMRnbndzH?V&2kDvB~SDd zUq>3T<}9`SER!gSvDq_2V4eL~UgVPNpRt3j}DCW2%R{ zq=kX)hyYj|HzO5Xo_gTWctj&}L<^Y%+_MSR0cFt=JpmZgj=a*xqTirj&(-O-q5^B3 zlegENmi)vg4{~TKzFz6qqg<;h%Ei>Pvl9GWmmM-Y^`D|a4F&R?&~PO@9f8Vvn{X!@#}7`Yawei&CIm}}2RJOT0gf#lQthAX!8 z{|WvWF@bm?PH6Ph5O79M`3`^xUrzvL)Ht#$lXF}wREDJNMG>KUIemO!G2@L`Ft_Tv zU-i(jj%)$N{98ZPZ&GiMtZ4>BE}y=95>(&5lUX%E!(1NR0_t$16@w-FGc;L(awKS!p?U;ME4gMd4polZX zIXUQ?FW~v@I@iqA3AB|kR(v<2Kp%x3V*b^Q>2kFkeykQ*Zma=Sn7vYV9u~-9<)mIl zNcyo)C`sbXUIDg+Hu_sl=!lZ&x<@p}`xBx7hqa(kI>RRVQjBIP$9(`hmGiUlRdu z!|XF}0NAQty(YT4s08w6o6FdX_+-dJmVSZ7**9b#>}&D=2(K^WK|{N|Y`Q2la)NGm;Wds~ltwCvJoYgV9=+*qo7t?)8T z@`{EZ4RGkaq2U4kCx(cXGd@D<31>?6W&fg!RJeGcvneLOfBN|S%vP{G$Rr9?x-POeAT)Da+>!w{JMhMZ|DG4l7 zF@`Zjt?<}Jpn9?*UmZIhK3an4y>0#3#73aXH_ZDl(NlY4l#TOYISo={aa;$6FzSSG z3Y(7XTsC$35b)_umJjH5Tb+*-&}K!87j`Q;d3h?~P1zgo6J4SkOG?#*+P%vdx@1K- zBlFcoJKm*RU^Cusm{6mpHlrY0wqNTn{>N$z3MK@ho|}*4#Onf7*~Mqj#2YOPn(B__oVyk-QI0>P(F zzIFyB#$A@*hA4j<2tZ|oxjc)eXmU{;vN+JP4UzmrnVuaK9Ww#JBmC#c@5HlQ92?C@ zt{@VO!Y^fQrr*94NT23c#&MjM5}-9i1Fplb^Lg8gxw1BA9`U&=&f45aBks^e1Z6q z-?a!@)`jRRHCp~3K4xeJMzn&kO(O2WsN4K|fv0UJw^{D$ex0E+dS*F1{V7ftQhX{t zKNuGe;ioq_rQf|=2XG$ka%sMYR;-6w6NOqI`>^LOG~~$q;E_9~z`EVBsj=z31`dwf zC@9>#t|MdDmv_AS1yGscSKjw+MT-*%Fc#P^jwkL*)4oMPtBBH)gs7;q{mCs-@x121qrIKz2?X#c&DG`OxMprw{w6`#WI(`?%+3b2rTz~LwZx5O(`(f^o~N)g%W&>?GPf`#C4}C)#70xj&Osq_TVo$rM=do2oS3{E|YxSqbOcR+^JBnexkJQ$2k_%Ji??z8&PZQI6Jb#Ev>ZoNu7RF=%agBtfe%IH=(f8{t}-%7#Y> zz2Lt0KFHH5DT59E1fJlqc{gafKXit&0dB`lrdOr)E`Z8eiI9wdtQ;hmXh4MRi6(Lx zWdqs2rW9<8#zzPJIVviS=JY(Nj(aoZ2e#8#z6OAUkNd9@Ov8>{e!gvl;3(<-5EsS? z)1y<$C35PGkx?EJ=f8P@5gO3I%rJ#r6d)`BBq5&|a4{^Tb?E&j*#_O%$qf8MUz^U! zr9fYzTrEJUPU7AQ$cIMp`2rmL#I7FHehrbZ?o{qQ47b2YUZgJdrUUUCjum;5Gv26y z3_K2IbbBxx1yK=3NgX0{(@LAGp^h zRHteOHNwHR4;w<5_rEvM-ELA5zvRdHc{hUa5}n+sN8(J+%lL)K1@&-0uwYv2c^(o* z0_sOgDKmQQfqz}3>n#5DLXg+r-39xbcR95kS^xrF8pPuM9G7uY)QZ)G1(}T27wh-!KNfkeHH>x?_0a!wE7IXf zfojXktQ`hq1d~elgY{-sjm-#!-2=TuhY%q(gzlUNGo)e4{QWhd86^&%JmKZSQnYLh z3K}x8Wo(d{?)JpxJ5OdFzjHRn2)R)KFI6`x%+j2@ev;G^QBwW1+VC4Bqb%YsnL(i5 zO-6-#kh5y`s8`TA^gr1@Nz8jLYLx#+a`R7!UJ3@_>9r{nBX|sST-?i$DAi+%p>DTP z<7l^wiY=Bq%}8Z>>Kgr$AFWy*e=s9(=vLukA^D;nbt|1bTUp1}GLxWcC5}1S+OH6y zbEvAyL84Cy>}9UJX25JmsLuld0Uukj0kO zOS*E#I;EADMRFPAK3yo=ke5^wn#(Buk(L*^a+-L z$7U0tq2YIV>KAN7*D&qbPs6B(uY2G7C~^+DprsOAHuh(T)fBBjN_owA90&jRmA>6Y zq^^&)#&t9R7=GW@Obg#iK8)k&w(?B`Vs=bq3B~1v{Gh)UZsZdUhk=pEOX}Gbl}}uM zoh?a>%fkFUe${>sTDab2fdyUCj{bB=AR4b@TZF{=TXugHVyfQFE(o>(WvH!Jy zP%;54R7>wLJluqTDDG4BL;hE$#gzL7h2QItLwwxlxuZh;XSK63qZzi6d!NbNr#{;z zPr#fe&&`{9gj{O*Rk6&1#0TTllx^q^KG3n*v+8O*ivshLR3C*r(W+>yUURarx%h_M}9d5l%O z7?uLQWkYG=@SmOn)J^W%b+wpFa({4vkN^=kfO@YyQ^?bOR%?p&PhXb83>ZR}l~WQMpiNLj=W|5ig(f-0P3rYE4%$H+dsWWor_qp$Yt80 z`M)dxOUCh5SDt*na{0J7F8DImw|F(@fea3thPKkuw`ok3eG}yYcCVp7<7@TNqgmhk zWKVOq40C`HSF>)OuxiSl*Oz;HR_spEeS2&e<#5j3R#d5yGI_)gU-J9kB~N^O5gW;G zNDx#6LmqabL8vzU`4RD5;vGK@W6W!pjy&w8+j3&yMcjp zE{)HBqS90J-mrq+>{lsqh*SW{P&OIrhU%q!FG-iJTXVHgl(v0rVb)MOjG!AV!-Bu# z9ihNL?`RhsJf{ourAm-+WXQL3xx?(%2uD}-&_6EAw;ILoIagfxIgw6=e=y$8A1UIJ zsECuY-ED)yQi_VuWRY%Y`(bFPCRGJTJ-d*Oo4WRrO&?waX{CwVKioQrXOe$aP=gqn z-U#G_8t__`{UkgG8J`vI(PSd~p>WHMhr?;0uS(^9>7lj?EVj_{V6NrVG5+Z?)_8I= z{$eHSCArDN^<7QK`Mdv%1V|Bl8>Qf^zmyon3xLGkP{O~7_-g6vS69@L#!X~kH-)=# z#ZltGFmqEd&oKmW?WeYu?pCF-IX!132Grau>J zI|P1rIbTOEt*Mum7z?AMAo5d{Y z1!>RJ`nm{Bpz(gMFD)Z9ul>0UnCYf1$3rny#$zxbISoMCAeAQ<{S_N~MHd?&rfdS?_wxZ}tuQNsPg zU3A*3h1ta76aiWlIRyO&nQT6F0lF5q@U0dY+Vi?>G?k4!P>g;a z%B%E71GEjorhWsOlV~!MhZ&!_*Nm`jQF1hhQ4E0asQmTnah2kULU*1=Olfj8+6>B8 zLf$ZIS%tf;{3-uenuMJNtgD9#)2}ujdQf4dfWr;3pI9(EX)-&Z?bfM#Z@x)YNZ({t z_#N3+QZhrRSs7aD;CF}sJjns2_e*-AYn)JmkLj|dR%(Pm*m~#7YdR05u#M!y6A!lE zrjywR?nyvvo@L*B`j0m;g~1HkIA3(-ROHMTygP5iQq5TVh_dcf=J+MZ8jULN;05H$ zMav%&hfL9CND@$+NR2JC z4HBXk%luv#L4Cr5ltViDEs%w5(DDPv35HT=`)&s*@1`nM0i42lG8%0OGunTc1opQC zkW}L_!T1gdG?w@8dy9TtE^~pjNR*DX`7-5$wudj7`ll2@Q73n*AKvZvw^;Vrq&09y5QWPeiA4azY>xVghw z)0JFMK*^Ypv3_+(K;zA@4nfxfnNjMDe0;rA4gC6DnOBAmA;1Y`G(m5M*&XP$!Ky@V6D!PgW8%1F}I$81W!=18Px`9L?Ip z5m^g}2J6R@L|H0kweRG^HL zUc{ZAN1^BTS><`J5X-4eez(gc+_%m|B2XoF2p4%e6P=EcBdn(E*yMXk`jW2WaVB`D zX4NHH5LMl7uF2HM1XE&Fl0cQXs%jU}4%T0#j#E>^U{cB_>9G1Ly2Ove`}~cP^p4q9fW}bO}>^p>X@fe^@Sp>iZ)Gt*>%d}3s*1gVUG;qEhhDeVAPM# z!$}N3G+Sd6=71_jFlQihBaAGZU<-Km0~Y z`|PtsUNY5jj)a&&4W~_sUPB*JqX!G?XXq>9D((TEKc3r5RPO4-4KNpkt$bZ1M#$I| zwN2ydeIM6V#E+P8<=04E*w>03&t+9h4T#yG)KoFLnv@?+BLV^>xJr7=(mmhjnv+n& zrrADxtu6ET-d!Y$niO6_Z*AyBbb!#Ob&4#AxNq3Ez@5ZYr;3JuY9bhB;Dxr5P0=zn z=#1HJ^a+3f{r;Z`5cy{|q6i_#M4}02_~KDpirikDF_Uzo|`UM z+xx|jWmx*m#cUW9dxY2<+0=V4&)8-1wk;X#7P(%QD}=*QIL1baixs7-pZM{cD?CTi zT{vLr)A)DEjJhceKnxW-TjwJ&p{WN9q64C-P2R>2(J;A$}c_J!DW=?{Wq)%WHkgq=-Toxc?B0^m8l`(~g{M)%0 z1OKdmeff;xA+)CgpMr_zK zcE@YAh<87W*o_{1-g)={!b3OX;t?34lt%ZnP)k~Qrof=W*zjeylXexLCtUq^9csTK zfndqhOXte?sTu1WH>&XK#98avZHEs2$tlfCJO1ctk)Jnq5BYubOmYQMMR9vd!O924 z&53X!ZS`#?ZY53gb>+Nb&!t^9*Jr2e*DD#qO9+7%8m&%WuWZ*Oyz=(KI39}k2rR6> zr=0FDAHMTXusk|6jBWh1iRLSB=*dl)tM=N4bqw*)H}YrM9CFcGBuCM99M<6ebXO~1 zPGhjw70#G~NtgBYX^#FIzO~hhEB&uN2SxGsS+IJ%#x@1d@apUJ{g*954YuuN6+fy` z5=W}`^SKO;lYbj7iy`SurS&q81yr{+vIh{vq_q&68zU^d&IplkA&=(AU z$>#pqk9<|V=GsYmk?~G5pX1xxwly#ZZCgik;Vk42yGVU}PbuX^MSDLhgx{6?;NNLG z7}NafbR@FDs(IUSd2_E7F<2S#9(M<6c?%nIKK1^HJqa8#8eOY-M%yr@KHaHK%QR`R zPZ+U880_(9x+K|!pX7lh?LLsFMy)q2l4mgcAo?K??sHX zqaBGqFUy!e|470(9oOYpwv4|zv$_s0W%~8Wh_glP)wfGZe+4?aM|pr=Yf-{RX4Mvb zCE6HL`N!T##tDCavQ#&Jc>QFWS*-nG>m#fl2JT0LoJke-m{d64XE_A<^iQcjGQ~wZ zLL|-!eq1Zo>nv9GYb{n8^TlMSq7eQpE&aR*AgiIBzNhBE7=3~XBNqhgdIZN5+)!gj zpzw0*wz^5`g5WpS-GvkD(>@Y8&ySm!*GTwX9*!}~9YX)q%Dc2}ey_i$g_!Xg3VF_; zeb&s2At6Z#+Ss6?YWE3k@9FS(AoHUcKO(>rcS+Or3JbGlH;pfX6kX54pr)_;9Axn` z%hAU)k|q=g(7?yXFU514?aRA3v`Yl{z&;ZSwUKZ&b-CwZZW-vppw_`74c8kZMy6v>=m`unk5c9Ch5=Luc` zZ%p{*s}Cp_^Blda$-KP0G=#)IP*?85jCHy^ZgF?q(j&B+M*4cL?bxZq&&szRS+r7a zHqFlBY@^=Q5Q0?Oh4naVGAC&*nZdW~VNWf<^6f3J{TU1CcL6R3cC*{Tmkg8ZI& z1w8_y`)CK-0T_(hZ2es*OJ+f#onV<-I|sj$YHhQim>Y`3VFj};9Z{#!94<>4U`VfI z^mU)~dLJ7cA2|4tKiyWKL3|9|oeFwI@F^*(`Xvp1xsp1?$c%K}i5@5~ib)9!=t;rU zJ7m6;W{7Xcd1_7N?2(>dF|w=iV!mQXLK zsh3*k6eo~4P#)hSmfyB|q`u{snCimiBvJ4y2D7mbmVl9T3A7R+tr8Ebvf5zu7~YX0 zbGjM#=VjB^fPRE+L)YqYjH~W}aIO*p@hDPXD-_NJ-2iv&Jfu3ry=vT^w250I)J{cS z&Mc_373JfO@x8<7^?qZS>>C_aAslAU=xYxB?i=gM6KSlt*3XSln*saYu6d^Uq%|kB z^!2kFSnd5}3D1zO?{2ovyKlFL{GWhPCO%%vj9!idpf~eE>~YE9xhlkLhNXk~zyLzj z5?<$YsCg6g6;*XUFWn3J-t$OAy%dIbNy6GQP&nG=CoMhud-b&~jS~g+wR1sVHp_VV zlO}tG5mvg5Vy;#Pe$^77p}!)kv@6Zb{9ddj;9i6=#aU#vAfv<|H@ohHP9vHPy{mBU zp}hBCMQ!JfzC-F!+{@~nxM&@(#w~r7sk}tKx)obL{rV{2Y5Ya zsOc6eBoW}}zd}1u{LTomm}9o=P1|Rj{3zU!ZT%EH>q`0#cw}hN-XO&P;2SO~0F);M z8-WjmMI;5RC1QT#ZeMmi+?W(Qa~^C$$J#(cKLrG746nYLmNA#UClochoGV2fMyPwt z`N=0BE}C~fJ^?f-i6z`jtx`xJvBdT>>Z-L)rVVD z`PdUpYmC3wiMf;rz?>gj8v{ZL2jc8A+~_wb6l~npbCw%NZd_l#%)qBCynf=eIU?_E zEVC^)X2mudY;rJ?lQ$%yPI_ktC|^;W6#h zN`PE$v9GpR7olAFaI{Ci&JW1{Op!EzjDXvJuP4NyTspyRJ{kR`4V;(wl+lbgErn%v zbI<02qXeL`ef?;LmcsmJZWMi!qO9cq&+76x(qf zm($7u4)^G$GYPe%kI}Cm8Hy-tTON(WXd*#GEuhR<+BukkVtDvnKF3 zdC|GS>*=yWm3mQ3&HGeA^V~~@m@ZT{2ob{X915FkFcN-GUra(G>3P(p-!cDoy7$t? z<3!8qW!aX0Yix)0id^vCAVXeCQ-F^1!4Q485%=+jEcOCtFFX1K`qQhU;=IsfMQz2x z6pccP!cCX#l?A5SLYC=><8!V2|5s5!$TFJ<94-4;(_9{wuoX6f*cx#WF@j+BSkRQ% zrR5S@qS$DbSzFKgSCKCPlQlJ?11h|$igN)LWpEGq6?52zbY~d6aV08qBX02e88uE# zFxE?L%}BYFcgqfkKGO?Nd%kExdEJVB;4H%IKXrvX5)7f8^(qxTh3gl@dDv(y0UP7%uw)Q-;Qt%XdA{PO98Qb4RKW8AKHaYYOj$_&OB}$sx49O9h`pE~R%V zgutlvtzpevFJm}s`Kgd!S1n*w)OPHCor}NRJ}XkJ255OMB6RCOEnjuYsAIG%M<*B1 z`>tba^;*YeZZ<|ZZ34Bp%gE0Pd3qk-HIj?mI^sW#a<&U>2-q=^!T}{9NEia~ls?s3 z)V0cyI%nR20l`faa^(>UDnk3D<(KwHZR?4K%3hp)^ne~_QNMhpuFu_l1(M-M^N}SN z$1B;XXZR*qOWg7qaXgHInnJHb?zmrI(T)NWx()-O;tt{me(Mt!Fum1zKW}o>l@>gS zXrpc(+lRZ5!Sz!ZQEfqG<^7wOy9bwmHZv-Zp+a4#)tr*MOo$=WCtR9z+%y3>_zx%s zf`?t<42STTXz2pHUw$(rV2Mrx`w+%-IEqtNU7v(`8jCv%l9N+89jQ3_{Xcsto-;h+ zhf&;zz#$Lm&V6)u`nud;TzLNaq2x&FA+Ru(AVELp&#Cyop3yH84HbNWUGjhw5Jd-o zh^Nv)&175|W?%jlxRw~_w5@*JM@gkK6MUOUjZ6zs>`vAq>L9@1AxI)a4*H7v zr&sTukRZKPu$zxu#@|{iTur+cLwm*BqFb_s7ps0DU*$GHcFSgoOFxzCt0gGvW^4TE z*q?zU7YkG$5+W;jvtHvfvl3|xHn)|xxNSl|cop_PMo@H(+{);eYQ~MPFPU2ta8cMq2-oex7)KbK=d%iWbXgu|g^jLI>eXKQ?BX%dfonig*aKPPNnquxM^>zypaush(lJmYjJpoWBF6^xu@2RZ3uF%5C zZmCTe#5bf&<`Xb&3->zR!;hAYXXa9PUP@&!jThl|YTTp&by>YzH0vE9*RM0qOmRIng;)wdPZ>WN*{1kk_QK76O#yp4j#_p;{y}iex!0j) z$h1(%Dtbq!(#Ju|AcRHy5B2fy{D_MHieAWJ{bIww!fXMcYw)$LoAE0(8i+HoLHErx z#=~E>wIVj@3{VL3Hy5(yXX1+mkte7cZF=cwHBhiFx6lMGEbq)4N~g&f_E5QpcCu+E zKcF4BKnaLdZ|2kf=$UGG9D0ziz5pbG(lFwoL+DMz)^in639S_U3t zD~g70P|9FWOSTlBNoJCpgnF+&nNgCs<)%wc2ZgBWH7I!4y7~4>lOnYD9qvs%)a-B1%4K5mrB6&f|DW_Dd%ON-==p zX)lpxAUXuqwX8=bed!Fu7LjhoP{1b(cD{^kujC5of0Pc^-J9Gzqkp=4&>oh^>D=oa z&CF0m(|%2bx;(+9C|xq1zdeLYq~J2UMf;vIM?Bm)|)^lf+MV;sX|y)j?dkI_4tsExzMHg;(Kw` zro+598tC6%Aj!(2yea?=3ZgGIpl_%9cs;1QYyNdPFxp9AhqGJX8;x?a*rgtgZ0 z)swttvlv^f3Qan?9q{OSn1SE#keTstKgAxDXnwz=AU_3SBHfQxQ97?|g@!u3Q;4Mq z8gCC1LoH?c7^{{-!J6_<9rh9-iGYyS7rRsvK#qHga?5TJyxqNb^`R^TJMTXyDhfL# zj?A-z+nCyhf!eniLBWsqh5Zy=HBvoV_fHeRj?R5zZ?S40Un_8xEfIIM_9;Jx7b-9g zdsk)znO#6V7PGHWh zh`eshKhO2gq_Bvg?sdpV#CXhDgDT@V76i+Z|CCuz*F#*3mP&JViAP`6HWxfGQx2sp zdNX9L(~SKt(Tx8FN^Fog7diPS`s(V>j!Pt0MEgki^X+3@17*9BEv$PU*QKd9CTg9> zRjTS4c=|jzYR38odUbkbC;^-TP0dk#JLG-MYJl#-`SG1FmbXb6 z=js`~l^ZV5{OJ|Vi@#7DC({&-UbOZOOtqf^3wFsT zq=)C(x&|Y^qcjHj6q=fy`(QxAvL8C||ppH|F2mr>jYrbQ}hNJF_dc(Hd-q zV+>FNkb%Gh<6)JHCT71Yob4RyuuI*J$*=_2{$LThFx(1bMgY*{5hqA!$mNmmKaU~aTrpq!DF&vP+vyJjfcb^>Cg>(U;T%cc`Ytz zer!Ch=L$VNO673t`E|yjBlwnHb*1Ob&$f)QDe$4)V{SaK5LTXPqvmk07%;lLE_bEi z=8wm_I-3qC1WWBXYQ!?r%VH{DY%+>|$+ykTUQ47&;uwk!&ale9ZHp3<`KZ@Ep)Zzs zq!v8}jha)Dt^HJ;Li0&fH1w~%BMA>-fg1{96Ppx4#ji=+SKZp;CloyDTO5FUU5bftX^zwV$9KNTB#;W3w1uM}Mn3VRi29FTra%vxA zvT^E(no|f|{5I9qFkE59SSy;PmdPSS#cVCys^swpH>(ly9*%`cE zw7|SNwgPRET6CzG0=m#guMSIX?PtS(@WY>1C$?530cxd{AI+VN#SFkueZTbnRQ#~d zRSaRh5Ie!$@`O9=ed`giqN1IH8>_t zb75Y8@>;afEM*kHygrDR!ZC>T9-ahqMV3W)wQ96|ylV9pe3%CuBRGZRs;kSN(Pha- zo_qC^A@;f$9dtpgmOorid>5leM^UI3O)kLp6SJ{ssXkxpfVblan+LUk7((0?=cV_H z(IT>1qz{(<@M9UT1kB`*dnnhia!FW08RfrMNR$R(qg-3+@OZxsG(Who0Nw98xS!Fs zD8f&~L-e^<5<0(ZGCW!*~3G)&ShcK z$TTN)jo;~tziKGW{#{c$x6|rT>W%0~F*W83_WkU2MdA!r&~_5|0rM>L!b<}lB0&W% zzfeMv}Kjkk0~g%nCVStI-P@VjfrKebMitU0F?baG*~z{ zT`W;Ytq@R`Pc8-mGBny?k{jk~vnH|mQ`B-lSC8f_nl~iLE>l<|qM@XjM6h;b1h*I? z*97>HIp>H8Kc=(y#pdjT)G4dsZ>114k7s%k5Mk&k9C*D}_SWyj!e|dIt@Jpp%d263 zvB^ES=GZpVwRv#V$VrfZDXw>&;!lsZDe_J2JSVC6S|)W)|yToIq~&6T~Py9Sy=Q{9gGDtxk8zyaM! zG|LWK(_ayPC+52y75skV@|hn8xG$kfHGfADJ>eyN`IA5y17qz=6Y6FG4|g-`mDFGD zdN&%VTkk$E-v1WZVL%4fO~+$G_kKblfSblXPF-WNmF*Y{%0SH|)2R*$hgno;JeVPD z-1+^dM`tTh1Wc^?noY%@xs49iIIxp?EF;=e5!(Z>sOWx`ja)D*OdloIgJIU@FLql^ z%1PkD;{?E>8Qe`KbP{+cH06rV5_pZp` z&|!>h_IZ7%9CsJ3#xikASNQ9l*6}MzTX|i; zRl5?m(u5J$CcONDC^F@TgkHX)9R5rl(YRb1k(LBoR)0o{o@0D$w zr}FdjSSNXcHF!t}cSrR$k83ZWiFM)Im-5H~q^3iqJgk~JTu*Zf5O@J0{w|#js2E>y z^A!P1WrRi28hh8HR>CpC^soNxg$m}(=u#W;`wu0_D4Lr^K{0&}?d#6CcZ(=(bvt zkecH1=>pX^=W)T$T~Kp=WK+{xr+EH9y52gh>S*g021MADq;z*lgLG^W|9dOKzih=#u{{n*d#ogkUKXlLOf;$Sc&0#OvqF3< z)Y6dK6mR`{Vqb!kEBNlW&{I7@WaE0_kjc6B{r84)3w_)n~1uRMsiFxl+ZV_3qBNKyvRJ z;WKbW{qZtWYjzSiM;jhr%Rl2+2znR_L*C;l^GJp99#F*@CYK8h2bkVD`8m#vmEH$$O9s!hX3%dQdb$_H1wEo_Z zd5|_XBb{*Q9ht=8UTAGj@kJyxN8<8WhQVSa&)etgQo4)8yW9oo@0tG&Z2Kw|`FEew zZ2xjm5V0=EFA_C(u!(bda7f2ozmdq9!UqLc3*j^VUu)SQsanbt(o{va{2=`W6N;J! zw%so|r(Ha;-@j#QDnlJ?s7w575pIb$F!=*tfslH%~> zzNc>iM>G!H4`6tyN19zMv#4p6P^RDzr5yLx05Ge+VU(!J49l;iFK#R2_fA>d8@kFf z^Iu}9R;sSq^pQ^Lk><3Hg@483x!a=ig@kaEH{MXiz&u@4H8y=eeFMh@H^CXcu07?d zz<}-1MfA>bevbt;YqMCaO$DEJW$@}IpPq*o(GVK(v&5wKgZRfTSc&(C9r*W_mdwZ; zlkSjCfyCFb=q;vXfP})nHEtsd36*N|rS}opfD8PNcHYuUImrXu?tu@$- zgFy{0rrZZZUXys;{R7qiN7ISQ4)mLEsu1&rFyW+8;9$3xC5&N1K;IdtGAfvm*pul$ z2a#riqU=1k-a=#C?I&`E7+Z#LDO~4rH#Fv>Z&7{{28n;}U>{GjM9IlSHa@CZPrYJ9 zZScgby_0Lsvt{c4lznno|KLz=Fp<~1SkrYIwY&gvAEPy;LL`iA4)~#|$Gsvg50vZC zZ-BD;eFM>&m%RD17arsA=}O2D6W06w!8yHc|A*nj{me}P3kLYR@b(m96x?K9lEE>z z;C`a!xWJ9`cQ06u<9R8rk2|-ZsJ^6SN?_G5Hg8m9_urK%cu2Cle{^DV^NuapoP7<2QGjQOTvuAki){U3rx4-p7v>zf=)8Q>q9#{{)6aKI}i z(Ki=Ug1Yk8roWBlpk%p3o3`gzY1d81<^J%JH_RGij}bD8$bKsROX8Gaqu)Hd-*mPb z*{Nx)9i-gW1KIeA8q_27iTjERyx{F$n5;Y;6va2b?1x+*;Q?}?qv+J}v(Aq*v-vED z__?+p{W=mz=i3&tX1_$M zjf6z^c~IV|{_AEiW-f+A&-JRsvk^gY#uv;@Bs4osyR|>`>u;4_b>#)F{kGCUF_?&f zeEULSW{w0s_rWM^TEqMt)tvo21c-OfebVezyqHyzD=pxbz z7DtEC?@eaCfj>9+{jsfOWwm&UeUR-NKWo#J2Z`NU`J%TCwFJwBN?c-J*fPP2_9|G5`Vkm+^V1&5tj<_6%F!A!9Ek$|XvSWkW1y|z?F1+ep-#U8 zXRv_;RXZgrTO&AuuFo+@y(?nNz&km%UN9FWjt=Kt5*_sQCejkz7r2o29466Y&RWY& z#Gfg3?Pbd3lyIy`AWStwMS7QwsIAPhK$sKX^l@h zwG=}?B=8B96pkf^u9>HFA@36?CPdiyE(PW(`wMfxu|xzhe^7ud(JDb>N8UXK1T)!O zEA@Bvqar5AkNczb+Kh#Xb3bWepUq&jqQ$A7?+@1S3 zN=Qg@I~!=+f)`fbehwGRweD#2`{1|3H94Rlp{8f^EK!5=Z~RTq?&SzCgRH!OK=M3r@=Z6M zuRV#{(-_egfqpEWZBQmcK#)`x_*vC&LYjh20#nA#cl|Eq`4`+t!F_^E{so;<4Qg|n zRPG;@wS0;@9O-3Kc1uba5({gYM=V2tK?|P3ysr8B_tJ0cep%nm&E!NLxS z3_e}IfGk&gvYtb>`74yS7fB&uV%w8j1;*CvRm<+tiV|_O@wxfj4!aans{gKn+(_IznV>oQtx*NkFX&HjnjIFYjIqJ9U#LPlCIq_qWBo>+f32Xc5OVfSo8{GJc!hVVz0Cc*)oGwRp_UF}w`{R6| z3lq#`QqLRoN|ofXlN$N;hrI%iI5uiA``6X6!_PNaOj{7_pIE^E9f0-g<()2bfH?oKvb1?i|KGszLJzGn*^~oUJrljv&)46FU^Fg_)j7=BhEQk)q@{MZ!b496(%WG8)PH{(cjzvz~` zoE{T0-M6FP?&hEJrL_5ae@caB=jPdB82SX;Pn)b2HxN*yqvs!}234KdTGVpCv#CX1 zW-nyBqU5FP!4qqApCM)PKkHlcgQ8{z*msw2!k{YgkPfkZ#wv!sF@;5f47x(nYEk@W z$Y4GCCBGg;esCqfHY5<7B87$B-+~KjyAWu^SkA_xU^~Qs!5qZ+8a%)`4_BT+E+H1Y zF2#XMu226Nr@(B3sX`O_I|{etN2OmjT6;VMv@ixYD!d`cRXm&mp)~AKhnT)bDD9{+ zya^+8KGfOma-WPiEYs@83-6R4ZIAFELFVD^j-bPA`n)TqFB+CzDxSO*^=nH;zH5W1 z(gv?=;Fqb2M(==fQF2^4wb6efyJ`>6%O3TirES1T8ZG2>%}~}| zeGck&TDtg7Jbf_MmdiF-qCDmfnGkno2p9Yi@Mbx}9JLn1nWm+3@2y-WAFkoWf5IyY zf+D~^TJECBy@++%V@3VeyjTiiq=+K9rGxLlZ!F4V-9-YKl&TeKKzHmCS$t2FY`YvP zAPYidiYo5I6|;Mu$g1k9hG{>YFL6i^T*)V8qW9%pB~TukhJmjX;DsU4AlsmsA^8&8 zHWZ-)vFTv3@MHhDS~I)UM*2i86$1ZD+Y&r%d-NqSs%g^qHG;e)lE&Rh3Ui-SG~F*_5nEJGial3Tjbnx} zZltJ;d!>_FGnYsn0fn3b7&8|3JG zQe6ZNXG~fu%@TWGK}u@u1jy}J%SJ9;pYb^xUhc)z-~SGC0YFxAy~TVXzzue(U;^o2 z925RvER#aca9(#16rN2(%*-)*A_4#eEV!4g?NohJkcuKac7~lk=wpD+7i+{a#vhEH zHGlowewfvhpF^tcC148ZQP^lYw+szFyv7{-%7bAH8~O*4$l69N@F++dWk8wG&2)C; zYPI$8jm9=76{Z~6Sy)gmVil66xF6@67^TOa%9}hygiKWlvoW!n@KSG8f-2#fllzeA z7DlQyZ|+UJPNF4WA3$ze8KKiu1zY2xqA&@53TTrEdF+1XVKB~`Fn4ix+(Aat*0t;K z5rC3SwGq!`_YF!QPE_yYA6^AfJ%q0(Rd77kp`G8M?T2&f+FV_haS`_zlJn-nV=*Kx364h*3HTAHEnwOZX$}5c772riZx82ux&OD;WB_X@IF9unIA8NI zh`gHxCO`6ztg&$bs*C%G0e+zz41yLw1C~n?l-1d!Q;Byi)M*JfE6*xd^v{yV2~Ma4}$xcB+X9amZ41Br5$|a{6nJ?!xQbT&HG%L?l<@jA*+1T ze!ceY9zQGDuL)?98>D1Idu7Ls^t6#X=8#`SFBe(k1g}mcBuY~E8ZK7hi<7dxAq3!Y zZ>e@qKdH0K#+K40q9Zsr2xkRrmn{w0pnx1hl0~VjKz~eW!jO{HW7mKghg8DUT)oOF z(js(~6NR4rV^&ZJ1#%}!1Vfa5-3YgBgkQ#wbXAkf{-qD#8qzy1RIn3r=n%u$&Q6y= zjm;qSLL?O1iVgM#8Kffh$u0MK#{;9lS}#RB-2{m7XGg67Qz`05!tfAlvyyq_oTfbz zep3E~Mm4oiQoer=j6^cSBV@olP9(RQj0)fX&~@l}`}Oj##sXljB#oQbs0~!X2zapg zRzGBjU@Q#T67z;R5k{3gD)CX-ZRL|-%392rKOcS;0^f8txqh}^5gr0RRinmcvzXBf z8q&Y(i2bDT>GPi~U2;ijEo||lRV_k#7y@K;6&3F97-l+ec;@a?!p5}46WEiKGZwJjFu3)tVkkBaOXpw~| z2!mQtz#0UdTlr%an7hisLjosb)v6L-YLp}pFy~_1Jd>?B*X>D(6O3I>1%IXs{FmXHp<;yA2poxgVra{GB=GP?CU*)mj|^c?d5UPlDOafEI$hJpB2r+ z>$V!$ZA?j~j}(f}M_&0L_2B>z8~X36{G^Hbg1p8S4 zx1qj|lIDYoAUZ1@o|Q^YC9X%)n}pPc*O0GR6UrHId0hN1)!+P0mSA-o2|VmjpbSD1 zsnP;kgP0Z_S6<%MMuE9!oIsxZyS>vco;YNY!d$3lLO~;CycvLDDj~@t{Oi2=KgvMf zMqpkH4*KF>MLh1_N8+kU^2K7NLfO076Jr6}->ifRsK&N)qWt@iP#9F+ zbh?!Vo0+2_bCJLp&P!Ebd1P!xMYxA|YIC79{FUq7p0mba3e^rVR+u8NU)L3V<+e98qyNl>7|KX9gFu156p4m$4|+&d~1j)uE^4hq~Z}t7^=#eZ+}>_&zK(h<$ygq1WwL{T*r#NN>ji+`~xI; z*F=FaD>9>SYVYT*yZW2VI+Eo=ZLx6MfEhFQ2{+>8+>6$UhkW%t#hx%UtXykJ&@F3# z;=4}jZOvU(tQ=}^9~iHRW*JQwhx@v_{WgN}dORJbMTNNmb z^nz~$!@#{#^viNv59_@G>|48!8xL6SA5VPk(1mZv_v3?(TJMjYYy^Ekx<0pCED}!z z35tYpY`a*%zlt9RYR{47nWc33$rbc2qAxG+VOSj0(AWCe4MHmVT^j$XV^ASz*o<0mD01gZE>dis znnC-Ap1Z-bTU>fVn{RB3Mur#cO5=C6{8nPV+ZZ$y^O2WFY^OZa{*^QXgm}T1arl{V z8#Y*>Gq;+GfhzA2f*8864%TwJwjn)MqhFISi0%qQ+C?c=or!6m*%7zDJmY=K1KQkt6-@|576s0-aR*r1|g!@7y+VNxd{96Nz8C1RVapP zzja9KchR<=nym<#T5%sW!)Cg}I-^c8OyzO+drHR&)iBSMz*FzUq7^&M;&aQzR#eR# zg^2FdDdD9k;bMgEN_1V9T&^bMS?98ZJo!c2`<1YAlXbtQ>IRyRWjX#80sM=<0*-!P z@r~JpQNJDFM|#fiLE2?RUos;4y!nGY`M`jBG#LlrM)r$JoR9NkprGfBzHf_F1nxUW ziQT5j?^wLMk#EhG5R|l?rFu!(cxvz7mXj~-Zs`x6~-7d-mNwsTAjKz<=wZ9v`@Q%FccJ$>y`v9D&8QeVZL+%5dz^l8&gs6~M{U#jJrKX&P zy?2cbg?}t_PD;`f3Q4=Igk;8NR71y6H+Qol7xUPSa5m|o^fP81fzDX9YIgi~Dv?T= zrHgq3C6jk8zT>#%TCX4ojH4cY7>IaNEb?g4 zbTJ}u#z}PN7Hd+X(?`D4s>lQ{g$gIVnf!>e341qHEW9;8|J&+l8t{f|dpxqc+W%fL zAB;wH=#+h*;e8gNUvn`kva)=?d`&36`AgMtF)D>FkLWE>e*-V}tTut(`RQaKIUr~c zD_?Y%Xd8LLbbm4LMV;|_2BEO$x?y$g{Hh%8i_iYa^Q9)PK9ayjD;oc&^sEOJZMfGQ z=q7L88vmn{rhQm2F={=(0>`I~#>FFHSL}1MK z1}$$NTNHwJ6fk>JLmiN-^gwo>>bgC&xh5?Q$v+FQ?mHI7!P8I-DvO@5584c}?yLND ztf#=3|J|i~{wtwKlYvQJJUou+bi!g#`?!6KjN;UT5=Pf_%wo!FfQuSw3>7Ok8Ye0t zpZ+7RSfV5#s-|uY?xSy(+T$ba`6@^2ympiHWo8j!bC+k6%#@$t{TWErnbsMq8vE_T zDQ*pSRP zQrJmnxo%TrRwDGx&mTW(inBZhSsFI6Ja5)Q1eTToCvMZGy`ppkA)#gEhG<0?i$I|B zdllY}Ngh+U9k!aWVJ+DfOkwo=cgP%EuT0F}Bw;e=54$74 zC6Z=@7(~Y53#AZHNuhU=Tfb(P>j*8sWo$&9i;2aTywRr#9hM{Zgh9}F3x2ljtili{ z?Q_ujlT~2lWZeIi@?;Np5xwsB1bGCJY3ix-Sr9!MUwk1_JWjUm?^`$h=!3|ndjGz` zxqJ*PUSwm(QX?!08IF&bprge-mZRhgJUW)Mv~{gs9L$#*4*}7?dy|Ta;(^9*HBz-N z9^aH#$k_-d_eiB6lY#+2Id*0C$JsvP9Ke3eA;Y$HvdR1HnG9J|+3(tIC{Q#>OQm(Q ztu>_FBzLwz6O^*5)I;iYCWG1&EF*oYCsbsv?Y=CEsBOO_5nU6jy<}+UcHe^%oA$Wx zPX!z*q{pU4+21kaee39tG6#2o8vXVS%Z#vo4+JU#o| zE8s(a8OI1svo!8f7A#s_d+mRXEyyH#Z_9=2JdgB>0Z~Lb>7DkMoeZbwKUqQor5b&I zTZtzf;JvYcqt8o;;TIRgV9hxk?8-KKs)yXp7T(DT=}j}VIGsg3&xEI9;$jkFC%nLo za*y^fN|eHL3ahphMB}L&p$Q$KeV`{nr2N)e<`fU({At}{xWlssxmV|Hc@=*oW7BR| zr9LZaso5HUXG6)i8U`y{v8icW40eU~kpfCLlpG9!)4fYJTW3XMm5!fXmGkuqJA5&&uHc<4w`K_I&IK zytA(S0zQp_VLSq}@Pz z1!ES_xzHl1dPJ+UAu?S0_I-{BQ&VGA^;OM|r_T&n%!iDn_Ve1RReF=wzFi|M5t#EB zEKb4KWC9!bVdg~r36M@y@><7M&j9<-W=BUWykzPl=%kWsx)xxZYR& z#z+Ld&XPUfNbf8wIiWWC(g3od_|Mw^W;Dj#{*%|-ny9)63VuI%wt7f=lDj9#?DnyIR^gI6IRv$PYpL3 z|Kk;Rl3ECIqf3e22@%uWjC%gC4Sa*BMM+QSI6NP6K0`5&H3ad%#&+UukOR95TiL$B z*!WP*8w$Tma_NIgy%^&+?C&iACkE<5hus$aGK0-;8yOFWvZ^PTUE;C<9^l083Epdh zAzsfyJ5BZU>kVSGx;&@a6qqM<%uX}5`H)X^aho4D3ht1?(G%|^bV-u#CFLkGh^zM?J zj(F|N9VXkDpQD_~p(Vo{>hTNH!&ZrkNKkunP#t9rzgS7EmIFCZukbbX2ZBP_PG_wC z$$_yAHH_di6vS&H{#NJc#T7Uu6}ptb;?< z1_3WZ1&x-66Uo41gPykb0ps$oTBI3*CAknLJtw=MJ~q1jTAKxH99k7czAcj!){z>&@J#C_CLXL*#|2gYqJVuMNlHk?YBPh3JZfi?8H<;1&HK6$9Z`^YN# z!Nl~I)q79=c&i2+J?e8x^0 zSqLQaCg5r!+4!OAMUuy%p9+Co9K{P%M1v1t5ENTW+ z*1GPdIK#kkF1#Zl*#$T#bzvNyXl_`}r&V1gbX7uc=7-HI!+|Fpk9OX8ptAGwVn*y^ zAKkqPXB>j$CPy6I)(3kr_ia49o|U3vBl4~uOE$5}#Q;hWiRU4{{}t)(U@?m9TjwAtV<{mwmVRLi4Fq1yk`yzN|(rlt!F| z`WBdlx%cQEx)e7;6av-Ple9&M9KNnkdrH35-^&}zv_)sH0W(UcxWs{I87S)G$@9}m zca<|&9-30F*%Ksm`Gt{=L{p2i$dBKbdjW2$JL@VWYx?ULSkLbP%d?7btfB9{9*`goNY;Snf_ zdK}Oc*xUPOKcw?CsM!yIJ`MtZ%0WHWHN_t?{&??)4Ki2H&?e9r7rDh`sMJgQL;HRk zFS~$7-(yz*?LkM3l@#^@yht1vZdo}XoHnJJ}AbkPt= z>5!rcaETkAh0=dNq4df^AanZ8j%Usk$!%8O;pe#K#kM^!(#6~LB!&N#YZV_$6CFL- z17B@eYra3YvID^_k7xBR0fn`&sV|iCra3ocvG_(u-Z*B#VG{GlFFDo$wX3a0ItxD1 zi=qa(OWtDbbdt}Vbo+hfN~)m8DE!D+Y~uR7=5Uzd zjm=n}VKLc8BfG40JhRIkN9&JQrP`A+y2y{c)0y3_W=7Wf<5}1-)RKKYE_7iZalOLh z%Cr=HqkDe$gi{)wW6ks#^W~PVhc5XGd0khj_l3V@(rM(I@3ixnOebcjbKDwav180! zQbJdBe5JtI2X&>E<71)8G~0V|mwmHIcc1xn-U8|TA~%GzTAsCK3NymFe7!Y39!tY4 zNMaS}ono@% zUTikVBIg)wcYiAu&!}w(_VzDICWxTGKC7XTPpW?$@EzelU}oyQHD4(|Y`6No)6gB4FQG@fIYFmSbJXSQ0&zfg^ADQo>%UVR>FM)wh z3!Sl@k$?@;06K^Zm-nm)JW;z*+3K*b4zZ zh6ha0JqDTyVa0Zezz2}|V0yi27hm#LB*iwdh{TA(%n!D#Q^6=^#=(s8OpFGH3i*&w zZjuJkC#kqKjx|kj6FQqu!>r>BfgICs?cGi!_k{$Ev$A0E;+0;N9eUdMkE zU+_M&{Ay;mwok+!_A)K%qG54m>gP=cU=a2}vA17!S9w

qi#XpANI~$ya~(*6S6& zj7)BJhF6UaQ6@Kbr z^;w&kiYqvjvu!?;dQS_?6B9pekUh%;gI6aqUV1iuDz;U4dL^;s^A}58 zmjb%v<(}q?y{6^PvQtzGGZ`;ePB|a=?1Rpxf%_w$J*2VURx?*Jezvof9WBl`s(At) zWz0JDWA1SrzKHsc%%M}s#;SdP&45o1YCnjEiNe8%gXLn)LV(J*wM=52m%bz{AWSD?Op*S*|GX6g6|&R~ilU@n2Q2pVK)h#Y+R{x#=B@P=z+NC6sZ3wjs8 z9rj{(n|rmu4h7CCb;U^K*)@Nui*_(qhz~U-i{83-^Nd9>$Q%4<5M^p8%~NZd&nkYo z#4j_UQtxXvwzh0>)1d5QJ-W_+9CeSsJ(@&dr`6u!-P_8ngUFP0&jnNEV%^18KT>JC z9ZfZGkhD9B1G|_tZb%}*UmHd3URl~Br}gnQ`u^tGhV>lp=9@&|Py<~u$R7r(Wq{Rr zH0MDvjB>0xV*pnT%fO$ON`~3JyKCMp#Tmi?-NTP9OjwJ;y*z?xR?0Qz3lrbV@i@K3 z2<@>{2v;p8wI3%O`#wgY*G3xPV_JhiM@=Lu!x0eJ_)Ue3IPUHbYv~SFLNqV7OWkCy z<~4t;Ntymm)v)EfpU*dHkF&WW`C~FNvL&PSzZgpUGmou_HqPH1Cfo1>8(p}P$L#VQ zgs}>yvZn%V31%u%W$(Uu@LUnxy-hJ$TP_XRdwp=pn7l@7urguNmrEPeoniBI_-m9O zJRV}fc|_bbFh@W9NwzaCF1w}+5iAoBZ&Jv%Q-miK;c*U4UwQaKcb@Q=5|Q|6k3eq( zq~-yk1s(&Et26P&4F0L+rbO+hwsk(l>YA{8;xW?{xm*HZs>v-|@9boqjs4A#sZ>rH zcCr*3WJE4YmH3o2rwL@~X8guB@u{U!jo3(F??OZ#I!FdXXS0Y~Co;tsg8qrGjsYI< z&-Sds9zIvnnpKal$ZkX{LRRIIJ1w2|y6vuC$4^|IEqD+{Jo1v7rSg&@jkOYQj28c+ zJ}}TE_q(pH>d-vrk@X9kjZhZ4XM<4OkNm=+xRIJjlwfR;zabkL;gp>(5S*MIODh(tY8%hs%Qy@i=;+MMbKXXoN$7#(12XbS@>7Q1o9f)0rHFJ zGpUH*4!CY)997Lr8fa$RaUb%JB+rIsi${Dj<3e7U}UJPN>W(KLW`KZpU z*@SpzQH+yyKb}NL7zn%J$|})VB9Cw3-LZyAm>Q4x_DO372GR zQn8~_OyI@#p$smJpK(76yD(GnRmLPF93Ie8rJpvC-X&ru3|sVXUotvlQE``u=z@fy zOtPwVXdcrxU9GgcbROfJ0fBdk433Lf?(KXIb8s#4-qM+p=&E(yjLB2lI2A+*bOfnS zv|>g4ONichCq$EhRilc1Mk%;%73=%h9y2yE2g9>`H}8EEyqagZmF+x8((>NVQM5AP zVb>{T5TlH5%TVdc$;lmVCF_567nSWFr%Etf*OQ5?eu3mim=YPz@37-Pfd3|BKN0&3 zc%>?INuS7ImFpt)ojS(0z4v5LaOyKx3a8_kB)J=LZJpf0TdK3mKEqe?<9J72l{Y59 z^&2CB?0_@-Hd3A@C5)Lnn9;XV*w4E9ZcP)FU0<=Hxn8Y#hPnJ`6wzs4&l`t4j$AoDRt|W#leLlP zV(E!Eh;z7;g+LLCsf^jvsN~%7@9ux~MYcsRH2V)QnOjB+LHuv@PNn(t1@9OIv-s=U zP98rTOwXkuVJ~sgcV6g|qFs9qw=PeL62He7{X$-z-)4OJ{>_3tu24ko1ypMDmtSjD zRf%i+G~SuV-@Ossp`Q50zk7eXIm*qdV)(gMq4c!W4e$+wIa!yDZza<|pQiHv#?1Nw zjLLYtE{kmFE|u4Hkcuz4SyT`Tqd}2vz!4a$$S81b7${W|uql4C%|7({(xCp3XPa`oHyA#tp+{%+Fjc%~fgTL`4VzA0a$3&@C0 z!fshpsS>D3gg9Eti>0aaxzxI*(RY%XsFVnr;U=_EMR)ByJC02LNo9laE1nFhg zix=MfFh@oLCkhOcdpHvGZ$nBQ%A{Uy@wP3h1R=5ubxjRDm!As?bk&* zt&FlWT&r(8Z~cDlx=R-1ZRP>rztmoH8Y~$KjuEF<5{@Ha| zuXmKA-5^vhof_Qtv+vU4{t7pOt1CJ{cb3>rCuUEr&3dLvF?ro1V+8WXM!(mYxnxOG zrN3T}WHO`#8I|2AXmizlCv%6;u8~l(V>hl}?gVOXGRqf%#HmCRfkRS0oGDPJ5fTfc zu4{+7DX7$$4kfYJ9O$8iDxxMAj~$=bBr>(A+-irELewktj-q$6*Vos(JsQ8o$H%iE zDdW1|?MjR6pQm7H^T)WK^~4k@WoIRUt{wp~HxkKh7{FAMf~VC*Iz2WH^jbI9=K63Brw=YZBi=e)~0|tMs~*O zNF2C{|<|-McwW6WD^| zWtAy%_IDE-mEI?QR`w0rYwpJ6nhP_=3$GTpuX^qay`tMsBiCEq^%c*Vx5bVY1$NI+pzW%nLN|eRjl1GK6XQ*SUQ(6|0JtmIXGQ zqBFzAL)9j8Bld~Ow{r9btnI4@0R9rWzh8Mi4po*a>UiLmhN3Q;2}N{t311FNPyR9p@&le$R7z<>yA#(tcO!&->`A=q+JPYo(Ww<8 zpgD47tC`-7>l_$gD<8UK4(@&azIpQ6*KXgF-ECBI)abNQN$5gU1m#3D!6Pg8DK&SA z8LoT(G(buH(3>B zo)p&g-8{8uJpso`E;f6UE+qo`f{3!4WHgAksNvGhv=EQD+&tt*>X;s@OPO93xhv`W z3XZneEb#auStJV`6>}!CZUe;d<3N-cTw>=#9y5prknH7H!QXAcHrlU&o1X3tbmNVE zk#RWz+BZ3!B#qChXeHz803Jw)gO3U9db}K)V1d{vb%NBH+zUVwi>2wlsrK8zNi}vr z2$rz5jNkwKF?tInkyY&(d@#f=q#JStNWXyA>R)eJ{m&Qp*6{c(4tIOSO(5l9*A*ZS zT}8V#=l2a6)z>C})Lu*oQqUgPFdR`rrlJr&3sb7UboR~izMAXao}!dl(naJ$M~*T4 znIO;7K+Udrj-1(~85uV8LRt9$SY6YS4&7*ndDP1yT`ga?zQ3*Pz60P4Lx90j@H)1H z=w&ez{~5fF#PDAAN7T`9_x%b<0R-G%;qN5<98+G9q!g4PdkP%5%pnPkAG^F);8 zwt6i&n>7G{605t$ypXCWNg%8_SwVcfw|N^Vp|;0f<8&wqOit6GpN_qkCt8UU1dIOr z4UqrD0QD36(?Q!4Qyg5(-xUI%r_@oeq9G3~=IM(F{fEw_ks6Chof41pmMS+a$nW2JzB3q0_0$0M@85+WA%iO=JKzEb^{;)4qh>--tI z?`7lMnR+Dy=Y1kilp0Ki7wYUzDjxxG)IJ-)GgoNOLmXc@ zWi2`my|X1_nxm0_sT}nV{yjtb-`gaZ5^i{9Z$N+_-uu`o+i5(R`o^~P_KU`1ilVFt zz)2keJoPW8L2;va(xe$(*UUm(YQ<<6)7d)!)=&w|eB_ChtAEpp_oKO4(1BOz z6G4^SNxuoJnujs1yz30R`?(38>gbbz>#_4DZI}XUB-i7yYg>B1Q(AVm9ohYC--B)4 z1Hf_F4%5ip`aYjL^PvLG9H|Dy(@fjbZxY{!v#U-N*80bX4Ix!*qImUPHh`n|Sa;EG1+oE6fX(k^{K>r|Q^-EY=PeZ$dkHE(qUw3bCe7sH$^f|+4S4I{u4 zk-yZOKd>?}IrY)f(F+6L(aqaHBjTyYo0>DrK3ZIE`OcTXlC1`k5-=Exg?MKY-+md6b@$_A)GftZm^ z@6rS2V2NT@rG3T+p|=!(P227!V+2JAYdB5a4mjc-W*9ao7;LSq+q8RYk75*m?vqFk77RT+H0R zD4qYDxqoF>{6Wyz_T%3QA5#r|@87?^k$LRqO8BpC9u(104-CIhNIb5`i<`P(F7qcp zXq6nRYig!!ieF8vfrOZz&tiOeJ+mL+WjlbfbQvu!IeBMd1GsQjIP}nwRB>gExNlxS zdnBcyta2(Us;(tYcbMRg3&K4B$CM32$`^C5(0xm3gbcHNTkr}uOc5)$Sv|hgnbPk?{33VRajcTB67Uul|28!smdeGjI89>3-`?nTw8w;+$% zvt+gb2;u*X!~bK`00Cl~!Z47aCc3bgfGx|*>nzO7ucNxKXVI!IP-9a$Ipf$d``!`T%b9UhFj*!ioRAgoLCKUyZX3zh(AP@(NLnnZecU@?So>ieA=nymP^1C;oq}I=fVSn^_v)+6<&pEA~wLX=6J&2 z9m3uLoW`u0lF;AB-H?WB;4b0j`F{o}8UbD=pxOXan;NUC#q*jUF7fa2WFf9Xs~N5i zzuRffx8&pn=-y<4tHp~kMwe*LDEw|9y0H06eSp64QOI@6jZ)OJS61kM2j^dRk^fo< zjQHZ-fIDw|TpY#ug*}X#K1D17ETp{{RxG;nv zzVCkb-rs*M)>)|dGVjcuJ$paTv-bp4Rae(@J32Ws1jd4U%R+<@BaM}%~NHeBJrOMG6P$Y ze(tAX1=^|*TJH({0U$22Fo(NkKfapT_BK!Aw^<)~3isHm>%Q@g#!qCcHuDT`iXHXW ztjx2@&wtBF<^RvB`z?e@s;LdzH(KjgAcABg?VnFiy*kf9xL`+tK~7@9<95`{@$f$~@>Y9!Iw<7eWRKQkgF9S4(@Wd{!<>6h+@6e&+0C9|Zn#`eF<*lI6Dhr}5s zUZ=*DL%yG`x>lF)`t{c_N=nM2qN4sH+O5=O2fQQgAOGrt|GU0RxhKBxPicAmeI zga0uJ96r=w^KaYf!5PJWZC3OT_wGmkqLqMI(hBsapu)qyOXsZik01G~rBQi78jDfE zUFISiAOsFg0a|W;?{gSCD{JX+rouGPk!>!9b5?`dU-cRwbw3cpnu-0F(fxO**CJcL zzr6MeG#Ed5uQCPrQ7bDOo6W;SQVu-}MDV5$h{c)15oN^0|Jiu|9gvYd6w-43X+F6Y z-yE0PLytD<6FnF|RX?PW0~O>$r)r=y!~0F^=7ec<72sEsxJ`rk`1mFVQbaaKwG6Ao zJPvEc$;rtN!M{%evw{oCv;X`1w+=~y0^T7QUOg5V%4skH4E4F(Pu&D6>E~KbV>;%O zt}SQ!kLw(t<`kH63?ytH8V$FBrmSo8 z#w9$6niwBcbWx18qMF(D;%8l9#0b;PW|o+wC&HQp5e2=Rj1*%|Il~z#)U{_pYO)2Ox*c zVs`nKl|_2QHE0w*dp+Z|+AyXcsX1M#d$-|}tG9%unfcZik5wN9-EY(C8T3#sWFq5D zhg>Xb@1|vSKao?!Qf}W_=_3 z*PAte-LLuIPd^xeVx*Qrb!rE`gdPAYSr8lzmDj=WqD#oQHYEiGYEYaJkNH65j6G!_ zl6KQhJGRuo4ixj(g{w^XemB*#dzatGnI`V^akkv5s;c7F?+gEZ^)np-1c&^NES~UM znruAQswCnCi6ur=P4m*2U`!SRM?P*h#BP@5j0OE^<3_|?~I&-ePeR+MbS zK8XFk7*Np@<*k9cWe$tm`EF02j(N)HH}47L-;?GnfGh|Z{`D61sgfU2SUU&7(IT`M zQIMa%?1w=EnYIOi=?X3Rf$NBpK#)ye9D}8$<;#SGyF|a&JjNMSk3%bGH90Zy!%>^E z`V)*_{~#ts_b{v$9#7cw^A71|Njgvefo4hHBgMbmW%qL#em@9&KV*g3xq<8sFy`K; zZ}myExF#nj9X6v={(9s3U$55%-yhhZaAT)I*>7qeZia-os{G5tn_^oV-S(F;B#!;A z##EP{04jW^1+N-}kmm~Cz-uI_du&#{J>RtR!5+IsAMGYeqf=v>5N+GP7d={q?3tAnn#3c{mh*Bd*u0-(2d=>Jv;(+77F^Qzk1u!7NSaf+1qP}@mI1(^ z)NA?q^XFcmZ6Vt8WJrY_3@bcs^GC}Te9kZ!=fvnY0A0I|e6SQP@#cF;i@~hd(F|Vy z=Ht0j&5)@I?%#&;w^?^4_G=5kxUZ}n)tc#ad45uUH0yyuz1~G?`0Di3Jwf<+zURr; zw^GwPbeJ)cg9M=sNV;lj^f}3nVY}xWy*9XDC%R1A)I{mESO0hQ|7r|F?@$Ls_aB^m z<@C9!B7X8j5t(RV$+qV4fZ$R@DT;-Nf-qZ$b3c^oa{*<}?G;*2y# zr`-Hw>@mW5dGxjf#kda$t&;7I_^7&a%>oqAEYNV(0XBd~!P`ruFAS=fls?ENw}4`o zuPhQLF9^6e8+W8TCf%qH0f)uj+v2{{hH<7Mw%5z(2uf>{&io)A;2b_3G+qT0R4odC z82Oy`@1U?re?a2OlQCNwhG%AfDAUO^cDj}sjW-t9b3I=23z{kuBzVENRmOM|5Ue~V zET28kWW2!PmPl*j>F*apVZ8tI4&VQtO22N>_3LCFI$rs{SWv8-ifS>aRfk{MF6S zCubcb{>-8*4m7&>s9k_)(*xpQW_wSW%OMa!Uor1o4%#^-r?bi%(;E=ZYfGcnT?%2(g-V^oPlQrX?PzOFkI7tQ=8}1IOzZ*jp1l*n}|d5za2* zY`UM-vUSfzCo|L)*%_3ltk~|_3vK#d!{b6>rBP+~ZLc~8pcNNBzjw^;eI$v-|6`7s z<~7pa=a@W_PQVJ(w;W=p^Oc6-5J?hTMnSDZ?VVOWS_pN)-oZm?L+3Ayi?W)lmiZv{tT^tb(45?rO_=sz?%{B4&YbnL}}?&`-k?6BBXj^Amsbp_uYW zD6BYn=gGmI@jJbuEAggTX=$x)A6oDS6_%@Hb3M5HcFiTu5**M2WOTTYe*{z7?u4gS ztFs>z88l7N+@Z-zA@7+}YsG7%Nm*o3W>i`Gn82fSKTEI@9skkyn=jljC_XVtf0h+j zuEge*MVC1bOsB@FlUb-bL8x>&y%PyfaqI3ah+$XMJ?IaAW4A*Tl#??ue-dWzO%|r5 z6QGD2kQRBKJ4Uyzi+tcQcLqoJI)cyw3|2=N1Lo~lu1x#Sby<;#Y|?6h_RWVs)mgA) zK^)=GoH39eIe7x1_ycc>M$1K9Ji9f;Nk5gF1bfoIBTK(oX`_qNA5u;e^r=*raD^&U z*nBcbjQke~ei|KwDMtK^m%y$-k_gt!V$k4EH1=Xubbk3Md!r9f1YdGFzyhUjbzI@5 ze_VMbyH@$bz`43>d$t2latap= zt}DaMNH(i5uR-xK27|~=Ne$Z!RD0PZqEccTA0C&Z&l&|~s6y2UQ1G2rSgLH=7*tc+ z>jLnEdwKoLi}Im8RljDh0ZR?&z7 z`@_3&Hxg!deU~sk>!kP9J22k;byIVFxTK9u(cP6zPt)L?7cs#$)71$o%)9TIq?uw0 zMfHL@wi0=I=zsgPf$J6)7SN<>wnyOZx6?dP;Dc`rl1twqsFM3RoA+C17SqiQk6zzy zl%$WBI*^6%;mrCE@hXd-EXR-)a$%}*z0$d0n%G`7qLtTCLK~nQ6BdcuWU;sE8f#rT zykcCX+Bhee#ZwbVraurLLMdpT-iLx#LYFH^b`%REH_y=t1vj6!Veu|}*;(jNB3$N{ zOGG{s$G}&CEi>TEVx@t2d`{^x(cKJYPCpJpNm|%Eve5ZX$HkMhmrLn^4ER63B=8;r z;`PqveDDLrko~skd|y%L+J*|ucwd|*g!#Y%DvU{>3bAjl*|{jjM4k)wR>}OwQe=_T zw0A`P7ihJY-DG%CMV+AKCsS4>mY>MgN`$E| z_(+%bVE9{*4)h81LPcL0p)PXxGSJ_8*}X`Ll5er__8QHDbGr7F;~CdEz87Axq$sH- zlINWrs4^(}6u-8MUkHtRZ)hqRcAX_k0ia{z5#*7-iOq2q!D|AwJco%=;~8$TD5>9J z?2I%LoPR95SBNhEs67GMr>|og{bV$zI!qq%16~P8jl61~$qn`(0U^c6ZJc^PTzAU( z9>%2dyqXkE1f_53iL#=z`b&XH5b0iHd;2p0##0qreqN19B3xVr{D@6{wFugMXOsOj z_oi7FuZ%mMTopaBhM#ZlFwEV||F|S0>A9e%q4E0WROA(|<@@`s)Z2MW(eG$4cS5os zL(o}LoL~Ata`(7uWjkibP*?{TKNF$vzWO;heI+EJXV$}fo?)b{-oia0YcEuZ&y52_ zx&E9rYwV~yD(SdK9U*=@c<&G?g`^>iQ8>NSwxkqOs0r%^)j-Vtb#zX-@87;zx0v?e zC5X5d3!y48Ylc6k`fj{_>wIKVh;EPchcOB+{O#FxOAD2KEArmqv>gWA_J~3a6PuKS zw?|DwT=hKg?F~nlcK@$Pa!-$qdWskgO0UptxS7reEyV8YocGcxuiH70N=jWTo_}lT zUH3lSH|+7){H1k)om9eH`O|PQW(1rpfmw7#=Zc8#)371){bY9jy`!dcU)ph<_WK3; zTCOha7Lp7E?};k>T)QDkOe#Qi-H0R>7p-YsrEUFRq6CMb?_zXm;Zne#g0|Avy{C0jxJT~PB;r9tj+ z1aUrY1{$>3Ku721TRq3SB5kCxXgg^aASUb(&X~B|l}PH}CAUnG);JM+HEZZ=TVll%;&urdYR?i%qT~fwhJaE=n)jg z)p{)~1QVY;pZ&3nzuq#re|wZCmLg3=scS7aLf2kRUW0fokRR-rVawdHGq07&l_Ec+ z3B!qw`x)~dxS*WquA$PuW-avdiLkLSwQI4PrA6$9Kn7n6%`7yjEfwvo9oI^~kgp)s zF_$hVll!^{;!(|cvcnJTCA}cUMn!uU>)5%n+xeWjO1q~l%1V(CRCA`1tL-B2owc3I z!yzxaKyi?rmya8GBwX4Xg4l$BsHIypR+ylQjDg(IuKoLASq}}*BKIe@41baMCP3Pk zdtbQ6xe(M92v|=H6^pE?+gnJL*+MFyvR6M!e8|g)Xd2r@C1XQ~gh!i>X01*e05p_& z+EO+%FB+qYS}%)K2g6y^PSMqLOKBi5fT= zOCa6_e8sGx0_41(h@k%NPsiUZjTAI*Ew|t5AwNU-_>5_XDnE3IJ-!Y3TF`T7IqsnQ zxi_vB4#hVG>4YM6=!U04kDu#}MtXqwD#eShSG5+`ue0qFTjNljAv8Jr#2ccW#>)fCHz#XPxintWKtlCaKWFDn)_Qe6NY@Pie&m-G)`jmC|GHe$rC{skt5L?o)CA=ht zRiXylAt5G4qs;E-eoO~FxyMa%`vlWD4=D`Gyk^e3!6CgGU(~yFCjg7oY8kZ&P+k`& zR4-{+w!vPfBZo|=VqY@8#VP0#`jy2U@FG9%Gy$2cm>6-?Pk^JzA4$4U&7!!VD04~j zAl&s8l_=j5GhN2~XK<>Rw4*DC!cyg}Aj2Wt`lgkg*IV@8Leh>u&ja?q>>Om&1WX6d zDGX$Ax=6)OB#Tuay+(wH$^ErIpRF(7dU>r!wV6ZaW7OTB(1JUGUnHeKVnZ2?RMXlR zjQ#ft%(EbEKMd_w#Mo)QhEJw43@Gm&AGpYSC2-f*P%0pPZzu%F1*q`Puj6Q$KKkmv z=IaT4_wLGbsGHw9>juBsdoMQ_h4uX<^8sU6;4!v3~lgIqO^FzN`StOx_9b@`WJ0RCH6av$$-Uwk5X_XJ-|syOdlkuA94LKXaa!@a7~&X5R6hm}fF3JW`3?wb z1nVC52`M6TMVvaz%geo>gTwM}4=w)G8w|uU(7{omC_cc4Cvt4JTzTge6ktz^<%@U& zP;W)Fan$nnq6}$4tY;f_Hs%glQa$f98J&J;-xo*x|1?r)>y<6YoLZEsjx6jjf{K71 z&?W=V7^(4${Updh=a? zr9Bvx5z&G*%wJ2K16qugA&?%39DGaP40-&={oFtyCMGTiD9n{aWat|*e(U&T#}7B( zJ5wgJ|6WS4nw9j>ajG<$Xa(z+2Tg}l1^to>1{+|5tQ?d-?Q7>-5<)yJl+v%gUAhkB ztjjl(Tz(pn_} zqkJ!-$^0}^5JAOA>Y=Ib$b8aRph-d(@m^B1s0=CKT?Z&3J!vFJdU;^ooPBDx% zu)}V#`W}EjARiEH9%_(kK;{ml-4>km>~UIaB6BCox!I!1h%xEkL!yCS!@zs>K8_`c zGb~jw?xMS+%v=2q;gd%p~YRvLLrbs_6d*0Y-$|&)ww|MOBmJ&JRbl+&zzKouCe=phZjrHr)`SV#u7^;I5 zq}Ht%YPfq?}hq~*?G&?<>Ip-%D)Xlk^u_y>Z_KGmw)6WnE%A+(dI-bSlw;t z8b5MncDB4`vEEHUh-I&P0Le+MUU9QbqU*i?dt5`CsP&bQPXO6`GFGT_$Me|!eP3T3 zh$SB1P)cM93JSVoKQCFgmJz-T*pBVPq!r(PeYszM=@lRePL2Nle#e>T&6rXd{^{p%=KKuXok`HbP?#gGjCYk=5( z=#6E3S6r+-R;%R-MxU@W?_AK060u?`brN1ra_CH>=|(2^aom5S;d(zCg1#R z36)SuS;ZFb3zHRYz<%cy77{h%cSppK{yy~S{ERz--w&nBc-W&;B_d4h^akQrL##q( zoJW+5yHvs{{b$ig_2>ZX7~ObrI87vZJ0AL&=uO72k;X{gAr#yw?J$8~=4%u@8Oc`S zHXOwMb)+k%W)2I2mPo``5gCc~x{{;?SSMhUCKI;YH>QGv(ev{2 zU(W6r3{uD>m^V#YvMJeWm{3LKjQhlC7=BlUdO`Pq=;Uyz+vB#QqvI5QJL9j-ps(CYJxCCf=~y%4 z!lUmE@M|^K3()5&sTxE(3ZKVI>Hp0q{4sD@1_bvSu-S5TY4#Ctayl_N>@S;6*9)?b z38s`%g!vvi!C8I*Isf|Umy6k{kJPjfOvxl5pcTWj5hH^a3(x??7)stD=so+s+cmwG zIVm^9Q&r|K$fFyL{FZiphTH%2k)7d^fN-J1#Rc|4=LP|fr51&iegSw$uq2=~Za(y- ziUa1wW3qaeC4~uSMRMrO8QHoRSJicO6GZ|1{t^y{i!KYQ|Bep+2VD0PmzoF1^BWL8 zDu9^BEQ|$6=K260^c488JrDu0+zJ2`3z=AUVd1THAsV0sTuJOt5>W6`YNX^aY;M$l z`+}>)CGMFJQcnxeFxCiN4EkJ-*q?)(YGy^y@Bjy~Yic6_X02!7vB|d|j$9I(e1+h9 zfy_l7;VKY`yBxi}E#MPK&KZo4MJ)atw7+-k^q}X_CEdQ2>LYjfn{1i=S-uK#1s_})b(kr)VHv|tRA8l!ET^Yh)Yg%{1Bu9UJ=iZFDnhRuSC|T*~bQ)F93osEDl=8Xy&up)1yycfhv47Q(Wsra;YX z=yOr^#pmk6tN}E1JpoW!M9u2)?h=~eG1?xG3mYz~5mcM+I6V5-$(~h2%Ajw=qi+P^ z8RoMT$qwHm7M#Vk)XM`Fr*p8=dA}iT*Dgh6DLA}meoJP%CjhHNVgm;H$ZcOa&D&?K!p(LDX(uFG6u~0?=Q2M02?-{@Cw?uNVJ>rGL)N-={=O zFRJPi>LQDC1cJ(l65Z?NAK4r)Nksthq*)gM7;A|3H-3s3Atg_$5pX}I(1mP@mUR1K z))azF45PxZ$Gqw@ykzim!(;WXUiv@I9{_~sozKr;Hh|`aQ!i2nzDXaPBTk77!}J&D zdv9Bu;rwSqcjuXFZD&9uFx%w3NR*OKk}}20%d=F%&bbHMb8|%T-d7h#@0y!ui4r+D zI1UXz4GkIXMKc)QkR4CpF%O%j79OEza3Lse0>N&=ou^@U@5aZ*E^H5O(YceRuMysO zz@^qGXuBG;bLH=kvUC9snAF5GsT5Mrpn~uIX9_LsnuReSz~z~nahWda=L3SPml_3{ z36`yXwxG(v%m*ZxE6x7>eY`_ceE$51Clx_?dQErR5~SlO3r0J%E&;7V zEb=w?x>nHgU0 zS3(PA`JF&{-TPS;ou{-e6UHy&;zm}k7g7t(Cv=JWlf$h<6=1=;=~AwbLDYNe5(Kgz z#i*Qzq;PDuw?>y?{~9x-+}Hs!N$fycnA;_2fE6||2SgE8((~84m>_wfQHWxIh+GR1 zX|`6-kmR0i-Pbq#45L8B$biraJFq=&@oqX_5oob=qJCBkSQfE8K=b!3-;gu5w1SQr zxPN<>ws^f$3Q{%T`3kyD1N3D=z(LuNnl)Tqo>BS{+Qfb*GFuHg%PlD2aSH4;!!=n1 z6Kff$;DPg)sHYidPQC)#^f?Xx8FU`_B6=YpZSJ?Sm#2^;Lsbf`EL~rJ*1zTg7h_^edA>i>D3S&C{O7 z`YCx(y7EQ7QyW{`Y8i4hLX)o`PP*tRy=Jpn({fRi5Ktkua;@Zvs0!4rs15likz*Wk zRwY+QcW05@A94zQNC_dqj(=V?Zj4QT3^Y;1uyM1ePft|*2vBW4;zuN-mC+GN03h9; zVRg*F=i<2#5Wq&JVeRt9r+_TY%oaw1l3Hi9x#U&&TARwf+ho4Y3t*0Dsi^ZI^j~uD z9mrnl|MQdpf8)&PmosJ;^I;|1mWq2%l{A9l=9Xagmot!Se@xpEcMelZH@kjogSvej zoIGZ-Ot66qpbNWW5VG^I@q2MY#2V@*Dua3&3mqyC>^|ImJr4zg2LeC#&$$U!OcnTw zMk=R|+}+swz{YumBZwbgTXkD}!ICIFy7Me3WEQx1>&_koyQ$jy4REfIMy55Z2z<^S zkfMyhvOseIsY`%Jla)wEAmb~uQ?&;Wz|*0A_@jFH&?Y+f3@Cr#)N6LF#IpdcTb%+# zNfaP$B{&)CNpYRET$w*G!AbrM8u&3HdS^wlu&e~0`I7uN-yiG)R>p!s14~;ViEl<~ z-;E&W`S=_vLQi5s4c_4tTOgnryJgT7$0U1HAvGI`AB>YE~T9UKsd$-UdAoa;DvrH zms<31Z1gZ0)Si9RRk^}8z=xZ|Z-$maXHzsEKeigf-LC~vVI*19D*E7!PX_H)Bu)Vf zLLq|q(b}+Z6YHD>IQ_@PNo+lA^M>Lt(FI6hFBmqcIhyQ!g20ur$f@xMfs-oC=pY;y zn1z`>P;Xsaz(7j%Z0&LdiU6*zqat9vS}1B>%RVU&zjngx#Ag^jBt>yoz|F z(j|zM2AK?2f{O~ROb2R1)NnKA6llGkpx$=ow;H>(bQ>dD9s=7bmUUpxp}$3$5l8&a z`?!OaA`%Vm$yys&>PduMjy{Zw!q7+`KrSjouRRG}pR6#aVg|B zvnDpibp%qLX-Jt}<8Cm)QwKb!pkqvGw4Nd^EWd4VT=Fk`hRg*k(8beLQuyC&q#8a% zbOn+@ooVMGG6gr_Xk+0kSb+=#epk-dVSXWW$0jER3z;Z`*RfG|m%vSresKoOwAt2x zN9BcZ4b+BP0Wu1eq|Zf9;^|a@)i4+q@_obxdl*F~Fz_ui4ue3a%9=LB$2Isaf>)9-dv z8j6C&{98twrU0R9BsY_3HBCURBp|EAG~ZID3MfS-z%dh}M$o*FclbnE*13X>a`e}Q zZ2<4N!YnhlLYG51n2+qB2ziLyhPZ zA61>*dnyikm`2=Yji=wVj~ z>~0U)+SO*Dlex5A*gFS-$++hUe(*#a$}|R@Y+Np0$0%w35fJZ@X`C5F(j~VykqP0` zeK9dOW=0WsM%BZ~@Teh#&F=OatBtBJ|Dw>5yaTPZ-mtKAUiNtqD3oKt(E&>&#!TU}{t) zV2f&u)HK4iUiFOBOZ8hg6~s>`qrlnA)|G(?G#4(=(BEorAT|iv@eV{IwBF=fsOBGM zlFm=r*x&;KzwQDe)2Zr1s_I?>3Jc90mstg)q^M>*e!SCC_v@lMMQxeagU|e2?b9Zy zDwG$7t1+wjVNnZ3-*{nPr5gi-gScSYxqPMYyn*lB0+6kFEc__|9nB;MW8A_rPq)iq z;X<(6ixeNra&B#*Sg6c|Qj-$*+kz&TacRU`(*s$J)k?gNVeZoJc<5v zd=-_@^d0mnl0LpeUEAM@?4Q^ik*x#lJmSN9kJ#o4|a4g%ezPKHjCfL`s9=M_!3^9(k`EETP>)~K9K73&`o7Z+~ zUrf2zpv%p#J#EyJY?3P$Egp+cOz@E_t02&H4y0vI$58k@tl3aJB4>nGf&BY6FBchG zy9ll+{TEui07wgl>6$qVJ_!0|nUC1cMK>T(#8ak;Y(s>wV_HyJV_OBrk~5CcLV~(k z=+`)h=9_U`_dyrUc%~NBNsx4v8aChfw*$)V2 zEBjM}6Hbyj&seX1(sd*^NtuMEmR735^r@p|heCYEfOc%6(Z!#Yq2DCqLyj>^j6?hR z+b3Nj{O^|143_B^9J(ZkWc~(LU#<4czmUpmI=e zx048l{ZO2YaAPPE#Vhr$Dy-Y=9Uj}W3b zgHKPVtxB3S@dW3QGDnS|upl4?tc3#%$AMEhV0~SY6k?2Pk}%RiWKf0>ZAf@A7(A}f zm~^0~>QCbgDhDamWN09CS2dRct-z*MXbV4sXBJyE3nS=d{fEL@t0pB89=tGn8KkZX zWR-h4_TiWG-Ami=0G-B|Z(QgS1hM3LzH%;$>?AtEwMrov^f$gW(V-yf$aS zLhJ+fwa3R^A-(HHa<$1z&OP3GsH(Yx$LTe7+*}UJ`?9R3pa}Hw(z9u50m2e)YmNZb z)UAoLjxlkjGCjhQlXBXBMNf{eOU5SG%(~x?64??=knm~FBS0y`K9QoYY(p*tMW{IL zLsA&zc4|3{6E|~`P4;e!|M;w)jl$pVtHw+@^&X+wP=`X${OKfJ-nC)l@5>Qy=~S2{ z#F#-EYFQu=;K^2BLP8WH&b&a17$JzMBFp6;A8~p+G+Rzw71G9^K@ulPb01>HDvpb< z2(bgA15Djg2>dOg0)+spOzx@MquU`|xDNQH^g;AkT+hE^;G6bD@twz@FMuSPXR)|Q zWkjvCHzCYq8eeqB_kA3e8nuvm%kd(%8wg1s+(Z7{=$A$vX^d;za^Vu!OfYVI(MI+i znPcZ1jYaTBTON05D@pqN40t9H6_x{UARSLgRYIq-F0USnQ@TUh$Jcfm zlqBYT6O|_d?TzY25v4<^%P>_(4n03AK?@|gB`dQ|Ty3f~9AcRAbvW**qiXH$sZGy{ zZYB6r$#3_iUq+JRHP}YmkaE!!tx{rRL3~MfKTzVs)XV5C;w|K(mmjyk>V>b3)r;r~ zjidP3O~o{Ck%_=f52{nKi{(T?U*;p??Nsb@z2}EsrAw}D?7}ig^v%%JEl_r;E=Nfd z1kEUhI&Lyt)C8`TJrls#619A+_e*;m& z*xjZVyGr!Bt%2=|m6!etyL1(O;39U*NK7*nh1oGv`hk2~GNM@iG8Oj`Zio!Etg7tA zE!|4iT%<5czg%p3iad<8w?;)qtcOmZvkwJ+52&7UDlBOBQt4=*UuPgj$96vng2=vYG2NNQqbwb)+Ysk&1_gr+{MC*zwUk>Gj6P0puk1Uq8IwD=@sc zmp0Vns8s6RLCwz2&NEuKgGnInw;o{}Sr{5_=yJ^bDgu4n&hsJif}l>MRjeR;Pu09` zXA+XP0=;xv!PQ#WQDe$Q*V!tHPgZ*$vQJ=_3FtDC-eDo}1CI}V44rnYJ*N4_7h3!D zQ^mlgV=f9g+?N*stB-}K#Kp%57o&|cz)_Dy^2Pdr7`m5Awrb}|#du98C+=gdxid9p zeRWmHfB?YX_B_1WlXyx&e(ljw1u+d)NX7kuH{Q%XAf+n1OAj%f;lubWgrGBrsbfe- zg~I&MUX#54mV=M} zE2ytN%rPuo&74`B?5e;}Bc~~dHx-nT<-NS&8^Y5V7|Tflug|5S?Z?Q1zS~Xi_)c_c z#Gw5#pVV3X`&O6=-uySqf~-6a}?XvZ(X=7Q+QfTY0#vmulV675rE!1-*Ynw14m(`PV2qN z^zKnqLgkEaWv#JykAT=CVuIu0HPHsv`N$G&9C+$+*x0-dP6L3(GMlpcrt3_p7c!&8 z@1#=91BF=2?6RPJ{oaz4+~qaN;5KSOosS%Bjt}V$K$JF#6-|~qyVMgckn{fu(hzKc z-xEe8PQx{eA(Yq-BG5c3Wv-er!h}R}LDPB@wDqfLu&K5$s3I-^l)=b~9z%;{yFSy0 zpoVS44z%%b+ve~>|zXje$wf^b#O*}+2Zum4l=MGL=0Jju)eD;5)cfXh& zNfp9)pwUlu=q9JW&ZTyRA=jq5jD=PbbAu=L)(TZxwY0GG7n^vGQp$>x+wcHmT8)4?P3RmL9c&1yMDgs z+FitjG8Q2r3b7;F*eNcsQZf!eWXnMWKPDXUOdMC&!%FoM+oo7PaEI0-XvK2?m%RcK zxBlGmc*>|$&?wgt#A%fUp2U58FC1HvG5${m>p&h+m-Qa$+KIv(AS+7Mz4qR-$iNlQ zxQSr(fW>+WT!Q}xAjZUho1C0nYd~1e1OGo42Ko9Lk)OeBUzf0mB@*IV2>U2C$CFp+d_oJl`89h4NgvbdOI`HRh#N?>$|RR|s+!ar*D4#52X z1u8LNPP=*=hLzZy0u^yxul63d61!nw*I?i!XrMl`C5ZnEJ7psEQ^ro~r@(`9uw-|Y zsn6*#R5buT=WBM7ZTZ_@;=cbGD$qu-Af=CBC9jWSrjLZeBO^1qsQnLDt)3BWSKBPk zl*@TOu8A*t_4e+ERufBjz1@)2ZzP9K2;!`Zq!cy1eCwadp*u73k1q$K1+96pFF0o0 zt*j!n%xgV@%yV47YZ8lUY&>csJQF7tN$R>iN|B)b`V9x)w{H(M9x?Q95lxDOx?0p; zdMwpt-a|j^o?a9K)bjxWV&y?idMsm3kOzPv*2^yn?oh2z&F(i&Sn_1s6#QWFP@?%PJE{DN+r z@%`e_hPD!D6n^XJ|Kdv42vJ38h=r^oU2}3>SsF#hue+F@47WM zhzb|E=7gnGbytjh-Tv@8X^B)wpUt^mV}}UW?a1A$T&`njykK6pzUTq}S~ZuL4BoyW zC_nt=4w)p8dC9dRo&KKI`aNMmYOBy?FZ}+N^!1pUgP1>go4TB`on;>E(Lvz+O{nN- zQUwE}4BH+?FCOu&q|v33M0NqKh~CMTG`WKfmW)N#k67?`d{yPmwhdTtk$3NNgyepl zqW53uyp6qSRkhvCWsu`S^&5S;{uT*ss8gSEAmZ=4z#Si%h+TSiV4Z?kyuif`CSTj2 z>Tm9KZ=Iv9!qD$k!KztY*z0cI)SQA`#fJxd+mm^-sbh`0$lh_9L7YX!s%;g`k@Gq^ zUSoNox&>aJbJt#!p5W9x?tc38-qQ+&FdLTN_!$$m-tJ+ z$zVQE4e6Al|2`itgM3TM)UxEoMESC2YP=e<9%+VgN*H?HyiJpcEU>8V_bEK)kk4u= zu#z#b>7mRdyW{Cyp~gEJS-jcOS@(mKmlpC2;9K4{XeyPH*)(t~ULT7*EI7nc;n+P6 zD*-BMRbKv%}U6jne;j`|K&QBcf7I`?Yrl*6Se&RXW5rm3ilLL?dI@=eg;?+3)1 zF$$2C%L`Sbx9L(?OATAY}Y?u)Tz^JWY>yc-!gEp+!}Ds>mA^O!!_0? zzm1q)51Pj$AeLUm>R)R!BaFn7)gygnhbor_?zSl=I-IIdP*tL_{gGJs@*c6#_k-t=Vc8n% ze7f)CqOFxxA#g&zEHRQc9(B19l>V!Pl&N)su(IDY^?Ld(?` zHS}z|Kw~Fxt*$;>`I!bMkX;V5{h*K=W?0?{j2w9X#tZwjAjBLVfe8V{ZS9`la||-C z@qBt}WEQ;q&AV#lc2S-&)RXLqs2oD~byo_E@Zl{&mVc{J5(4voi2G(_B=?<8*)(QuxZ@{tcOtK*T21BkQb>$dYMQdR`o$xRE&N8TiyweSmGd$*M$9DP7V^x zCH*=L_8OU;z#8+s;QCC+Lk;)r&z7);D(fFqDfz@-bMoG2zr%D;t_1VPFFpu^A z8cblAxf#9r%SXHn2R>q1mvi9F-+hFzilp3teujA+3D3h}b-8}Um7?P7gBfy*C+-Wy z`P=TuvV%5k?z$=}shU~};wt0wZHu40qBVnDzb>N^OMK^T^eWSgHdNKP{k>|@pGnMgbVM}_oC)w43)%c%x+yiSmOn-?^nTA|3RgpFutOiB? zrYTyxkNRjiaN^sDv^-RT(ajwtQl&|zNw=!WZ2@JsSmzzaBlgjy#|z)`48Q6=pAj(f zt;N87Lj+0Hh4m)Z)J<+x+iJk{wX{kya2pi8|X}9%jpFa_}lLjzMc-j*BVwA6Vrf1KH*~v=Fn6b9OV1x71?*2 zm@S!q<>@IWB|L?B1cQid!AVi`%@vGSB=?KBYzb}F!ltE~d#L;f=k%|5?SKZ1#bZ!+ zPxbsNG!G){5E1r^A=J1av7}fA+PaKIKkH znNC5gd8j}IAFufRvy|b*O`(TqLs8m(B+}9bLsD#>*|G`vm5dh^f2M>5F(oD3+kc}i z)81a^l8dai8Oh(!{mhEBZ#~U68??T8(r*H0rAD*ZpSe-2FC-dz)H|&EUO${dE>^`! z!8G%uR>|Re}2tY*pq~`Y#V@eq?`n^_r0~ zEbPhZS{a*X)XUGTjXVpRbq_kGC7xNl8sx1p7n2?2#dWeP)5DERP4{1wy4Zi0nxnI~ zzt7B!waYWQ`SLok$-*XPiOL@-Iw+Yz8-m3?`BN?PpRj1?H7c+5OpSh%Sl@@v9F0*% z^1#8BAj21{6@%xmzc?tRXeueFW%jeVpYnFjL{g=)iOb)K#!AJ0bnKZ|ZJ=q%>`L1{ zOz&Mi!F2njv$jL}gYSlPmg`lMxwGkfYD7lT#76|WT}`r0ax*v%HG*}&0*_Ee5Tq3d zx_l`9uf6_B*A`{ds>fg+r)=yQ>qFz8S^Ew9Ne3I<77GgY4b|Vo-h9W*TyFXW)gh9M znFE(Q`j$tS==ycefffm6s8t72o* zz-1n>Qs>d6_U`{IP6E*vC`<)LD#Gf7^Y<+i`bP5UyR2e;MhY2@mvM|4ZrmLSD*L{t zK`IKynQZ-ib7gZgVsJ2Juaau`lpJjnR8v?R z44%Lr9C)+9=AG>-$+g)_X8or`4- z5671@4+Pc;_?*1MdZ%=GpF+c0PsFm^deSLh@#<0yJ1TPh=4mljD(5SWdrH1LLOR%# zZA(rMTSTt}hc}jU-aX?x!nv#TyifASY1T10E?@QUQXDqmtQx{@o&7S#|AGvD_NbNG zv83(hZf>=jmrk?wy0@(l5=aD3@5!<^KHj(Jb^o?J^R&Z2y(?*Mjs~w~)QaK`&)5DG zrM*wpck-~I;zz0~DlkuF)xrKK?Xoa<@Ajl2Hs$8$p>wPho?pm=FRFVY*oklhdKUGoBUc>)K${jqw*{bN1ILdE1%bp<49diiMM7{#%Qt zl||2{X0;XDFKj=ANpm*e$wivsk~1wiW|E4lO<3f*yG0opGqT*<`X!oYf`TX;vyvwoiZNzNN-uRqhfy9?NnU_xmEW~vIRH9b=YE+T9SbbIbhAq@vA})tC+OFE?>~HiKhAk)&dfY>=A844 zP{M;aM)xSe!-Sfmp+ynmxI&j#>qaAmcUPW zE1Wz`l5z`Qe!am3#PgE?^1#%>Tg$)l)xYF5$2??YWSi8+CQB)CqXGACmj&~OMJ#qT zFK<^=sC+GU?R-$;y2h5%={G2N?*WK0tT(41`h{wdIla3K(~9-$DbXc^%Ak91Z~t7r zZ+W{&*4bdu9-{#xz+;gv&Tlj#WaWED%Kf$lu zNZ@I~1Q~M5yPD)aUF}?_r^xm`ZmvpQ_=c$xxhVkhI%_%P~+OiEKmN~LTO zGjmk@@5zAJHNdFuy}A1m{lEU~GcEjG$q`y5?>?+7KNVx#Wn}8Qi?eO&T(_UIm$NmN zuRH~jjc18x>z*W1mZf4L5PxA{)J09o(OD8VLM61~bnJ~{Te5UWmiyXB)Z&Y)ThOp4 ztAott8)h*MVA^oYmY2H!ObgJH{OE-7;W$~Jtz0kI;L}aGX*~aJw#4@Qx`F1alqzuQ zNo7xDEbZ60uH9J6#hvjQg1O;@v=KT)SNN_dZuCRObjN0l^sHEkEti~uo)l`2-{=Zl z>LWsF2yh@W4Z3{9^m|eJ-LOm!x%^3mHOI&)~#dl}5SLA59BvN5)bYVH6Xt|g8nG`Mq*$h8nk4(_bqFCBCKGWM&G_tkfZb9}kYGkZ8aL zb76Un<)C~slMxMZ`Mt<6D%g9J5?XfY2ilc74TZS}h|lj3X%ha6Wp4#(BqxN;E4~jr z#c^?Z)@NKcI{LrKD1Tg7)fW03&>ZeRnL1vCx2p-2ahB??lZZYaPO$aap}y&vUecGe znp*SGr#ZYxO!ct%C;Ka4Vr+_h{b>ITlC|l4;h-dFowCj#C!3pQ(^P2Q?n#=6m?&dy z>38|3{XGBko_nUc)H)^}u2HKxwWaegMM;7utWD)J+H#XA0VjUMBM`fHFC9oX*mW7{2%XKtrbkhlCRSDly1<1(CK%1B ze%7m|z^%`E#x93%_rBvoQg0-Gz8_|O7gg#u8Sva=b~xc3l6?61&a4whOtV=P&_JUY zEVX_1wxFRwC!sHaH|drpH{8JMNo9UDVU$pxMIt}Zp&OV_mzOnuRJ42-HV6u^oM3&8Z9SQzpgN%-v@8yW#{-}XH;r2YR_#h7JA7YC4|M`IhF?NXAwJx1&*l_ z!qBXN9K!9h>TBT!r_ek_N|#@JS9?!^*vVUWt@u{(i&?H7)*)iMvt2*_kTCYf;M0h@ zQXa0Rl2V8*GneF3Plz4!Tw6&g?|xMIH^dL9e1R;Qr}YGMsz5Xi+EjkHTwE%(IsV1~ zGgn%A`>c}x*lkTpvg>&)w5ysh(FJwRt-y|X3qj|}5jIe>xBNOb-C~j_vi@Z>W@xoHPOM!Bui&s zK22EXLszY%bAGOv3Vv=6)mUMm?bH)C9k&Wpaio}*Mh`mV?rhXLOANe4ko!B8pOT^E zR4%e=YexAeTRNse*}sD*glK^<4TIn4oW9FAnyPFmwNQ_bdQ=N%dBl1*yLd*hwJtCx+RjW90bc+Jm45+Bc;Bk!LvaM1O; zj6xG~3yr~>u8Gkm=nr=O%HF4BzGpocX~4;t3%H|VHE8F)UCrjTmqG`&kj|MM6w524 z4vJX&A!uL|E^0F_XE6CU!k>Ty)cclDbIt z7Z*>PP<&GckFuz@b$q+&->&;mQv2o{k4Si2tewW7YB_~uFnb0vtI z5{>1cJE&ot%j-m`XE)ZhJw+iGmL!?hv7eIygQfaVF}`ORm#0M(Y3K{dP0WVUUSEtz zeQg@_IEakpJW}5{hthRm zC3L3%1zjP_H!|+j_%%E|4?@pace`8c_S6^SD|LTFWuQR)CynA0{+{)ALj*Zw1_yMl zs9(n?{Wd+SweZykwuMu)?Y;FUyXF1z)%5P|wUXFWuaF-h#u*ijy4d8*g;qE&6p7)? z(ZS?XDX&Fo8NbVPc7b!Zy7!4qaa@vqFDirg4)MR(&Wh$98A%7t66yr!2=SO}zMdYY z25YSGortY<3#RujMgept*(_R-PW$~rRou27KO3ft1zY;q6&CO41V4Hd#rpZS@3J9EXt&vf@Fqb0+QZweeI9$EDF?<392I z3?lkELHeb|_L*)eFN(>lL%9Fn++laOOslG`dq}p2%a$f@IbbKT=Vfg37kwEjClryG za*9iTzy@}9O%GQpOy4R~y0{ z$nwxn)Q&GD9_=05Xul=hdkbXfCtGv7BmeQIWYv2nCjEM%Ov3i0y^g+hUX(o`_wn&t z%G1HTG77DM;dH#m%7WoaA3I3R>iC8Q@jrzij3r)lCP>Hkn7em9dk9Z60KGpu)1P%z zV1q1==HD8V7MyF?1yU<$u60kYp@DdZiR7KtltqE1|EO_;cs zCr@?rrb`h3$WdIInqge?ls2X;eVY%8+zaN6+x0|ESHqr*we*#-*_|Rg$#;_dLF_>x z2kZbVtlCu zZpNAo+EA#Ki|{cQZM}#OiDrt@T!u(L8suRHM2wGDb2nBj*F_SjN(SNd;q!Wiklpm12%_O zaFApv1IZC{MeIbDsPS?XB))VGN#K$aoTmN=oj724R~GH+x0JQ5%i8tvlIr8j3zUYg z)_8`KE9d`TfLq~1WsaQo2xY8?K*XEu`pm5QqdGU@9X#0g zzjS+QXWAeOp(NmUz@8WwqN#kM&0@a3)?@B$qr>Bx-p#1&Y3x;?!H>d@&>NzsCEdMkHEw=YnVt)}Y$lz$d|8e^{$BQ-&zpBJ!OVOo` zn`yzLF*sKM#w88J_N==l*Z%=A7|qNSP=L^EBI5q(&O<7ZvF*Domx?t6doXQU^m+y^cY-o_Cg3BPPkH zcH!!{Dw*`c*)iJ5{tYy^XmYC_n;P4j*@2_P{#5LJ9_r~CaojwdU}dEJ@^ZX6QO;rY zGocKpg(Lguh%edSO*iLDY!x@Bz;iZSQo?=a%E-s>%8D&NSlTHl8p4r09V5(_Wi+v5 zH>|CxN;em~%Tma181+%Z6hLZKQ|V>z-mjmC)f&@j)zj+>Qc&zP;UuTN%FU*fouQ6h>tD#l5P*`Kr2;yn7MI zlODSLnZh^wy8OTwEm@8~NTlPu(X$r->WbqjVG;X5t!gw< zs3#iq?qGb$=%}fLpd>$Y#F2jaD;bMuq=`;HL%sl09Xu8)JE)BICrhHi0J2fPV%!Au z|9Z0b2XCiD6}O^Wx|>s~s`NO0{7el!@lM(~!k@H0lRxZf5(H-?c2X2mOwC& z_mAtZa6bL-3gj0Dxp)8of!t z4?(-fZ@25{2RfS;;uwMC9ag{YR6_GL%&c1Mw9z+o+FuH6Roh@7k+8U2WnMwR!usAV zXaA0)mnTH8Fb_E_?&zwU&!$jHH72K4cILM`r)*Bk?DyKb7z$<{d`;=(Mw9y!H1&X>85gtu zQ~e4(FOF9QK-9m27NgZCR@qfoQ}=C`&_tX}V#dlWkZW8Ob^tQEKz`OP-ai)R&67}w zJR2|UhvIVUnx>?=spN-Y6hX)+VnSH}$<-6aH1KnJ&{3YyEi87nZ{&NY{o(Qn5d#d2 zZqV!<%YVe~7sa9s&az&Q`sM^zFV-T_brDZ*Q3SKU@lNkk*)G*Jk2xj@hB$j-9P8Jd z5MhpUZ#Hcj?s(-v{8jUMnna9@^Ol!icF0QFabQ>TKYJo7B8R-y@tOWINXTDQ)gmHi zc&_~W;+N!Lye-Qq#_68;pVJ<@5bTWAeN|zN)NyTqSiLo0u{m;>(VP#a2~Z(Ew||Gee+ zA*tLv!=~_4eWkN!=0-()I=Nj!6Dq={?KNz69>W@NbI~nL}0*}PA>zP=b_*p>04wksSTJ&=WhTSFlg(@oR_sdIWF zF7H~i8FZP|Q37|NXJkr_Y-|2wJ-`hliSMe~%+u^7@2BT{i8nahxt($kk0JyZfEQc4 zw5Jh@EszkcnWP0I^BZ z(agV8z25_w_VMXa>Md7K4D{v?m#!m(b{`ca6ws3MpQKVi0WUjmiq`YfE>h`j>2?ND zOQk8eP*dq5smhsSVc3H17~xpzPGWAC4$I+8Yd;rjo_*+F%3a<#5nU7^y(WKiy=Zn9 zB27kr6!g^9j;j5|$5uh)V=-B5+o;c1p&{%mu#Gd&eAXNNpDlZspZ5}UFF$bRDCegM zIM1%7R-mtw%Ccw-Wctj)+%jZ$kxCwN;V+KgyfW8`&j{u4!;2{|O z)>Ga8@OKHosgtLplL)$mhe%ly*c(qOoG%(h-iwYggFkBeeDY*0=lv3D0fseGRs6k2 zITyFn#JuHJhbzuUOOy#A8+iO+<^DPeJ5OB5yJNAa+$7LR_OdaXPHO1!pNl9y%@6zY z(E2+&C93xZC4q0<|Fm^p>I8jdtA$8`d-1-t`?;jOa0|mb`Tq9=n((lEvG{}i$PGkp z6786$;Z)Wee${H4NQ3CkV^Kzl3gzIa&XGu3-Oz?iROYZ$DTVa==tiiNezz=oQAp|c zP7;$=$FO)yYe(5wPk5CBRsmIHoq$IZW1+}i&i(y;fLb#6B|p&9CEtBjQa3CNy?Yv9 zMJ09tsGx}Kqe)0e6i}1-fmTZk0EcyF4IF6Xgz0TG{PsF3CZ+eaiys05m>-+s23*Xh zXN*^TmuZ0RE`D9OXCvBk&~ol#t0hY>?b`Zy%52Si0ReZ**dMZ`r#Hy8cg%PWP3fK; z?0+5B@8#k+C=|41UJh)7g+H8k0!8*fU;gM5{EXW{)Ubt~GfYg)&CR`7 zbbeg42(-Hiv-aGL-vKDtJ*T;yXTGBx1vDN8%$gNnjGC^|9x0haw|^I&!O}k7g;E50k%tmEGt3d zr`TML1l8PQSZz4>0E<)?7e8Ed)?v!<#PJ*%7GFd)`h*w@H1eW5oh^FrUx6RBs+-h;-RvqgU?AlU@XpYEFLLi^q2%orG4 zrv*(wx`X{yD8#b^*wG)vur_32WnXa5NKG$1&blGc#03k*icV#;L z)y>c`(3x5r278*D|Dk=LJ1ZAC{-9gPWIwvl5KkV`NI% z$~-$70kIa#H`c^l+tpT9r8fHHeDk-z;8zMPL9)y7ed71#DiwTk9OYVXmwly@k< z$go~Io25AWVX8-`0nkd4iA~{n$9-Qm=G!H90cd*ce|CKVFP_Dj7uvL7={nFOnei znD-SyZq;z#qj+*@lCCHFV@tZMC-IG^lBcy4H+}JBml_)A7;l>Vs1G70Hy+StNd`N~ zupiY`6PddqqTHW!pFC%scSG{#L54@UOmj{t&I>K)8b8Dhn3`|dsUU`RVWa0*r@Z`R z{_PQbT*B8QRV4F1_PCD6q|(#TEyicb1jHUqMdUePAfVQkQDcYjk4J8^%1GDl>wrb> zS^rQPgkCSIFqf}fa4ROH-`q}sIc&W!!)>c8JZ4LleRZ+QYT7=jSl%*HQZH?-l35dr zGc#LrY$r0TLDo|SUUA#}(${tY^3zJ=U4Y>B<^BabSEBb^gU3n2ZByWFd)Nh|7cK?g z+hPI=dllKkABf%?}hb%WfB$x!5WU%7Ig!{Q;zyNc~CfvKc5faxjH9v7`$*qwb?4RV65~nEF-^fM zu)Z_l?gDb$u)-;uo)}iqGsPJX-k-PzEu5zcVTRkx33`J8Y(*741B!@Q!Z2MeZ!3>E*nI8ru_r>wKWn(@H)x#befDYEF2@c_n;bD z4NZDSVPOf<(&yQ)(!Dm;LUWu&j-m$fj^--}u}t3CHI$^~E0KJNg)c8p1aJP1&vI3$ z!rPtJHFsjhpwdoN7AP{64K$n5Bd@IzSI5r)dV6m5M2h47&pg2Oi1Q|jGdTaKdwdyGnrafuH6vw8gh++4D8l_<1_LV&=LQmL*$l)EVCACVW?AibR+FEQNqT4ll zR0NO~gmab-HJ-rMKmTB0W@a9>AOqJ!+_AK76)jE%2%nl2ip?)T14>Fte7)Wlg+DN@ z$ykY3g6}?`>ZdrWR?AEz)GU3n_<4q`C z-N7@~Ze!9WuzSozolEKN{5~2WREgKQXBKJQDjtaWnkzsX7yw{XCUyhqwVtTewQ`VZ zA{I#vik$vj1u-z+Bh4;ocJ;Y**WGLE zx<)LYbB?@>{EzU!NPN9#KV)|mS1xphJc2GyW8Yg*Z-@I9)F-D>lZ@`Q53;Em))e;f z9e)#YbJZitX>81JIxjqo*HmC}Rr?TS>@b+7;FZr^YuJ!%Z7FH0uR_PZa)ik074vdW z7;xpC*3~W%C1D#~B?O#^e~Bclt=f8!M3)0aPH|Gen5&K(7RViBbbMg3z9f8d!S`8u1h_@-h{daYfQ?p$`(Atc z?Rt5Tpg)o& zFn}g7HIv#=xW{O9N*k>;SI%jkxLRn{im(6^GZt0N`2$%I(GlDO?zRIwkTtLbLm?(G-My)ghdETgO)uWwq&E!VC4d~)XRH?5dszeo zdl8Dn)Ts}N-Lk>zC5;8KAnUJft#_D*H$l)+&2gfTdhJ5(`J$n**?^@Fvd<~IemnCw zN}p60ikaT-Gk_iO`TWnGmjXtFj*8uf2?39+8;&x;?g-cvrxWSOv#$-iiSYKyhJ9vE zw*w22iIkLwoVE*J6urg_qYW*xP%jf6fondm?|yN#|FU|PR3Idi5b})|Y4_SQaRFOW)Ie)EFdvEU1=Zc#^5-QPy ze7udBO^4fydWqTf7rMP5&L69=jA!lW2f9`z*v@+boiM+wIyL|_g`D;)a-9_yT%Etz z`ZgVV#K#zWyEikJz~~;_{j(4Z3@C2z1iED_@P?bc$3(r7M#=gSP@gUY{m7JWc*?guvr5M-HNNf2 zGXGw&B_KXLASMsL!NqlN-K)+cPfJHLUahYpSv9tL_YFu`K@CQ_?EvJhsoBZrNt^%kPPS z$>dCl8OCUDaNx!VHkx?RD1G6U0yVI}4JttK1oCT~6wJngL1*AST=bg9 z`o&+b=PC6pHid5MVryXRk#ZhWxSMHS`$*@USs;*nHt8IGZf7zHj-qU#x5!rR;6PR=219~D?WV9Xv!L6B}%V<0)%T#c3 zRMpVNmpTUG^3MkMLi#B)EnTt9Yps)B((2ECD>XGX^mVt8T)UsTdQ!EnDJhxGZffLQ znIw9dwu4ZG39hkMZUG%W?CGeUYftxq_@5ZrrjLxMY-iMO+gc-lUVd;EBc3>qB=?<; zz+)`}Blo@&SgovxsA%GHQDI?J57($$#d4p++nrtLK7mkb6XW^MP^1&~tYP z?lXPYUe)m~i41lguiB%aCdc0&&G5h{72CU}GV@iVI0wJEvYCSx)vn;LB<+8uoZOF| zO1^9->pBFcPN6?zvWfN4w7Tb2D`OI&B-M}=yY0~#;RZm-^~Mb!V?hg7ZOw{#e*|9@GPY#!*2* zb$x|?7C}E&gDFKua0Z-zKA9d`@ zJV2LrRU*ID?lT@2`0V`N4Ly|0@LF&UrQgAAz53l>8yH*cQKD8LUgHntWfW)wZ1YFe zqg;x$kD?B~-#i;hB$n>`Og%^p+Mc=@^`)^~+CDHl&{1)GDV|=YONvL%)lj>@{hPG$ zXHZszY4phl66D_cf++sw_>|g|yV|0z*du?11c5hGSj}|2T#0srXN`CE-0RybE6owFE%7pIt8iiEy3Qvlm0b3? z>sfu$=gWPanl37u00eTnWcJr&F}eb*+=;}{dG&Z5ZlT_UC}2i9j>yR)ooB2y^wE!S zJRA5(!?I)d#7bE`m`RYgFv8<$qRvZeMR9MpwZd+h$?bY=j6aI>nN%cHbCm*b)|(|c z`q96=QfbD8N?j(ZO)fa{}OdP z(J{RXp>BPV)PH%#!x>T%d^dxaY10MY*qQ7P9wzRYFpels4USD;e*v8XqH5AjVFMbz zsH~zC`R;tUj-j)o%GudMj#^YUM^$}}yp^d$56%9#wWxB2#bX3+wrhD7?xoEX0Rry3 z!K-cm5>Nr{aiY%@1QM+f-GJx1V^i$rnqZcfO2q*)9zTpv8{a>*e@SMps7bG4X?grM zVgCyT@UEcw5iwXwUa`V_WSPQhY~%6raaKXTk|PVHARFY_=%>rTF6w=4<+vK6_2{p} zd<>oK{wjQ}n|F0!fEo{=&>4MZ|8sxF+DpJ)S@}^L8<@O4XWla|2(BV**jkm`@%4Pu zJ*&>{?a!Rtn}>_wT1PQCWBWekHHVUnF8fK@%WW*>H#wh4I18aG{`g;k##`Vmw!Krp z$Jgpb51;oxNjZ0YT+I6p+%pHQqb-K9)D?^+gd0uDCT_F?0ba$JhmZdyP4GdlVf@)x z@8+PIM%(~Pb*jRLbk&tWIEh=vRSJboNO`U14I>HFpOuqKN1dOi;yN&{(Cd?uTruW? z8Q102$Qqm6*0EzqbJ3+?OnlpeQtmzn5f7#*Ft4i>TwaK9U*1!9Qie^B(w&srhdYrT zIkX}^#2pd$i~f=8J_>4bz(6YVUpj1|OAaRy?oI4+Q7#2)N(b*O^k{%4n-Va0U6B*w zo@1UR3^CF*;SvKonf2YNym`;wW_q#b#EBH9)?F`kTw5~OhZXK(rmL1-#i?#4fot|U zR3ujt0!d;Ws$*8r32|JHt4~VP6N$a%?NP_g_YaWYKQ}XWC3u`>168uTS8|eVXys;g z^t{h+9MSh6H@oKvU4pzAg(JG0DUoO!eQ!5a#!T@=FiocA|U*P)# z@<@}QLU`!zGA(jF>z*o!J0fwsyq*(SrUZx;{CeD&X{-%DX=(7u(L33qx|jG}G&skP zkM>}R3IyGW!|EC8Ah$fmsjRuXmjJ0&223OzUi_yrcnwvUG8v_79w>x2e8LFDgIldpmp(j-0{+Eb+GJi`0C@8-=%}+{jWMVFdFY_~ zNy!!ZzXO6t4Iaf^d=lOdhmeyP#c;MJ=-B?<`0Fu zT#dHV{B0f`p~~~q4#nprH#sMy0Ux#;+Zf}hE~&4R>b`xMz~52J1@xD3vxNUI%R{Gm zV@~;e+=g$R_@poC?#ZyK-UPGsnmhDHvuZqN&VU>ZqjC9qEy3df8~##byR+;xMJa@n zBBKLd20ye~`~BG@g&-sQsH%56Cuuy?;hH z6*_R)>!y}d{b8kbpNSb)FU7!ytY+I%H#=+@`$N;tJ6T=5bN;Df?Kobe5Lvc}OyNXB(hwc++~y#@eWOS+Db1!~dq@$c47wVMQaoc*d!EgAus%-<&Lqn&2n{Z)bYYjyF$X{6fAJh_rj69A>N&rU!!ya&Q6UV=5 zcr|gh8U);8ItaW@r>wMcm>g|#K>x=OtbmESm^-&GWdtN^XGL*)Be1b;(^py|Sc>Pm zAwO^wlz_+gPA-3mu&=_p_Y+Mzvj=klms%D=)X$FO5NV?-t9O2GenOaY6XV{x4t8%9 zh%2B@x9$AT&a{hp=CshT8st0wcMSg?hdIj1inqDeeLOjv1;TS<<6to<^ZfAfw$|*} zL16D8U%vLZhcpmyqr5X$k3FF|ta!uQucWI1G(2$YrUad#xv}q-aQkIS}x#1ttVcoMty<}#y*(ggr;C*D@*uTZ!1`a29x zbn+9u&#u@(A-!KexZBbVqOA9Vt-GKtx|>bJ+H4E#*m~MM34^_J)Kl;b8`?j;H&H23 z4TW1Bxaaq8MjOS7b=cityfkB@1R1pr^zm!f5KI6X*x$yN-3pJzeD$xE|+bNqi^t^Ea7<1blY}x`pP3-LmlBK%Nj6Hx070t5}+A&uFU<;kOuC zFqWYWb=$wMVi#ZxaYl@0uAdwMgU8$T9X}vNlxm<7&U@B;x(wytJFx-7bkR?`G$O$1 ze1JZMu^_tZLwU_L4eR;v?NO7Naq|{v+}OT{L?s2}S~f8Pd=sM{cX3Et3gW~HGqw!U zH}ZJSafL;afQi(g@i+?qI+FPHgVGv}j=m-EG`8R)=%{?ONn9IxdunEEuOOQ_G%gO% za;B4q7Xr1WA-*H@J^~z`Ao1G<)ZPN&fazBE?48LF56RSP@uJN`mB}juNuhqv?1ivP_(~e`*{M zvp+Sczgd(QnvC)4(PeD>u7SD5V!%A)BJ-CeM~Q!Lc=u}K>?~gtoBt%x(SyRVLUVr~ zuWGt>y_Oiks-#Ij7fQVLU=oNbYh^<7;g9RHs>N9f3T$4u@xgNQL%3Q`_sUZ6et)UT zMZUZCa`*A&q(DG^-01D-5uC&yyGl8n}X%DGgg*5{fK!D(VbqM5Bjt{I~)G^PCzpz;H&bO0eTPKh6mL|K>23?1* zhZEzhNj(O34Fm&c@XU3$AX`_x*ALcK#mf$x4=fdyfIwRxOG9YZ{LYv4pC zQQY3+yda9MI-L%YF5;G(*DU2Jbmbf)8s1Jc<{mekJ^vb<RbvZ$1Gs@dgaKHaLKe$L0 zzJ0pbC*A>NiTttanN+myW>M3#L;ZHWTI;FpRSr)>~j{?%F;jkr_ zztB_-;{r_$TswEgLX)hpC_JB4^d#<2QGW+=!(F%u#3_%%34*u*u4#alc5sIe@y5mp z2vj^@ukHFl{o5iI`~)avdq05HK7L2v++ln9vg!-(^Y~Bp{qB{4fq`E#v?Fg8e=HA( zF0h{Yzt7+_o=pX0%aHy2RRV9N#`Eh|er(&Z!OzZKZlJh(JR7yWZ4;Z2V6Te3HsBni z02ISnxDY=h+qbhD`OogPwbiHH-II8i}y% z0QkAV;GN!;Y@{(_l17pOveBoWh~>#KZ&G{{P`|bGaF`c=DVgVw;g$opx6>!PFU)>a zdX16NHR?QWtB*-cjP&-D7x-mYYT77gJB5Xa65a`g>rqBt_g=EKJbtq8+~xQ-fG5Ol z8a7Ym)EBU{G+9mLV#AaGy$%!hoD5hfecDLvd&~_>syvBt*Y>SSc!JR5?4>?Rdfc`> za_mXC!~X6GA)>KuZT;RU;g4N-Dm*f}nIz~dwi#sPi^qO>(k9&=^PA5yt`U5X!m5d2 z>u=OPtV7{zk!oMG7Y>fO#&~6omMi>ib2`($%M>GZV$L6SG|~s0P=r$yR1+Q9P`tzj zeBcc)5Z{@<*Xk-O(*n00B)N7Uk`tL38MShX6gYY?4p~2x&TRi^MD=)BUpz)~GS;nF z=)2)$qFZrbj3kfJvnlNG@}`Srx&!o}LQ38z?cV@;ZB<;X(%a`QQqPL=fbLy(^y&hV zbduI;*L8f)G=P)AQ7WQYbgJ9~?qDI2J;UL}&xJTom15&nVC+G=f zBy-$ptV%bOY1gYWLR7RZU6XZK%9MpU^1Q(#K4S53iruI00}d$-#WDW2+w5dT-Qj1q zP#01O35n-SIsaQ58h_y_F`a0X2Bi_eeP2d8>4i zWo=oHiN}X!j>?_My+EG9togv!E_$a0T+cY8=wX#*;Vm$)@gGz-0HCLXD+a;T2$FQwxZ)ty#mT3MuBT=y6{`ez88c4Z+q{3w?EmStc&}Eb#Q`u7TRp= z0m=Nq4ez`^0T$+&y)EIKqz6-W{{#SB)=iZkC`+$#nuR09E)gTB_1t83gCrr{5^JG3s)OMKGe_!*pS4O(S3n#cH zZ>M8(BAu6J;dwQ)^A^r<7?k(MCXHA7kLJ$OkjlDg|rKZ7pq@#%=0*$xt>m}X`P=e4f zy_djKGg^OzEHS((c+xwOOI-oF|H8}1a8(q$zE$8rFdK)M-1UU&PFnWixjvh2 zt&QyJg&l+SfNFFg?x&)HL1fE`bBa?%Fxx3u18r?}-Q z+IH5hEWyTV-^YV8sC82MM*wWV@Xb?t?f%;Ehe!*QK1<`}P12+m4gNw7HBwfUoLMvH zomfqW?>Jh*0Q6})cY)Ojv1glgTH@>vM(L;0fS1DJa$RcM;vN4KcK;vCJ5n!DRx3$^ z*YWO~u@TKhL1zDlrgazB_5df2rwkgZUxf~KBQZ%-n*MxmrrD;%c}f&rk&$0bREYu* zE~N8IuL@)M{s7#)u2xK-zw#%60qoK6aJ}2Ot~$ntzRS%(x8e%*cn&|WKAk>Ca9oQ^ zCKck=W+};=)7k3kXChv*!Kz96bJVPl1{>BUr&f7oKm9Sy+ZU$!LwW}PaxH)*e~LN~ zw3~EiVD>FmHOb)N?EL996I+z*a)z+6zBL zp{o$jpo8LOq_J_v*N4qE?>WwOmi0(Q@BGveia>`My~NCJSpXP2~j^-p}q&+#_$7$+xBA|hH{4Fy_~ z6lqS}{{ZMx=_6b69!6!&2c5qury1JxL1$RZZLZQItkc%(-k_(`ZdC_o; zL-F$F9wvSQ;85>t4HKUJaj0jQqo=2Oecc9H{sle*{rAozvu|qw8v-+4)F>{&lcO4( z=V|(GON?1hq}qX)oJfZPqP@t#EQDWlCEWxrtGWNatai=kC$Q7dh(UZU{gYBNyg|{d zO$Y8-*QC?yN-T`Z7uV`f0B4L?T5b$3c6ywHgN(*Ttj0kXb+>y`r_*D0ez}Nsz8jU4I)pgB)P}=&tiGE&o_QD&bQUnc!||%{Jt%^W&UmUv#T#<;RI% zdDB*!MYFn?Xvlh|PES{T9&&O%TiwtpI&uG2PdhJXC7iQ*T!QLU z`g^xItRqUu#uB+5CS@BPeg2?q3mxEDqc8@@hbzwtfuPDKhP3Aief~^~+?gGv!x6%; zZfwenRU00*?W73FcC_wOez36-R(!azb&{L6Q1j_HzsH=ttE@b5qc30J@}h&!d*Dbx zqrWmZ{Aycvhk*vg%|OVa=h{AFN>&oSUo#!s4xFFz%qSSP21MfYp!qR&_DzQ`LfVj_ z2%hl?cz%3(FaKl>XK|1DTWg05RrX7S1q#tXxr^C#j!?-Ccvnnmln zV_*?)q=#W$)lr%_JWA8nbhpp4Bl6s}kUI3?F)WI1OhG=hj?=JO8`wqN`b**Y7Yg6+ zwblNka2MJ)w{iRmVq;fZMlMi~FEkrh%jtuf4oa7RKr0D};hUsle%-;sJ?p?A&bB<@ zRL5z2;$EAK?v}a8YN134KUn*6Z|XA#6jL>{5aD@O8_GNFqA27=tZ}p_tjIMcz4vLE zYB)G}RddP91c&x>Y4hQ+C_@Xa4>w<*U++1y@8XRg&tSa&PL=h6lZvuL4k#Fn_^Hxw z9SF!{ZCPbyThSB!wVuJK6|wssMs)K-8y(_U{6ODJKIaK?J(unwtLLw8+~el-0WePO0{ITRD?kBMWTPe-M#-O9^@AYurUrK>GauQ z^6r*lh!`hYHfVFVG&|MbaqOi?@XrO5GXL2*B}sHn`pPKEF#nnbO_bo8SFep4trZ8H?yHebEOswMe4GOg1K3F*>a*wdtiauZ3NQCvBdj>9UG+}b zngrHi%rOJ@?f+eh1CHX-QOBU<1gq$F>VL$*%&+n`@*2NBe#%=QBq~A!^x(=@nGg9f z0T(mB3sq8NDpek~b)<4SSiVlK{hSy?w(4J_!C93uvg@2#LU$+RprqtKg5+HYQeUb? z^p_x5LJpstw(nQ6?hW0GBOoE|1A85LZm-`55x11Eo!F%%4EhC^1S22Gyxa5hOnj>o zpo=&@pBRpP!)tHo3DL|AH_Y}=_tPw3Q`jt*V<4}nzPEA0L5dU5EI z1ml4CUs=+(In;q9EBHp=f?qJM_7fFh;nr%kAIXOr6AK?5Jt`d>{|{SV85U*Ng)1#0 zAfZS~NFylSLn%skN=Qm~$AExHcQXv#A>E;LcMRPf(hcX~HNSJtKmIV+v-e)T?t86M z)0Ki^D7Nge_TXcoI2n*> z8-6VX#jd4iR}BNwc~5!_gEA{R)=;iYjb)d$&YLS)xsx@NDfJ8?uVh$PaMuIc-dC^W zDz1M!TR24j*BPM_szB1oqB()3=|GsMY<9Uyk-%?MBy5O=XC9 zCBnL9wEZyhK8PPX$UI+)pn_0*8OSAG!+s!@w^m4 zk#~%Tw{+7(Uug~;gIWe5f znIh{@hY= zXgrrZ z2Ubou&R@D;ompk={KsH<@uGL4sHTQHV>gIn)-tIXQC0HmF|ao@`fJ1fHLwpO(Qy}8 z!|FSfYhnFIfPij1xi-~{gV-vXtI0f__S!OaAUOfI_w}HSoyIEZMkGSx!#m$Tw{fX6 zP^-Ky%cCy7KIJfzV8~*M1n?LT}M-X9j4Zxs@@rHdBl%2XvL z4o7_674`bn__{djGtBbcl*AiX=T}e_zi|5ygA76-iObyB`5o-s5l{pMB|aT3M@Vy@ zZUYi5i3plW_*nB@%Vnor%E1-~QRA&uDF^_g)`qS!=PdEZ-OHzSn6XoC<2RoX#^hVp z^D$se{tH4)o<}rg*}7a7X>;<`;?6&om#VjzV&Auz1yE+pOFqUSsiim(%y~-yxM?XR z^y)oAviW&E5e&+9v2xjVphE63^#=+lpk0rKD|zn-NMZWwS!d?z{ZT#Q!%gJUmam5p zk~8Tq2L@!UBO+G-_2Pm-6OMb$YVEDWm;kV;4YRc9vE)Iu&Ds#hU%Uyaj0gPS?rvad zMHW>nsz+NTqInf ze`5)oE^uk$8Gro8T-n^@-Ah;BlL}pi(c}#M4^QQJsxDL9?&0`*o1j=LLCI6c%AV7! zObC}RP}Md}95QA3o^Y=quBtPy%4=16BSCMN4@}$oK1Q8!Jywz54M+1!s|2(DXJVDy zb~R0&pKO%`moR|HV$y2Mlo~(U`m=1_eD&V+75K1s?j|EXUbm{XuJAK!bK#?a%i6nw zd7<04nSep8;?KyCVU9f<$yej^E(=`6i?{9D0OKt4O>LfuMM~t*2*XKFVB%Afu;QGJ zNXbrmVE_=4A?2>;AIHB@NyyI07n0~-23E7Go0oI94QF=9 zk`dpGH1nk(+>{}pn<^rH*NM0txA8&^DW20RSkfG;S_fC%@tFDnz+xHaGPnQ{vbPvq zBo1#^MzNsJRwsy|XK>D8xnhq!*k#^f%slgKmotoxh6x ze_zRnXmH|i*sD`g&0N#gYT)lUB?kM;`)-obRY z8&i+Trktit&&{fxZ4UT!W*k#jTT>bD;6|HFGjFD?Zttbm^pj>2C&3xC>w}ry!D7&o z*4G3Zj=emCi_7Phi;mlko^#aS2+bqI)7krw-JnJQn;-8xR@{Bj47%9s`VES#e*Xl* zi``6QSgd|_)&sX^m0;n5v!hd8TA#UqjP)0L4buTD$M3raqc1gqOr)lXX(jk86A)%v zTEV{t$Qv=U6BFx!1xEbYPLC^R(G{G_t+sm5!sB{6t{kJHqo3SG0>kLDwtgG{J?biR znIOWHY)u07Epj;T8na2Y?IpAk}h&QoSWZbNZcRJtq4coOsge zd=wS)8l~Efk=n9V5S6`U_F2uiA&J{mML@>hQgBe9*h{n9WqWGdz!T*+y1ET9CLV_r zgi{&M#VcXmTgDkeECMlVXZ0$%^9MWq^(y?`86-y4d#!Q}!2Xuli)|{DaOjqAT%4JcdPs;K z&27jRJEJRH6!Aq4w&?|`ai^?)l`>_0$-;r9Y#CgN-{<)o$t#eoaz}sMM1C&$CR5pJi>Z1UyQr7e&Z6YsyuLA~>DcO*TV*Ns;zOf0(vanCz@rp^8Fhiw{|d3hP`)~GJ4&s{_(^0~05 zRio;Wa9OwHG;Wu4CyEZ6&Gg`c1#9W+fsJnXm%RHeIM{w7iR_QLxBlX&1d^K&kLnYT zP7M=nMW{Q=q@ls1F&hIy${LWtwxNBV7jBrG+08nE>yP+h{spCVb91j^uFu0x*lD|& zm0Z&7Q;NF7@#A;SM*aYXw`A{(`?!V!#B{2nH!gO4!S|U|bqwh28)o7@OXa4=p)W>; zy{PkQoZ)hDJ4wXwZi3RxFygyj@5ig?>y204jnCVfpA(&Qzdi^)kyt~?%A z!-oATau(X})cpIX=ZkoaG+Ni2$}81tNjSr!S9g(RbZ_zn_&f$2_iFz;5Gd2n%*^yP zOjTMq1e;PRoejk_Jcj~Yu~xmv27t+K4J@SRE*-nuCpt7ZNS0X;7sgjT4RC~YY9^lX z*h}1;h1{y#ZSACXG!FjI_J9p;m}j?&!U-?>FS6x_)ac#97Hv-@IQxpmYk>wcPLI0?-){D8Dlgd`aw~_*f0AD$*RNcv`dip;K2x~ok zZQNkLZ34&Ij{vPJA&J!-+RPgV@TtR!JoF-x1vy*36ZAJKMWT4Lvm@E_Bgu6r#n2t1 zf^PJql(k;G2Ki+{v{F{DD-7C5@H8@M13zoW06kMvOAE?rfAA@Tp2tpz@=rz8qv<=M zKixfJ#`FJ3AWY2oF$O~9EL^fCCMIJ|O91}^s@N5{=QQVb>KE74L6oOvxHDbnWHY84 zhY+<{c?mvN22(q|V2l8W8XNSd?Ar39VEoEkKsYWBSP81T11Z~_zvfLv9*HqSQ+(tT5g zp+SA|!fCV9n29uxmfh`4rVV9^QYA z3v0CWdhza=*n+>8e?Old4pKZrLef>W+)+%SgW0;-`xH91%ZG2qu;Y`iOV$~R%x=?r z#h(7VwJ{ys?hii_Fma~%hiD-O6i+fOzy9e@0qogCK<}7ze||j_%s5En zxFo5QfkG)R)-pb^T*l?}f&4DjPTAE2d~2el^4O?9G ziw`iq6(t&8P&o8Yf8a`Rh)<{iaG#ZC+21CD3N`#7e@HRNPAI(tt)Ck46US_#iWEa+ zR6se|&%4I7x9c<=vA3@;PSIpH?EJwn{4S*D9|Fof&sLxmqfv+N@=yF#>mg;WhpE?h zMdE(0UrbF6$vZWJk>3=kvpjbt)KyJ947!H#@V$^e!Q{S7=nUGO%~nlQ1xhhu`$mNS z2>Buar)HcDI%stPA*Mvc%6m=BGInAUbmq{)U=Xh!; zjoV8=dmqgsS4&D-iX-soY{@f%R%WNdr-}prNZR}hfqNrUK5Bm(9>;XaFlgVhsL}Yd zo8AEOVHT+Zi*c#_1S|ew|55&!Yp(petjv&_YUTWl;UI(CbI5m!B?K78P+K#kvHS<( zE-+6HNQNdNLG3?RQslvxJu}q25Q!G3Sw*^_tQ9gbMo1{HzJ8WCxN4Bq2ZY1)mxmJ*RdpoL*9~;@F{dW?Sk++~8c}9A$A1+B z?GDi=(STI zmm(-rG?d()$F$$~@YqFG zaMxfbfO2u3MXf*F)6B@oS1>+0|I;tH2QQ8XT#2o4@Sox`sSF~#y@hm}qIVJicjr+e z<~(EESIh|N%!$A~UxYy^xUJ+0xBKxvf6mY~p5#(dSB;yV9tqudHAjL$3laZ7Pnw4A z#Cs-`;vXrg^5A2qI~h^vQWUwkc<#LDm|s{X0zdvfp=2iiEWUipg*V`E`^dD-zVf>s z_@*jvN+W6xHo@o);q1&aWnhZ%XNXJD+{yW410Qt&Le5%=ss7h@j(GwNrUa{~4=g}{ zhB#{%r4nBrC-+IPc(L?bLNQ2`4S`kt6%4Aezb%b3Usu6zMl^U9%#ZHBo2>`5CWHTm zId~KT_l|T%CjLU|fuS1PXiGZMEmTz2OkIK3_-<>EiCifq+^PdQ2l>Up@v%bM2x zo*UPj>Xs*oOnAt$JYlwo=d(WV;_dNfMy3RVXVsny39n zgFX_Nzq2Zub;K_JNDOf8T(l==j=;71B2ympO%-$RteV#}X5mp^vS>PO3Z2BwnS}rb zws$7aba>uS#_Ml67x>6KP;*@p(H~rhK(Ys7oq;=M;}4;sp1t^M)ueb}ocdFUl39UG zR6y?yufiVWS~XQ;qw?da!$0=!;Leh|%Lo5KC6dbAhaVo+2?@?jG_rM6A-5n78bw0D zuG-v;+%af1o{P;^2P6%^OQ;(9Nb}9RU%c)e!_rOFz`ase`N;&> zN5x3jI=)9b0?xP#@Tg1}L2pHeCZMnqP`BL>1`TLWPYes==6WZfAfRWmZDHWv&wjGjN4Fyf{n+`N zD-F0x>$}66%Le=YG035(@PnE~faZv#79UARTyNy%idAD6jyXWCXK-1zYCG@sDun}2 zYYR1nK@~kcVt4+LlqekF5v(hS@PB9R0;JE7SRIWpv-H8iL1N46VG}Upn4>_T3Z72G zb!pR342mcPfoh7`OitF!dvE^o9q#An15KBxwA1qq` zB|~~%VSl~6v`y?E+3xRHnCsC8YdTk-85_$E8wnJ zvpb%JL8%Uh_5LE6fgjccez=S9@ZsMxpOYqgO1o&2u1KT;tgN%3aUDB;arp5}eX#NT z@OV1hfFl>%tA#tBaoikuXeJ`taJwH7>5mLH#RJGX4SDyZzgU0ma>Q*<_hP^aX4#a0 znz6F_z?~V3c?FU%^SHu2Zn$vT;_ZH!u9M}cacU~O$ZMFACKXU~`A0)G@Ax6ahDm== zI|AA-kXSESCbMSdBK-NdxMHr2jF?VMio_@Bbd2!qpmA3;nycM6u6iF^%;=sOohF^n zlrX?#Rl`2E|09e4FS=hl2x&WoEUaw1|116_3vT!QIC0=_9tcZiK= zYkN*jiEf)|3c=qUm)6=O4Sx|ti(D@bHROgL^;fFsQGXT^Nh?7(g&R@5YpW(sB%V5vT zGt3wAquZD5l7&4f3%tj_Pl?1MnEUM3GKG~txG^N>P^2v%A{Y7 zfqtxVP&~EJ@7Vc~3YzD~z3HCq)$^SVBh*Ih{+p5Grj(Il(ChY_GJ@3-ko4B_8<>CG z`SrKLU7*`rUakjdK(0iY?w=GwK>PU4a3rzyd;d}KuU+sbU!){hSt1Vd7nC)z{S3V? zPPehNnx{oXGgC$WYG`^SKPX|R5R{4KyImw;}g)#L9cf(=rdEZzWD3vf>d%ViGQYf zDQS^@y^!u?A@>ld@vzYO9_KfPf!H$UhnHy8-ZP*Z?JoMEX6T&3?aPa)8sqaTorK=L z3P-{aMu(05X6~PX+Scnk-i7JmMZCiIqKy_&^OLm`;zf(WV*UvK(V=ja1b>lqYEeM+ zJ7G|R)l+mk=C~8rRqX_zU;s_@LS*=Ri3E3AieQ_B?dk2|&Upyor)-ejNzH0aN({}O zgHF_~BjF&9J`+dCNi7}cj^K{f|E1wu5~RA@{R5`PCu_$>FsObG3~ET7*#acX1y`bL z(>({z%q55NBw`ztkd>+Z%A7gGt5n3#fCOydid zOxTOTJ}Fdyb*xo;5XD;gMXOZ4j>&Gq&VJrPIYG8&cTvmB6T{DFe|k?%zop1R z18w5sL)TbxIv;llIfn3s=Z*pmdGkdqje#rkbuk@XD=>=}?)v2_p7_;d!7MNG(I6gJ zBHyG;-OBci<2V zVHz2xRRkLB2>|{t&dkqK}l-0b^3wzcjfh=zGlJo_oV9wEc=ANeL>|6j*#V%Vc z)s^P5hnkP%bbz9O=Da+^Y5kW?JXS1X^?jWH7+9!X_5J2t{PGqv(2Ur#AbWDk?Z03V zy$_7d`q0}4jJnQ=oZ0|gWFju{F_Q1>FlWL^55D4^!OqDaNy0BaRr=s2>888=DOR-T zpv?!Biv_Fu#&lr3Rs2}|`v<;g7pCd)gBS7m+)vS6h{-LsrcGW_R}r1+6flm|W4JSo zZDREB0o;SW?;7!gofGg6Wml~Kn;McI4@3isX_^hxbi=Q5>5hXgRXhnUR|(z>GN*wsogA zME>3z?#542P*(7(XX|aO1G!5w45)T+kwH;X41h1Fx#Yvw+Rnq>BJr>CXz!;9lgXr)0N z*B?b=YV%x^Gv3i1xy8}0mQUAZCfrI>Uc6qIMK71K@E2KvS9yC=|8kE5jqR^6n!o}O0K8~&F`PwIOh_KZ_6LSO;d8DE>!mb&V?$h( z+?C^t7TBoIbDv`gZlJ)D?dmf!Ln zYZq7!>(t<_I`Hn@ym_&B^WwDr<+F?EgNy`NqllM$_Xp#`-bm|SWLA)mio_3KwA6*k zdC3f#oAjeN$h(b|N6yFksO>E51(FHru4{#n>N|7AEGPQOWAQ~4C|QOO4{NbBU-<&Gipvq3ov^2r`|>QB4e^y`i^UZIpC!}S{n;g> zm_tq6gEdA>f?S}&;rZzfWc#wxALqq=xZayD#%)%eNXyvSYz&g9@dp^*AL0yQ&W{d$i8K;ODHLGIS*=!J|F0aqTwY?aN zqV$HZ$f0hDpd@Kod4Ab^n~?bQ)1Th^K(mQtgo!+&!J8EZU2rM^6{K(XGww|CLntl} zzX5+^mg{%kjF1sjDB!MIK5zniaS8rThP6?FxGlz`y>Lst$(FFPmhDH5MkAiQDa3;g zfYf5U?>xpr$90Y(frKsuA`?^FAo0%-v)L%5k`*uM!m3BZ8r=Qr>?E$HDg!JorHWba zSFfkGm$8+yrIu1$#>ca<*zUJK9n_75R-Cgvy&^ac`&sR5DnX4+ddhubZ|z6K zM}9I9E3Y3{%w1_FJ7fggxakZM{K2#*uJz+*>607o_ST;10-B zZ_(v9815I3f|;?j8Jp53#^o}IM5MAU`DZ-nxSy}>r_YYXY{JZS;6{;2j!s9TxWwV@ z2-JvgsS!ly6cnSSuMQejRvC`c{k%h{~E z=+&xl9I}!dRi1q1Ih)N$mbk$tGL$hPCnnMEdftG)cH=5?aTqB=TW#~O+q^e0$Nb|X zt$O$M+Ao+6ve%ZAdE;j8xoq^uz(A=rS!ndHO9)3yZ5B45s%?3t1EkQ;0o9rdRbfX{ zQjGd??uU0c$!esjTpAg<-Or#4j{Yd^XWua2P5zeccU|mj1@L_1V^BjT5l=tA*GM2U z^sCpiPU5w-pn?eWR%Y46Zg#=;YZN_$+s8viOL`>ZdNO%J2$2_XbnMXWMaxA_jBh>{SCb_~>6*Ss^`L0O6g^46JkXM2K+D(DI{Rqzwv8Ye>t<~v_lzT?YYtRX1H4dM|adz#XA@_ z6&H7!uJ?)2STDdJlG94njrmp3R!P6Bl^HnVS58b4e%~;|D!NZ1mT=DZ;?hQ@@hPiT zk3K!-rb&`Ieq9{ zBn_ejfRA#Foq8z%FF^8&fR|NY%*(%Mi!+((ta`#)5R}hXfDdhF^@xA3W?Pan#VTI< zx^^Kh?Tu~VKu1rY_wjTy@tSSbK&Z&EmzqKZW7=fz3?mbs%P>{gs#0=IqwXpD~o2jlJ19YuYR7-9m%`ho(*Rc>r5Xzx{b1c8G4JHxS=Y~9G+ zoI;({^z+V-|4roxeXgAqa(!N50LH-NdBNbuceFDdH=^e2Vq_w%+8=d&USkjzk9sX( z1XKqSN&>HqoLOVZbpNSto&(inR?r@rVU#cYDnfK8bEl#yM1bt-suY-Rd>n}y+`~}n zNoaw6+aPXx)-yq!u8Pjwz>9-@SQU#s)bj-AHn>=EMVTgA2mYab0${>>ZgWW9CGyb z+QCb#~|Wiut92uJ9+z z-V=C?&*S4}A1vTSwB3wsOP|_}voaU8E9vW}*2kq=II>?3yvor9SBMET(nv66QnyfD zV_PU`;rs)7`+E--@E36OMGt>^3e-+J7n)?UryH5z=6EgtAO7Lg*7 z*6WPzq#!5gVYO@MRWF+{zfM+xOw35^TCRLn1DGan`lO^f>7|r*H(lK59=*#hpKEHIb1U20d{3c{;YjQq$ zc$yjKbY~V)k~caOU*hhpQ@G{Yg{;AqS|m!%D8V^;EOlZi?2KbL6L{UJGM2>pm#gGN z@e(hrGeLe)&|gXlJl|8!S%4c3&OW38A)oeF{KF=93W%Yo!<F+8>FrL_-F)Cc&a?%T1O0OF!&ICeV>(Z#5U*`FC>&vd&|U!k2ze9*m=LY#d%)M z4P9yzD=g`ni{ybxzt_n4XiHyBXyWzxH+JIhZA;?S%mq{F|BzMy_XLyzdiO4yVJn^< zZa0Y&BrI8?eUPXD`Nr_0Z5$}_Lobdwa{k_~BRWyw++UmK*QyNk;{*FE!@csH2P)1u zp=RBWh*-Z#9_ntXq*6U17~L9$6M)E9y0|>g2SsZJ+#Ue6*35K)?27Esz*u#=IYNTX z70I3G+j#r1Q>~}t7pfZrrUUKHmk?+-Tf|YNTKtN(#?Gfwz3&8;{JCo*Y2Orvks5Wv zVCkACBmZ@5aKi*HCZJhTnN%UVFeyTl-?R;H zP$!hvy45KVso09~e0aTK{AWhu{pGunWLNbZiK-+3g^7DNE7&RSy(xM}FECCuIL(Q< zS}nhV9WuYJv~Mwj>QfRK5pOn>5pk0q=`ZHD>8MwMeC7&_d9E4k{WyP!$5(%F*7PM- zC8se}?1~$15P9d#?M*ZlS2D|lT8+6uP5DtfFSlU#+4I-5?OOBx8@nV`2O2tB);S{5 zknoNT`S?+MLcCrHDlA}@|Nh=A#f>=QApUc?+RTLfSn)hM{Yv}uXZzaWk0hD@R|mgL3dEWF?C%nT-s<&hVcGG0}c@)TcqjwzuU_Xi)z71|O0Z%M90 z+|%bZ>A}GW7cHWaQj-KW?#`x|6ZY_m8!%kz-H5GU2p+JS9G+u_!L$Byny$ZH3bJ#p z;K+Dp6cz1<5AAXN!UD>SRhk9JsDemQfguAa`SOqITo34mNL zzA@$szI72QRpzbPbxG2$8=|Afxy^|22lvkgqDGLn&xL)|3ONIG&Xuh^$@i58TyOsl z!M2-ZJF1%+Lfc+s^k)8H^pOeG&g~Ft1jz_y^r@+F3=bPUYqd-28eI8M&s*^O_77Ww zn0WxrTikBQF&B1{LULCNOLj!scJDo6Epph`an_%hDNfGF(4x9H?yg%7=RaA)C+_A} zcM@d*g|xY7j;zPnRdP6mTvrP>o>5$~B#{4UG5!tumIFKtj3&9&ot=NYyjnc8`6i7n zg=}6W3y)HSORp>=SCpgW1#O>JHL$C&ha&U8=v}lG<2`c{L+}A@Y=#|S>T_&7g*H|i zrI`+FcRB^&Z%OoLD0#*56N6at1v_Bh;81~%%ZKaas_gUC8PtdhvxqOjHUwN&-<_QA z2yLKqZ}+M?32L#)3|R8)vs+*jK+r|XKKG6k*wk3eon(-4#|nW3HrSZ76^d6rbwyoi z5%q*`&L`3eeY@c$E4eyblQFYUoUIO-aoHV~(O;seE@{I=-5NDZ53HGRP-FTtls9sd z+#M<|FO5RsS!=kux`nymG;tQv&D<*3nUp^<(PDr_GcbjZI|e~~p6kuMz5mer9zxGc z3z9<*&kwQq_qzq=i+iBw2rLK|B+d6=@8B_D_c2th>e_2x>_wOF)tbGV(b(e|Rm|9w zQ*@XLH+``tFg>bRX|8;oTPt=ty_aZXC`(3J!@I=wLfk5`pRuSv>{v+P{(U$yQbcc$ zM~~oRf;~6|n21fz)8C1ZDyWOUfB(%~Fdwb19;O!M*j-Q~Hl(&tJb;ce@4kFJf*a-d zC{yT_F(#Q|*6h-Khq++~_EcXanP}dZwUH_xu!|@uh%Cs6pyx90)uKb{^v@n*tUddr zV{pGuqpt{G!0R4w>Qz_G2$|-ZN>QWl#nIJ}KC6t3$hVo=FHyxr=FfN?J%oCL?k8wT zibW3CJ|jvO|LTUypN6sUgRmccK#E2frlNnCf#7S4Vo_%-qqLB;3le{U;H419?^VTI z6bl>&IGe;Xcs0UAXoW5Fx6jVYMTMhqG;e8_)O(~8Kgd4__&lebgWUhB2=&{No11gj zlUXu^LFDnNfP0h9w9*lg4D3WDG(T_V?wUMNr3HwRac;0X;WUlmdA|mAH9Kzi92AU*~2ft$ou=#k}*NCE!|Vty)nMi5)@g5esHrQ8`Y~%a6a8F zMvhlg=RvPO5ml(ul2VX)R9cIz;D2i?hP~5ClW}YOF`a~{nKUscli(;q#_LI2PWHl; zS#2nzNAnI$b=!INRmfX{pKsliPVTXm+t&Xw0X7P47&UqOYN*{32dQ6EYDb4a`ebw} zse0A<(2gQWGIQqh2uml)W5s^|FKsM%jBa9fF76$9S2TE%g_(U5Mc2)Dq^xbN5az`y(P z*0}uGWVuMTe;y%Gn5Tn1);_pnAHSQGQ!oa)m9-b$n@L~l3#rRTzGfatW)ep4rt5bs zz;K%8WN5gBC#u;80=54{UPxO9Yd+OAen%C@)I4<{M}EPpMI;#vEBS(8zk0*#W%y=b z3Eu|0I`h>2c>OYKABlAq}0t@-{E|510pHw^)cG(6 zcXq?}?JGTM8oYPdIRCw>XO1}O$Pkg0wWrxs{^?CrUh)l*d(y_H^b|`hF|=IP|Jmzo zkgV;1dDq4 z9E%3F#O7hOB^80}M0npt_vP+w@v@Z#Th_eSH^oqBVe!~(RD|SV%6q?Y|M@R79*@q$ z(I(oahzyS%U4KqS)^R)3;~} zTV>Msoen$XJ_%Ms??RZmrWsI}88$JX`$}BE@Br@agPYQ+Jq(dyRPCP&{(4ony&tvoWP0NK$cCUQJpu4NO3YdAZ!uC*tn_qj8bzx99- z4S03UC))=pI6(sUNAFSZ zr`#69htS>jD!tHtIjS2=>1H=y!o;NvsgS$&fmw{6febmnO-~?-e@Db>UOu^OGuLo!=+@PhBEHnw~n#P|bCAW1+!OH93LGFT$KaD6&UyiyVX5#bmM0 ziyGP<$1<}7xCYQ2C&stWyPA~c$WPm!^+#kC$e0z5-N<@3v@<`(5M#6~Bja$H&P%e% zaGv}AIz2ijt;WSw4Daz}IyqZYozvk*VHdRC-7U&zL>->T@_e&#uUO8g^SGar!Men&> z!yOH{(GIN^(ma!EFCx;I?d|P1E=!jVpU8qSN+T)%5oZZ3_{Dn=ay&VyC2~{&%BS2x zi>Mgo+S-TEw|RMJ4*sr5eh$-YYm}8VX~vGeB8uk$7_{2XYbmW-nXO8Jv&ovHmaQ+520GGU%Lj&JmmE1j**{4 z;eUIGb6o}}263vNSK6oCYfTQW%~#})a`Q%ozwXQ+0Z56a;GMDDmML>;Hi~BFel#M+zCLMr zGT5h=yA;_vB)yK85#{Jg5h1;C!Z!0GMdmiJQwwNoE8u_$frRs>Ux3`>676ka&(7~fE@o}7H+XrGIO)Vx-$L`gPbdpZwj-&-58+D+9gXw^o$>aUEsxWS#jI_}zNo0M9wtb|;7(25`x_@Bs1z>O)&P{DcnJd9N!vpeMycd-JCN- zZq~I~?8Q8<#TK8VPiQnLy4P5vrO6hr|i)()-4g-<(uc~l#a8ObYiw`rR>LRy)xiTvi_`LW-J}33*2jX$Yr<7Kazdl zv->WPiMb`{zbRi^gO^g%F z1)@<74rzDI$cQ`&3ke4{`&lG`Z-@%G%b=?kuJNwuVNS}TM_FWia6%-)J3D3EJSKot zA3kN+^T0Whh%Tw0Dvh-IA+Q~Kw#)c{ja7!5e_R~>qF^&^1U1T#LHBwmY}S1eC5w$Gv-Nj;f^)yT zLWGz4Q>{HVK8B>$icWfVXPK0T4LE$M8dve!i1)DCixBOArn#7@sNb7B(#9$P74Dz+ zmN178(~RwOW3)aJB4vapTaH=o+OtO^0lfj@CLd+8Uj6T84kF$o`^ZRKt+*6;G(7$G z4#xxMPy6~cvgIy?bW4wA!5Zu_2eJyCMo}zs*^c4pUn7qm%Hb)7hja0s*(SwY*JEqIPUYJl*DgH zj|S#L%P=z?#~4#6BMW{=JpX{KX9dN3P+LPv2idBiP3H z9n!uGJXU2XI9(NbXN|2dX*-(b^tDe}3e|%s$9Th3Zjojry)RbK?{pUL)klbr&=Y!g zeqkfGK}ZaNBMKqmk6` zwnc8gpHJ7+LXlivNjAP_c~-+OYa|hC#r=ehzMK0a+p~qOV8e}^8+Vghn*|3)rC6(b z9Lp;kv>y}A3>H-*bn~e8rWyDCCq|*dyo*eeOejUU#o#CJ9k`Dr`K`b1zdXOPk$Hl# zzBLLf?QAr66CuRN8a{LU>NPF><#Ah>3(>&$kn&Zy2wsGkEEP7i~|-z>di_ouPStR9W9Ekwi5|g*)bTx zm`sdw#w^1PD5(arLQm!GaM^T3=j6Occ*YZ;EE}3B;i%-IxoY=CiZ-$3VocvU_>9J4CR&_UM4X$6ESif@ zxioKQN8??LaQ9oGlAiSN?~ z%&9N>Cg36F>lHHDDWN_fM=#QS{VpQq$uW2OxSHBVL4{=>v%R@{ zX6iIsnYo}dK}K5D_ekc%A~}_}n^)%|mJ)5uTAKWnZtKlTnlk1f-9zZdmCTrI&Zm5> z_t4ShLJzA`b6wy^Kww#RUw*$^mEPvDds(Z>m`cO!%NK#|&Y$7S^0Gp>f6)C3K4M}7 zvMG7FIz>|!+$~l%+FWjH?3z(f+Os|x`pulV&+f`TEppRta(;8EdQDpm+F93A{ga74 zf|rj5Di>MX8BsW;IQQ8o^d=3+mf?x7n!CpcZa?RBDytdSuOU&=C*UT0_Z}yHjkN!< z?Yzz1E_2NIAY}+U_pNy8Yyw)Mys5o010)URd{AF99D7*0@z_b1sNvIBCqAb2O`rz0 za%L5OdW2K9ZTgJzjeG<-v0!#i7-uE@RJn|&d$~2%3O?5q^4IG*sl@d5{*hM2W4HS| zm_pSqRG%ZJxMfs-W(}UNsFUp+gm<{Kj|nFNzKa4IOqQ}ALJKCa@<5M8cl&YNX8SkS zLN6oG@Js6HIeber9y}^Jh)CmT`RNdTCE%GDcS;R0p*36-6xiCL;O60dI+VOf z@|jx%!V`04VrGne)GWo-hhki5@UX4Mbiwh&tZo{mwKF6bk}g2Q`?Uu9oYsOiP$9#T zM~R-XPhzb#$Ba{&F}US*c%lz&B6Z;!QvC_HSk)57y$5+ZY~IC&yf-uj1t4Lk;DI-5 zHz|D>57^Squ!v%94AX}=o-n2A$KflL;w8jy@Qk+e=E13>7bkzvb|t_VaO*%f#t7e; z(*N~pZ80A1AP#>-7gZZHh*q+q;2Ax8v*|)Qthw?P5 zRq@F`$A39`r`Wk#AGkocJ@^=Dx$3S}Ez)CK>3z$EoXuHH=O6PQ zxce~nIrEV)_RK85TbDb&`sswLL=da2dE^pep8zGTvzxw5a=B(@-tK7;Iy`=BiBQPP zNwpwTG$w84p2YMzR7>d?a$%bSaXpt|HzU*ZnWN)Gv1@%};X2ZKEy zFDazoXQycPqjjAt@6pp{HYR8Z=2#|`*rkyF5cUljq7oHDWIDO_OimIm>!qn%7?t{M z&Yog@B$iT5`v6b(RkwE3%KHMR|8f0USo!0yr=Iu9ou`Y=Z*Jpd6_cQt4v$|e1##5U ziVBMSUnds@;#K**_%3|d{fSX%%85OJvwlSRDEPP>OALE9`3Y_)vrXpcXFOyqCAY53 zH+~aeI8Z|Kw&QBvn6S}<0#$M2EK|3+;#caiFw6-F z$|$bkB1J5<4+T#1qR=+z9y&`-P0tG6x}M!Zwro=&VgHRX`Y|=gXYZuvI*wp4 zs}AdboEG<;)3PhfNc8jaX!9a|m^|_Mrpk|%MgYdysJHl{a-(|;)N$I{8k*RKkjtHR zkZ+@Y&+YbUZi{BwhrG85f~PSUYgDC`P~Y#oapCf&Pf@>5REGn8i?gmzw^K{?wcj%l z&4Gs$Of-poUVy`vvp3DKdxy$8Wf@$vQ+RwVmtE|6zwgI$8q3?7y1*M(4C@a3;9<3N z_Oq^U#?DzR`A_V&nj3;qfKggJVHkqb$@YvT42Xbndiny6@}MO zLG(>U(s#md^}jX*cgQ%;-jCY#ibA77yZmZpJ`U!y|CCf{Ldt0WxSO+mnw>_LCsbPM z*B7^a_89tXSdnwnV&Z*aW0rm)-y5I+wl1zQz<@dwYi5JAj4<%{mwWgA4_R;F7v;9T z4+GLD9nvK&h*Cp$2uKS`ONY|k-5mmwLw5;CcOwjqbTf4K5bwj@=X`&ApZ6c&!G*0#S&#~N&T2Bcc{q2R;}=l)v5+QRqesVgtHK;_7TMwQOtGVZ(3#F+?EtUi~IS6 zMDc4X=RO$AE(C3}e#bi>dVRvWb{>#NAYav@9Dh&-eg~cF{7!`n?t~G(rt-U`|G`w0 zMs)7Eq1P~OLCEJ7flT3@aO&2GgL_nwX!&;u;)0y zw7MqK-$gQvUnyUIh%8%A!1VlgbC%3E_Ty{X1iSp8X_{9;M9g{`T~{=SfyyDOwE8Kk zxOZLIHx9vNxW``@-P*7i%*uk}`e%<~mxS;;#F7PX@%}UX4X4er=?+@W-JqbMXPYVa zZ;LM=`I1&n&%o3C_0FAeWP(0=A=6VF{#KmBPL$NixypSm(UmPeYfI>q`Dye2U8CTs z*||_%T+s7-2L?XNDsk#-j??y0*IZw$V5R>d9Cd^ zwBV*$H#J?74hWsn+Y%GC!_AZ1bjAyA{>NZJfAHzlCbZN{Ty=}F4&M%-y!Am(=OaLd zm(KOxIsL-;QiREVNa}fRJ-mj8u=&O%*%EcLLvdTClXGBK;pf>K;lZuvEUI54 zrG9*ieBCPvhxxuXmosf|Lha6Ff4fTrwBs8cqoRUq)!_;NpGagquh79&n08Q%OGR^5 z8W!g#6vxU~B8L~qz004PL(n7H@8Iy$AFs}d-D>i```H%qyT<@NVOmF7U8U#}y1#KY z1tRFGfb}Az&c11Cv<-PzabgwiqQK9_rOEaXB87Fqz#={iZ!sqb5I}YVW8Y}v{#BvO z5md_;HsQoHG66b%uECzc(aCAtVFc;xV_i|9{IMiJMf!7n*RlHB(3-z*m&TJLUFvnMk*yOvUec)G1W%5cjfY%x+y?K z$aJg>aH@Mr;fuZFS-F^;{aHx3r@IN`xrkIBnnyol8lEnwg~yOHF>&Ev-hW{{o9N?* zEzQN0BF3oSL$J^iapp(Ac;!Xnb`g!gK>4tUghi2y?FDOaZV8^w&mGqd5V>s9f#|~# ztdYJS9->16U-d9Co%dAFo@3yP&zXYSy`wwMh&?ie1S@`@X{Pq|E~n_&M7@!i^H)hC zvp3=YFE8c~cbE#UTSbs$J{*{w%(Ygtd2=CpVEscA^`gMgU|((dvk|AQqG;!cBAZ=b z9iJtA;+pGfjbH97+4zM_h1aLL|GXRH+7s-2oT$%lfaC!Dsp;umA{NqtL+?k}^4??1 z{pI|5ZH4CxNA+wX&9@cr_*EVsjvh0dj+V+Q2K!N=E&_HNsNd#lLcOk(#By1`W;4+GJTc61|s^MGrTTzajwNqh1Qmb3x4G7^smX9?8MNu?~a zSvR<16X`-EaB_vmfmngGRcTmiX9Z z9zC{EJ*&U6l?@Gqxsf$j!GkZfKX~ybw3}FsCG)?3_SBKKJJ7&g6ioB&=c-li8zM2v zcSvM=52%C(Wxyuhf*sIhJ>OhF=|RE}64&Mm5sEQhwl9Z){r9}$lbM8u|3SU*fBhjf z26kgR9|}H)kUHz-a+Y_LZrjw0Hv>f0)1{x!{VDH6u)Mu@@{(34G5Gtlb^iQB^)(4B zK<9eN&;J3+aP-q`jOKRwS8LwU!`&GK?-y4Yl5BNG2F0K+-q!t*x9}nm1x5SN^l-a3 zymiXgizpT}M;l;CCg8xnz|ZUlQ!mGSW0;jQBDtj4M>QxA!M0)h^|GJNi7aZYMMuzW zf)gZU&Gzy$Z-%$L9LI#9>``(iO$CuW+>Bh0M1ab6`AXFlRNTTp7bYx_#(+5p>40YD zfQP-a-+2&yr+s7Q*LPu>STr8caivu6Lx}a`u>svx^sS7^x){l$)7mS+K$=e{YMw`1 z%~|B>-ODWhK{z3=vK$>^Oqzu00bo<2R5Kcp+iq0t7ediwJJt|n4?QR;H6FRY5mN;b zfhx|`y$6l*kUsni3co;=Cq~%@(iO|~%AH&GbQ1Opn}@;2mNVsfqKwUd&`5?ZkCvl7 z?nIecS@Wvjfqs+q?wy{VI&5|f#50tY<=;*mve$ACc3$K9Obn`_B0C>Qhz8nVDlV_( z<-`&E#w10utrLpXj4L>Fq{_A{jFZq&mdzn`2QXHMEGNeBLuW zRud3n22Kov4rdEu;!JYC3pWp8M0V{$5|jIxokIs1J5Bq~=EPoO{Bt5kh?pPo?3@Zc zv(omuDYfcZ0;{GkjlHnd5c=p5aJOz1hs?l{Q7)pU7^j*K(d{_+*fn7?l6Ch2jG}FT zG)V&uM5UcEs7pouahK~hm_;VIJ-(0zTwLKkJu3=z@79WL!Oh52)k!|M ze;!UB?BqDae0fZzk&6KnXPLQ*e+GfxFb4NY{{&uy_V)swd2Zl-n^@*)(_i~5IGv;X#rh{+ zBl}C+6GkEEfZ1;iajv67)q<5(zxCbp^bL*?R?Z5)+n8QBS2UTSK!X{7VoS+oQ5XK; z$%ff@R9Ef;sf{4A{Ue6?%E-7C8J2S4YVnO*KqNw$6AH zCGBczZ@lfD`^J3FCU&i>6Spk+x{#>Bv$O)~Y~^UXhvdi`Gaf6)co1SS0TKk>a*1s4 zde{68t8U=@wMI0pUZZlmjPLD)_ic?Arltzb3;j4t>?=9$vRsGmpRhl(_LB#(fzY|##v>*Jed zh`gv>DU<;2mH!Y)E91Dy@0^#sKzV=Ur3>(KKKCqsg)UGW!I1GgZw3^-vUwg$+HOlw z`&MlkfB&KIy8a>vWrSjEqRoPy@0@(M{wRE^;Hci#s| zim(!D_?dR5G|oqzke-vKn$l zV{4u>FSn(Cloz)Anvd!63%PPxagSsh6Vz)x=r;)Jl@MdNoDC9v%$~#_RCxI~e?-?r zNMB>?mblyl@@)yR>f6!wBX zj?BC@3Y`G7!m7JNn>K~i*E5~23nNzwULyzbtH5UrwNKp8e6EM#6%={N3tC8h{Wmux z(u^ozAhN#bX;8Bs-0uo1{?lAsT0rbdh4Q?oqlb*Qfvl5nDD9@f^pE5*k>K!&oZ{U!mJ>l^*W{FqLaUM&z98>DG;BT z*+kJ{5;AUQZ@x0p#1yd5^0o9%exW)Ax32XAuVzTOPc?_8u2Ld-S{FRtUqJkJhdXkyX7=>lw|yxSYjP&c$@B~)^CQMZg1Zr2_MynmgJL}t}a6Stdw|)Ok~6~|CTl!phmPI?%+~w zS{yJbq1ni!CHYTiXPRv!HjZ}j*>636>hrRCd!-gr2BuaIdNUvY7)C2=X_1zFs-h7? zQ`gom)rj|A@kI&g+W5TM@TChGb~ycx|6>btkNwgZkWA+PY4}H`a5Z|v)O#`vGGO_P__fcN7p5^s(Y=Cp+tI0|22$d3-**GiZ|3ON+zO-2h!q6w>dMmJ#m7l&XPdD9 z2{PqQrMx=0f~0&%!9-ycF45RDqDBgZad=A<5LFqd=wL*j#V-e)K_URn2|ixC%Sh`R8vcF3Ve;E%p3IS)Na426`%b#$ zXsSeG0*L0781Zru>HWBVBh14hLNJjhm$dukm<;!56K)eed4?)N*gsNFn!{y7Xdi2_ zYC9U%OqMT_I0Vfm7YeZ1v?**1vs}sim`xcKZC~T|+{tZq+d^!+iD)RY|H5o;u_rLW zpo9wYtZ385i$22l*1lL9zH1I@&6q@t0oZK;rIl9;vMe+WD5ydke(~D=OMYu0%2nJOt%kQPx``Nk# zH=zZVW~lyWe4yWLx}FISL}^)ae^yDl_@0T*g$m>hjd6QE7hij`R~uqGa$@y;S+Y5| zw{7v09>yVL3EGneen+H#eu&;Ha}KvYgX)hSU5 zP~&&;E2*xW>GLg1_J<>F_w#L2u+34Y^5;hmUTz0_+SbhLQd;#Lvquh|6&F-9L;MW@>?WywQzF9gH7U}k!sSl4=wsM1tQ;fzz!`k)uCt4B$*PK06?VcvPDeK!l7#`RFU34^baC&02LL=L!Qxz1lY z$IU%>-h}6TO1Bqwee$di2ePuE+m?DT5{td(V|~I8 zwrED#AteKnai0bTdT)5wtID)~?OG^eDz6-0DSE8oDLe1&?@wY-++X(z=l}Te1Lqrm zmIBCyJuS!XWNr1;BQ42bSUu|YF28uk&T`w#9K)9bKF`mTO2mN`rIEb|8ae_7h5YIB z*Y{o^)#%HMz0^-r&goz_?s-Dm2rB{mkd#c?3L-_3;h|s>hYZu(XFp8|ZUo|m&cCoI zmLeP(c9y3nT=!Zg_$^dh#^W+#5%{;f4p?>!NzAD4JI;b-60&BIezjO-Js>~F7F(MK zlSKAe#Io3-pfjgOcbStYdjCD>WXC}@Gljj^@F79rIKQmCRgIowrzhqDF{3p!9dHMj zau*QU{#?K2*XO~%`hxv0$fPMEKtG==}!eb~Q+M3x9kw6OOQ2tZ^j*_K5C3*ONKW$#$MRsjy}I&5S7gCy!qV= zhFJ7?cz$MmJ+8O-P}C!``xPH$g2OR#+Um&nmZM`8*g)FG26&lp1tQ1PZy4#M+jP-S z59ycODQ|cYC)gMPtu_E5yhSg0zaqlD~E0D2gYp#R0*Qo zP*(uFd^mfI9jX3j6lx5-aJX`?bq4}^Ixojvx-G95Qq)kuo^_V*Ij0KGV<`T{=H%R* z*~*DDpnom9e4xuZx zy3FXhmz6KP4nSY2ozcyCyn{SWre~HuzfbbMA=0e-`dNF%`u|^6mLx+hYa@-cnM(KFj_FbM6tnVT+#bhbV%fK%xu0OF^vuow-4{<O-ok8L7EMa@f?QpSdaV5c$3rWMSm(4!n z1#Pp?_o^Rhs&DtZsFvmK@k_8~*`A6;{Q@G*WM*GN}2jkB?nUzJ52fT{D&| zcjt|dv-aMt1@-+OHwO}z8`$IiW1so&@AdRcovG;B_7J#H(_p$*l$Fr|A`wjGsdcOs zy~o>`t@7!;ry;ZDjTjNAje%dc3FmMyGczMYAulg4BhwpvuL!%ek+BT&XR7o}4eUaJ z{`AAdS}w@iz_?>unM!M%DPFjuv?y9*)noV{ZZW(qX{B${9CaFteHLhtOT zN5;aB{F2NYTsduK1QV)nK8#y*N%A3R{c3NRmEc$*(+NN7@G9h+2je<&lr@F>DKWv= zXJgY7Wb~wTNBc%0`CIBOHxz{tZGJJ=5Vhapx{ZG@FHAsRU@Po8@TQ{00k)0QNsa*C zkXJ&OYj-U?YC6wZe^2hd>HOIE;`?oUN$cK?@Hc3Og?okRCBKXAd=6JR1X6oItZVeE zpio#^KbGL1qoz8<{QVji%rJn+7+G6e8(Q=D{rh(=o!s|rSLv3tiit}to-n)_%~grc zEXI&djz(GxPxJoOZ9TFP>fVnlE&DGg9hT-4vzNL5*n1Q1+>jl3CKmKKsn5<2odW7W zBN5vgeQKPv47ESu6)LNVP@UujBug5{1xCBKo-t>@DH48h%&{f|SgqQZQp z+@2P?DR~kWQ}~S zb8CZ+TxHb>Z?IW&NINHM?_wcV06*~EE`Yp9aRv+e(Cn;^!<)lZucts6xBd2p{6YUl z;~zwZm_8p)ay91Yjss4?a0AVvNFi%!Hpc!-b#G+CH`GYHx@fyXPBsn1EZdq{xo6eR zG1V&x;D_J-7Dqo8_CxX;J|!KhXj5m$Hm8swt^kge!Writ0W?67RXd}+mhPOxDbT9N z{$lX3gMlKi9qeC2)uAZ2g31+6o&x(^(E%5zv<*;}%Vm&({0a@$Y#Px1EQp}XJG~s7 znO+R;U<_0HDUA7S5x9G#-C@~$r^AEPDin2I`~kfY;LM|Y>041Q?w!$v5Td-);EgrG@5Y*_rz7&5?g&$vPPk9DH@?nnVD zor!}m#kI33P0jNhto?;3(L`GsIqcggk91jCW+oZ_6_aN+_*_!9YzJa&Z=IN{Y%?Gw zWEOE>L7zFmmw2gei1qAoe81!=1El%+37%a7vvzr)?3aAjhft8X?7o-fcmF<3N2l~1 z0&^<9xKOSXBE!vm4&5tY8Qc;7DHpocqf8vObdXH0FAf@brDm@OZBcV{~B)$OVcW^dlmg=C(M(F6Ng z%;)z8w{pXdTP5WWHJdT{+7!Rm@iBEkXYi4`tXBqCY-m!Xr5#_l587l!jVcGVj*7u z7s836i~fdQwLAn?)bTdYXv9IXopo;o?xOktm5#gQHz5rVE!4bbUjF%LwIQf?|G?Na z4p&Bj8gTna*yi%dd};y=5il>0t4?&<7rVn91>PS3MY;-I=W0n7;bn7Gr+HCXWwi-79JTEt_WJfq!w8jZO4AAX3S@ zJot9PMw2;EMsy88SGy37`hteplI@?KHcHE!SXa9&)lx!N_Q%azkKO;!FHix+GTsLY zTAWo{!tfoE%4*9*RiJm`M8e$1phWJB^ppy;V&#FHjqc*ijr5o-L`-Fpf>$^WmOck& zKX87L@6#yc5x3up`z~0W6@(XTJwhhywvv;Jwc<+KYUn$oSI|L}Szy2_Q-Y0qltP3* zT>JAL-_IgpCCwCRTUZRut?5IL2VC(;1zBl6p6B&p-xuwT@RIHscM#^Tz$7 z2q|+*>MrCN4D5C^2HS+?S?fw(bGWRHZ|cs2KpVvm0?}wW%Ut?v{guKjOx#4YxY{Jr3IK%An45%^kO4Z#hbm`q!rXEGBN&*Thz96Ly`93RJ&5DTubhxN(5 z@BH=)9QV2e`X>y$)dY~u_bKpw_!{0ssZxkp^5}&4N@nTf7=ta?GS7{>{fZWTl=ISLdm8!!HuM)w^8ehc*eHb_DD>RE zzVL<8jTd*B+741Y5h-^iGeJo@r@<)BzPsxe$XK3;d-0p#c^#g0V<4_oBeCwy@qwD9buixf7g>-if~)Ag8X4;w zMzY$jbDaP#+pl>VjLE%UAvTgkC<+g_ZLIj>5+SWIQTt{${;p5tR2i-T| zBfS(yASn@!gj(r83aug7e3gA$Tnk~vcth*4SZ4M`a5B^X$*0%ScNPg4AQn1c+LM`S zO{pwx|2BZ2!oLBugcmfldkl#1a`z3+Z(_mES7jHp6SaJANtE_ahu_D3>t|hQr^F{;x7*{v;NiIE+)w#tYi@pR&Cb*-h7(_Jz%qP9zZo zFuQ+-DEaV#>1{dGy&v+uKuLdEl=340#}u&M$i*71VE9@-cg&)^pN2u*u$W~wKy{NP zI7iaEn(0sKEiNhUU&B41$U#P(X~KV>C@!5!azXQc;CkVaw<1D_uV8XeZfJ+3HLN*{ z|0Fe}4}|(4V=a8(OEOuOZ71Lhrpb11p7)oH}+4M$1j~ALz+3$p})(IBu;yG_B1Z0>J0lT z-k*lW-Cpahhi*jd;7_u)oL%xMBMxxKUbi-v1GBZs#0Y>qw5CTLA6K>R=Nu!=#|%@ zwu)R#7hBk~c&vrsJ1`k8AWsd1JX?pg*L~wW-dIz^qeWIimsHw&H zjMu!a-bTo_ifuv%LVcnqt~;8P8q_5&2z$p(7M$}%!!#+cFlTU6#w9kMjKidyC+5sB+r@I+EV&*DkBX=Nb<;<4j%fC z!STOdZ^& zZHt9@qm%8^frm!}ecp=<42%2b-bkQ*2`p$>x+yj>Rvo{mq<<`)nx-#9qfU>02vaG> znrO^LWacdH(|SRMp=tdlFbnx(R(T`VR5WF&woHSR%pG>?QFK~S-Vr6QR2s`03d7*8 z=}QF=d8b90%VjM0U%k;A1WgU1F#Mn#HiwIXw*!UW6XUV)lD;xaf77SwGnWQOIsc@W zKW?6yF1METvB+TXPWv7%7C0L;xYY3B-&-6No&hj)@P+onh7{qV$jS;*#rmRg_{C-u z7PM7{pdG8-04#*( zQF-@HP%}V-;^LI&}=rtPt5NHHL;lCpfy_QKGxhd2T@7r=v)>e(9isM|SE1z;w^s7+iV4mJ`id2kH3K@OqZ}yr7X*CKUh;A^vDCd1HG3xdK;u}qLEk{N^yddqAV|3fNTqqZfeN_Ry z=*_RpPT;RciqA-if)!IzpKO{J^)d{E|DOW*se{H2i3TpO_4QX}Yy`eS8_)dL=_@@; zE4(vV`cv_X%5+~wQ2yJc8bU4uItOOHkCOkb5A(QU^LSP5_`AkP9kS51` z!Ps9=FxsHLWYv}t+IVj=F_npF48xwxELdfC{5gS|vphdhVlCVAYK&ZxRvbFS9@`!N z{eQOeiWq9Y-qv!j0UVqN+EG=EFxbnBg~K#xX8u9Yw^LfT zp)oU*ktHzPU2O?zs@j`2+(X1!O8uI;(#xl`=pG!VER1w*tV5G zBPN?jNrbJ6F)W${iwlk4SGO{7qM=CGm3=dO)rExi&`Oh0RFI}!02wke18>@7_V6Ro4FbD=lHkW9cjQa`K!HNV!&!EiwZekCP!k`= zJH*|}6=A0Ctf>~G2kDF1$QORE_5#~E8L%zB)iSrCI8*5Xg1qF|is~@693`-dfF)wy z&*Z{V@hC+ee!t3V!``~*g)7#~?D(g3y#$ONM$C)fN{s&+g8j9VfqO~>wfhwy<4+FB z?Z~6aa&(O4w#o3D3htuZQ4OBTEJQdm2DGvzjZHs?F$2!IGbu+ke&*Px0IhTsWIVwL zY6Y~o5kKHGf@ey)b2L3(pCu+LuSo+Zvmze}i!z@I{S~W~B{3~q~z7qjh-gpCkc~GtW|MiF0w0VDobY)i=p`M zVYL5TT0|(mpx!TA>*J5C)LblbZ8JL!!sLaG>_y|du~CLRfU)^yd?IfsnuF&<|9c41 z4lQ|cRCLU*f9+VY_IQ!=u^OmVwJ2%?EBwJigAKv#J*+g!Q!c3sW0$svF<<{%C9BQb zK>BF-DKAj8LQXZ87R4!3`9N>`ZFn7%ez`$-WYVXXB70*#tPlrfq{X>K=` zXEMH;%M5F-^2)(*!rku*^I;M1-!U#r!7T4eA|pBRka`}zta>zhU%p-&9QeBLbG@~i zb3T8IjJ|e-7kxq1>?r3Rtji>narsndeVJt8t`d~AQoZb`1q}Ux)B{C)9MT6zjol%) zDd!~j#jz!uj#muC>Uociw)Lm5MWh>Adit}mWqD~X7gOPXsmXz#K!O(7!)6Xf^LezR zV_+!z)2;kp-?#8RY-2-EqNl*!+n61w+pTvPQEFpx1lfnR4rMw^l^+IreGH7fnQ%1CFORG9z zQXc=MghZep8f5`jm+jm)wNPiP`9GsFw{1Ma?dx|Nv4PyR)1*5Op@~b4OA51wn zXwdDIadhQTwN(&oFbnDO+}u#;-oD+HwwJsKAhpWvc0Z(hmpFLciq&!~uzB|&sY&`W zyBRdjlbpCF2d=(FPRqLNr0F~1zd7ybswrv+{KG^NeUTM1puVxoF*S9vPx}PslMxVO z_{8^irWIdJhQ1&2N9kk;)b4ty4Z6i+vbBVrM(x(@@(pE}rG3U-0%#64q+_eRKoU>DHTf3N&(U%EJ!Z?TbZ2Sk!&2K<-lF;foI@tJ z$)G9bW~K1mLsOlMDUw6lHjroU503_V4u2E=Ff!YyL%Wnx2kgXuA7LVKs&N*Fw}Smb zVt_?UjL{1pbtHTiXKMEE8&exiRpRw0!Bzk-V0GPyy+OxU%T_y?J#~<%VjSPj4dg=Thekj#oE zZ0|1}tx}jz?~y$)pnj-0;<*3%6DVLHDPD@riXnwZb3U}B^QW;s4Rj;}cAjJegFgrW^zJqF00XiL1Nz~eF#NXU)EPVPor>7Hb1TgR^f zi;}plf7~;A0?|Ookjk-kE&8v|TJN2`s_`bxTU;)M%~cb;;9MKy$>Tv}mm{l3|I;k# zzXlfO@S0JUlIVlaN4kz6t}Lp_3vF5&%EW~;5DOQq zm?U1dbPbyH`waR@Ea{pcq+zM0Ylse(#^IndL%~JFDG3foo(VR6mKXO==G^2B zkS{*Sy&E^@>Re9m4UDF7cc_bbYPSB1XrJ(o*Aw2!OspY8L0DY_GP^Q7hmzvo86qo( z{fdwKt5*L^%6fQC(xY8IFbtOG(Eekl9dc(-^tLl{@ix&>Y()iIDfZHrr6+#pPsAzs zAuuuFyh<^~+WZfE$A{V|OqEOzKlQJ!;e^>Tal+qWK8ZR?e)L@2qtd7LJ9~_Aob&nh zbK`)CxAKvyWobO96H*oas%kyHOUD&A8|h=_UuSg#apn3swKKaGD&BYB#imXMM2Jym zll60=*PZvW18Ky%*08IWWO-K3k-HhXZY5>kQB`?UKn3aM3ioN&y8H;QQXPB-WeUa|TQJQ8*VYy= z&f}=>?)xuW`upQse9f(H)dgznE6+*?Lb}ek9n)Be(@(^o^+!s&kqXl2^k73rYYi-b ziqc)u1Nb4|f%jCCN>~t6P=M$5ltxI_uF=^WJ<(Q#g&ch$^>Y3Qg4JQRcq;hwi|`*n z@Kh7_t*#F&c@ebOhfpZ|KNGt4Pfw1(PvZ%#Y#~0=@LU66e~BRfZXU{yRUGfKUJaDtk9p$0y=Nz>&deD@l|kS3sx3 z9UaoR{aAWn71~(V64!O*ROf1*54@VEnbN{mFzut+P}Evm-_}iX9&@vymTa2vasISeznAd9qc#RTJaj#?l!JyW-?qkj!wHM zIE6y;;5kjt6jkeM9&K{nEhv}tH_(2v@a0zD2_TEZGM z-~X?WDd6vcf%@bI?#{hW#vaA)EeH3@N}SLJw3YBSD1WrBRmy4zoP2@kUsJ?Omfwfs zVmc;s!JrXOfQJwt3RmE-^pq|XzAX8CUG@}>w0xW#Oj08K0wZ1Ey(`wEpE{)wwuY6` zI5r_xfi(B&gXb4@9Dv*=HhvXQGHLG?8u=pfJkLLPhrRRD`diF&;b*3=gL`;f`KGb;_v3ZLF^(epsl z$y_v`IDRlUOa2QISrzrNVOtQty|NL#&Sa*pnen{@eY-+XoXRO!2A9`vu93WMunp|K zBy`(H7a1jxF0g9*J4a@>szd z0^I6Rho^hly~mK!o!OSWN;b%47feI?4q0(J>XZIQ5+dPu!>qYh85?*UC;fXybTJtU z%;>L}KF`{wivq?j+1u^D_d1pB78P7v${p)i6T3R|Oc=5d7kPf~sJK*;opp1j^Ta>` zwKO1v=IDdGfCr9b95*S_hqaw(`)5{`N4erkG1aP1Jg+aB6kMB};1vTV%b#gkU%{gQ ztQZW+MSXlBKLj-{JD<7eR>IW-Q=QX?LkZ>xL7hdn3H;1xie$le)@Hb;1W7daLI0zF zd5Hzw+~ior%FL9T$w(78Y1Sd{@!;KgAb+dtYeg}+eo!nT1kim%a z5c&qCLZgsd4!_Xmlr%)pZ=H0GU=IM<(a~+aFElq#aDXl*xIhE}E+uf@5SRK3j9U{B zcBsJznrd4p?GH!}F*owoF)cT(54qs!3lQJ1d%f0*~L!o}tT%yHhpedi=6MF3xNqmuzAI435| zx`ReVxh6zKnoxyZxXMupq(nJgJ#>V@>k);Au zS?-8m7;_Z!QR^wX?CqIijQ6H2h_%zSP&D%6^%GVitx#^2t2N=lBq1ZVkVaV~MjwBa zvKnPBRQbSna@^+1~fao-J|V1F}g1=>_a5tNFVRA3UFVEqBEy%rwb65gKH_>+9z#^$xkCeHvN) z1@!+~bZ;`S=)RYQ4n+QNVteOznXs$RFL>|L4e(uyI@$+^FE7uPK&N zEIN3wpXJ>@nAIl5X!%39ly(!JKru)5ZkhJ=L*z6!YTttF_moV5J0K2jsC7lXW{@WB8q*Au-JgQDwY|dW3U`!qjLI z7{Zh#rI|^f*55H`?(ZgvE$@c`AdHT0m-$gf%P@a(p+Nai2pink51F?R$3CN=H z-mz2RO=FXNX&o6}93Y%ULj8*&;@hlJUMVEhA$v=IQ9en)Faw2wG@$(zeO7K>JeFkK zY~*lRIrN8`D+D8&KAGkCit@st%42JpDr;u@pfj90PO4b zrNvb&Nl7O*kM3TiKM4^L$8Pg&ElM^#2B)3o6{1L9orS2v9B+LOS%{0sSmhP61tcs& z{%5FjLJxFcC-krHvG7C+8>i>QI`I0s_Eyp+yPR_Z+U!5+{FL_Uges5C)y!j@4IJ5I zhuu!((pF3a!3Fo^$b@ox!<(~3p);Ux(Z=IiH>KlJUKi?W4=fz+?dLYM&C9T4cm`D~ zA@Q+cAe$Y|3fR}m_#?l zT;d-Ak7L;Q*B@s-89N4DRV9Sa9n zv*9h6zq=l^nv%V)lck0xU3LG@xls{%GA7IEof&Q4 zGu1YeC{ssLbo@H8orC|6v9AD%Yg^g{0zrZdP6#r%LvXjjLtyX(CnUIA7~ElkyE8}# z8X&kk1VSLVyIWvzhrCJt|K4-XeRW?wYFDufncAy&%eTJn)m_tPv!~YZ+yL9|DQMx9 znR(>ZA-lhx!RkKr*M^=DrcvAFg$*~N71nn*#w~cKcdJ1kclvOW4L7;x+h+Lh<0D`< z48gA#&`*Lie`1Iuf^Je1WF>UX5;pl6pz*PmUsYR&VtXv0%vT}nLu%RJkd-o&|;ejt%Z_u;|(XGhd*oq4)kzROpV zlQp&DFPTTT3h32CWmHJ)etr0&;!E27Dj?hJqtlC{j$?|p`M2Nr?kKxcYzgLjm#Qbq zRW*e3PzV{m`FW0e%Hx$AeRP@f_Qg7(4_((C0&lr{-WD%&fAGX^KOg$KOpKYj(!#`+ z(A3;a*S~ct8=4n4Z_G1(ttTs$j^uC!ANpobdfS|8xy>FjJbngZt*5yYQ67{FimQb^o;{orNC_V4^dj}bOm z*lqz=3GE3+@wz}Qne3;AD;qs-#dT>?eOfKpbXP#Uy=H;a(lC#Z98-khi4C`^$&Wmf zuSn%8Iuyj;`gQ^GiwB9VBx;bcg+|#ZoP4)l^+d~F;(QL>7r8!OO7pcr7}xjSNADA< zXA~IvN_IC&+$LHq@#DDa$a&FTZK~c~9!{j4tzFZ2J}o%}QkVoehPI))?U--ec!S>{ z8%9vHNo3F^wu1{2*!lSgsQ$vSnlmF?H;Zg6?1|7dRLU8 za7(daWk<=bkR^_NT8&%g5L_hYJCF@KpoY3oDjS zoQ4J4+d6=I6F zL9ct4*Xwy*w8L=n6W^5g!m^-I|ubeR~<$o9-G<|uXLztp=sAo ztsFfcbmOz_Re$6~r7l2P;h)=z(P?V&=w7WbYGprE$LrQGiW&??%5=X(Xa3ceh06|}Imo==kWUiTS3tPRR z^42_u9K9^xT&F*;;a@kVydC2iQom9x4@o%M^m1q?d3$`hwJDM?ddJB*%`GbV3h|5n zO)tr*X;9ZP*K%zl7?IaA@g#&&xeK+<_Rq!;BmZCvH3-a2=<62mTHVucNhJ#2@|;y? z7DG4ea%=q3m4OjtrpayD%q^&a`F#)k;}^^CGajPIXfL?j)EQXhUE~}yVE+5am)QbA`+L`tq{_~|?$gA=69I;#caC)D6 z6Mjl3mJ-uCm$tgj*(yZ0)Js=4O|G!r-z*7$8d=DnKmY5<(Pv7}^Q6@-g;Je_fnkdx z()0PRMjnl75QW&anE68*;xY--Vp<~R#blMEw00ZErBe-J!I5Q~zjQFDnzhDEQ zcD{P@ly$}jz)Yr6dg*`&CZcd4B_1iP;YUEi$1ky6Rn5{b)S z*=b53op?aJ6afZiGX^cNK?_i0UAov(+uAy3_^EcT)M(Dn!K<{s>|C#9(&k=Uf4-?k zTfbIYKdwn@fs{xtmOnTaNtcEVQ`h|359Z(;n)dDG)Hfd9>}|vm$ZFqVVK_WfSUCHX zJ=(8P%V2YJ?0iqTz_jJj^c!w0CY9cnM@O9ERVKMJeDED4TgIN@UoXK)2b6{<%+6R; zdyV<^bgy>8$W)b;lfbJ@Fe-X%!Plp?M;McDu*6SF}@>!>P5hz!PmiIf2yiG?u( zEnC2_Wg|An$C$B5e_fG4WUL%Ij1(1_Y+ISmgvI8baZM$yZhJ~Z))>)so*M6!9Wt>Q zNA$1Yt76)sISs9(A>|i5O5e&hdoiXz#vx&12OtIGHhX$WpE-U6h8O1MF3VE@R1e16 z>B|j8d#}IlZ;^wYh0P>RFzQy*xwv4MG-vNB&sPuZ4rEpDESpLk$X#ib}}Ij z8QmK=N1b?AtOap}k^eqPy5q@e=Ku=#v`SsdqMD+Hwe=~y$kkvWj z&P5gd4=Layr0^D1Lg)RyupXDHR(u`1{V0p+@Q+HydkKEqbEM!Hlc`PT?~;HA``KR2 z)vUF*H(&ULbBY3oO{3!xLJ0Q9$_?RN!549p0u^DUIAMHrA560erxzA%$Kn2D7_W`N zF-zCCN~{OV=5k}orq$EdyE%Sv2C_e36Y+@<_Z<_od=tpq(Mb4sEAUBJ!fhl_fpqo8 zINa1C9lb%awCi{$7M-RBN|k%Rx^_YaG%<4A`L0AHbs}$(bEJMJY%@43S^$P8QO{s< z!4Dc3%er{lQmG`hSfcue%LLf?<$ip-h&f92uiNfS$dE^&m-_2^tK^S@;51_{eAWZL z`Fx@1jd{^L%;$vCwG76Wj;=JRBh+3F5;xdws>lVEkw|G|#2i?l1$UFk5alA?e((pW z7{FeDLXm>Z`t31BZD%u&@2nMp`vA+@LYtw1{u2M>)+C*)kzC`TkhdahmCL8CcaUHb zk`HGICA9c7HKp-{W-KS3_94@=j2ri5HFRKRRPedfxT#X(mH4%g7)UbEpab&}hRbwI zm&nG-LP0kj-*Z35)c(q+1Z2yov8UsjAp`J&qtVl)w^5>YAv3(Zz0pLZ@e8-UO@zf( zpE0oE@?^sP`^VSmhbaj+pfwxbf_0vAKknR1+t9XKjdh{FOd+=8H6RB`gk<9FyKL5U zA;}rfysDpfXr|MSGr;!%Sax#9(}vflf%Zk^$3Grqshs_8UX%9_c>3KStuiA7;jiOg z-%01%0kZTC?BUYIqH(QkxcEu3o$TJD=`KrvPb*s3q=B`MI0Pb>V!g`0#{%Q1%kPnh z;lqTzwGbkBXQol*4K)ZEriRD*2s8_=D(MA%S-i#5>Nj5TdH;||AuK4WuMvmPhHB^F z0#HpBEHfD30cm}VA#h*Yxk(QwMqO~LU6Xd=-mWWS^s9h{aror*WYsO?7gVYY?Ymir zMb~HSb2tVpWxyrX>@~D^s-8Gko9X#O21YKWtTP#r34s5)$kN%dmWBoe`z{o0y1HgWWe4K ze)t-TPZ|zY8aa7IQ0eEFnQ;41!7J&s97Jx96f4gCi|H-8hPZ`lIG^V_S+`?^K#6UrsIhT=(AOLdi`|O)uc_+6Ka)x%A2KdJ|eH!Fu~MTXv;!+ zy}lhmOB2_(OH%OD8>&e1ZG7rDUT z66;BKQZXga?D*v5l3YNeL5k?2ygv2A7pBM~+z415rpmQjN-fOg= zC)1TrKt-kRDqcD^zY`GaDQrJY*7`X|qs?m(F$m+hVE?rc5MYrHOHWr;3&Xbm@XTH{C`}iBzPt4{s(j%Yeey>*F z*x|!?)~176!Lf_g+6r%?U_vIqAVij=Arb~XLs#SX9UyWtxKnbnv5vf4BA$-MGec?l zZZ0UOw2BT_>TM$^o>*fm0#UMRhr&m}VGSNnLr;&Nosh_;Psg#HDR!X}QvG5Kuh{(p z$ZgdV_C=OJH@)%7mmvr3YGar|?xc#GvJ`L^l2NCe%nk zI7M!si;CreRInOHO}yiwLa)nL-8JuD;v1O0#sol-AJzc?EQgUGVq#}z!B@6>R_Q_B z?f1c-EV^w@2|vtMl~gM)^*%|MjsrIkfQmx7%k#9yLqi(aG_G@%=EA{yr|3*(N(#t5 z#KF5SUcAU=62^bCZ&SteJ?)5f5vyWM5TDAJiqxTNoxkP&aE!;}=gJ%1AP|QI_(!}e zhTy#}snJ%V6+|pJhXRsgGfGeFZAlMcK|Rwun|S}qYB~i6MyD{CJI}*QZ$E3r?2n4f zC*BQr<>R{i0pMbSb5$HIr7*2*_B07s$MVp2I(Q-C!s-g@U;ZGrLx@HS*)`KumX9*@ z37&pAp5z%!V?hJrr+S7bZ70PH+mv22h+~Qte zldGbzid_lLnk#%|N!#TKR;a8sd-}BM(RIjjDEf{>v&mtATx=?co2S~fq2w)(rm*jW5Ygo)7=j-r zH;^W~zYM3*;pV4)+03=}@>i&qH{t_k$>y|roE3>3YM$*xgp9v)D)@C?mCUT+<0H%_%P;VSx1@aZS z@~WSqCFHz=fsktJlXt@6Y}K#iY)ZxVHOJjwPp*%TB!}nTH0QtRDE@i0yiqT z%<`iTj_LmjCy2OWB8HEMC@@bttmxj>cCCfYMZ6d3>s*I5#J;`6HF=-IAN$NUpW~cZ zgHsJ#78yxj9{Y{tU0N8QHAJF*WkJSBe19$^!w&%yo)FLuE{;zfuu$zm#nTebe zPh~O*Kte<$7dVfPP&jPuxMjrWia(df)1^}r+09R~A@D*kjpyTP17a5L%ty!0)MAS@ zQW{Fz+;qVLUA$V`yGMQDK?Npb!u>2IT~>YtM!Bp6yRjjs6GsqY8FUTD;`X}6Ow`B9 z9h#a!HlLNmu3+vG_k7O&Q)!H6u*{=0KkjRHev$5JzK#2CpwZkU#Olbe=<3kbYj96_ z_53Qed(UI7D~6cF*IVL!fK9EjuP-gX;(iAqs$HA)1piMHJFx_q0s zYzyzv3I6mTNyJt;rJIm-@ai-D)72wZ&JjeYrcFc=qqhSHxB>E;9Bm-pGBVk634CHp z+!p0e7y5Db<{D)TL?sfHu3A$kMme)sgY#yTHv^n^6@x4?5Ze@t1TZMK23W9qe9Bmm zmWv517EO;y`*QIDqYwx1JcA~*&2yvZ>Di?@Lkmcf$3vsSluJ52`J*K`%$ zK!Fh76HFH{3d|{HT+mK3fAu3pz%8tEpD2%cP4ayAums-Nz}0;_JIz-n)BIf$Zt=v~ zZMn`pn_~aN%t%ViGCSv9(rt1syp3L1^{<@>{c|HsH6+Rkal(MUPk4r`FJ6<>uestA zQB{6MNlMIU%A9&vF-&>%^`^H?!|5D~77^DDH!j?~EqnDjY1s-9&PSz@gar@A3nDC7 z#bjs#OUUMoN9n!hWGgF#9!;);NiTDQm>cR02>^_ekCN8igvKHba&gWz+AUHP1x{iw zG=*3w1Gm!vg2PfX%@6X$qzgtH&&m|gJaFQsHEd(3MA&1f4>>byctn*|DWz8Flj2L} zMc?`9KkJHZe&;*DeR^UJoZSBS(Zz%0qe@3`%O#Qco~J{EP<1R|&XhrcC@`R!8rj6QU{ zJ;~9lcKI%fVReA(P*2B|?P;r90AnE6BOkb+b=9J-Yx3;+sT$}$Rc}EO;c2a#94Y3d zsqPqz@GbyDRNIC+w526n>;lX9EzE;&x@3DV`RfZGtrg?KCbz?FhV7l!&;~d5tkWX* zrVkJYf9h)#i9s8pO!VO&v%8e$U^+EZs|JvkgD!{o@Z6*_<_^>tHDa08kOapT3u!*> z;Sm(AVE8bfF=B?Ni0iLtmFK2Ko;Bl&Kz7u#UYao($cod?xv<#t6hV)SDE&r)%XQ%+ zlqg;<)RrgD^Q^4S9`2Y`_eTwOm!fCNQWKU()Kh7%!LiRZO> z$QU?n`!1QNiog?$*K!F9{jYY4i;1NnUVi4?5phBY(xelOOGB^)MCg8yn8u-^5t&N@ z+X>U|XXhZ!2<`7C{E;lE{qt#>bcU*I!%=a+TagS$i;!fn`wb@~Mc~6?2AAtwx3V2I z|I9pPUS2a@^Uy(m$Xs@Z0lBD4vDDu;)O0>7uluBH4WL~PXFbs)`YPd8 zGEa(qoB06+>niecwACv6n7P^eeo`rbWNLXB(mSKBZq#Mxq)Q;P4EFSDY3zwM;|h zO?xUN>hDQw>Oa>-B^iH#EwQa_^J`r%KFOi zoOTuCAApbPH{f$0fk=ws)bk=h-BpKIk<$~>aPh)s=Vn3}gWgl&RP8*ZfwE@hBcv^! zh{Epf4N({K49Dh<^U3NLGFrd=5w{`|=#yut$jNDM&fPnh`K-+#l!sJfYdPn7gC zMszJ-Si&gXP}PG3EfyBCZB5>ym?DK?2z|>H`g$UzcpKJ6Uu8sOw)VP;25K&YZIZl(#22eG6UL_$5`SBix1>B-`5F z(j~gD*+8Oog~xYpFa>{40sj_YiO__<_Ds*CG0*O`1fuaW!vrKSAl4=74I5CQ`b&(Q z5oSt@EIa3@UHs&l8c&&dpO^*`;hP5w@CUw23^EUlNB4&dv+?Az%!(jpZ>m;FpJS&? zRMTXC4t`=&)e^-{z62&Mw*F_sBUD@`CMKV|l#Y*Hx_94>YirOJ-@c$ny9U)dBAibx zK*nY@ZmC4>9njZ0m&}U_Rh@(Ek%N|T21dY1Bs)hYpVFSXm^rmw387$(WF)bKY0?MzPq?mYz2ZNL5<10aYJ7Zf6?zLqDnS4copPrl_DNtkrMp*vaD z+e5Jr8BwNBp8NPqqH+yOiDH2Ze+6t?&?T2{uFxEKA)F%d>yb@4Nb1MS!uT8ncX5Tk zgD*pghPD={@9a)e%eL&Y(or=a;4F@<2Lpg@Gw=!&Yp&gVAaQpcCqFB5>ZlZ5#YZEOfRZ%_@A z728=i}+^U3;LedRpH|ooAn#cUON~?jB(8p)QjsUCQPA?J6;pkU%-wVoF*w zyWdEyyEeSSOVQ@1%F*ITxbQnyI-gLIf>y_0T)r3xc=TY(>|3E{V?ENNQ>hZ@pPsUu zCCUdZ%=7%hC2u54IY$FIbg)YhPv<3ibD}DY&)Pmc8?webD(Q46MHc60$7?;Ej*h< z*cU-c|C`W~B}WiiaZ_*CHzCJr6{Z2FWr8gVM1Bt;rr8pRGMLp)WAPO4vZpZC`or1% z%cJtOLXuw?W{j5>MIo8s-!>nBI>DyVy8PTSwFO-+hVSLlM_8_ekwdFM&=Fz1Nx)Mv z@!VeAb8HT_q`W3`;R0cqgjimpH8Xwjt{ru&rGh{}i;SIK5sOWg;CWbGS<(jh1vnw? z1Ew>@n-c{`16GNHFtRzPmZ9JQ!H-Z4?N^0&m{cEQv?v6KH%V6-z=2^qV$kaQkO9Fp zA-9qD{MyjXry`hKiDazQGu@-S2c4J0MAWi0T-&7;p5VdC6AGun%n*aRBdsa1I*Eb@ zSaZf|+E2oy+{tA)Cr84g2IEqzr8zejb6R4RmbAJDOPVioSLLUoX85a2f5|LXik=_` ziTp#Wz*Jvk7~?Cs5Lc@mFK~b6qmUR144-1-1#ybc+LOoyUjY7yz_6f^5qy^39$75tgzhT#ryYNGq zFnXV5BS_d`5SiX1Zp0}X%lHWpwp|EQp%k;WTYtgoc@f9~RUHzp7m~>5>O<0XZb4Hs zgTtjtocqEN`!&nTMa=K}*YO;1*-x4;xKnf(rRR4Xm&fC-hz{6zd> z{feUww8=kc)Vg22QY(n*A)~AKEk!FhX1#e`zUlERZ8P_qG}yS)4@o5gl5FcXuR0dZ zJJqTu%Rc8qntvH^Mmg={acjUjYJ@=3apPa=bh8$DouEADeeEzUX&oJvlHSF9L85REPE|8PS z(Pv@s!lZV*dSFYgxT$5~B8|do@6vM`b@C(qX=J!qfCa3BZd?G9ravvsGz_D(SQ@p@ zBy5C_va0}hdkd0}wfa*}O35WmpoWUqV!g++450hWDRNl4=3ca9@EJZo196}wt20!U zZ>>nD0MDGWCBH*C!E-(zQX<}GedE7RzM!$~4uBiXeCf!&5baL!3Iu@~i z29jf_Aw`86Q4G0s$RwV&y7xA2W*xSIawFT1-Y&k$5f*j^&=!Wi)0+witH_2gGJO)K z$|XDkK2So^h}Kpm6B$54BLKB?lf2&~cLx;UZ;?U5$CKNY&P1gBB??$#lqd0~tL=8m z)Hb@oLI91cxgJ$rQx%& zVaQo?q^U_;frog#^xYW;B>(%e%nPVw*X?-KF!9FsaKxjKNZV!o`_Xra5@80hT zvc$QpGB|q7qX9QjX*;xVm1DBCyW!J!diyDg%5KpY;^LDL->Nu+-z_ydwEvX%6|F^u zNcdPB+7e<}ml0kG6XQ9Z9*q!_u1QNc`l$Y9N;|$_7BQKDzf(kbVM}VbkdNFWKD#lB;!LMbqJ6=xQ&NH*`6f`3&lT7(Z#i9>c7rEy7C8=I^ z3sz){Yi@-?ipBscK9FhCCx1dNnI;MVsdlc}#BoJxg04nXi>lq?CSE~YJ}8w0#M2wA zYDi1Gg=V@)Qd)Wfj&RC2|(pk2Do*~)Zn0)qoK}cxhZT4IftkQ#}6V&dHa!qEPuUN!5;c5p*<|zt!Ng~G5uVXSo-(wg8umu#tZ=Wa}aQ0hDb9I!+H~s82YhuFm#zW zjO3$q+mCV&CM7Hhem_fE)Xx+UXcLY1dx-5<-WsFRuqMMX%nJ2UX`IoY0=93zOTA%$ z<^`g-%%=VbRcpdd(0f_@I5620noQTL$Z!%;a1Mc5tAEnD<4oy}a1)f6U!f`Skb45K zt}xvhoOs8gs)~Pp_9R(BOQs;ka?=yr+76&oeN0dagL%3 z&~Uvu&T*89!CpSj1DQn}t+i&kj-PK})ZbL0DK0#(eaV^$KBl$i2k`NNh-?9YYqXNlhS9FVmxFO>nb1TFp z!m^X1T9QG*gBhkTuF=*yfQ8v_rj-24k^>Hqkc4`F_X5C3cP=tGS)thr2s38GV&(s+ z825C=rHP+~5t_wPKNFHwm!9#a2{t5d3qRH`*WTvG`LHE&&hJw^xpji~>B8{G zaU7zFMY(F()#1W%uFx`Mu{i-kwCWpnKnNGQZ&Zw6psJ zWS8Y$P$2p`%wR!Df~vFnh*FK~&3|}GHUv%_ovU6I;}UyaD+w{a>=vJ^RmsAFCxBA{ z>O*i{kCw&edh`LdjY&F%)^VE%i!;{poF}y%qe{D0@jL52Yl~32KjLIT)dHx5kcD@$ zJ?o+yCXlgX4kTg^E5e~o)JdMEd0=l53p7bX!slykB8TNo0%S{j+-g+yz8De8jwT0V zUPY$7-|%xi-|**yJ|7Mxefb^|z9mD%;m_=;b3s4)I!5{-&b`G+;S za1fG>O>_cYw9p0zv0U6MKHR(_&oGkGlp=gN&EU-)`RngcElgH#>z(HPBbr@LM56Bz zRKM=!n)>mggbfMm&qKde^5)E{+E3c?Od>t2bk_EEw(;n=S{-y}vf2Aj`ViR=B@zwp z*9Ba(g_m)>Q+hG~0d~Iq20PC1sIkiZ?i8EL!^2h@>V@E8U)7&Q0?pKKQuHi4tHK8G z1|n(|J}X_%MeaIJZLlK<-=c~ETM651?o~v}Nh%S);|bNX#{s}y>6xY{G#j`Z`daD} z1oGk%d5N%AoWR7lBB<)v$&HnUw6de@bm4gzxL-v9OACQ@8EZ_6yb}6i_D)^R!<*}? zg_8#iF7Yqr=HkTiNx5^+8XLbVtBK3l0K;{pDJF*@!9%x3I0vTe?@$XH^tia>~C51%F&o7ZdxyF`gGR&e!_rKips@!VNORS5{%L zXT%Yr4>zO^yyTvw_3w)DC6v`wbX2Yr`xW{Mj#3Gq4 zA!Sz_8-G%CvTVrbu3@h}sf59@

$!I2mw!c^_GJxi<=Co=RB^(P`B!?-+sXBE+B_ zB4{4x2GRM%5 zk>(IqUlg8qB&)4n!<1cQUy)Z7UQ~lfU5$=Gd_dk)@KBfU&1>Ek)ze-xkB@m)6S6@| zcHX2w#nOXb@zr_8uHaaa&^b98wQ))^%RGZ*X7)H?M`&r>|ziWo4_ zvy)iu$Nv`0jF5wU@DEX&()?b0hpugX{3-j#E&k43nj=A-)A1`fuTgVrU(2auqo{wT zVL@0~VDM1&L4f7%LalOUMQbFj{^I59#g8~k-GmRD8cu^* z=-^e^(=qi{i85O4CRM^ZAEe8*OZ*FaU9p>=1X|EvGZO|*s2gtLPgyMamf8NPC#E3? zJ$b#mOOt6a&HJpU)**XVLxz>fpi-wMGLS@EV;&crb=D0WQ``pK>w zQojunv)pC+pAQ3~B4cP3jCq5Fs=D=XPznnvi{<&Id3x0Guqe}ZsqiER-SnDMQMwV1;%VK@5tQ-GnkG`r=jZ`AzQaj3!Cv} zQA5syech+p!A-w*9Y5841SFfC##(Rf^J1C3@DD2-n0Z1|w8aSPN4B=>2LFxKMerZy zC|aNL8r{R3?9*+^Pq|_|^cAmuej0;PIL{xgY<4}mCJZax_%=X@XI;#kSxGXvL~4=Q z&sa!oHhG%hD`OP-hfN9S#2-0w3I1V&(VBjFNraM$mCFhx2diV`%airOt62~S9wkRR z`xvE=RDuxoU454I&*$!EAL9+{VC3zw%;n`Vf1MX|DnM!Vj4Ua%D#0?h$kQX_U`G{W zezz%K4i?lP7c%zRl!YC@rfO$Q?n)-tq}uGfoGNG3VoUG?chMWr-Sc85tw*!l*-B%x zI2O4W+>B_0_x-YxxooPcdr4N#(;{!r44`ufOkX}HRsGqz{0r0fkRZZT<>c}s5q+Fu z+h3c8jCg7GIWVTaSJ_PZ&&XB!Zx?R|?x-5f*E)aB=`xiJSVx10&WnX5|DYx*6gnxS z_G&$!C9RxGJ0qDbnn<^>5qDsKMOg9k3OTolgylG6U|vv+id2J7(=oxp1(9})I;Ebv zV=?iqx|3FwDv`y>Y~sfthLK5@!|+o7Y&i7PaO3*yL*AE87v4iUw%!0D^!NGn zU$#~F*Xl{i=05B^8M-VE+FJ0v8+ilGWonTC2Uu*sT0KuA209k38W{xNWuM(p*jhOG z{OxTASkg*`u%>(RzT6%=KU?n6^zEoI52}amm(-q6^$g>!VxB1nw+MDV>8*O=;<9mi zB=BLJmio)?H>HAs+YjW`Q|Wi~f2cDuCloy&svz+%mT{gqC1hAMD!_P8wKGM@AtbOg z`ibw0PQLbTW!bobml(NyY&Erx;$xeBr!l2Pqim0BQyFAw z6^|9!>P5mC7TXOlLZ*@lO4MNXg$tP@bM&~hSw;Je3g5PbS5ga1(9nLl79A}{RSS1&0X99z9#0XhSwJJ^ ztYgh1{;~y8H1#_Q;LA(7$d#l3AM5L-3qybQUoRzJTH%MbbD`LcJYf=D%mA?@i% zd0aQKPioMC1U$-JnVj9euA{wOI2rE^;p5RZO(BOkPSw|RJ)Mphm)ld% zw0`Rr_3hHbw7h@s=Qn0oEuh|n;uA`h-tYac%u^w|Rs@AUa(`Dkg!b*o0ZSej*tAHnl$G>MB#jPpJ=hwTG^Mp;Nl*L z*%+Hs`VL8m#_wk+B8hgM{5+3&4ZJQDr@!K0gamIhmGnOh7fLd1SrT2mWV0c$4A=VF#QZ$FVK+k0&OM>n1%yJo@5V|&R&yOw7$t59?p6XoT@z(#)SYBTxd<2gWN9_U&!Gs`-um&qDzB@!4zBe=tGt+{&slnG7#>SZz} zgu8f(3YeIlb|>?bV`T+{CFh@FI}zXJ6F4bPPu_q@>dy(ma{SCPgMkH0V|XxraOXd z$2jn0nc1(+$7In$zwV%rJ?l?=c6r{%nNRYkobqBv(=%#WUKUqfkWjyw#&Bl;7!Sf(j6%OaajpSSv*v3^*h{W1eOO>&bDXfdc$ z3vbx!%1>T%=<|;s$+K3)d0cdb(sz;5fjWPE@E->nAdL?tM?U+MV#LGLW@%M*;cY6!#;;DQ+YYrc6N61FDFAt4IHH_++1N;Ibg~A8Uw`W2gdCwwYuHO zBeFrq9M-b^q?8Tn>+ucI!Yqcq^`_oo%%T$ntx>)CB6xl`i0og`%rwa#odw}z26`H8 z3L#nuXzcF(Hh~(K)IKRpLu0dBNw6OG_HYfE<#rU^ex>c z@!oVfiS!JkE`$-%ye~P@_3!NTasmqcVb=Z?*-Ue;-wWH#@qLp?nv%*8bbuYx+F4u2hlgWn0<4(lQR_ta#sQvL;0JAcqn4Lc z8Qi7%cRi1E@TqDJ_1~9#;}zKJ>!JML1W5-O!BNBFTDche;r*wzt`+(73tWi1zzW(M zx)VnM6QT5Enn_^~jf_-!h36D<6G|=yfek+einC*x-BPECyEJ(pjnemugSHJI?_t52 zu-O@T7rWAxwBTWSh0L^q9pO-8)Wg45Ljg-L5X&lQl$hk<;R$>n%fHMTOXT_qmUdtJ zsnp7ggM)){yGSRPdO*O3vfGlK&C5K<`&SL`Cjk~YJufjy$qY%kDJ~J+_*DwS5sz#Ec zsMGn^^q92to9s=z5LV)zJ#|G07Ne288IAd@)DGT< z)Xe$tJICPkWt!b?>Nj3HW5bv^j`WnIUR1(KRTIF7O3p1r-J?2+m5|rdYddjqkBQLl zZ#QE2&Zs`SjPUPuERp2CiJtQK!UeiKx8!(Z97(eKaM|6qGt`fRr=0SA(yJ3`d*{M) znDWpe>OzHuRzhNaip9y2SDAtb6PEQsiSU#751Kw2qY{jgMln$YFV|2EDpcnX7kaJS z+S5QnKM?RdJ?aIVcNNEzM@%dwpC*xVDbKfkxBu5{@go7eX`iW^Xu0ltnRb*L$!Pf* z*}O`wui%PGt{Ku*dJ~`QnjyzKc{f+rR3htDG5I>VKR#(hGN1?`b^fM^GHGxUL@e|| zMTP(SdFXhyl?1p}&2E>cRNSUrW>$V4vxZa4MLas}qlr23`!TaN!8Mcp3&COTkBatP zkG={v%bG*=r-rkcs^eU28xAh&m*e!Piu))d74ox{^Qqe%vSn`?Wp9qFxx6tWvyUCU zE50f})AhKH|6-9%-EAtazDU2im}$R^9i zIOa_ZSvQcWr4;hxyCiOaJE|&ld&qVRTV~7u`Tm8g>-q`rU0vvH7Wv^tx8$R`KjEcP z33%rB6Thb$w*tHDp}gx2bHI3dwm*8x13Ov>#)OFa88A`zK*irZY)-8XbLzdbbJfPy z)?@v5EvnA|C521$oDJERONV-rHy9F)etOu!mc_I7BzZX!KH2E7QaPB78#^Zm82k@s64D99v&V%4ohN$ zW*J5}F8@5m_V;{Gtkmn@k;C(5xA1%9Q{Mdid|3heg&p+W+%PJFp={mPubJ6&J|F2{ z-0W9;KOL6jGBY#tJ#9Y<;Y!`Tg|{1fckkKQ*lflrP;r0fY&{!$|2iI8#j~nH--e4`Br!5>m3BwkB^T}5R=gAUxgfq!do;;N=gRn z=>Lw9 z|8w&{>n+hjKo=JG1*b54DgKv+==|r{t%#}VW{AOS=AJ;m8qk5E`=Z(0?-7Est%{Sb z4;GTQy1zltH{`u}1PC}@g#8mW{u@62xmjfZXrmP637=1BZdUeD@fxM;^(aOVs-*FJ zN_lPd#v6f%%SMf|2^n9;!hPUgRrZS_EF2ufF@1F71u6-NiHS!0dc2`kQMDYBN`K2O z{ii6EfqSTGPry1l$vXhth=_;;#7H{alHy!j-Xcmzv;T)i`C`hd;OF}eC$jxR7yc=# z9ts3=z&*Vw_U6BS5|r}ycC@gd55$eAB(@xM69%XLBQx@M zxtfL{IJB8v8yL$!@}hq|3KC`yMLm>c4E)Y&MHNM-@sFRxNU>q*!cV43nX^iGpFC=Raa{=s%8q=GQORZ zYPbdYRfpMlUHptwp!RWnJ+yN#gI>F_@lHumkxa+yZEfY!_u3fRd!rw1WTg(VzkEFm zuTIInvt0DP9=lw2Z@A4vj597!&Dv`xRQ(?e5-Egue{|78ccLR~1Jfo)FQ3d&910tI9n_xel~$yi!+dY7 zSeTi~>=PDzp{V7JTjcPdyHK2G^pcW(P!6Z+RRR-Z>=cu$It%0Am`z9hlYtNgdW_kCo6mzM2gCc+l_um{G%F7U#fmZnL{S0Zh1z9=CVcV^Jcn^zWho^1l#=YlBa-CMk&0J zQAUZu``*4jyy}p{Aw$lT9hrPR1B1-q(#6B6Yxn1<3-sXg5S z;^49EY?a*J-rit9g+Zf6d_n>poIY;@>T_#4S)v`Em`KE6LE=$}5QwDD#bGmiAP#X{ z2>G~)Y~J_7RNG9wL-tNi9pu=Krx)d>Mb5Bm8k~F}43L+JH_(5t7JjQ-kdPYTb%J)8D=( zwR|n$zxta0-vZXpoIR=_wXf-81akL z#FP}}7cauu4I0#@KI_u;_V&`$uLgGFlPMb8&ebT@+0Bzwt$s^j`fiqDNa8F^8qAnu zYwUBbE$q4z+puV#CgIbR>GbvK$!Lr(--djgx8aclS(Uzm8!|lTW4P$fkH;tfr@iZn zYO34T(gdZcND-7sSCFDeMzFnl@3xv4Iouo zXd#3UIBR41)pPF|_u)RCJTQQdm<-;IR^@RJmowkK?a#MJx3_7~x2AmOZacsCjyw;U;?({xp8n0Md4ea6EA9JJTY z`^4KL=k#h>*IOzHQ*q{C#UOL)e3W?dkVF>(c&n%IwG7S0vN@5}L&tBaZ5E6|O3~9B znwn{E-*(YmzI@qXF1xrXoKNkE-6uD%<6@tyO2E}-%~8T!dfJMcwbU@W(s)_~({-*T zxe-Ee#m+X-V(dczQ>&$;<6DA$g=_vBfF^6iMcdu}&J0GF%C!+k`&M8x_V&C+SW`9E zQFoY?(UOAn{d5WgaTcr>g?SwgGKS;q`#w{gWp|hOAVZoEfS>*{9mp&+;RoMJd6|*X zJXYs7Ne$od1h2~#T6t=;b)h#~kJZGsjX}T@Ea2fr47F;t(B92Kj<0CqOvJG*PU->Hfdc1N&|lZKlp_C+#5*Tgz7;9vp9h}s z)Y_T02h(i{1wj27Z+wyH6A%zE)v~&}Ith4Ww3M%Qwf6Z0ifiM~U6mjV_fZ=AKwg^T>M)F5` z2d0c{e3H5#QA^;Wp(_%#v9(pMJoL7pfNtkNtHb(n8Y|e=f<20o!k z627XV1dTR3qbIuR3{e=aU3id( ztR;?)ZZMcF+^S$?@TNuY`UWfh8tzLSL*)w7>z!(%u_x)Xiaw~@=KH=EHL4o&UVlY% zg~JV;8^0hfmoJ~A;t?`b$E>4o{6U@B0ivkzT7l5uZOu2LbloB)qfcr+V%!87aMPCu zY)#ujYu9|f6nGyp6dfb3WLicyPlj7}rBhY1XdP&6Ym*cuTgQ&ryBz(B1_m}p-1dIK zhoP=hd~n~6ZmL+SGZ$YQc7#+&-NlMimsBeU&oUmgo1n2;gfol7Tt__>o|exX zoF6D`rhyAGOioVXWKZBiHPyeSr}JlJX3iAGi*NFa#Z-NcB8P3Qu@|hc;wuG2=#|f- zR(mMciU#k|mIa>a#TM(TsQ40|t<62bD|Q8qe__OWnRd~85f*~h=8$)37$4<&eWkUF z#MgShJvQ?^-xYTB1V@f4#J@VO0Kf+25oFV8K)!iH-T(L_hM?{FvSu=#0Y;Y1j{$5+ zP}r8CQ6SOvLNMn!n+qC+n=8=EkY#qZR_ltDip*$+P&=gXxH!4v>?8xx0e?d^Nbt!O z5qi%^^ZH>L82Ec;=Qgse97s5HTt(z;WhVEUbw+~4eqFG?)ozsXu4J51Ee|2abWtZI z3^i%V$DA9$gfUsY$}UVifnef^oV}PIW?;ZZyO?u{ytweyj_w?l$h6Z2&8?jeuck-s zQAE8CRLwgY@4$o93`Y+Lccw6ZvL53TyS}W6w=tkj$ze6{LqEI0_M9F(hEY3n9=5$i z&ExTkI>)!JxVX3(itZe4>dw(E_?JYAlC?f?Z$F2RxvgX197zc347gM+vMJ|b*i|+1 zB<$EU^1d0t*9Iw8T?t01jL}d$ay{_~3iU(<^$`G04s##fC=@P!ij0a}LwBG_duPOmynHflBq)P~wyuc*y;88mQ z9FliKyVT(ZZ8FpG#`6Jj43Uqg6f|j_D`(EkJ=ZuGEa^bG#-$It*nQTg7PVE6iZQt- zo4tx!djX5(wnqpLyJIW)3%x~N&<8z6ycoG(N1r+t%g#BY_fj}Kmli|qw~l1X$_8ub zWWkuNRCq3aNrMe0ysRKyMmms_KRZiQTVy~WSR*SI`huQV0= z&b!^)+dFSDE(piJG%1R@-F6PI-|_jz$X595K13s5G@Iyk<_kOERpvMLECktWO@UxH zhUDbp;yaY&-1=9H7}tcXzStx3&B3%R?%<;`CWb6C&L**jaNXLif@Tr%F;Cr+*%tBk zcPy)pp`gYaIb7y^;$l| z&aF=+C`cm?mkK}xhOL^CjU4Vl9%6ll{j^Z&Z76x91L@3?!KWbO{UJ+7Gah2 zL472uT@EWRep_3cf41-<2a>!Om-;d5!jatA>t^rejC_QIRh(zDG7w7nNPM!fH!Z8k z*|%?OCaaod6vW7N_h30jHhd4lqlNYOAvvEC!^ll4%L^VUbD?OwwMx-0cmLcq#-u%1 zuR2TVPT^Zq3Ba;>Fy9L==%p|mP49)5 zAeleX4Ec6}%7!M|G>o4%Lrz#flODZta`l{Y@wfz2AZDh$y|MAMTu+|F!00|J&YroP z^0vdc&a)J{M-*ac*uJ8bQT6S%h^3K)uz@E44!Px)T!`g{@AwSGDW+vQZC4be<=^K! zz>aeA-P_$$6Q&N3C2YixLY86<#)Av(918ceaJ!ae$!%x0Pkz16G<}xn!)ka!+@fWs z@a^;G?BeQAY1bmX3edjs2PL5a>`5*!E8|sAz~nAKaG0CRiff8F=Uy~&=9+hOP2xHC ziR@xbIEcMUL0-d4(oj#&c8c}zVM>EGumN<7ghfT;=FoA>RhvDzP3Wri!cZaAO=sg{ zaI6aXyzBr&Nglm0@issq$>y*V`*?e zA#t=J%&d90-AAB$8x=S`@kGy@W+gq~sAl7zf;<+o*fLUH`1tq~n0(w{Z5~ep_w0GC z6Q^?1Nkmu}H&`1btatI8X?N8&E|(}g{$b|O#(UPPJ1(v}Bz)ZGJGS8TiH!!HkeJ5n zoB_RMTSpbhkS8O;!nim?R5@Ko$$-#XX}~=-dq?cRMR@%N$tWs((QQJh%Ixp$fEUm zhg>q-u%?Mg>X1{{RfjryDYo^^@X@W+c2>VSuH1^1z@xHGmjVL=CxH4EqH#eS;q*1z z2mN3kx8C)Ind>fWf>q`6Sw6d0c8iXYj5(=?LPzPfQkq9z1x2F!qD%y>7XHoqK}&_z zBal)yqqNU%gK}>tEBlnkbi%{H{hd*K0IQoKg@UA;v?y63J(07tqCgBeKm;t$Mp|!9 z1zL#7JrSnFMu@bdD;XztDN=rhd>#*odw^YuMNt?CMM1Rv!#CPK0EKf!oX189nbdLo zMQC0;NV%T=sK2-5qz)|)8SaFZ=$lWUR8ESBM0Iy}ixst{zgIp_y!`P7(�(bKd^` zu>sb>uW^rAk~>F4gb$=6Q+zV5Ou_3lAcs^x<<3;1gQ2Dfc-H>)QZ=StpeN-RGuBIK zT=I|}hi_c3KQ3l71=rO7w?W@V$kU(q{%8%^Wgy?hwOaTB z(aPfb^n+R|R@Y)uqq@hLOuN!Ef66T0abVi-Gnu*<-X|in+swaz>DyPqPq8nQcRBm~u|9G?C(8c>`6jm3 zZ8||fv}z+&X7P1S*%e3}!@Du~F~Xv#Xjv|4YDVx12sDFu)O>xuA9=6N=ilzz2|SH# zr+2@?_x4{rPDYQ?v2W|{)^c2{CbZg{0nr(m7=Q?TwE>R$wJX-xnvY;%{X-vl2Gx`6!th*7L(71rf_$|@T5wJU=>jdLCU7)X7jQ4KAoMNCf*;VMjmNKi5e{qQt-4 zM8mp~LYiV7wut<;Xg;$Z!=v1vRyN!BMS2R0Kn8JuL@~bW49;?!dzYFyL-efJB zc+@sw1vtVuDthiiVw)XT>*3q^klVB%+5>HkzdAFwI1OQH zli1PQtIgz+%bl5fmMaCocd>AGgZ@kBW+JqOl@{w#~YDm zbyAtfZ@!VoR;*s&tgH2+Hpf)BrTgD7N{m8njl(8^a+GR4-6)Xsw>olG6C){FGD`jy z#;nJ9RQfoBUtM_%yT-DX@~w<6>!KdXQ?!TACn48wYiLCOeg+}$fniF74vqif>SX`V zY6H5k$G+)1ThnkUJ=Hl@C3luw1Z?}oOCC4&AGWPa`H1TB$a2*>YUXoF6$`aN`pxmeSMDw@MGp!1>K$b1 zYMWf;*cyh4O(XZ3aQNC%9iy6^T{bq=a6CdleNiGr{QL6wk+(~Y0A6BdUMG?HJN`Zs zB7YvOsr;F+I^8^IUwv-WRPE|jT+VCe;q3eF!tY;Fg+q4wltuUq^H5h;S8G|7Pe=ua zqxUXU{|%EJw)}K&=gBqH zf#lfz@U(j9_nv%8tU*_AJI8{u7@1IxdnJmxdU{@*-C#{x3YLPhpwQ4!gAAYy%jJ9w zdlI`p2@(P;o|cuBRgC|($sHiIzX6y>(_*aK)Q@?zp>%L`oK%HxhV>h{xs{$Ws?wLV z$IK7!Y?FxQq)nUyu16=%?Z}ZM4hw~7M0AqNfHlTuNiZCO0L+1hICdnq5=Wh4@Q#6D0& zLQ*m?Fza`#Jnq@Y2U1>r3+{>*j~$kH-glHMNv;4p?Ygl zv~pmK?uYe%r%Zj^cUEgm=Kgjo`iYX`p6B)S;;cGSo0_hSFJu`eT3cI_ICb6K%dI#+ z7ZjXn1X6&B4S_#}LjThpt@6L04e`@eLo4Ab zv=ZKqxb^;Q)Stf`V1iQ87o}Et|5v_#ToSUbR*&i1zaI7XYZ5s0{9oYgM|P&+mWV(2 z!auF??E61;iQMWhCvo-g-|hRSp+BZHfe@r`{$ZlT?*{uzvit)j0L}jn%x@F4J3Rky c1GA0dO)4$VdR?Ko2YfCoU%Qxd;m(tP0r%h{SO5S3 literal 0 HcmV?d00001 diff --git a/docs/articles/icea.html b/docs/articles/icea.html index a588b270..f39674dd 100644 --- a/docs/articles/icea.html +++ b/docs/articles/icea.html @@ -6,19 +6,19 @@ Individualized cost-effectiveness analysis (iCEA) • hesim - - - - - - - + + + + + + + - +

-

Individualized cost-effectiveness analysis (iCEA)

+

Individualized cost-effectiveness analysis (iCEA)

-

2020-03-15

+

2020-05-25

Source: vignettes/icea.Rmd @@ -97,7 +117,7 @@

2020-03-15

-1 Overview

+Overview

Individualized cost-effectiveness analysis (iCEA) evaluates the cost-effectiveness of treatments at the individual (or subgroup) level. This has two major implications:

- -
  • - - + +
    • @@ -111,6 +138,7 @@ +
    @@ -139,19 +167,23 @@

    Authors

    +

    Developed by Devin Incerti, Jeroen P. Jansen.

    -

    Site built with pkgdown 1.3.0.

    +

    Site built with pkgdown 1.5.1.

    +
    + + diff --git a/docs/bootstrap-toc.css b/docs/bootstrap-toc.css new file mode 100644 index 00000000..5a859415 --- /dev/null +++ b/docs/bootstrap-toc.css @@ -0,0 +1,60 @@ +/*! + * Bootstrap Table of Contents v0.4.1 (http://afeld.github.io/bootstrap-toc/) + * Copyright 2015 Aidan Feldman + * Licensed under MIT (https://github.com/afeld/bootstrap-toc/blob/gh-pages/LICENSE.md) */ + +/* modified from https://github.com/twbs/bootstrap/blob/94b4076dd2efba9af71f0b18d4ee4b163aa9e0dd/docs/assets/css/src/docs.css#L548-L601 */ + +/* All levels of nav */ +nav[data-toggle='toc'] .nav > li > a { + display: block; + padding: 4px 20px; + font-size: 13px; + font-weight: 500; + color: #767676; +} +nav[data-toggle='toc'] .nav > li > a:hover, +nav[data-toggle='toc'] .nav > li > a:focus { + padding-left: 19px; + color: #563d7c; + text-decoration: none; + background-color: transparent; + border-left: 1px solid #563d7c; +} +nav[data-toggle='toc'] .nav > .active > a, +nav[data-toggle='toc'] .nav > .active:hover > a, +nav[data-toggle='toc'] .nav > .active:focus > a { + padding-left: 18px; + font-weight: bold; + color: #563d7c; + background-color: transparent; + border-left: 2px solid #563d7c; +} + +/* Nav: second level (shown on .active) */ +nav[data-toggle='toc'] .nav .nav { + display: none; /* Hide by default, but at >768px, show it */ + padding-bottom: 10px; +} +nav[data-toggle='toc'] .nav .nav > li > a { + padding-top: 1px; + padding-bottom: 1px; + padding-left: 30px; + font-size: 12px; + font-weight: normal; +} +nav[data-toggle='toc'] .nav .nav > li > a:hover, +nav[data-toggle='toc'] .nav .nav > li > a:focus { + padding-left: 29px; +} +nav[data-toggle='toc'] .nav .nav > .active > a, +nav[data-toggle='toc'] .nav .nav > .active:hover > a, +nav[data-toggle='toc'] .nav .nav > .active:focus > a { + padding-left: 28px; + font-weight: 500; +} + +/* from https://github.com/twbs/bootstrap/blob/e38f066d8c203c3e032da0ff23cd2d6098ee2dd6/docs/assets/css/src/docs.css#L631-L634 */ +nav[data-toggle='toc'] .nav > .active > ul { + display: block; +} diff --git a/docs/bootstrap-toc.js b/docs/bootstrap-toc.js new file mode 100644 index 00000000..1cdd573b --- /dev/null +++ b/docs/bootstrap-toc.js @@ -0,0 +1,159 @@ +/*! + * Bootstrap Table of Contents v0.4.1 (http://afeld.github.io/bootstrap-toc/) + * Copyright 2015 Aidan Feldman + * Licensed under MIT (https://github.com/afeld/bootstrap-toc/blob/gh-pages/LICENSE.md) */ +(function() { + 'use strict'; + + window.Toc = { + helpers: { + // return all matching elements in the set, or their descendants + findOrFilter: function($el, selector) { + // http://danielnouri.org/notes/2011/03/14/a-jquery-find-that-also-finds-the-root-element/ + // http://stackoverflow.com/a/12731439/358804 + var $descendants = $el.find(selector); + return $el.filter(selector).add($descendants).filter(':not([data-toc-skip])'); + }, + + generateUniqueIdBase: function(el) { + var text = $(el).text(); + var anchor = text.trim().toLowerCase().replace(/[^A-Za-z0-9]+/g, '-'); + return anchor || el.tagName.toLowerCase(); + }, + + generateUniqueId: function(el) { + var anchorBase = this.generateUniqueIdBase(el); + for (var i = 0; ; i++) { + var anchor = anchorBase; + if (i > 0) { + // add suffix + anchor += '-' + i; + } + // check if ID already exists + if (!document.getElementById(anchor)) { + return anchor; + } + } + }, + + generateAnchor: function(el) { + if (el.id) { + return el.id; + } else { + var anchor = this.generateUniqueId(el); + el.id = anchor; + return anchor; + } + }, + + createNavList: function() { + return $('
      '); + }, + + createChildNavList: function($parent) { + var $childList = this.createNavList(); + $parent.append($childList); + return $childList; + }, + + generateNavEl: function(anchor, text) { + var $a = $(''); + $a.attr('href', '#' + anchor); + $a.text(text); + var $li = $('
    • '); + $li.append($a); + return $li; + }, + + generateNavItem: function(headingEl) { + var anchor = this.generateAnchor(headingEl); + var $heading = $(headingEl); + var text = $heading.data('toc-text') || $heading.text(); + return this.generateNavEl(anchor, text); + }, + + // Find the first heading level (`

      `, then `

      `, etc.) that has more than one element. Defaults to 1 (for `

      `). + getTopLevel: function($scope) { + for (var i = 1; i <= 6; i++) { + var $headings = this.findOrFilter($scope, 'h' + i); + if ($headings.length > 1) { + return i; + } + } + + return 1; + }, + + // returns the elements for the top level, and the next below it + getHeadings: function($scope, topLevel) { + var topSelector = 'h' + topLevel; + + var secondaryLevel = topLevel + 1; + var secondarySelector = 'h' + secondaryLevel; + + return this.findOrFilter($scope, topSelector + ',' + secondarySelector); + }, + + getNavLevel: function(el) { + return parseInt(el.tagName.charAt(1), 10); + }, + + populateNav: function($topContext, topLevel, $headings) { + var $context = $topContext; + var $prevNav; + + var helpers = this; + $headings.each(function(i, el) { + var $newNav = helpers.generateNavItem(el); + var navLevel = helpers.getNavLevel(el); + + // determine the proper $context + if (navLevel === topLevel) { + // use top level + $context = $topContext; + } else if ($prevNav && $context === $topContext) { + // create a new level of the tree and switch to it + $context = helpers.createChildNavList($prevNav); + } // else use the current $context + + $context.append($newNav); + + $prevNav = $newNav; + }); + }, + + parseOps: function(arg) { + var opts; + if (arg.jquery) { + opts = { + $nav: arg + }; + } else { + opts = arg; + } + opts.$scope = opts.$scope || $(document.body); + return opts; + } + }, + + // accepts a jQuery object, or an options object + init: function(opts) { + opts = this.helpers.parseOps(opts); + + // ensure that the data attribute is in place for styling + opts.$nav.attr('data-toggle', 'toc'); + + var $topContext = this.helpers.createChildNavList(opts.$nav); + var topLevel = this.helpers.getTopLevel(opts.$scope); + var $headings = this.helpers.getHeadings(opts.$scope, topLevel); + this.helpers.populateNav($topContext, topLevel, $headings); + } + }; + + $(function() { + $('nav[data-toggle="toc"]').each(function(i, el) { + var $nav = $(el); + Toc.init($nav); + }); + }); +})(); diff --git a/docs/index.html b/docs/index.html index 761fe0b3..96a22f98 100644 --- a/docs/index.html +++ b/docs/index.html @@ -6,28 +6,32 @@ Health-Economic Simulation Modeling and Decision Analysis • hesim - - - - - - - + + + + + + + - +

    • - - + +
      • @@ -90,6 +113,7 @@ +
      @@ -97,14 +121,15 @@

      Overview

      -

      hesim is an R package for health economic simulation modeling and decision analysis that provides a general framework for integrating statistical analyses with economic evaluation. The package currently supports N-state partitioned survival models (PSMs) and individual-level continuous time state transition models (CTSTMs), as well as summarizing the output of probabilistic sensitivity analysis (PSA). It is designed for high performance simulation modeling and heavily utilizes Rcpp and data.table. hesim is being actively developed and subsequent versions will provide support for cohort-level state transition models.

      -

      Features of the current version include:

      +

      hesim is a modular and computationally efficient R package for health economic simulation modeling and decision analysis that provides a general framework for integrating statistical analyses with economic evaluation. The package supports cohort discrete time state transition models (DTSTMs), N-state partitioned survival models (PSMs), and individual-level continuous time state transition models (CTSTMs), encompassing both Markov (time-homogeneous and time-inhomogeneous) and semi-Markov processes. It heavily utilizes Rcpp and data.table, making individual-patient simulation, probabilistic sensitivity analysis (PSA), and incorporation of patient heterogeneity fast.

      +

      Features of the current version can be summarized as follows:

        -
      • N-state PSMs and individual-level CTSTMs
        • -
      • Modeling patient heterogeneity
        • +
      • Cohort DTSTMs, N-state PSMs, and individual-level CTSTMs that encompass Markov and semi-Markov processes
        • +
      • Options to build models via mathematical expressions using nonstandard evaluation or directly from fitted statistical models
      • Parameter estimates from either an R based model or from an external source
        • -
      • Sampling parameters of a statistical model fit with R using Monte Carlo methods or bootstrapping
        • +
      • Convenience functions for sampling model parameters from parametric distributions or via bootstrapping
      • Parameter uncertainty propagated with PSA
        • +
      • Modeling patient heterogeneity
      • Performing cost-effectiveness analyses and representing decision uncertainty from PSAs
      • Simulation code written in C++ to boost performance
      @@ -112,23 +137,23 @@

      Installation

      -
      # Install from CRAN:
-install.packages("hesim")
-
-# Install the most up to date development version from GitHub:
-# install.packages("devtools")
-devtools::install_github("hesim-dev/hesim")
      +
      # Install from CRAN:
+install.packages("hesim")
+
+# Install the most up to date development version from GitHub:
+# install.packages("devtools")
+devtools::install_github("hesim-dev/hesim")
      - + diff --git a/docs/news/index.html b/docs/news/index.html index 81e28082..802c6d4e 100644 --- a/docs/news/index.html +++ b/docs/news/index.html @@ -8,21 +8,29 @@ Changelog • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,12 +38,12 @@ - - + + @@ -46,9 +54,10 @@ + - +
      @@ -79,28 +88,46 @@
    • News
    -
    • - - + +
      • @@ -111,15 +138,46 @@ +
      -

      Changelog

      +

      Changelog

      Source: NEWS.md
      +
      +

      +hesim 0.3.0

      +
      +

      +Highlights

      +

      hesim now supports discrete time state transition models via hesim::CohortDtstm objects. Users can build a model by either fitting multinomial logistic regressions with nnet::multinom() or with a mathematical expression using define_model(). Furthermore, $summarize() methods now have a by_grp option to facilitate subgroup analyses.

      +
      +
      +

      +New features

      +
        +

      • The hesim::CohortDtstm class simulates cohort discrete time state transition models (cDTSTM). State transitions in a cDTSTM are simulated using the hesim::CohortDtstmTrans class, which can be constructed from a multinom_list() object or using define_model().

        • +

      • $summarize() methods now have a by_grp = "TRUE" option to facilitate subgroup analyses. If by_grp = "FALSE", then estimates are aggregated across groups. A new patient_wt argument in the patients table of hesim_data() can be used to weight groups during the aggregation.

        • +

      • hesim::tparams objects can now be used to store transformed parameters used to simulate outcomes such as means (i.e., tparams_mean()) that have already been predicted as a function of covariates.

        • +

      • General cumulative hazard functions can now be simulated using a discrete time approximation where the probability of an event during each time period is simulated from a Bernoulli distribution. This is more efficient than the previous method based on a C++ version of the sample() function. See the random_method = "discrete" option in params_surv().

        • +

      • rdirichlet_mat() has a new argument output so that multiple object types can be returned.

        • +
      +
      +
      +

      +API changes

      +
        +

      • The auxiliary argument random_method = "sample" in params_surv() is deprecated and random_method = "discrete" should be used instead.

        • +

      • stateval_tbl now contains a grp_var column used to assign state values to “groups” of patients. This is distinct from grp_id in hesim_data(), which is used to define groups for subgroup analyses.

        • +

      • The public field input_mats has been renamed input_data in R6 classes for disease progression and state values. This is a more generic name and will allow for potential feature enhancements in which input_data is a data frame rather than a matrix.

        • +

      • rdirichlet_mat() has been modified to better facilitate sampling from transition matrices within the context of define_model() and tparams_transprobs(). One implication is that the number of rows in alpha must now be less than or equal to the number of columns and that the number of columns can be greater than the number of rows.

        • +
      +
      +

      hesim 0.2.3

      @@ -136,51 +194,51 @@

      The input_mats class now contains an element time_reset. If TRUE, then time intervals reset each time a patient enters a new health state. In other words, state values can depend on time since entering a health state.

      To illustrate, consider an oncology application with three health states (stable disease, progressed disease, and death). In these models it is common to assume that patients begin second line treatment after disease progression. Suppose the second line treatment is a chemotherapy that patients take for 12 cycles (or approximately 1 year). Then drug costs would accrue for the first year but not afterwards.

      State values like this can be specified by setting time_reset = TRUE in create_StateVals.stateval_tbl().

      -
      hesim_dat <- hesim_data(strategies = data.frame(strategy_id = c(1, 2)),
-                        patients = data.frame(patient_id = seq(1, 3)),
-                        states = data.frame(state_id = c(1, 2)))
-drugcosts <- stateval_tbl(tbl = data.frame(state_id = rep(c(1, 2), each = 2),
-                                           time_start = c(0, 1, 0, 1),
-                                           est = c(10000, 0, 12500, 0)),
-                                  dist = "fixed",
-                                  hesim_data = hesim_dat)  
-drugcostsmod <- create_StateVals(drugcosts, time_reset = TRUE)
      +
      hesim_dat <- hesim_data(strategies = data.frame(strategy_id = c(1, 2)),
+                        patients = data.frame(patient_id = seq(1, 3)),
+                        states = data.frame(state_id = c(1, 2)))
+drugcosts <- stateval_tbl(tbl = data.frame(state_id = rep(c(1, 2), each = 2),
+                                           time_start = c(0, 1, 0, 1),
+                                           est = c(10000, 0, 12500, 0)),
+                                  dist = "fixed",
+                                  hesim_data = hesim_dat)
+drugcostsmod <- create_StateVals(drugcosts, time_reset = TRUE)

      hesim 0.2.0

      -
      +

      -Highlights

      +Highlights

      hesim now provides a general framework for integrating statistical models with economic evaluation. Users build a decision model by specifying a model structure, which consists of a set of statistical models for disease progression, utilities, and costs. Each statistical model is used to simulate outcomes as a function of estimated parameters and input data. N-state partitioned survival models (PSMs) and individual-level continuous time state transition models (iCTSTMs) are now supported.

      -
      -

      -API changes

      - -
      -
      +

      -New features

      +New features

      Economic models—which combine the disease, utility, and cost models—are constructed with the following classes:

      Disease models are constructed using the classes:

      +

      Utility and cost models are constructed with the hesim::StateVals class.

      +

      The economic models are used to simulate disease progression ($sim_disease(), $sim_stateprobs()), quality-adjusted life-years (QALYs) ($sim_qalys()), and costs ($sim_costs()). Parameter uncertainty is propagated to model outcomes using probabilistic sensitivity analysis. Summaries of the simulated costs and QALYs are used to perform model-based cost-effectiveness analyses (CEAs) and represent decision uncertainty with icea.ce() and icea_pw.ce().

      +
      +
      +

      +API changes

      + -

      Utility and cost models are constructed with the StateVals() class.

      -

      The economic models are used to simulate disease progression ($sim_disease(), sim_stateprobs()), quality-adjusted life-years (QALYs) (sim_qalys()), and costs (sim_costs()). Parameter uncertainty is propagated to model outcomes using probabilistic sensitivity analysis. Summaries of the simulated costs and QALYs are used to perform model-based cost-effectiveness analyses (CEAs) and represent decision uncertainty with icea.ce() and icea_pw.ce().

      @@ -190,34 +248,31 @@

      - +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/pkgdown.css b/docs/pkgdown.css index c03fb08d..c01e5923 100644 --- a/docs/pkgdown.css +++ b/docs/pkgdown.css @@ -17,12 +17,14 @@ html, body { height: 100%; } +body { + position: relative; +} + body > .container { display: flex; height: 100%; flex-direction: column; - - padding-top: 60px; } body > .container .row { @@ -69,6 +71,10 @@ summary { margin-top: calc(-60px + 1em); } +dd { + margin-left: 3em; +} + /* Section anchors ---------------------------------*/ a.anchor { @@ -102,37 +108,135 @@ a.anchor { margin-top: -40px; } -/* Static header placement on mobile devices */ -@media (max-width: 767px) { - .navbar-fixed-top { - position: absolute; - } - .navbar { - padding: 0; - } +/* Navbar submenu --------------------------*/ + +.dropdown-submenu { + position: relative; } +.dropdown-submenu>.dropdown-menu { + top: 0; + left: 100%; + margin-top: -6px; + margin-left: -1px; + border-radius: 0 6px 6px 6px; +} + +.dropdown-submenu:hover>.dropdown-menu { + display: block; +} + +.dropdown-submenu>a:after { + display: block; + content: " "; + float: right; + width: 0; + height: 0; + border-color: transparent; + border-style: solid; + border-width: 5px 0 5px 5px; + border-left-color: #cccccc; + margin-top: 5px; + margin-right: -10px; +} + +.dropdown-submenu:hover>a:after { + border-left-color: #ffffff; +} + +.dropdown-submenu.pull-left { + float: none; +} + +.dropdown-submenu.pull-left>.dropdown-menu { + left: -100%; + margin-left: 10px; + border-radius: 6px 0 6px 6px; +} /* Sidebar --------------------------*/ -#sidebar { +#pkgdown-sidebar { margin-top: 30px; + position: -webkit-sticky; + position: sticky; + top: 70px; } -#sidebar h2 { + +#pkgdown-sidebar h2 { font-size: 1.5em; margin-top: 1em; } -#sidebar h2:first-child { +#pkgdown-sidebar h2:first-child { margin-top: 0; } -#sidebar .list-unstyled li { +#pkgdown-sidebar .list-unstyled li { margin-bottom: 0.5em; } +/* bootstrap-toc tweaks ------------------------------------------------------*/ + +/* All levels of nav */ + +nav[data-toggle='toc'] .nav > li > a { + padding: 4px 20px 4px 6px; + font-size: 1.5rem; + font-weight: 400; + color: inherit; +} + +nav[data-toggle='toc'] .nav > li > a:hover, +nav[data-toggle='toc'] .nav > li > a:focus { + padding-left: 5px; + color: inherit; + border-left: 1px solid #878787; +} + +nav[data-toggle='toc'] .nav > .active > a, +nav[data-toggle='toc'] .nav > .active:hover > a, +nav[data-toggle='toc'] .nav > .active:focus > a { + padding-left: 5px; + font-size: 1.5rem; + font-weight: 400; + color: inherit; + border-left: 2px solid #878787; +} + +/* Nav: second level (shown on .active) */ + +nav[data-toggle='toc'] .nav .nav { + display: none; /* Hide by default, but at >768px, show it */ + padding-bottom: 10px; +} + +nav[data-toggle='toc'] .nav .nav > li > a { + padding-left: 16px; + font-size: 1.35rem; +} + +nav[data-toggle='toc'] .nav .nav > li > a:hover, +nav[data-toggle='toc'] .nav .nav > li > a:focus { + padding-left: 15px; +} + +nav[data-toggle='toc'] .nav .nav > .active > a, +nav[data-toggle='toc'] .nav .nav > .active:hover > a, +nav[data-toggle='toc'] .nav .nav > .active:focus > a { + padding-left: 15px; + font-weight: 500; + font-size: 1.35rem; +} + +/* orcid ------------------------------------------------------------------- */ + .orcid { - height: 16px; + font-size: 16px; + color: #A6CE39; + /* margins are required by official ORCID trademark and display guidelines */ + margin-left:4px; + margin-right:4px; vertical-align: middle; } @@ -222,6 +326,19 @@ a.sourceLine:hover { visibility: visible; } +/* headroom.js ------------------------ */ + +.headroom { + will-change: transform; + transition: transform 200ms linear; +} +.headroom--pinned { + transform: translateY(0%); +} +.headroom--unpinned { + transform: translateY(-100%); +} + /* mark.js ----------------------------*/ mark { @@ -234,3 +351,17 @@ mark { .html-widget { margin-bottom: 10px; } + +/* fontawesome ------------------------ */ + +.fab { + font-family: "Font Awesome 5 Brands" !important; +} + +/* don't display links in code chunks when printing */ +/* source: https://stackoverflow.com/a/10781533 */ +@media print { + code a:link:after, code a:visited:after { + content: ""; + } +} diff --git a/docs/pkgdown.js b/docs/pkgdown.js index eb7e83d2..7e7048fa 100644 --- a/docs/pkgdown.js +++ b/docs/pkgdown.js @@ -2,18 +2,11 @@ (function($) { $(function() { - $("#sidebar") - .stick_in_parent({offset_top: 40}) - .on('sticky_kit:bottom', function(e) { - $(this).parent().css('position', 'static'); - }) - .on('sticky_kit:unbottom', function(e) { - $(this).parent().css('position', 'relative'); - }); + $('.navbar-fixed-top').headroom(); - $('body').scrollspy({ - target: '#sidebar', - offset: 60 + $('body').css('padding-top', $('.navbar').height() + 10); + $(window).resize(function(){ + $('body').css('padding-top', $('.navbar').height() + 10); }); $('[data-toggle="tooltip"]').tooltip(); diff --git a/docs/pkgdown.yml b/docs/pkgdown.yml index d36e47f2..7c72d02c 100644 --- a/docs/pkgdown.yml +++ b/docs/pkgdown.yml @@ -1,9 +1,14 @@ pandoc: 2.7.3 -pkgdown: 1.3.0 +pkgdown: 1.5.1 pkgdown_sha: ~ articles: - ctstm: ctstm.html + expected-values: expected-values.html + getting-started: getting-started.html icea: icea.html intro: intro.html + markov-inhomogeneous: markov-inhomogeneous.html + mlogit: mlogit.html + mstate: mstate.html psm: psm.html +last_built: 2020-05-25T16:35Z diff --git a/docs/reference/CohortDtstm.html b/docs/reference/CohortDtstm.html new file mode 100644 index 00000000..cfee9f4a --- /dev/null +++ b/docs/reference/CohortDtstm.html @@ -0,0 +1,533 @@ + + + + + + + + +Cohort discrete time state transition model — CohortDtstm • hesim + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
      +
      + + + + +
      + +
      +
      +
      +

      Cohort discrete time state transition model

      + Source: R/dtstm.R + +
      + +
      +

      Simulate outcomes from a cohort discrete time state transition model.

      +
      + + + +

      Format

      + +

      An R6::R6Class object.

      +

      See also

      + +

      create_CohortDtstm(), CohortDtstmTrans, +create_CohortDtstmTrans()

      +

      Public fields

      + +

      +
      trans_model

      The model for health state transitions. Must be an object +of class CohortDtstmTrans.

      + +
      utility_model

      The model for health state utility. Must be an object of +class StateVals.

      + +
      cost_models

      The models used to predict costs by health state. +Must be a list of objects of class StateVals, where each element of the +list represents a different cost category.

      + +
      stateprobs_

      An object of class stateprobs simulated using $sim_stateprobs().

      + +
      qalys_

      An object of class qalys simulated using $sim_qalys().

      + +
      costs_

      An object of class costs simulated using $sim_costs().

      + +

      +

      Methods

      + + +

      Public methods

      + + +

      +

      Method new()

      +

      Create a new CohortDtstm object.

      Usage

      +

      CohortDtstm$new(trans_model = NULL, utility_model = NULL, cost_models = NULL)

      + +

      Arguments

      +

      +
      trans_model

      The trans_model field.

      + +
      utility_model

      The utility_model field.

      + +
      cost_models

      The cost_models field.

      + +

      +

      Returns

      +

      A new CohortDtstm object.

      +

      +

      Method sim_stateprobs()

      +

      Simulate health state probabilities using CohortDtstmTrans$sim_stateprobs().

      Usage

      +

      CohortDtstm$sim_stateprobs(n_cycles)

      + +

      Arguments

      +

      +
      n_cycles

      The number of model cycles to simulate the model for.

      + +

      +

      Returns

      +

      An instance of self with simulated output of class stateprobs +stored in stateprobs_.

      +

      +

      Method sim_qalys()

      +

      Simulate quality-adjusted life-years (QALYs) as a function of stateprobs_ and +utility_model. See vignette("expected-values") for details.

      Usage

      +

      CohortDtstm$sim_qalys(
+  dr = 0.03,
+  integrate_method = c("trapz", "riemann_left", "riemann_right"),
+  lys = FALSE
+)

      + +

      Arguments

      +

      +
      dr

      Discount rate.

      + +
      integrate_method

      Method used to integrate state values when computing (QALYs).

      + +
      lys

      If TRUE, then life-years are simulated in addition to QALYs.

      + +

      +

      Returns

      +

      An instance of self with simulated output of class qalys stored +in qalys_.

      +

      +

      Method sim_costs()

      +

      Simulate costs as a function of stateprobs_ and cost_models. +See vignette("expected-values") for details.

      Usage

      +

      CohortDtstm$sim_costs(
+  dr = 0.03,
+  integrate_method = c("trapz", "riemann_left", "riemann_right")
+)

      + +

      Arguments

      +

      +
      dr

      Discount rate.

      + +
      integrate_method

      Method used to integrate state values when computing costs.

      + +

      +

      Returns

      +

      An instance of self with simulated output of class costs stored +in costs_.

      +

      +

      Method summarize()

      +

      Summarize costs and QALYs so that cost-effectiveness analysis can be performed. +See summarize_ce().

      Usage

      +

      CohortDtstm$summarize(by_grp = FALSE)

      + +

      Arguments

      +

      +
      by_grp

      If TRUE, then costs and QALYs are computed by subgroup. If +FALSE, then costs and QALYs are aggregated across all patients (and subgroups).

      + +

      +

      +

      Method clone()

      +

      The objects of this class are cloneable with this method.

      Usage

      +

      CohortDtstm$clone(deep = FALSE)

      + +

      Arguments

      +

      +
      deep

      Whether to make a deep clone.

      + +

      + + + +

      Examples

      + 
      library("data.table")
+library("nnet")
+transitions_data <- data.table(multinom3_exdata$transitions)
+
+# Treatment strategies, target population, and model structure
+n_patients <- 4
+patients <- transitions_data[year == 1, .(patient_id, age, female)][
+ sort(sample.int(nrow(transitions_data[year == 1]), n_patients))][
+   , grp_id := 1:n_patients]
+hesim_dat <- hesim_data(
+  patients = patients,
+  strategies = data.table(strategy_id = 1:2,
+                          strategy_name = c("Reference", "Intervention")),
+  states = data.table(state_id = c(1, 2),
+                      state_name = c("Healthy", "Sick")) # Non-death health states
+)
+tmat <- rbind(c(0, 1, 2),
+              c(NA, 0, 1),
+              c(NA, NA, NA))
+
+# Parameter estimation
+## Multinomial logistic regression
+data_healthy <- transitions_data[state_from == "Healthy"]
+fit_healthy <- multinom(state_to ~ strategy_name + female + age,
+                        data = data_healthy, trace = FALSE)
+data_sick <- droplevels(transitions_data[state_from == "Sick"])
+fit_sick <- multinom(state_to ~ strategy_name + female + age,
+                     data = data_sick, trace = FALSE)
+transfits <- multinom_list(healthy = fit_healthy, sick = fit_sick)
+
+## Utility
+utility_tbl <- stateval_tbl(multinom3_exdata$utility,
+                            dist = "beta",
+                            hesim_data = hesim_dat)
+
+## Costs
+drugcost_tbl <- stateval_tbl(multinom3_exdata$costs$drugs,
+                             dist = "fixed",
+                             hesim_data = hesim_dat)
+medcost_tbl <- stateval_tbl(multinom3_exdata$costs$medical,
+                            dist = "gamma",
+                            hesim_data = hesim_dat)
+
+# Economic model
+n_samples <- 3
+
+## Construct model
+### Transitions
+transmod_data <- expand(hesim_dat)
+transmod <- create_CohortDtstmTrans(transfits,
+                                    input_data = transmod_data,
+                                    trans_mat = tmat,
+                                    n = n_samples,
+                                    point_estimate = FALSE)
+
+### Utility
+utilitymod <- create_StateVals(utility_tbl, n = n_samples)
+
+### Costs
+drugcostmod <- create_StateVals(drugcost_tbl, n = n_samples)
+medcostmod <- create_StateVals(medcost_tbl, n = n_samples)
+costmods <- list(Drug = drugcostmod,
+                 Medical = medcostmod)
+
+### Combine
+econmod <- CohortDtstm$new(trans_model = transmod,
+                           utility_model = utilitymod,
+                           cost_models = costmods)
+
+## Simulate outcomes
+econmod$sim_stateprobs(n_cycles = 20)
+econmod$sim_qalys(dr = .03)
+econmod$sim_costs(dr = .03)
+econmod$summarize()
      #> $costs
+#>     category   dr sample strategy_id     costs grp_id
+#>  1:     Drug 0.03      1           1  3755.059      1
+#>  2:     Drug 0.03      1           2 10407.712      1
+#>  3:     Drug 0.03      2           1  3866.673      1
+#>  4:     Drug 0.03      2           2 10380.370      1
+#>  5:     Drug 0.03      3           1  3857.659      1
+#>  6:     Drug 0.03      3           2 10656.812      1
+#>  7:  Medical 0.03      1           1  4369.837      1
+#>  8:  Medical 0.03      1           2  4916.838      1
+#>  9:  Medical 0.03      2           1  4461.001      1
+#> 10:  Medical 0.03      2           2  4898.309      1
+#> 11:  Medical 0.03      3           1  4436.859      1
+#> 12:  Medical 0.03      3           2  4985.520      1
+#> 13:    total 0.03      1           1  8124.896      1
+#> 14:    total 0.03      1           2 15324.550      1
+#> 15:    total 0.03      2           1  8327.674      1
+#> 16:    total 0.03      2           2 15278.679      1
+#> 17:    total 0.03      3           1  8294.518      1
+#> 18:    total 0.03      3           2 15642.332      1
+#> 
+#> $qalys
+#>      dr sample strategy_id    qalys grp_id
+#> 1: 0.03      1           1 1.609056      1
+#> 2: 0.03      1           2 1.795590      1
+#> 3: 0.03      2           1 1.720567      1
+#> 4: 0.03      2           2 1.862527      1
+#> 5: 0.03      3           1 1.634413      1
+#> 6: 0.03      3           2 1.823913      1
+#> 
+#> attr(,"class")
+#> [1] "ce"
      econmod$summarize(by_grp = TRUE)
      #> $costs
+#>     category   dr sample strategy_id grp_id     costs
+#>  1:     Drug 0.03      1           1      1  3204.808
+#>  2:     Drug 0.03      1           1      2  4520.140
+#>  3:     Drug 0.03      1           1      3  3841.830
+#>  4:     Drug 0.03      1           1      4  3453.459
+#>  5:     Drug 0.03      1           2      1  8760.761
+#>  6:     Drug 0.03      1           2      2 12700.582
+#>  7:     Drug 0.03      1           2      3 10657.367
+#>  8:     Drug 0.03      1           2      4  9512.139
+#>  9:     Drug 0.03      2           1      1  3350.944
+#> 10:     Drug 0.03      2           1      2  4590.430
+#> 11:     Drug 0.03      2           1      3  3923.935
+#> 12:     Drug 0.03      2           1      4  3601.386
+#> 13:     Drug 0.03      2           2      1  8878.986
+#> 14:     Drug 0.03      2           2      2 12490.930
+#> 15:     Drug 0.03      2           2      3 10532.352
+#> 16:     Drug 0.03      2           2      4  9619.212
+#> 17:     Drug 0.03      3           1      1  3321.574
+#> 18:     Drug 0.03      3           1      2  4590.881
+#> 19:     Drug 0.03      3           1      3  3957.257
+#> 20:     Drug 0.03      3           1      4  3560.924
+#> 21:     Drug 0.03      3           2      1  9056.857
+#> 22:     Drug 0.03      3           2      2 12847.193
+#> 23:     Drug 0.03      3           2      3 10947.037
+#> 24:     Drug 0.03      3           2      4  9776.159
+#> 25:  Medical 0.03      1           1      1  3745.103
+#> 26:  Medical 0.03      1           1      2  5247.860
+#> 27:  Medical 0.03      1           1      3  4446.936
+#> 28:  Medical 0.03      1           1      4  4039.449
+#> 29:  Medical 0.03      1           2      1  4148.056
+#> 30:  Medical 0.03      1           2      2  5998.254
+#> 31:  Medical 0.03      1           2      3  5007.632
+#> 32:  Medical 0.03      1           2      4  4513.408
+#> 33:  Medical 0.03      2           1      1  3888.448
+#> 34:  Medical 0.03      2           1      2  5269.681
+#> 35:  Medical 0.03      2           1      3  4506.194
+#> 36:  Medical 0.03      2           1      4  4179.680
+#> 37:  Medical 0.03      2           2      1  4210.850
+#> 38:  Medical 0.03      2           2      2  5870.963
+#> 39:  Medical 0.03      2           2      3  4945.831
+#> 40:  Medical 0.03      2           2      4  4565.591
+#> 41:  Medical 0.03      3           1      1  3843.504
+#> 42:  Medical 0.03      3           1      2  5255.145
+#> 43:  Medical 0.03      3           1      3  4530.152
+#> 44:  Medical 0.03      3           1      4  4118.634
+#> 45:  Medical 0.03      3           2      1  4256.178
+#> 46:  Medical 0.03      3           2      2  5992.152
+#> 47:  Medical 0.03      3           2      3  5096.951
+#> 48:  Medical 0.03      3           2      4  4596.798
+#> 49:    total 0.03      1           1      1  6949.911
+#> 50:    total 0.03      1           1      2  9768.000
+#> 51:    total 0.03      1           1      3  8288.766
+#> 52:    total 0.03      1           1      4  7492.908
+#> 53:    total 0.03      1           2      1 12908.817
+#> 54:    total 0.03      1           2      2 18698.835
+#> 55:    total 0.03      1           2      3 15664.999
+#> 56:    total 0.03      1           2      4 14025.547
+#> 57:    total 0.03      2           1      1  7239.391
+#> 58:    total 0.03      2           1      2  9860.111
+#> 59:    total 0.03      2           1      3  8430.129
+#> 60:    total 0.03      2           1      4  7781.066
+#> 61:    total 0.03      2           2      1 13089.836
+#> 62:    total 0.03      2           2      2 18361.893
+#> 63:    total 0.03      2           2      3 15478.183
+#> 64:    total 0.03      2           2      4 14184.803
+#> 65:    total 0.03      3           1      1  7165.078
+#> 66:    total 0.03      3           1      2  9846.026
+#> 67:    total 0.03      3           1      3  8487.409
+#> 68:    total 0.03      3           1      4  7679.558
+#> 69:    total 0.03      3           2      1 13313.035
+#> 70:    total 0.03      3           2      2 18839.346
+#> 71:    total 0.03      3           2      3 16043.988
+#> 72:    total 0.03      3           2      4 14372.957
+#>     category   dr sample strategy_id grp_id     costs
+#> 
+#> $qalys
+#>       dr sample strategy_id grp_id    qalys
+#>  1: 0.03      1           1      1 1.375800
+#>  2: 0.03      1           1      2 1.934901
+#>  3: 0.03      1           1      3 1.642369
+#>  4: 0.03      1           1      4 1.483155
+#>  5: 0.03      1           2      1 1.512953
+#>  6: 0.03      1           2      2 2.190878
+#>  7: 0.03      1           2      3 1.834263
+#>  8: 0.03      1           2      4 1.644267
+#>  9: 0.03      2           1      1 1.494186
+#> 10: 0.03      2           1      2 2.038979
+#> 11: 0.03      2           1      3 1.743161
+#> 12: 0.03      2           1      4 1.605944
+#> 13: 0.03      2           2      1 1.596044
+#> 14: 0.03      2           2      2 2.238000
+#> 15: 0.03      2           2      3 1.886450
+#> 16: 0.03      2           2      4 1.729614
+#> 17: 0.03      3           1      1 1.412300
+#> 18: 0.03      3           1      2 1.939659
+#> 19: 0.03      3           1      3 1.672018
+#> 20: 0.03      3           1      4 1.513674
+#> 21: 0.03      3           2      1 1.554219
+#> 22: 0.03      3           2      2 2.194891
+#> 23: 0.03      3           2      3 1.868326
+#> 24: 0.03      3           2      4 1.678216
+#>       dr sample strategy_id grp_id    qalys
+#> 
+#> attr(,"class")
+#> [1] "ce"
      
+
      +
      + +
      + + +
      +
      +

      Developed by Devin Incerti, Jeroen P. Jansen.

      +
      + +
      +

      Site built with pkgdown 1.5.1.

      +
      + +
      +
      + + + + + + + + diff --git a/docs/reference/CohortDtstmTrans.html b/docs/reference/CohortDtstmTrans.html new file mode 100644 index 00000000..1700e3b7 --- /dev/null +++ b/docs/reference/CohortDtstmTrans.html @@ -0,0 +1,286 @@ + + + + + + + + +Transitions for a cohort discrete time state transition model — CohortDtstmTrans • hesim + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
      +
      + + + + +
      + +
      +
      +
      +

      Transitions for a cohort discrete time state transition model

      + Source: R/dtstm.R + +
      + +
      +

      Simulate health state transitions in a cohort discrete time state transition model.

      +
      + + + +

      Format

      + +

      An R6::R6Class object.

      +

      See also

      + +

      create_CohortDtstmTrans(), CohortDtstm

      +

      Public fields

      + +

      +
      params

      Parameters for simulating health state transitions. +Supports objects of class tparams_transprobs or params_mlogit.

      + +
      input_data

      An object of class input_mats.

      + +
      cycle_length

      The length of a model cycle in terms of years. +The default is 1 meaning that model cycles are 1 year long.

      + +

      +

      Active bindings

      + +

      +
      start_stateprobs

      A non-negative vector with length equal to the number of +health states containing the probability that the cohort is in each health +state at the start of the simulation. For example, +if there were three states and the cohort began the simulation in state 1, +then start_stateprobs = c(1, 0, 0). Automatically normalized to sum to 1. +If NULL, then a vector with the first element equal to 1 and +all remaining elements equal to 0.

      + +
      trans_mat

      A transition matrix describing the states and transitions +in a discrete-time multi-state model. Only required if the model is +parameterized using multinomial logistic regression. The (i,j) element +represents a transition from state i to state j. Each possible transition +from row i should be based on a separate multinomial logistic regression +and ordered from 0 to K - 1 where K is the number of +possible transitions. Transitions that are not possible should be NA. +and the reference category for each row should be 0.

      + +

      +

      Methods

      + + +

      Public methods

      + + +

      +

      Method new()

      +

      Create a new CohortDtstmTrans object.

      Usage

      +

      CohortDtstmTrans$new(
+  params,
+  input_data = NULL,
+  trans_mat = NULL,
+  start_stateprobs = NULL,
+  cycle_length = 1
+)

      + +

      Arguments

      +

      +
      params

      The params field.

      + +
      input_data

      The input_data field.

      + +
      trans_mat

      The trans_mat field.

      + +
      start_stateprobs

      The start_stateprobs field.

      + +
      cycle_length

      The cycle_length field.

      + +

      +

      Returns

      +

      A new CohortDtstmTrans object.

      +

      +

      Method sim_stateprobs()

      +

      Simulate probability of being in each health state during each model cycle.

      Usage

      +

      CohortDtstmTrans$sim_stateprobs(n_cycles)

      + +

      Arguments

      +

      +
      n_cycles

      The number of model cycles to simulate the model for.

      + +

      +

      Returns

      +

      An object of class stateprobs.

      +

      +

      Method clone()

      +

      The objects of this class are cloneable with this method.

      Usage

      +

      CohortDtstmTrans$clone(deep = FALSE)

      + +

      Arguments

      +

      +
      deep

      Whether to make a deep clone.

      + +

      + + + +
      + +
      + + +
      +
      +

      Developed by Devin Incerti, Jeroen P. Jansen.

      +
      + +
      +

      Site built with pkgdown 1.5.1.

      +
      + +
      +
      + + + + + + + + diff --git a/docs/reference/CtstmTrans.html b/docs/reference/CtstmTrans.html index 400da4e0..6968b60b 100644 --- a/docs/reference/CtstmTrans.html +++ b/docs/reference/CtstmTrans.html @@ -1,28 +1,36 @@ - + -Transitions for continuous time state transition model — CtstmTrans • hesim +An <code>R6</code> base class for continuous time state transition models — CtstmTrans • hesim + - + - + + + + - + + + - + + - + - - + + + @@ -30,17 +38,17 @@ - - - - - + + + - + + + - +
      -

      Transitions for continuous time state transition model

      - +

      An R6 base class for continuous time state transition models

      + Source: R/ctstm.R
      - -

      Summarize health state transitions in a continuous time state transition model with parameters that were estimated using a multi-state model.

      - +

      A non-exported base class for continuous time state transition models containing methods +that can be used to summarize fitted multi-state models.

      - 
      CtstmTrans
      - -

      Format

      -

      R6Class object.

      - -

      Fields

      - -
      -
      data

      Data used to simulate health state transitions by treatment strategy and patient. Must be - an object of class input_data. If params is a list of models, then data must contain a unique row for each treatment strategy and patient; if params is a joint model, then data must contain a unique row for each treatment strategy, patient, and transition.

      -
      params

      An object of class params_surv or params_surv_list.

      -
      trans_mat

      A transition matrix describing the states and transitions in a multi-state model in the format from the mstate package. See the documentation for the argument "trans" in msprep.

      -
      start_ages

      The starting age of each patient in the simulation. Must be the same length as the number of patients in data. Used alongside max_age in sim_stateprobs() with type = "individual"; does not otherwise need to be specified.

      -
      death_state

      The death state in trans_mat. Used with max_age in sim_stateprobs with type = "individual" and in IndivCtstm$sim_disease as patients transition to this state upon reaching maximum age; does not otherwise need to be specified. By default, it is set to the final absorbing state (i.e., a row in trans_mat with all NAs).

      -
      - -

      Methods

      +

      Format

      - -
      -
      new(data, params, trans_mat, start_ages = rep(38, data$n_patients), - death_state = NULL)

      Constructor for the class.

      -
      hazard(t)

      Predict the hazard functions for each health state transition. -

        -

      • t: A numeric vector of times.

        • -

      -
      cumhazard(t)

      Predict the cumulative hazard functions for each health state transition. -

        -

      • t: A numeric vector of times.

        • -

      , -
      sim_stateprobs(t, start_state = 1, max_t = 100, max_age = 100, - clock = "reset")

      Simulate health state probabilities as a function of time. -

        -

      • t: A numeric vector of times.

        • -

      • start_state: The starting health state. Default is the first - health state.

        • -

      • max_t: Maximum number of time periods to simulate the model for. Default is 100.

        • -

      • max_age: Maximum age to simulate each patient until. Default is 100.

        • -

      • clock: Currently only supports "reset", which implies that the time-scale of the - cumulative hazards depends on time in the present state.

        • -

      -
      check()

      Input validation for class. Checks that fields are the correct type.

      -
      - +

      An R6::R6Class object.

      See also

      -

      create_CtstmTrans, IndivCtstm

      - - -

      Examples

      - 
      library("flexsurv")
      #> Loading required package: survival
      
-# Simulation data
-dt_strategies <- data.frame(strategy_id = c(1, 2, 3))
-dt_patients <- data.frame(patient_id = seq(1, 3),
-                          age = c(45, 50, 60),
-                          female = c(0, 0, 1))
-
-# Multi-state model with transition specific models
-fits <- vector(length = 3, mode = "list")
-dat <- data.frame(bosms3)
-for (i in 1:length(fits)){
-  fits[[i]] <- flexsurvreg(Surv(years, status) ~ 1, data = dat[dat$trans == i, ],
-                           dist = "exp")
-}
-fits <- flexsurvreg_list(fits)
-tmat <- rbind(c(NA, 1, 2),
-              c(NA, NA, 3),
-              c(NA, NA, NA))
-
-# Simulation model
-hesim_dat <- hesim_data(strategies = dt_strategies,
-                        patients = dt_patients)
-fits_data <- expand_hesim_data(hesim_dat)
-transmod <- create_CtstmTrans(fits, data = fits_data, trans_mat = tmat,
-                              point_estimate = TRUE)
-head(transmod$hazard(c(1, 2, 3)))
      #>    trans sample strategy_id patient_id t    hazard
-#> 1:     1      1           0          0 1 0.1898667
-#> 2:     1      1           0          0 2 0.1898667
-#> 3:     1      1           0          0 3 0.1898667
-#> 4:     2      1           0          0 1 0.0534576
-#> 5:     2      1           0          0 2 0.0534576
-#> 6:     2      1           0          0 3 0.0534576
      head(transmod$cumhazard(c(1, 2, 3)))
      #>    trans sample strategy_id patient_id t cumhazard
-#> 1:     1      1           0          0 1 0.1898667
-#> 2:     1      1           0          0 2 0.3797333
-#> 3:     1      1           0          0 3 0.5696000
-#> 4:     2      1           0          0 1 0.0534576
-#> 5:     2      1           0          0 2 0.1069152
-#> 6:     2      1           0          0 3 0.1603728
      transmod$sim_stateprobs(t = c(0, 5, 10))[t == 5]
      #>    sample strategy_id state_id t      prob
-#> 1:      1           1        1 5 0.6666667
-#> 2:      1           1        2 5 0.3333333
-#> 3:      1           1        3 5 0.0000000
-#> 4:      1           2        1 5 0.6666667
-#> 5:      1           2        2 5 0.3333333
-#> 6:      1           2        3 5 0.0000000
-#> 7:      1           3        1 5 0.0000000
-#> 8:      1           3        2 5 0.3333333
-#> 9:      1           3        3 5 0.6666667
      -
      - +
      +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown.

      +

      Site built with pkgdown 1.5.1.

      @@ -257,6 +236,8 @@

      Contents

      + + diff --git a/docs/reference/IndivCtstm.html b/docs/reference/IndivCtstm.html index 70f862f5..a831d9b4 100644 --- a/docs/reference/IndivCtstm.html +++ b/docs/reference/IndivCtstm.html @@ -8,21 +8,29 @@ Individual-level continuous time state transition model — IndivCtstm • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,12 +38,14 @@ - - + + - - @@ -49,9 +59,10 @@ + - +
      @@ -82,28 +93,46 @@
    • News
    -
    • - - + +
      • @@ -114,136 +143,237 @@ +

      Individual-level continuous time state transition model

      - + Source: R/ctstm.R
      - -

      Simulate outcomes from an individual-level continuous time state transition model (CTSTM). The class supports "clock-reset" (i.e., semi-Markov), "clock-forward" (i.e., Markov), and mixtures of clock-reset and clock-forward models as described in IndivCtstmTrans.

      - +

      Simulate outcomes from an individual-level continuous time state transition +model (CTSTM) from a fitted multi-state model. The class supports "clock-reset" +(i.e., semi-Markov), "clock-forward" (i.e., Markov), and mixtures of +clock-reset and clock-forward models as described in +IndivCtstmTrans.

      - 
      IndivCtstm
      - + +

      Format

      -

      R6Class object.

      - -

      Fields

      +

      An R6::R6Class object.

      +

      See also

      - -
      -
      trans_model

      Model for health state transitions. Must be an object of class IndivCtstmTrans.

      -
      utility_model

      The model used to predict utility by health state. Must be an object of class StateVals.

      -
      cost_models

      The models used to predict costs by health state. Must be a list of objects of class StateVals, where each element of the list represents a different cost category.

      -
      disprog_

      A data.table simulated using sim_disease containing the following columns: -

      -
      sample

      A random sample from the PSA.

      -
      strategy_id

      The treatment strategy ID.

      -
      patient_id

      The patient ID.

      -
      from

      The health state ID transitioned from.

      -
      to

      The health state ID transitioned to.

      -
      final

      An indicator equal to 1 if a patient is in their final health state during the simulation and 0 otherwise.

      -
      time_start

      The time at the start of the interval.

      -
      time_stop

      The time at the end of the interval.

      -

      -
      stateprobs_

      A data.table of health state probabilities as a function of time simulated using sim_stateprobs. See the description of the output of sim_stateprobs in IndivCtstmTrans.

      -
      qalys_

      A data.table of quality-adjusted life-years (QALYs) simulated using sim_qalys. Columns are: -

      -
      sample

      A random sample from the probabilistic sensitivity analysis (PSA).

      -
      strategy_id

      The treatment strategy ID.

      -
      patient_id

      A patient ID.

      -
      dr

      The discount rate.

      -
      qalys

      Simulated QALYs.

      -
      lys

      Simulated life-years.

      -
      - Note that the lys column is only computed if lys = TRUE.

      -
      costs_

      A data.table of costs by category simulated using sim_costs. Columns are: -

      -
      sample

      A random sample from the PSA.

      -
      strategy_id

      The treatment strategy ID.

      -
      patient_id

      A patient ID.

      -
      dr

      The discount rate.

      -
      category

      The cost category.

      -
      costs

      Simulated costs.

      -

      -
      - +

      create_IndivCtstmTrans(), IndivCtstmTrans

      +

      Public fields

      + +

      +
      trans_model

      The model for health state transitions. Must be an object +of class IndivCtstmTrans.

      + +
      utility_model

      The model for health state utility. Must be an object of +class StateVals.

      + +
      cost_models

      The models used to predict costs by health state. +Must be a list of objects of class StateVals, where each element of the +list represents a different cost category.

      + +
      disprog_

      A data.table simulated using $sim_disease() containing the following columns: +

        +

      • sample: A random sample from the PSA.

        • +

      • strategy_id: The treatment strategy ID.

        • +

      • patient_id: The patient ID.

        • +

      • from: The health state ID transitioned from.

        • +

      • to: The health state ID transitioned to.

        • +

      • final: An indicator equal to 1 if a patient is in their final health +state during the simulation and 0 otherwise.

        • +

      • time_start: The time at the start of the interval.

        • +

      • time_stop: The time at the end of the interval.

        • +

      + +
      stateprobs_

      An object of class stateprobs simulated using $sim_stateprobs().

      + +
      qalys_

      An object of class qalys simulated using $sim_qalys().

      + +
      costs_

      An object of class costs simulated using $sim_costs().

      + +

      Methods

      -
      -
      new(trans_model = NULL, utility_model = NULL, cost_models = NULL)

      Constructor for the class.

      -
      sim_disease(max_t = 100, max_age = 100, progress = NULL)

      Simulate disease progression. -

        -

      • max_t: A scalar or vector denoting the number of time periods to simulate the model. If a vector, must be equal to the number of simulated patients.

        • -

      • max_age: A scalar or vector denoting the maximum age to simulate each patient until. If a vector, must be equal to the number of simulated patients.

        • -

      • progress: An integer, specifying the PSA iteration (i.e., sample) that should be printed every progress PSA iterations. For example, if progress = 2, then every second PSA iteration is printed. Default is NULL, in which case no output is printed.

        • +

        Public methods

        + + - Returns an instance of self with simulated output stored in disprog_.

      -
      sim_stateprobs(t)

      Simulate the probability of being in each health state as a function of time using the simulation output stored in disprog_. -

        -

      • t: A numeric vector of times.

        • -
      - Returns an instance of self with simulated output stored in stateprobs_.

      -
      sim_qalys(dr = .03, type = c("predict", "random"), by_patient = FALSE, lys = TRUE)

      Compute simulated (mean discounted) QALYs using the simulation output stored in disprog_ by random sample from the PSA, treatment strategy, health state, and (optionally) patient. -

        -

      • dr: Discount rate to apply to QALYs. May be a vector in which case QALYs are calculated for each element in dr.

        • -

      • type: predict for mean values or random for random samples as in $sim() in StateVals.

        • -

      • by_patient: If TRUE, QALYs are computed at the patient level. If FALSE, QALYs are averaged across patients by - health state.

        • -

      • lys: If TRUE, then life-years are computed in addition to QALYs.

        • -
      - Returns an instance of self with simulated output stored in qalys_.

      -
      sim_costs(dr = .03, type = c("predict", "random"), by_patient = FALSE, max_t = Inf)

      Compute simulated (mean discounted) costs using the simulation output stored in disprog_ by random sample from the PSA, treatment strategy, health state, and - (optionally) patient. -

        -

      • dr: Discount rate to apply to costs. May be a vector in which case costs are calculated for each element in dr.

        • -

      • type: predict for mean values or random for random samples as in $sim() in StateVals.

        • -

      • by_patient: If TRUE, costs are computed at the patient level. If FALSE, costs are averaged across patients by - health state.

        • -

      • max_t: Maximum time duration to compute costs once a patient has entered a (new) health state. By default, equal to Inf, - so that costs are computed over the entire duration that a patient is in a given health state. If time varies by each cost category, then time can also be passed - as a numeric vector of length equal to the number of cost categories (e.g., c(1, 2, Inf, 3) for a model with 4 cost categories).

        • -
      - Returns an instance of self with simulated output stored in costs_.

      , -
      summarize()

      Summarize costs and QALYs so that cost-effectiveness analysis can be performed. - See summarize_ce.

      -
      - -

      See also

      +

      +

      Method new()

      +

      Create a new IndivCtstm object.

      Usage

      +

      IndivCtstm$new(trans_model = NULL, utility_model = NULL, cost_models = NULL)

      + +

      Arguments

      +

      +
      trans_model

      The trans_model field.

      + +
      utility_model

      The utility_model field.

      + +
      cost_models

      The cost_models field.

      + +

      +

      Returns

      +

      A new IndivCtstm object.

      +

      +

      Method sim_disease()

      +

      Simulate disease progression (i.e., individual trajectories through a multi-state +model using an individual patient simulation).

      Usage

      +

      IndivCtstm$sim_disease(max_t = 100, max_age = 100, progress = NULL)

      + +

      Arguments

      +

      +
      max_t

      A scalar or vector denoting the length of time to simulate the model. +If a vector, must be equal to the number of simulated patients.

      + +
      max_age

      A scalar or vector denoting the maximum age to simulate each patient until. +If a vector, must be equal to the number of simulated patients.

      + +
      progress

      An integer, specifying the PSA iteration (i.e., sample) that should be +printed every progress PSA iterations. For example, if progress = 2, +then every second PSA iteration is printed. Default is NULL, +in which case no output is printed.

      + +

      +

      Returns

      +

      An instance of self with simulated output stored in disprog_.

      +

      +

      Method sim_stateprobs()

      +

      Simulate health state probabilities as a function of time using the +simulation output stored in disprog.

      Usage

      +

      IndivCtstm$sim_stateprobs(t)

      + +

      Arguments

      +

      +
      t

      TA numeric vector of times.

      + +

      +

      Returns

      +

      An instance of self with simulated output of class stateprobs +stored in stateprobs_.

      +

      +

      Method sim_qalys()

      +

      Simulate quality-adjusted life-years (QALYs) as a function of disprog_ and +utility_model. See vignette("expected-values") for details.

      Usage

      +

      IndivCtstm$sim_qalys(
+  dr = 0.03,
+  type = c("predict", "random"),
+  lys = TRUE,
+  by_patient = FALSE
+)

      + +

      Arguments

      +

      +
      dr

      Discount rate.

      + +
      type

      "predict" for mean values or "random" for random samples +as in $sim() in StateVals.

      + +
      lys

      If TRUE, then life-years are simulated in addition to QALYs.

      + +
      by_patient

      If TRUE, then QALYs are computed at the patient level. +If FALSE, then QALYs are averaged across patients by health state.

      + +

      +

      Returns

      +

      An instance of self with simulated output of +class qalys stored in qalys_.

      +

      +

      Method sim_costs()

      +

      Simulate costs as a function of disprog_ and cost_models. +See vignette("expected-values") for details.

      Usage

      +

      IndivCtstm$sim_costs(
+  dr = 0.03,
+  type = c("predict", "random"),
+  by_patient = FALSE,
+  max_t = Inf
+)

      + +

      Arguments

      +

      +
      dr

      Discount rate.

      + +
      type

      "predict" for mean values or "random" for random samples +as in $sim() in StateVals.

      + +
      by_patient

      If TRUE, then QALYs are computed at the patient level. +If FALSE, then QALYs are averaged across patients by health state.

      + +
      max_t

      Maximum time duration to compute costs once a patient has +entered a (new) health state. By default, equal to Inf, +so that costs are computed over the entire duration that a patient is in +a given health state. If time varies by each cost category, then time +can also be passed as a numeric vector of length equal to the number of +cost categories (e.g., c(1, 2, Inf, 3) for a model with four cost categories).

      + +

      +

      Returns

      +

      An instance of self with simulated output of class costs +stored in costs_.

      +

      +

      Method summarize()

      +

      Summarize costs and QALYs so that cost-effectiveness analysis can be performed. +See summarize_ce().

      Usage

      +

      IndivCtstm$summarize(by_grp = FALSE)

      + +

      Arguments

      +

      +
      by_grp

      If TRUE, then costs and QALYs are computed by subgroup. If +FALSE, then costs and QALYs are aggregated across all patients (and subgroups).

      + +

      +

      +

      Method clone()

      +

      The objects of this class are cloneable with this method.

      Usage

      +

      IndivCtstm$clone(deep = FALSE)

      + +

      Arguments

      +

      +
      deep

      Whether to make a deep clone.

      + +

      + -

      create_IndivCtstmTrans, IndivCtstmTrans

      -

      Examples

      - 
      library("flexsurv")
      #> Loading required package: survival
      
+    
      library("flexsurv")
      #> Loading required package: survival
      
 # Treatment strategies, target population, and model structure
-strategies <- data.frame(strategy_id = c(1, 2))
-patients <- data.frame(patient_id = seq(1, 3),
-                          age = c(45, 50, 60),
-                          female = c(0, 0, 1))
-states <- data.frame(state_id = c(1, 2))
+strategies <- data.frame(strategy_id = c(1, 2))
+patients <- data.frame(patient_id = seq(1, 3),
+                       age = c(45, 50, 60),
+                       female = c(0, 0, 1))
+states <- data.frame(state_id = c(1, 2))
 hesim_dat <- hesim_data(strategies = strategies,
                         patients = patients,
                         states = states)
 
 # Parameter estimation
 ## Multi-state model
-tmat <- rbind(c(NA, 1, 2),
-              c(3, NA, 4),
-              c(NA, NA, NA))
-fits <- vector(length = max(tmat, na.rm = TRUE), mode = "list")
-surv_dat <- data.frame(ctstm3_exdata$transitions)
-for (i in 1:length(fits)){
-  fits[[i]] <- flexsurvreg(Surv(years, status) ~ factor(strategy_id),
+tmat <- rbind(c(NA, 1, 2),
+              c(3, NA, 4),
+              c(NA, NA, NA))
+fits <- vector(length = max(tmat, na.rm = TRUE), mode = "list")
+surv_dat <- data.frame(mstate3_exdata$transitions)
+for (i in 1:length(fits)){
+  fits[[i]] <- flexsurvreg(Surv(years, status) ~ factor(strategy_id),
                            data = surv_dat,
                            subset = (trans == i),
                            dist = "weibull")
@@ -251,19 +381,19 @@ 
      Examp
 fits <- flexsurvreg_list(fits)
 
 ## Utility
-utility_tbl <- stateval_tbl(data.frame(state_id = states$state_id,
-                                       mean = ctstm3_exdata$utility$mean,
-                                       se = ctstm3_exdata$utility$se),
+utility_tbl <- stateval_tbl(data.frame(state_id = states$state_id,
+                                       mean = mstate3_exdata$utility$mean,
+                                       se = mstate3_exdata$utility$se),
                             dist = "beta",
                             hesim_data = hesim_dat)
 ## Costs
-drugcost_tbl <- stateval_tbl(data.frame(strategy_id = strategies$strategy_id,
-                                       est = ctstm3_exdata$costs$drugs$costs),
-                            dist = "fixed",
-                            hesim_data = hesim_dat)
-medcost_tbl <- stateval_tbl(data.frame(state_id = states$state_id,
-                                       mean = ctstm3_exdata$costs$medical$mean,
-                                       se = ctstm3_exdata$costs$medical$se),
+drugcost_tbl <- stateval_tbl(data.frame(strategy_id = strategies$strategy_id,
+                                        est = mstate3_exdata$costs$drugs$costs),
+                             dist = "fixed",
+                             hesim_data = hesim_dat)
+medcost_tbl <- stateval_tbl(data.frame(state_id = states$state_id,
+                                       mean = mstate3_exdata$costs$medical$mean,
+                                       se = mstate3_exdata$costs$medical$se),
                             dist = "gamma",
                             hesim_data = hesim_dat)
 
@@ -283,7 +413,7 @@ 
      Examp
 ### Costs
 drugcostmod <- create_StateVals(drugcost_tbl, n = n_samples)
 medcostmod <- create_StateVals(medcost_tbl, n = n_samples)
-costmods <- list(drugs = drugcostmod,
+costmods <- list(drugs = drugcostmod,
                  medical = medcostmod)
 
 ### Combine
@@ -293,74 +423,68 @@ 
      Examp
 
 
 ## Simulate outcomes
-head(ictstm$sim_disease()$disprog_)
      #>    sample strategy_id patient_id from to final time_start time_stop
-#> 1:      1           1          1    1  2     0   0.000000  3.759731
-#> 2:      1           1          1    2  3     1   3.759731  5.971252
-#> 3:      1           1          2    1  2     0   0.000000  4.687091
-#> 4:      1           1          2    2  1     0   4.687091  6.211248
-#> 5:      1           1          2    1  2     0   6.211248  6.572338
-#> 6:      1           1          2    2  1     0   6.572338  7.618977
      head(ictstm$sim_stateprobs(t = c(0, 5, 10))$stateprobs_[t == 5])
      #>    sample strategy_id state_id t      prob
-#> 1:      1           1        1 5 0.0000000
-#> 2:      1           1        2 5 1.0000000
-#> 3:      1           1        3 5 0.0000000
-#> 4:      1           2        1 5 0.6666667
-#> 5:      1           2        2 5 0.0000000
-#> 6:      1           2        3 5 0.3333333
      ictstm$sim_qalys(dr = .03)
-ictstm$sim_costs(dr = .03)
-head(ictstm$summarize())
      #> $costs
-#>     category   dr sample strategy_id     costs
-#>  1:    drugs 0.03      1           1 38164.878
-#>  2:    drugs 0.03      1           2 57546.053
-#>  3:    drugs 0.03      2           1  5308.587
-#>  4:    drugs 0.03      2           2 65741.343
-#>  5:  medical 0.03      1           1  8643.137
-#>  6:  medical 0.03      1           2  6216.491
-#>  7:  medical 0.03      2           1  1204.330
-#>  8:  medical 0.03      2           2  6919.485
-#>  9:    total 0.03      1           1 46808.015
-#> 10:    total 0.03      1           2 63762.544
-#> 11:    total 0.03      2           1  6512.917
-#> 12:    total 0.03      2           2 72660.828
+head(ictstm$sim_disease()$disprog_)
      #>    sample strategy_id patient_id grp_id from to final time_start time_stop
+#> 1:      1           1          1      1    1  3     1  0.0000000 0.4036627
+#> 2:      1           1          2      1    1  2     0  0.0000000 5.5078506
+#> 3:      1           1          2      1    2  1     0  5.5078506 5.8230571
+#> 4:      1           1          2      1    1  3     1  5.8230571 6.6985224
+#> 5:      1           1          3      1    1  2     0  0.0000000 0.7563655
+#> 6:      1           1          3      1    2  1     0  0.7563655 0.7578541
      head(ictstm$sim_stateprobs(t = c(0, 5, 10))$stateprobs_[t == 5])
      #>    sample strategy_id grp_id state_id t      prob
+#> 1:      1           1      1        1 5 0.3333333
+#> 2:      1           1      1        2 5 0.3333333
+#> 3:      1           1      1        3 5 0.3333333
+#> 4:      1           2      1        1 5 0.6666667
+#> 5:      1           2      1        2 5 0.0000000
+#> 6:      1           2      1        3 5 0.3333333
      ictstm$sim_qalys(dr = .03)
+ictstm$sim_costs(dr = .03)
+head(ictstm$summarize())
      #> $costs
+#>     category   dr sample strategy_id     costs grp_id
+#>  1:    drugs 0.03      1           1 42596.065      1
+#>  2:    drugs 0.03      1           2 44645.168      1
+#>  3:    drugs 0.03      2           1 25228.525      1
+#>  4:    drugs 0.03      2           2 89673.803      1
+#>  5:  medical 0.03      1           1 10533.655      1
+#>  6:  medical 0.03      1           2  4610.220      1
+#>  7:  medical 0.03      2           1  5296.932      1
+#>  8:  medical 0.03      2           2  9632.893      1
+#>  9:    total 0.03      1           1 53129.720      1
+#> 10:    total 0.03      1           2 49255.388      1
+#> 11:    total 0.03      2           1 30525.457      1
+#> 12:    total 0.03      2           2 99306.697      1
 #> 
 #> $qalys
-#>      dr sample strategy_id     qalys
-#> 1: 0.03      1           1 5.3989021
-#> 2: 0.03      1           2 3.9859904
-#> 3: 0.03      2           1 0.8963122
-#> 4: 0.03      2           2 5.3724371
-#> 
      
+#>      dr sample strategy_id    qalys grp_id
+#> 1: 0.03      1           1 3.572185      1
+#> 2: 0.03      1           2 2.606998      1
+#> 3: 0.03      2           1 1.440414      1
+#> 4: 0.03      2           2 2.703469      1
+#> 
      
+
      - +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/IndivCtstmTrans.html b/docs/reference/IndivCtstmTrans.html index 0850faa4..75f3c3f1 100644 --- a/docs/reference/IndivCtstmTrans.html +++ b/docs/reference/IndivCtstmTrans.html @@ -8,21 +8,29 @@ Transitions for an individual-level continuous time state transition model — IndivCtstmTrans • hesim + - + - + + + + - + + + - + + - + - - + + + @@ -30,12 +38,11 @@ - - + + - - @@ -49,9 +56,10 @@ + - +
      @@ -82,28 +90,46 @@
    • News
    -
    • - - + +
      • @@ -114,98 +140,178 @@ +

      Transitions for an individual-level continuous time state transition model

      - + Source: R/ctstm.R
      - -

      Summarize health state transitions in an individual-level continuous time state transition model with parameters that were estimated using a multi-state model.

      - +

      Simulate health state transitions in an individual-level continuous time state +transition model with parameters that were estimated using a multi-state model.

      - 
      IndivCtstmTrans
      - + +

      Format

      -

      R6Class object.

      - -

      Fields

      +

      An R6::R6Class object.

      +

      See also

      - -
      -
      input_mats

      Input matrices used to simulate health state transitions by treatment strategy and patient. Must be - an object of class input_mats. If params is a list of models, then input_mats must contain a unique row for each treatment strategy and patient; if params is a joint model, then input_mats must contain a unique row for each treatment strategy, patient, and transition.

      -
      params

      An object of class params_surv or params_surv_list.

      -
      trans_mat

      A transition matrix describing the states and transitions in a multi-state model in the format from the mstate package. See the documentation for the argument "trans" in msprep.

      -
      start_state

      A scalar or vector denoting the starting health state. Default is the first health state. If a vector, must be equal to the number of simulated patients.

      -
      clock

      "reset" for a clock-reset model, "forward" for a clock-forward model, and "mix" for a mixture -of clock-reset and clock-forward models. A clock-reset model is a semi-Markov model in which transition rates depend on time since entering a state. A clock-forward model is a Markov model in which transition rates depend on time since entering the initial state. If "mix" is used, then reset_states must be specified.

      -
      reset_states

      A vector denoting the states in which time resets. Hazard functions are always a function of elapsed time since either the start of the model or from when time was previously reset. Only used if clock = "mix".

      -
      start_age

      A scalar or vector denoting the starting age of each patient in the simulation. Default is 38. If a vector, must be equal to the number of simulated patients.

      -
      death_state

      The death state in trans_mat. Used with max_age in sim_disease as patients transition to this state upon reaching maximum age. By default, it is set to the final absorbing state (i.e., a row in trans_mat with all NAs).

      -
      - +

      create_IndivCtstmTrans(), IndivCtstm

      +

      Super class

      + +

      hesim::CtstmTrans -> IndivCtstmTrans

      +

      Public fields

      + +

      +
      params

      An object of class params_surv or params_surv_list.

      + +
      input_data

      Input data used to simulate health state transitions +by sample from the probabilistic sensitivity analysis (PSA), treatment strategy and patient. +Must be an object of class input_mats. If params contains parameters from +a list of models (i.e., of class params_surv_list), then input_data +must contain a unique row for each treatment strategy +and patient; if params contains parameters from a joint model +(i.e., of class params_surv), then input_data must contain a unique +row for each treatment strategy, patient, and transition.

      + +
      trans_mat

      A transition matrix describing the states and transitions +in a multi-state model in the format from the mstate package. +See the documentation for the argument "trans" in mstate::msprep.

      + +
      start_state

      A scalar or vector denoting the starting health state. +Default is the first health state. If a vector, must be equal to the number of simulated patients.

      + +
      start_age

      A scalar or vector denoting the starting age of each patient +in the simulation. Default is 38. If a vector, must be equal to the number of simulated patients.

      + +
      death_state

      The death state in trans_mat. Used with max_age +in sim_disease as patients transition to this state upon reaching maximum age. +By default, it is set to the final absorbing state (i.e., a row in trans_mat with all NAs).

      + +
      clock

      "reset" for a clock-reset model, "forward" for a clock-forward model, +and "mix" for a mixture of clock-reset and clock-forward models. A clock-reset model +is a semi-Markov model in which transition rates depend on time since entering a state. +A clock-forward model is a Markov model in which transition rates depend on time +since entering the initial state. If "mix" is used, then +reset_states must be specified.

      + +
      reset_states

      A vector denoting the states in which time resets. +Hazard functions are always a function of elapsed time since either the +start of the model or from when time was previously reset. Only used if +clock = "mix".

      + +

      Methods

      -
      -
      new(input_mats, params, trans_mat, start_state = 1, - start_age = 38, death_state = NULL, - clock = c("reset", "forward", "mix"), - reset_states = NULL))

      Constructor for the class.

      -
      hazard(t)

      Predict the hazard functions for each health state transition. -

        -

      • t: A numeric vector of times.

        • -

      -
      cumhazard(t)

      Predict the cumulative hazard functions for each health state transition. -

        -

      • t: A numeric vector of times.

        • -

      , -
      sim_stateprobs(t, ...)

      Simulate health state probabilities at distinct times. -

      -
      check()

      Input validation for class. Checks that fields are the correct type.

      -
      - -

      See also

      +

      Inherited methods + +
      + +
      +

      Method new()

      +

      Create a new IndivCtstmTrans object.

      Usage

      +

      IndivCtstmTrans$new(
+  params,
+  input_data,
+  trans_mat,
+  start_state = 1,
+  start_age = 38,
+  death_state = NULL,
+  clock = c("reset", "forward", "mix"),
+  reset_states = NULL
+)

      + +

      Arguments

      +

      +
      params

      The params field.

      + +
      input_data

      The input_data field.

      + +
      trans_mat

      The trans_mat field.

      + +
      start_state

      The start_state field.

      + +
      start_age

      The start_age field.

      + +
      death_state

      The death_state field.

      + +
      clock

      The clock field.

      + +
      reset_states

      The reset_states field.

      + +

      +

      Returns

      +

      A new IndivCtstmTrans object.

      +

      +

      Method sim_stateprobs()

      +

      Simulate health state probabilities at distinct times by first +simulating trajectories through a multi-state model with +IndivCtstm$sim_disease().

      Usage

      +

      IndivCtstmTrans$sim_stateprobs(t, ...)

      + +

      Arguments

      +

      +
      t

      A numeric vector of times.

      + +
      ...

      Additional arguments to pass to IndivCtstm$sim_disease().

      + +

      +

      Returns

      +

      An object of class stateprobs.

      +

      +

      Method check()

      +

      Input validation for class. Checks that fields are the correct type.

      Usage

      +

      IndivCtstmTrans$check()

      + +

      +

      Method clone()

      +

      The objects of this class are cloneable with this method.

      Usage

      +

      IndivCtstmTrans$clone(deep = FALSE)

      + +

      Arguments

      +

      +
      deep

      Whether to make a deep clone.

      + +

      + -

      create_IndivCtstmTrans, IndivCtstm

      -

      Examples

      - 
      library("flexsurv")
+    
      library("flexsurv")
 
 # Simulation data
-strategies <- data.frame(strategy_id = c(1, 2, 3))
-patients <- data.frame(patient_id = seq(1, 3),
-                          age = c(45, 50, 60),
-                          female = c(0, 0, 1))
+strategies <- data.frame(strategy_id = c(1, 2, 3))
+patients <- data.frame(patient_id = seq(1, 3),
+                       age = c(45, 50, 60),
+                       female = c(0, 0, 1))
 
 # Multi-state model with transition specific models
-tmat <- rbind(c(NA, 1, 2),
-              c(NA, NA, 3),
-              c(NA, NA, NA))
-fits <- vector(length = max(tmat, na.rm = TRUE), mode = "list")
-for (i in 1:length(fits)){
-  fits[[i]] <- flexsurvreg(Surv(years, status) ~ 1,
-                            data = bosms3[bosms3$trans == i, ],
+tmat <- rbind(c(NA, 1, 2),
+              c(NA, NA, 3),
+              c(NA, NA, NA))
+fits <- vector(length = max(tmat, na.rm = TRUE), mode = "list")
+for (i in 1:length(fits)){
+  fits[[i]] <- flexsurvreg(Surv(years, status) ~ 1,
+                           data = bosms3[bosms3$trans == i, ],
                            dist = "exp")
 }
 fits <- flexsurvreg_list(fits)
@@ -215,72 +321,65 @@ 
      Examp
                         patients = patients)
 fits_data <- expand(hesim_dat)
 transmod <- create_IndivCtstmTrans(fits, input_data = fits_data,
-                                    trans_mat = tmat,
-                                    n = 2,
-                                    point_estimate = FALSE)
-head(transmod$hazard(c(1, 2, 3)))
      #>    trans sample strategy_id patient_id t     hazard
-#> 1:     1      1           1          1 1 0.21087494
-#> 2:     1      1           1          1 2 0.21087494
-#> 3:     1      1           1          1 3 0.21087494
-#> 4:     2      1           1          1 1 0.05853501
-#> 5:     2      1           1          1 2 0.05853501
-#> 6:     2      1           1          1 3 0.05853501
      head(transmod$cumhazard(c(1, 2, 3)))
      #>    trans sample strategy_id patient_id t  cumhazard
-#> 1:     1      1           1          1 1 0.21087494
-#> 2:     1      1           1          1 2 0.42174988
-#> 3:     1      1           1          1 3 0.63262482
-#> 4:     2      1           1          1 1 0.05853501
-#> 5:     2      1           1          1 2 0.11707001
-#> 6:     2      1           1          1 3 0.17560502
      transmod$sim_stateprobs(t = c(0, 5, 10))[t == 5]
      #>     sample strategy_id state_id t      prob
-#>  1:      1           1        1 5 0.6666667
-#>  2:      1           1        2 5 0.0000000
-#>  3:      1           1        3 5 0.3333333
-#>  4:      1           2        1 5 0.3333333
-#>  5:      1           2        2 5 0.3333333
-#>  6:      1           2        3 5 0.3333333
-#>  7:      1           3        1 5 0.3333333
-#>  8:      1           3        2 5 0.0000000
-#>  9:      1           3        3 5 0.6666667
-#> 10:      2           1        1 5 0.0000000
-#> 11:      2           1        2 5 0.3333333
-#> 12:      2           1        3 5 0.6666667
-#> 13:      2           2        1 5 0.3333333
-#> 14:      2           2        2 5 0.0000000
-#> 15:      2           2        3 5 0.6666667
-#> 16:      2           3        1 5 0.6666667
-#> 17:      2           3        2 5 0.0000000
-#> 18:      2           3        3 5 0.3333333
      
+                                   trans_mat = tmat,
+                                   n = 2,
+                                   point_estimate = FALSE)
+head(transmod$hazard(c(1, 2, 3)))
      #>    trans sample strategy_id patient_id grp_id t     hazard
+#> 1:     1      1           1          1      1 1 0.17340092
+#> 2:     1      1           1          1      1 2 0.17340092
+#> 3:     1      1           1          1      1 3 0.17340092
+#> 4:     2      1           1          1      1 1 0.05617766
+#> 5:     2      1           1          1      1 2 0.05617766
+#> 6:     2      1           1          1      1 3 0.05617766
      head(transmod$cumhazard(c(1, 2, 3)))
      #>    trans sample strategy_id patient_id grp_id t  cumhazard
+#> 1:     1      1           1          1      1 1 0.17340092
+#> 2:     1      1           1          1      1 2 0.34680185
+#> 3:     1      1           1          1      1 3 0.52020277
+#> 4:     2      1           1          1      1 1 0.05617766
+#> 5:     2      1           1          1      1 2 0.11235533
+#> 6:     2      1           1          1      1 3 0.16853299
      transmod$sim_stateprobs(t = c(0, 5, 10))[t == 5]
      #>     sample strategy_id grp_id state_id t      prob
+#>  1:      1           1      1        1 5 0.0000000
+#>  2:      1           1      1        2 5 0.3333333
+#>  3:      1           1      1        3 5 0.6666667
+#>  4:      1           2      1        1 5 0.0000000
+#>  5:      1           2      1        2 5 0.3333333
+#>  6:      1           2      1        3 5 0.6666667
+#>  7:      1           3      1        1 5 0.0000000
+#>  8:      1           3      1        2 5 0.3333333
+#>  9:      1           3      1        3 5 0.6666667
+#> 10:      2           1      1        1 5 0.0000000
+#> 11:      2           1      1        2 5 0.3333333
+#> 12:      2           1      1        3 5 0.6666667
+#> 13:      2           2      1        1 5 0.6666667
+#> 14:      2           2      1        2 5 0.0000000
+#> 15:      2           2      1        3 5 0.3333333
+#> 16:      2           3      1        1 5 0.3333333
+#> 17:      2           3      1        2 5 0.0000000
+#> 18:      2           3      1        3 5 0.6666667
      - +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/Psm.html b/docs/reference/Psm.html index c48690d9..b566daf0 100644 --- a/docs/reference/Psm.html +++ b/docs/reference/Psm.html @@ -8,21 +8,29 @@ N-state partitioned survival model — Psm • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,12 +38,10 @@ - - - + - + @@ -49,9 +55,10 @@ + - +
      @@ -82,28 +89,46 @@
    • News
    -
    • - - + +
      • @@ -114,94 +139,181 @@ +

      N-state partitioned survival model

      - + Source: R/psm.R
      -

      Simulate outcomes from an N-state partitioned survival model.

      -
      - 
      Psm
      - + +

      Format

      -

      R6Class object.

      - -

      Fields

      +

      An R6::R6Class object.

      +

      See also

      - -
      +

      PsmCurves, create_PsmCurves()

      +

      Public fields

      + +

      survival_models

      The survival models used to predict survival curves. Must be -an object of class PsmCurves.

      -
      utility_model

      The model used to predict utility by health state. Must be -an object of class StateVals.

      -
      cost_models

      The models used to predict costs by health state. Must be a list of -objects of class StateVals, where each element of the list represents a -different cost category.

      -
      n_states

      Number of states in the partitioned survival model.

      -
      t_

      A numeric vector of times at which survival curves were predicted. Determined -by the argument t in sim_curves.

      -
      survival_

      Survival curves generated using sim_curves.

      -
      stateprobs_

      Health state probabilities as a function of time generated -using sim_stateprobs.

      -
      costs_

      Total (discounted) costs by type generated using sim_costs.

      -
      qalys_

      Total (discounted) quality-adjusted life-years (QALYs) generated using sim_qalys.

      -
      - +an object of class PsmCurves.

      + +
      utility_model

      The model for health state utility. Must be an object of +class StateVals.

      + +
      cost_models

      The models used to predict costs by health state. +Must be a list of objects of class StateVals, where each element of the +list represents a different cost category.

      + +
      n_states

      Number of states in the partitioned survival model.

      + +
      t_

      A numeric vector of times at which survival curves were predicted. Determined +by the argument t in $sim_curves().

      + +
      survival_

      Survival curves simulated using sim_curves().

      + +
      stateprobs_

      An object of class stateprobs simulated using $sim_stateprobs().

      + +
      qalys_

      An object of class qalys simulated using $sim_qalys().

      + +
      costs_

      An object of class costs simulated using $sim_costs().

      + +

      Methods

      -
      -
      new(survival_models, utility_model = NULL, cost_models = NULL)

      Constructor for the class. Note that the number of health states, n_states, is set equal to the number of survival - models plus one.

      -
      sim_survival(t)

      Simulate survival curves as a function of time. Equivalent to the member function - survival in PsmCurves. -

        -

      • t: A numeric vector of times. The first element must be 0.

        • -

      -
      sim_stateprobs()

      Simulate the probability of being in each of N health states using the survival curves generated - from sim_curves.

      -
      sim_qalys(dr = .03)

      Simulate (discounted) QALYs over the times selected in t associated with each health state based on the state probabilities calculated using sim_stateprobs. See "Details". -

        -

      • dr: Discount rate to apply to QALYs. May be a vector in which case QALYs are calculated for each element in dr.

        • -

      -
      sim_costs(dr = .03)

      Simulate (discounted) costs for each cost type over the times selected in t associated with each health state based on the state probabilities calculated using sim_stateprobs. See "Details". -

        -

      • dr: Discount rate to apply to costs. May be a vector in which case costs for each cost type are calculated for each element in dr.

        • -

      , -
      summarize()

      Summarize costs and QALYs so that cost-effectiveness analysis can be performed. - See summarize_ce.

      -
      check()

      Input validation for class. Checks that fields are the correct type.

      -
      - -

      Details

      +

      Public methods

      + + +

      +

      Method new()

      +

      Create a new Psm object.

      Usage

      +

      Psm$new(survival_models, utility_model = NULL, cost_models = NULL)

      + +

      Arguments

      +

      +
      survival_models

      The survival_models field.

      + +
      utility_model

      The utility_model field.

      + +
      cost_models

      The cost_models field.

      + +

      +

      Details

      +

      n_states is set equal to the number of survival models plus one.

      + +

      Returns

      +

      A new Psm object.

      +

      +

      Method sim_survival()

      +

      Simulate survival curves as a function of time using PsmCurves$sim_survival().

      Usage

      +

      Psm$sim_survival(t)

      + +

      Arguments

      +

      +
      t

      A numeric vector of times. The first element must be 0.

      + +

      +

      Returns

      +

      An instance of self with simulated output from PsmCurves$sim_survival() +stored in stateprobs_.

      +

      +

      Method sim_stateprobs()

      +

      Simulate health state probabilities from survival_ using a partitioned +survival analysis.

      Usage

      +

      Psm$sim_stateprobs()

      + +

      Returns

      +

      An instance of self with simulated output of class stateprobs +stored in stateprobs_.

      +

      +

      Method sim_qalys()

      +

      Simulate quality-adjusted life-years (QALYs) as a function of stateprobs_ and +utility_model. See vignette("expected-values") for details.

      Usage

      +

      Psm$sim_qalys(
+  dr = 0.03,
+  integrate_method = c("trapz", "riemann_left", "riemann_right"),
+  lys = FALSE
+)

      + +

      Arguments

      +

      +
      dr

      Discount rate.

      + +
      integrate_method

      Method used to integrate state values when computing (QALYs).

      + +
      lys

      If TRUE, then life-years are simulated in addition to QALYs.

      + +

      +

      Returns

      +

      An instance of self with simulated output of class qalys stored +in qalys_.

      +

      +

      Method sim_costs()

      +

      Simulate costs as a function of stateprobs_ and cost_models. +See vignette("expected-values") for details.

      Usage

      +

      Psm$sim_costs(
+  dr = 0.03,
+  integrate_method = c("trapz", "riemann_left", "riemann_right")
+)

      + +

      Arguments

      +

      +
      dr

      Discount rate.

      + +
      integrate_method

      Method used to integrate state values when computing costs.

      + +

      +

      Returns

      +

      An instance of self with simulated output of class costs stored +in costs_.

      +

      +

      Method summarize()

      +

      Summarize costs and QALYs so that cost-effectiveness analysis can be performed. +See summarize_ce().

      Usage

      +

      Psm$summarize()

      + +

      +

      Method clone()

      +

      The objects of this class are cloneable with this method.

      Usage

      +

      Psm$clone(deep = FALSE)

      + +

      Arguments

      +

      +
      deep

      Whether to make a deep clone.

      + +

      + -

      Discounted costs and QALYs are calculated by integrating the "weighted" probability of being in each state. -Weights are a function of the discount factor and the state value predicted using either the cost or QALY model. -Mathematically, discounted costs and QALYs in health state \(s\) are calculated as,

      -

      $$\int_0^T w_h e^{-rt} P_h(t) dt $$,

      -

      where for health state \(h\) and time t, \(w_h\) is the predicted cost or QALY weight, \(r\) is the discount rate, and \(P_h(t)\) is the probability of being in a given health state. The integral is calculated numerically using the composite trapezoid rule from the points in t_.

      -

      Examples

      - 
      library("flexsurv")
+    
      library("flexsurv")
 
 # Simulation data
-strategies <- data.frame(strategy_id = c(1, 2, 3))
-patients <- data.frame(patient_id = seq(1, 3),
-                          age = c(45, 50, 60),
-                          female = c(0, 0, 1))
-states <- data.frame(state_id =  seq(1, 3),
-                        state_name = paste0("state", seq(1, 3)))
+strategies <- data.frame(strategy_id = c(1, 2, 3))
+patients <- data.frame(patient_id = seq(1, 3),
+                       age = c(45, 50, 60),
+                       female = c(0, 0, 1))
+states <- data.frame(state_id =  seq(1, 3),
+                     state_name = paste0("state", seq(1, 3)))
 hesim_dat <- hesim_data(strategies = strategies,
                         patients = patients,
                         states = states)
@@ -209,29 +321,29 @@ 
      Examp
 
 # Survival models
 surv_est_data <- psm4_exdata$survival
-fit1 <- flexsurv::flexsurvreg(Surv(endpoint1_time, endpoint1_status) ~ age,
+fit1 <- flexsurv::flexsurvreg(Surv(endpoint1_time, endpoint1_status) ~ age,
                               data = surv_est_data, dist = "exp")
-fit2 <- flexsurv::flexsurvreg(Surv(endpoint2_time, endpoint2_status) ~ age,
+fit2 <- flexsurv::flexsurvreg(Surv(endpoint2_time, endpoint2_status) ~ age,
                               data = surv_est_data, dist = "exp")
-fit3 <- flexsurv::flexsurvreg(Surv(endpoint3_time, endpoint3_status) ~ age,
+fit3 <- flexsurv::flexsurvreg(Surv(endpoint3_time, endpoint3_status) ~ age,
                               data = surv_est_data, dist = "exp")
 fits <- flexsurvreg_list(fit1, fit2, fit3)
 
-surv_input_data <- expand(hesim_dat, by = c("strategies", "patients"))
+surv_input_data <- expand(hesim_dat, by = c("strategies", "patients"))
 psm_curves <- create_PsmCurves(fits, input_data = surv_input_data,
                                bootstrap = TRUE, est_data = surv_est_data,
                                n = n_samples)
 
 # Cost model(s)
-cost_input_data <- expand(hesim_dat, by = c("strategies", "patients", "states"))
-fit_costs_medical <- stats::lm(costs ~ female + state_name,
+cost_input_data <- expand(hesim_dat, by = c("strategies", "patients", "states"))
+fit_costs_medical <- stats::lm(costs ~ female + state_name,
                                data = psm4_exdata$costs$medical)
 psm_costs_medical <- create_StateVals(fit_costs_medical,
                                       input_data = cost_input_data,
                                       n = n_samples)
 
 # Utility model
-utility_tbl <- stateval_tbl(tbl = data.frame(state_id = states$state_id,
+utility_tbl <- stateval_tbl(tbl = data.frame(state_id = states$state_id,
                                              min = psm4_exdata$utility$lower,
                                              max = psm4_exdata$utility$upper),
                             dist = "unif",
@@ -241,55 +353,49 @@ 
      Examp
 # Partitioned survival decision model
 psm <- Psm$new(survival_models = psm_curves,
                utility_model = psm_utility,
-               cost_models = list(medical = psm_costs_medical))
-psm$sim_survival(t = seq(0, 5, .05))
+               cost_models = list(medical = psm_costs_medical))
+psm$sim_survival(t = seq(0, 5, .05))
 psm$sim_stateprobs()
-psm$sim_costs(dr = .03)
-head(psm$costs_)
      #>    sample strategy_id patient_id state_id   dr category    costs
-#> 1:      1           1          1        1 0.03  medical 22926.58
-#> 2:      1           1          1        2 0.03  medical 13773.33
-#> 3:      1           1          1        3 0.03  medical 15038.67
-#> 4:      1           1          2        1 0.03  medical 21886.43
-#> 5:      1           1          2        2 0.03  medical 13471.73
-#> 6:      1           1          2        3 0.03  medical 14342.99
      head(psm$sim_qalys(dr = .03)$qalys_)
      #>    sample strategy_id patient_id state_id   dr category     qalys
-#> 1:      1           1          1        1 0.03    qalys 0.5366655
-#> 2:      1           1          1        2 0.03    qalys 0.3598469
-#> 3:      1           1          1        3 0.03    qalys 0.2662830
-#> 4:      1           1          2        1 0.03    qalys 0.5123176
-#> 5:      1           1          2        2 0.03    qalys 0.3519673
-#> 6:      1           1          2        3 0.03    qalys 0.2539648
      
+psm$sim_costs(dr = .03)
+head(psm$costs_)
      #>    sample strategy_id patient_id grp_id state_id   dr category    costs
+#> 1:      1           1          1      1        1 0.03  medical 22604.40
+#> 2:      1           1          1      1        2 0.03  medical 11718.16
+#> 3:      1           1          1      1        3 0.03  medical 14536.24
+#> 4:      1           1          2      1        1 0.03  medical 21722.29
+#> 5:      1           1          2      1        2 0.03  medical 11348.15
+#> 6:      1           1          2      1        3 0.03  medical 13675.53
      head(psm$sim_qalys(dr = .03)$qalys_)
      #>    sample strategy_id patient_id grp_id state_id   dr category     qalys
+#> 1:      1           1          1      1        1 0.03    qalys 0.5288440
+#> 2:      1           1          1      1        2 0.03    qalys 0.3823026
+#> 3:      1           1          1      1        3 0.03    qalys 0.2696837
+#> 4:      1           1          2      1        1 0.03    qalys 0.5082065
+#> 5:      1           1          2      1        2 0.03    qalys 0.3702313
+#> 6:      1           1          2      1        3 0.03    qalys 0.2537155
      
+
      - +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/PsmCurves.html b/docs/reference/PsmCurves.html index bb44121d..d77548f9 100644 --- a/docs/reference/PsmCurves.html +++ b/docs/reference/PsmCurves.html @@ -8,21 +8,29 @@ Partitioned survival curves — PsmCurves • hesim + - + - + + + + - + + + - + + - + - - + + + @@ -30,12 +38,10 @@ - - - + - + @@ -49,9 +55,10 @@ + - +
      @@ -82,28 +89,46 @@
    • News
    -
    • - - + +
      • @@ -114,144 +139,225 @@ +

      Partitioned survival curves

      - + Source: R/psm.R
      -

      Summarize n-1 survival curves for an N state partitioned survival model.

      -
      - 
      PsmCurves
      - + +

      Format

      -

      R6Class object.

      - -

      Fields

      +

      An R6::R6Class object.

      +

      See also

      - -
      -
      input_mats

      Input matrices used to predict state values by strategy and patient. Must be - an object of class input_mats where each row of a matrix in "X" is a unique strategy and patient. All matrices in "X" must be sorted by strategy and patient.

      -
      params

      An object of class params_surv_list.

      -
      - +

      Psm, create_PsmCurves()

      +

      Public fields

      + +

      +
      params

      An object of class params_surv_list.

      + +
      input_data

      An object of class input_mats. Each row in X must +be a unique treatment strategy and patient.

      + +

      Methods

      -
      -
      new(input_mats, params)

      Constructor for the class.

      -
      hazard(t)

      Predict the hazard function as a function of time. -

        -

      • t: A numeric vector of times.

        • -

      -
      cumhazard(t)

      Predict the cumulative hazard function as a function of time. -

        -

      • t: A numeric vector of times.

        • -

      -
      survival(t)

      Predict the survival function as a function of time. -

        -

      • t: A numeric vector of times.

        • -

      -
      rmst(t, dr = 0)

      Predict (discounted) restricted mean survival time. -

        -

      • t: A numeric vector of times.

        • -

      • dr: Discount rate.

        • -

      -
      quantile(p)

      Predict quantiles of the survival distributions. -

        -

      • p: A numeric vector of probabilities for calculating quantiles.

        • -

      -
      check()

      Input validation for class. Checks that fields are the correct type.

      -
      - +

      Public methods

      + + +

      +

      Method new()

      +

      Create a new PsmCurves object.

      Usage

      +

      PsmCurves$new(params, input_data)

      + +

      Arguments

      +

      +
      params

      The params field.

      + +
      input_data

      The input_data field.

      + +

      +

      Returns

      +

      A new PsmCurves object.

      +

      +

      Method hazard()

      +

      Predict the hazard function for each survival curve as a function of time.

      Usage

      +

      PsmCurves$hazard(t)

      + +

      Arguments

      +

      +
      t

      A numeric vector of times.

      + +

      +

      Returns

      +

      A data.table with columns sample, strategy_id, +patient_id, grp_id, curve (the curve number), t, and hazard.

      +

      +

      Method cumhazard()

      +

      Predict the cumulative hazard function for each survival curve as a function of time.

      Usage

      +

      PsmCurves$cumhazard(t)

      + +

      Arguments

      +

      +
      t

      A numeric vector of times.

      + +

      +

      Returns

      +

      A data.table with columns sample, strategy_id, +patient_id, grp_id, curve, t, and cumhazard.

      +

      +

      Method survival()

      +

      Predict the cumulative hazard function for each survival curve as a function of time.

      Usage

      +

      PsmCurves$survival(t)

      + +

      Arguments

      +

      +
      t

      A numeric vector of times.

      + +

      +

      Returns

      +

      A data.table with columns sample, strategy_id, +patient_id, grp_id, curve, t, and survival.

      +

      +

      Method rmst()

      +

      Predict the restricted mean survival time up until time points t +for each survival curve.

      Usage

      +

      PsmCurves$rmst(t, dr = 0)

      + +

      Arguments

      +

      +
      t

      A numeric vector of times.

      + +
      dr

      Discount rate.

      + +

      +

      Returns

      +

      A data.table with columns sample, strategy_id, +patient_id, grp_id, curve, t, and rmst.

      +

      +

      Method quantile()

      +

      Predict quantiles of the survival distribution for each survival curve.

      Usage

      +

      PsmCurves$quantile(p)

      + +

      Arguments

      +

      +
      p

      A numeric vector of probabilities for computing quantiles.

      + +

      +

      Returns

      +

      A data.table with columns sample, strategy_id, +patient_id, grp_id, curve, p and quantile.

      +

      +

      Method check()

      +

      Input validation for class. Checks that fields are the correct type.

      Usage

      +

      PsmCurves$check()

      + +

      +

      Method clone()

      +

      The objects of this class are cloneable with this method.

      Usage

      +

      PsmCurves$clone(deep = FALSE)

      + +

      Arguments

      +

      +
      deep

      Whether to make a deep clone.

      + +

      + +

      Examples

      - 
      library("flexsurv")
+    
      library("flexsurv")
 
 # Simulation data
-dt_strategies <- data.frame(strategy_id = c(1, 2, 3))
-dt_patients <- data.frame(patient_id = seq(1, 3),
-                          age = c(45, 50, 60),
-                          female = c(0, 0, 1))
+dt_strategies <- data.frame(strategy_id = c(1, 2, 3))
+dt_patients <- data.frame(patient_id = seq(1, 3),
+                          age = c(45, 50, 60),
+                          female = c(0, 0, 1))
 hesim_dat <- hesim_data(strategies = dt_strategies,
                         patients = dt_patients)
 
 # Fit survival models
 surv_est_data <- psm4_exdata$survival
-fit1 <- flexsurv::flexsurvreg(Surv(endpoint1_time, endpoint1_status) ~ age,
+fit1 <- flexsurv::flexsurvreg(Surv(endpoint1_time, endpoint1_status) ~ age,
                               data = surv_est_data, dist = "exp")
-fit2 <- flexsurv::flexsurvreg(Surv(endpoint2_time, endpoint2_status) ~ age,
+fit2 <- flexsurv::flexsurvreg(Surv(endpoint2_time, endpoint2_status) ~ age,
                               data = surv_est_data, dist = "exp")
-fit3 <- flexsurv::flexsurvreg(Surv(endpoint3_time, endpoint3_status) ~ age,
+fit3 <- flexsurv::flexsurvreg(Surv(endpoint3_time, endpoint3_status) ~ age,
                               data = surv_est_data, dist = "exp")
 fits <- flexsurvreg_list(fit1, fit2, fit3)
 
 # Form PsmCurves
-surv_input_data <- expand(hesim_dat, by = c("strategies", "patients"))
+surv_input_data <- expand(hesim_dat, by = c("strategies", "patients"))
 psm_curves <- create_PsmCurves(fits, input_data = surv_input_data, n = 3,
-                                bootstrap = TRUE, est_data = surv_est_data)
+                               bootstrap = TRUE, est_data = surv_est_data)
 
 # Summarize survival curves
-head(psm_curves$quantile(p = c(.25, .5, .75)))
      #>    sample strategy_id patient_id curve    p  quantile
-#> 1:      1           1          1     1 0.25 0.1787365
-#> 2:      1           1          1     1 0.50 0.4306516
-#> 3:      1           1          1     1 0.75 0.8613031
-#> 4:      1           1          1     2 0.25 0.3576335
-#> 5:      1           1          1     2 0.50 0.8616895
-#> 6:      1           1          1     2 0.75 1.7233791
      head(psm_curves$survival(t = seq(0, 3, by = .1)))
      #>    sample strategy_id patient_id curve   t  survival
-#> 1:      1           1          1     1 0.0 1.0000000
-#> 2:      1           1          1     1 0.1 0.8513320
-#> 3:      1           1          1     1 0.2 0.7247661
-#> 4:      1           1          1     1 0.3 0.6170166
-#> 5:      1           1          1     1 0.4 0.5252859
-#> 6:      1           1          1     1 0.5 0.4471927
      head(psm_curves$rmst(t = c(2, 5)))
      #>    sample strategy_id patient_id curve t      rmst
-#> 1:      1           1          1     1 2 0.5964516
-#> 2:      1           1          1     1 5 0.6211002
-#> 3:      1           1          1     2 2 0.9943679
-#> 4:      1           1          1     2 5 1.2208820
-#> 5:      1           1          1     3 2 1.2105347
-#> 6:      1           1          1     3 5 1.6958283
      
+head(psm_curves$quantile(p = c(.25, .5, .75)))
      #>    sample strategy_id patient_id grp_id curve    p  quantile
+#> 1:      1           1          1      1     1 0.25 0.1788825
+#> 2:      1           1          1      1     1 0.50 0.4310032
+#> 3:      1           1          1      1     1 0.75 0.8620065
+#> 4:      1           1          1      1     2 0.25 0.3628697
+#> 5:      1           1          1      1     2 0.50 0.8743058
+#> 6:      1           1          1      1     2 0.75 1.7486117
      head(psm_curves$survival(t = seq(0, 3, by = .1)))
      #>    sample strategy_id patient_id grp_id curve   t  survival
+#> 1:      1           1          1      1     1 0.0 1.0000000
+#> 2:      1           1          1      1     1 0.1 0.8514438
+#> 3:      1           1          1      1     1 0.2 0.7249565
+#> 4:      1           1          1      1     1 0.3 0.6172597
+#> 5:      1           1          1      1     1 0.4 0.5255619
+#> 6:      1           1          1      1     1 0.5 0.4474864
      head(psm_curves$rmst(t = c(2, 5)))
      #>    sample strategy_id patient_id grp_id curve t      rmst
+#> 1:      1           1          1      1     1 2 0.5968732
+#> 2:      1           1          1      1     1 5 0.6216060
+#> 3:      1           1          1      1     2 2 1.0029980
+#> 4:      1           1          1      1     2 5 1.2374070
+#> 5:      1           1          1      1     3 2 1.2065710
+#> 6:      1           1          1      1     3 5 1.6858722
      
 
      
   
      
-  
 
+
       
      
       
      
   
      Developed by Devin Incerti, Jeroen P. Jansen.
      
 
      
 
 
      
-  
      Site built with pkgdown 1.3.0.
      
+  
      Site built with pkgdown 1.5.1.
      
 
      
+
      + + diff --git a/docs/reference/StateVals.html b/docs/reference/StateVals.html index 5b7e2c47..4ff7780b 100644 --- a/docs/reference/StateVals.html +++ b/docs/reference/StateVals.html @@ -8,21 +8,29 @@ Model for state values — StateVals • hesim + - + - + + + + - + + + - + + - + - - + + + @@ -30,12 +38,11 @@ - - + + - - @@ -49,9 +56,10 @@ + - +
      @@ -82,28 +90,46 @@
    • News
    -
    • - - + +
      • @@ -114,108 +140,143 @@ +

      Model for state values

      - + Source: R/statevals.R
      - -

      Simulate values (i.e., utility or costs) associated with health states in a partitioned survival or state transition model.

      - +

      Simulate values (i.e., utility or costs) associated with health states in a +state transition or partitioned survival model.

      - 
      StateVals
      - -

      Format

      -

      R6Class object.

      - -

      Fields

      - -

      -
      input_mats

      Input matrices used to simulate state values by strategy, patient, and health state. Must be - an object of class input_mats where each row in the matrix "X" is a unique strategy, patient, and - health state. "X" must be sorted by strategy, patient, and health state.

      -
      params

      An object of class params_lm, which contains the parameters for simulating - state values.

      -
      - +

      Public fields

      + +

      +
      params

      Parameters for simulating state values. Currently supports +objects of class tparams_mean or params_lm.

      + +
      input_data

      An object of class input_mats. Only used for +params_lm objects.

      + +
      method

      The method used to simulate costs and +quality-adjusted life-years (QALYs) as a function of state values. +If wlos, then costs and QALYs are +simulated by weighting state values by the length of stay in a health +state. If starting, then state values represent a one-time value +that occurs when a patient enters a health state. When starting is +used in a cohort model, the state values only accrue at time 0; +in contrast, in an individual-level model, state values +accrue each time a patient enters a new state and are discounted based on +time of entrance into that state.

      + +
      time_reset

      If FALSE then time intervals are based on time since +the start of the simulation. If TRUE, then time intervals reset each +time a patient enters a new health state. This is relevant if, for example, +costs vary over time within health states. Only used if method = wlos.

      + +

      Methods

      -
      -
      new(input_mats, params)

      Constructor for the class.

      -
      sim(t, type = c("predict", "random"))

      Simulate state values with either predicted means or random samples by treatment strategy, patient, and health state. -

        -

      • t: A numeric vector of times. The first element must be 0.

        • -

      • type: "predict" for mean values or "random" for random samples.

        • -

      -
      check()

      Input validation for class. Checks that fields are the correct type.

      -
      - +

      Public methods

      + + +

      +

      Method new()

      +

      Create a new StateVals object.

      Usage

      +

      StateVals$new(
+  params,
+  input_data = NULL,
+  method = c("wlos", "starting"),
+  time_reset = FALSE
+)

      + +

      Arguments

      +

      +
      params

      The params field.

      + +
      input_data

      The input_data field.

      + +
      method

      The method field.

      + +
      time_reset

      The time_reset field.

      + +

      +

      Returns

      +

      A new StateVals object.

      +

      +

      Method sim()

      +

      Simulate state values with either predicted means or random samples by +treatment strategy, patient, health state, and time t.

      Usage

      +

      StateVals$sim(t, type = c("predict", "random"))

      + +

      Arguments

      +

      +
      t

      A numeric vector of times.

      + +
      type

      "predict" for mean values or "random" for random samples.

      + +

      +

      Returns

      +

      A data.table of simulated state values with columns for sample, +strategy_id, patient_id, state_id, time, and value.

      +

      +

      Method check()

      +

      Input validation for class. Checks that fields are the correct type.

      Usage

      +

      StateVals$check()

      + +

      +

      Method clone()

      +

      The objects of this class are cloneable with this method.

      Usage

      +

      StateVals$clone(deep = FALSE)

      + +

      Arguments

      +

      +
      deep

      Whether to make a deep clone.

      + +

      -

      Examples

      - 
      # Simulation data
-dt_strategies <- data.frame(strategy_id = c(1, 2, 3))
-dt_patients <- data.frame(patient_id = seq(1, 3),
-                          age = c(45, 50, 60),
-                          female = c(0, 0, 1))
-dt_states <- data.frame(state_id =  seq(1, 3),
-                           state_name = paste0("state", seq(1, 3)))
-hesim_dat <- hesim_data(strategies = dt_strategies,
-                              patients = dt_patients,
-                              states = dt_states)
-
-# Create StateVals object
-fit_costs_medical <- stats::lm(costs ~ female + state_name, data = psm4_exdata$costs$medical)
-dat <- expand(hesim_dat, by = c("strategies", "patients", "states"))
-costs_medical <- create_StateVals(fit_costs_medical, input_data = dat, n = 5)
-
-# Predict
-head(costs_medical$sim(t = 5, type = "predict"))
      #>    sample strategy_id patient_id state_id time    value
-#> 1:      1           1          1        1    5 31541.88
-#> 2:      1           1          1        2    5 25135.76
-#> 3:      1           1          1        3    5 34310.12
-#> 4:      1           1          2        1    5 31541.88
-#> 5:      1           1          2        2    5 25135.76
-#> 6:      1           1          2        3    5 34310.12
      -
      - +
      +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/absorbing.html b/docs/reference/absorbing.html index b6b0dd1e..d2162c66 100644 --- a/docs/reference/absorbing.html +++ b/docs/reference/absorbing.html @@ -8,21 +8,29 @@ Absorbing states — absorbing • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,12 +38,10 @@ - - - + - + @@ -49,9 +55,10 @@ + - +
      @@ -82,28 +89,46 @@
    • News
    - - -
    • - - + +
      • @@ -114,6 +139,7 @@ +
      @@ -125,71 +151,71 @@

      A cost-effectiveness object

      -

      An object that summarizes simulated measures of clinical effectiveness and costs from a simulation model for use in a cost-effectiveness analysis.

      -
      - + +

      Format

      A list containing two elements:

      • costs Total (discounted) costs by category.

      • qalys (Discounted) quality-adjusted life-years.

      - +

      Costs

      -

      The 'costs' data.table contains the following columns:

      + +

      The 'costs' data.table contains the following columns:

      category

      The cost category.

      dr

      The discount rate.

      sample

      A randomly sampled parameter set from the probabilistic sensitivity analysis (PSA)

      strategy_id

      The treatment strategy ID.

      grp

      An optional column denoting a subgroup. If not included, it is assumed that a single subgroup is being analyzed.

      costs

      Costs.

      +
      - +

      Quality-adjusted life-years

      -

      The 'qalys' data.table contains the following columns:

      + +

      The 'qalys' data.table contains the following columns:

      dr

      The discount rate.

      sample

      A randomly sampled parameter set from the probabilistic sensitivity analysis (PSA)

      strategy_id

      The treatment strategy ID.

      grp

      An optional column denoting a subgroup. If not included, it is assumed that a single subgroup is being analyzed.

      qalys

      Quality-adjusted life-years

      -
      - - - + +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/check.html b/docs/reference/check.html index 6c8b1623..478a5730 100644 --- a/docs/reference/check.html +++ b/docs/reference/check.html @@ -6,23 +6,31 @@ -Input validation for class objects — check.input_mats • hesim +Input validation for class objects — check.id_attributes • hesim + - + - + + + + - + + + - + + - + - - + + + @@ -30,12 +38,10 @@ - - - + - + @@ -49,9 +55,10 @@ + - +
      @@ -82,28 +89,46 @@
    • News
    -
    • - - + +
      • @@ -114,28 +139,30 @@ +

      Input validation for class objects

      - Source: R/input-data.R, R/params.R, R/utils.R + Source: R/hesim_data.R, R/input_mats.R, R/params.R, and 1 more
      -

      check is a generic function for validating the inputs of class objects.

      -
      - 
      # S3 method for input_mats
+    
      # S3 method for id_attributes
 check(object)
 
-# S3 method for params_mean
+# S3 method for input_mats
 check(object)
 
+# S3 method for tparams_mean
+check(object, ...)
+
 # S3 method for params_lm
 check(object)
 
@@ -145,17 +172,23 @@ 
      Input validation for class objects
      
 # S3 method for params_surv
 check(object)
 
-# S3 method for params_joined_surv
-check(object, inner_class)
+# S3 method for params_mlogit
+check(object)
+
+# S3 method for tparams_transprobs
+check(object)
 
 # S3 method for params_surv_list
 check(object)
 
+# S3 method for params_joined_surv
+check(object, inner_class)
+
 # S3 method for params_joined_surv_list
 check(object, inner_class)
 
 check(object, ...)
      
-    
+
     
      Arguments
      
     
@@ -163,48 +196,47 @@ 
      Arg
       
      
+    
+      
+      
+    
-    
-      
-      
-    
      
     
      object
       
      object to check.
      		
     
      ...
      Further arguments passed to or from other methods.
      
     
      
       inner_class
       
      When checking a list of objects, the class of elements within
 the inner most list.
      		
     
      ...
      Further arguments passed to or from other methods. Currently unused.
      
     
      
-    
+
     
      Value
      
 
     
      If validation is successful, returns the object in question; otherwise,
 informs the user that an error has occurred.
      
-
      - +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/check_edata.html b/docs/reference/check_edata.html index 2573f100..97fa0e58 100644 --- a/docs/reference/check_edata.html +++ b/docs/reference/check_edata.html @@ -8,21 +8,29 @@ Check data argument for <code>create_input_mats</code> — check_edata • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,13 +38,11 @@ - - - + - + @@ -50,9 +56,10 @@ + - +
      @@ -83,28 +90,46 @@
    • News
    - - - - -
    • - - + +
      • @@ -118,24 +143,23 @@ +

      Create input matrices

      - Source: R/input-data.R + Source: R/input_mats.R
      -

      create_input_mats is a generic function for creating an object of class -input_mats. Model matrices are typically constructed based on the -variables specified in the model object and the data specified in data, +input_mats. Model matrices are typically constructed based on the +variables specified in the model object and the data specified in data, although there are some cases in which input_mats can be created from object alone.

      - 
      create_input_mats(object, ...)
@@ -165,8 +189,14 @@ 
      Create input matrices
      
 create_input_mats(object, input_data, ...)
 
 # S3 method for params_surv_list
+create_input_mats(object, input_data, ...)
+
+# S3 method for multinom
+create_input_mats(object, input_data, ...)
+
+# S3 method for multinom_list
 create_input_mats(object, input_data, ...)
      - +

      Arguments

      @@ -176,7 +206,7 @@

      Arg

      - + @@ -185,70 +215,64 @@

      Arg for use in a statistical models and the id variables for indexing rows in the input matrix.

      ...

      Further arguments passed to model.matrix.

      Further arguments passed to model.matrix.
      input_data
      - +

      Value

      An object of class input_mats.

      -

      See also

      input_mats.

      -

      Examples

      - 
      library("flexsurv")
-
-dt_strategies <- data.frame(strategy_id = c(1, 2))
-dt_patients <- data.frame(patient_id = seq(1, 3),
-                          age = c(45, 47, 60),
-                          female = c(1, 0, 0),
-                          group = factor(c("Good", "Medium", "Poor")))
-dt_states <- data.frame(state_id =  seq(1, 3),
-                        state_name = factor(paste0("state", seq(1, 3))))
+    
      library("flexsurv")
+
+dt_strategies <- data.frame(strategy_id = c(1, 2))
+dt_patients <- data.frame(patient_id = seq(1, 3),
+                          age = c(45, 47, 60),
+                          female = c(1, 0, 0),
+                          group = factor(c("Good", "Medium", "Poor")))
+dt_states <- data.frame(state_id =  seq(1, 3),
+                        state_name = factor(paste0("state", seq(1, 3))))
 hesim_dat <- hesim_data(strategies = dt_strategies,
                         patients = dt_patients,
                         states = dt_states)
 
 # Class "lm"
-expanded_dat <- expand(hesim_dat, by = c("strategies", "patients", "states"))
-fit_lm <- stats::lm(costs ~ female + state_name, psm4_exdata$costs$medical)
+expanded_dat <- expand(hesim_dat, by = c("strategies", "patients", "states"))
+fit_lm <- stats::lm(costs ~ female + state_name, psm4_exdata$costs$medical)
 input_mats <- create_input_mats(fit_lm, expanded_dat)
-class(input_mats)
      #> [1] "input_mats"
      
+class(input_mats)
      #> [1] "input_mats"
      
 # Class "flexsurvreg"
-expanded_dat <- expand(hesim_dat, by = c("strategies", "patients"))
-fit_wei <- flexsurv::flexsurvreg(formula = Surv(futime, fustat) ~ 1,
+expanded_dat <- expand(hesim_dat, by = c("strategies", "patients"))
+fit_wei <- flexsurv::flexsurvreg(formula = Surv(futime, fustat) ~ 1,
                                  data = ovarian, dist = "weibull")
 input_mats <- create_input_mats(fit_wei, expanded_dat)
-class(input_mats)
      #> [1] "input_mats"
      
+class(input_mats)
      #> [1] "input_mats"
      - +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/create_lines_dt.html b/docs/reference/create_lines_dt.html index 1f393aba..59058366 100644 --- a/docs/reference/create_lines_dt.html +++ b/docs/reference/create_lines_dt.html @@ -8,21 +8,29 @@ Create a data table of treatment lines — create_lines_dt • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,13 +38,11 @@ - - + + - - @@ -50,9 +56,10 @@ + - +
      @@ -83,28 +90,46 @@
    • News
    - - -
    • - - + +
      • @@ -116,6 +141,7 @@ +
      @@ -127,11 +153,9 @@

      Create a parameter object from a fitted model

      - -

      create_params is a generic function for creating an object containing +

      create_params is a generic function for creating an object containing parameters from a fitted statistical model. If point_estimate = FALSE, -then random samples from the posterior distribution are returned.

      - +then random samples from suitable probability distributions are returned.

      create_params(object, ...)
@@ -142,6 +166,12 @@ 
      Create a parameter object from a fitted model
      
 # S3 method for flexsurvreg
 create_params(object, n = 1000, point_estimate = FALSE, ...)
 
+# S3 method for multinom
+create_params(object, n = 1000, point_estimate = FALSE, ...)
+
+# S3 method for multinom_list
+create_params(object, n = 1000, point_estimate = FALSE, ...)
+
 # S3 method for flexsurvreg_list
 create_params(object, n = 1000, point_estimate = FALSE, ...)
 
@@ -154,7 +184,7 @@ 
      Create a parameter object from a fitted model
      
   max_errors = 0,
   ...
 )
      - +

      Arguments

      @@ -177,88 +207,87 @@

      Arg

      - + - +
      bootstrap

      If bootstrap is FALSE or not specified, then n parameter sets are -drawn by sampling from a multivariate normal distribution. If bootstrap is TRUE, then -parameters are bootstrapped using bootstrap.

      If bootstrap is FALSE or not specified, then n parameter sets are +drawn by sampling from a multivariate normal distribution. If bootstrap is TRUE, then +parameters are bootstrapped using bootstrap.
      max_errors

      Equivalent to the max_errors argument in bootstrap.

      Equivalent to the max_errors argument in bootstrap.
      - +

      Value

      -

      An object prefixed by params_. Mapping between create_params +

      An object prefixed by params_. Mapping between create_params and the classes of the returned objects are:

        -

      • create_params.lm -> params_lm

        • -

      • create_params.flexsurvreg -> params_surv

        • -

      • create_params.flexsurvreg_list -> params_surv_list

        • -

      • create_params.partsurvfit -> params_surv_list

        • +

      • create_params.lm -> params_lm

        • +

      • create_params.multinom -> params_mlogit

        • +

      • create_params.multinom_list -> params_mlogit_list

        • +

      • create_params.flexsurvreg -> params_surv

        • +

      • create_params.flexsurvreg_list -> params_surv_list

        • +

      • create_params.partsurvfit -> params_surv_list

      -

      Examples

      # create_params.lm
-fit <- stats::lm(costs ~ female, data = psm4_exdata$costs$medical)
+fit <- stats::lm(costs ~ female, data = psm4_exdata$costs$medical)
 n <- 5
 params_lm <- create_params(fit, n = n)
-head(params_lm$coefs)
      #>      (Intercept)    female
-#> [1,]    32854.05 1305.9017
-#> [2,]    31637.66 2554.5058
-#> [3,]    31105.35  659.2022
-#> [4,]    33760.10 2059.9307
-#> [5,]    31828.68 1601.3897
      head(params_lm$sigma)
      #> [1] 10311.3 10311.3 10311.3 10311.3 10311.3
      
+head(params_lm$coefs)
      #>      (Intercept)    female
+#> [1,]    32241.56 1362.5530
+#> [2,]    33182.47 -464.0904
+#> [3,]    29893.62 5229.2240
+#> [4,]    27190.07 4162.7643
+#> [5,]    31854.01  322.2505
      head(params_lm$sigma)
      #> [1] 10311.3 10311.3 10311.3 10311.3 10311.3
      
 # create_params.flexsurvreg
-library("flexsurv")
-fit <- flexsurv::flexsurvreg(formula = Surv(futime, fustat) ~ 1,
+library("flexsurv")
+fit <- flexsurv::flexsurvreg(formula = Surv(futime, fustat) ~ 1,
                     data = ovarian, dist = "weibull")
 n <- 5
 params_surv_wei <- create_params(fit, n = n)
-print(params_surv_wei$dist)
      #> [1] "weibull.quiet"
      head(params_surv_wei$coefs)
      #> $shape
+print(params_surv_wei$dist)
      #> [1] "weibull.quiet"
      head(params_surv_wei$coefs)
      #> $shape
 #>           shape
-#> [1,]  0.1601773
-#> [2,] -0.4105413
-#> [3,] -0.2323527
-#> [4,]  0.0533142
-#> [5,] -0.1131416
+#> [1,] 0.53510124
+#> [2,] 0.19730657
+#> [3,] 0.28343027
+#> [4,] 0.07687391
+#> [5,] 0.38388145
 #> 
 #> $scale
 #>         scale
-#> [1,] 7.367877
-#> [2,] 7.777375
-#> [3,] 7.597618
-#> [4,] 6.904363
-#> [5,] 7.696416
+#> [1,] 6.918687
+#> [2,] 6.999763
+#> [3,] 6.835843
+#> [4,] 6.983513
+#> [5,] 7.148762
 #>
      - +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/create_trans_dt.html b/docs/reference/create_trans_dt.html index 8e1e943e..1d465881 100644 --- a/docs/reference/create_trans_dt.html +++ b/docs/reference/create_trans_dt.html @@ -8,21 +8,29 @@ Create a data table of health state transitions — create_trans_dt • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,13 +38,11 @@ - - - - + - + @@ -50,9 +56,10 @@ + - +
      @@ -83,28 +90,46 @@
    • News
    -
    • - - + +
      • @@ -115,80 +140,77 @@ +

      Create a data table of health state transitions

      - Source: R/input-data.R + Source: R/hesim_data.R
      - -

      Create a data table of health state transitions from a transition matrix describing +

      Create a data table of health state transitions from a transition matrix describing the states and transitions in a multi-state model suitable for use with hesim_data.

      - 
      create_trans_dt(trans_mat)
      - +

      Arguments

      -
      trans_mat

      A transition matrix in the format from the mstate package. +

      A transition matrix in the format from the mstate package. See IndivCtstmTrans.
      - +

      Value

      -

      Returns a data.table in tidy format with three columns

      +

      Returns a data.table in tidy format with three columns

      transition_id

      Health state transition ID.

      from

      The starting health state.

      to

      The health state that will be transitions to.

      +
      -

      Examples

      - 
      tmat <- rbind(c(NA, 1, 2),
-              c(NA, NA, 3),
-              c(NA, NA, NA))
+    
      tmat <- rbind(c(NA, 1, 2),
+              c(NA, NA, 3),
+              c(NA, NA, NA))
 create_trans_dt(tmat)
      #>    transition_id from to
 #> 1:             1    1  2
 #> 2:             2    1  3
 #> 3:             3    2  3
      
   
      
-  
 
+
       
      
       
      
   
      Developed by Devin Incerti, Jeroen P. Jansen.
      
 
      
 
 
      
-  
      Site built with pkgdown 1.3.0.
      
+  
      Site built with pkgdown 1.5.1.
      
 
      
+
      + + diff --git a/docs/reference/define_model.html b/docs/reference/define_model.html new file mode 100644 index 00000000..16e6d8a2 --- /dev/null +++ b/docs/reference/define_model.html @@ -0,0 +1,376 @@ + + + + + + + + +Define and evaluate model expression — define_model • hesim + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
      +
      + + + + +
      + +
      +
      +
      +

      Define and evaluate model expression

      + Source: R/model_def.R + +
      + +
      +

      A model expression is defined by specifying random number generation functions +for a probabilistic sensitivity analysis and transformations of the sampled +parameters as a function of input_data. The unevaluated expressions +are evaluated with eval_model() and used to generate the model inputs needed to +create an economic model.

      +
      + + 
      define_model(tparams_def, rng_def, params = NULL, n_states = NULL)
+
+eval_model(x, input_data)
      + +

      Arguments

      + + + + + + + + + + + + + + + + + + + + + + + + + + +
      tparams_def

      A tparams_def object or a list of +tparams_def objects. A list might be considered if time intervals +specified with the times argument in define_tparams() vary across parameters. +Parameters for a transition probability matrix (tpmatrix), utilities (utility), +and/or cost categories (costs) are returned as a named list (see define_tparams() +for more details).

      rng_def

      A rng_def object used to randomly draw samples +of the parameters from suitable probability distributions.

      params

      A list containing the values of parameters for random +number generation.

      n_states

      The number of health states (inclusive of all health states +including the the death state) in the model. If tpmatrix is +an element returned by tparams_def, then it will be equal to the number of states +in the transition probability matrix; otherwise it must be specified as an argument.

      x

      An object of class model_def created with define_model().

      input_data

      An object of class expanded_hesim_data +expanded by patients and treatment strategies.

      + +

      Value

      + +

      define_model() returns an object of class model_def, +which is a list containing the arguments to the function. eval_model() returns +a list containing ID variables +identifying parameter samples, treatment strategies, patient cohorts, and time +intervals; the values of parameters of the transition probability matrix, +utilities, and/or cost categories; the number of health states; and the number +of random number generation samples for the PSA.

      +

      Details

      + +

      eval_model() evaluates the expressions in an object of class +model_def returned by define_model() and is, in turn, used within +functions that instantiate economic models (e.g., create_CohortDtstm()). +The direct output of eval_model() can also be useful for understanding and debugging +model definitions, but it is not used directly for simulation.

      +

      A model_def object would typically be defined in four steps by specifying:

        +

      1. Data (input_data) of class expanded_hesim_data +consisting of the treatment strategies and patient population

        2. +

      2. Parameters (params) objects storing the values of all parameters used in the model

        3. +

      3. Random number generation (define_rng()) expressions that randomly +sample values of the parameters from suitable probability distributions for probabilistic +sensitivity analysis

        4. +

      4. Transformed parameter (define_tparams()) expressions that transform +the parameter estimates into values used for simulation

        5. +
      + +

      Step 2 can be omitted if underlying parameter values are defined in steps 3. +The output of step 4 is used to instantiate the economic model (or a portion +of an economic model) as a function of a transition probability matrix, +utilities, and/or costs.

      +

      See also

      + +

      define_tparams(), define_rng()

      + +

      Examples

      + 
      
+# Data
+library("data.table")
+strategies <- data.table(strategy_id = 1:2,
+                         strategy_name = c("Monotherapy", "Combination therapy"))
+patients <- data.table(patient_id = 1)
+hesim_dat <- hesim_data(strategies = strategies,
+                       patients = patients)
+data <- expand(hesim_dat)
+
+# Define the model
+rng_def <- define_rng({
+  alpha <- matrix(c(1251, 350, 116, 17,
+                    0, 731, 512, 15,
+                    0, 0, 1312, 437,
+                    0, 0, 0, 469),
+                  nrow = 4, byrow = TRUE)
+  rownames(alpha) <- colnames(alpha) <- c("A", "B", "C", "D")
+  lrr_mean <- log(.509)
+  lrr_se <- (log(.710) - log(.365))/(2 * qnorm(.975))
+  list(
+    p_mono = dirichlet_rng(alpha),
+    rr_comb = lognormal_rng(lrr_mean, lrr_se),
+    u = 1,
+    c_zido = 2278,
+    c_lam = 2086.50,
+    c_med = gamma_rng(mean = c(A = 2756, B = 3052, C = 9007),
+                        sd = c(A = 2756, B = 3052, C = 9007))
+  )
+}, n = 2)
+
+tparams_def <- define_tparams({
+  rr = ifelse(strategy_name == "Monotherapy", 1, rr_comb)
+  list(
+    tpmatrix = tpmatrix(
+      C, p_mono$A_B * rr, p_mono$A_C * rr, p_mono$A_D * rr,
+      0, C, p_mono$B_C * rr, p_mono$B_D * rr,
+      0, 0, C, p_mono$C_D * rr,
+      0, 0, 0, 1),
+    utility = u,
+    costs = list(
+      drug = ifelse(strategy_name == "Monotherapy",
+                    c_zido, c_zido + c_lam),
+      medical = c_med
+    )
+  )
+})
+
+model_def <- define_model(
+  tparams_def = tparams_def,
+  rng_def = rng_def)
+
+# Evaluate the model expression to generate model inputs
+# This can be useful for understanding the output of a model expression
+eval_model(model_def, data)
      #> $id
+#> $id[[1]]
+#>    sample strategy_id patient_id time time_start
+#> 1:      1           1          1    0          0
+#> 2:      1           2          1    0          0
+#> 3:      2           1          1    0          0
+#> 4:      2           2          1    0          0
+#> 
+#> 
+#> $tpmatrix
+#>        s1_s1      s1_s2      s1_s3       s1_s4 s2_s1     s2_s2     s2_s3
+#> 1: 0.7003560 0.20656394 0.08062025 0.012459847     0 0.5744352 0.4112925
+#> 2: 0.8517514 0.10219732 0.03988679 0.006164498     0 0.7894522 0.2034866
+#> 3: 0.6961992 0.21960668 0.07431412 0.009879994     0 0.5712030 0.4182734
+#> 4: 0.8653078 0.09736414 0.03294768 0.004380364     0 0.8098899 0.1854444
+#>          s2_s4 s3_s1 s3_s2     s3_s3     s3_s4 s4_s1 s4_s2 s4_s3 s4_s4
+#> 1: 0.014272341     0     0 0.7332638 0.2667362     0     0     0     1
+#> 2: 0.007061228     0     0 0.8680325 0.1319675     0     0     0     1
+#> 3: 0.010523605     0     0 0.7436586 0.2563414     0     0     0     1
+#> 4: 0.004665713     0     0 0.8863493 0.1136507     0     0     0     1
+#> 
+#> $utility
+#> [1] 1 1 1 1
+#> attr(,"id_index")
+#> [1] 1
+#> 
+#> $costs
+#> $costs$drug
+#> [1] 2278.0 4364.5 2278.0 4364.5
+#> attr(,"id_index")
+#> [1] 1
+#> 
+#> $costs$medical
+#>           A        B         C
+#> 1: 2022.271 282.2611 10691.563
+#> 2: 2022.271 282.2611 10691.563
+#> 3: 1326.233 786.1712  5936.631
+#> 4: 1326.233 786.1712  5936.631
+#> 
+#> 
+#> $n_states
+#> [1] 4
+#> 
+#> $n
+#> [1] 2
+#> 
+#> attr(,"class")
+#> [1] "eval_model"
      
+# Create an economic model with a factory function
+econmod <- create_CohortDtstm(model_def, data)
      +
      + +
      + + +
      +
      +

      Developed by Devin Incerti, Jeroen P. Jansen.

      +
      + +
      +

      Site built with pkgdown 1.5.1.

      +
      + +
      +
      + + + + + + + + diff --git a/docs/reference/define_rng.html b/docs/reference/define_rng.html new file mode 100644 index 00000000..7131ec87 --- /dev/null +++ b/docs/reference/define_rng.html @@ -0,0 +1,292 @@ + + + + + + + + +Define and evaluate random number generation expressions — define_rng • hesim + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
      +
      + + + + +
      + +
      +
      +
      +

      Define and evaluate random number generation expressions

      + Source: R/model_def.R + +
      + +
      +

      Random number generation expressions are used to +randomly sample model parameters from suitable distributions for probabilistic +sensitivity analysis. These functions are typically used when evaluating +an object of class model_def defined using define_model().

      +
      + + 
      define_rng(expr, n = 1, ...)
+
+eval_rng(x, params = NULL)
      + +

      Arguments

      + + + + + + + + + + + + + + + + + + + + + + +
      expr

      An expression used to randomly draw variates for each parameter of +interest in the model. Braces should be used so that the result +of the last expression within the braces is evaluated. The expression must +return a list where each element is either a vector, matrix, data.frame, +or data.table. The length of +the vector and number of rows in the matrix/data.frame/data.table, +must either be 1 or n.

      n

      Number of samples of the parameters to draw.

      ...

      Additional arguments to pass to the environment used to evaluate +expr.

      x

      An object of class rng_def created with define_rng().

      params

      A list containing the values of parameters for random number +generation. Each element of the list should either be a vector, +matrix, data.frame, or data.table

      + +

      Value

      + +

      define_rng() returns an object of class rng_def, +which is a list containing the unevaluated random number generation +expressions passed to expr, n, and any additional arguments passed to +... . eval_rng() evaluates the rng_def object and should return a list.

      +

      Details

      + +

      hesim contains a number of random number generation functions +that return parameter samples in convenient formats +and do not require the number of samples, n, as arguments +(see rng_distributions). The random number generation expressions +are evaluated using eval_rng() and used within expr +in define_rng(). If multivariate object is returned by eval_rng(), +then the rows are random samples and columns are +distinct parameters (e.g., costs for each health state, elements of a +transition probability matrix).

      +

      See also

      + +

      rng_distributions, define_model(), define_tparams()

      + +

      Examples

      + 
      
+params <- list(
+  alpha = matrix(c(75, 25, 33, 67), byrow = TRUE, ncol = 2),
+  inptcost_mean = c(A = 900, B = 1500, C = 2000),
+  outptcost_mean = matrix(c(300, 600, 800,
+                            400, 700, 700),
+                           ncol = 3, byrow = TRUE)
+)
+rng_def <- define_rng({
+  aecost_mean <- c(500, 800, 1000) # Local object not 
+                                   # not returned by eval_rng()
+  list( # Sampled values of parameters returned by eval_rng()
+    p = dirichlet_rng(alpha), # Default column names
+    inptcost = gamma_rng(mean = inptcost_mean, # Column names based on 
+                         sd = inptcost_mean),  # named vector
+    outptcost = outptcost_mean, # No column names because
+                                # outptcost_mean has none.
+    aecost = gamma_rng(mean = aecost_mean, # Explicit naming of columns
+                       sd = aecost_mean,
+                       names = aecost_colnames)
+  )
+}, n = 2, aecost_colnames = c("A", "B", "C")) # Add aecost_colnames to environment
+eval_rng(x = rng_def, params)
      #> $p
+#>        s1_s1     s1_s2     s2_s1     s2_s2
+#> 1: 0.7330803 0.2669197 0.3250276 0.6749724
+#> 2: 0.7982298 0.2017702 0.3032865 0.6967135
+#> 
+#> $inptcost
+#>            A         B        C
+#> 1:  148.8518 219.40136 4423.237
+#> 2: 1305.4172  33.66907 3494.450
+#> 
+#> $outptcost
+#>      [,1] [,2] [,3]
+#> [1,]  300  600  800
+#> [2,]  400  700  700
+#> 
+#> $aecost
+#>            A        B         C
+#> 1:  249.4532 868.6456 2327.9261
+#> 2: 2208.7228 736.8902  119.4825
+#>
      +
      + +
      + + +
      +
      +

      Developed by Devin Incerti, Jeroen P. Jansen.

      +
      + +
      +

      Site built with pkgdown 1.5.1.

      +
      + +
      +
      + + + + + + + + diff --git a/docs/reference/define_tparams.html b/docs/reference/define_tparams.html new file mode 100644 index 00000000..4605e014 --- /dev/null +++ b/docs/reference/define_tparams.html @@ -0,0 +1,267 @@ + + + + + + + + +Define and evaluate transformed parameter expressions — define_tparams • hesim + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
      +
      + + + + +
      + +
      +
      +
      +

      Define and evaluate transformed parameter expressions

      + Source: R/model_def.R + +
      + +
      +

      Transformed parameter expressions are used to transform the parameter +values sampled with eval_rng() as a function of input data +(treatment strategies and patients) and time +intervals. These functions are used when evaluating an object of class +model_def defined using define_model(). The transformed parameters +are ultimately converted into tparams objects and used to simulate outcomes with an +economic model.

      +
      + + 
      define_tparams(expr, times = NULL, ...)
+
+eval_tparams(x, input_data, rng_params)
      + +

      Arguments

      + + + + + + + + + + + + + + + + + + + + + + + + + + +
      expr

      Expressions used to transform parameters. As with define_rng(), +braces should be used so that the result +of the last expression within the braces is evaluated. The expression +must return a named list with the following possible elements:

        +

      • tpmatrix: The transition probability matrix used to simulate transition +probabilities in the economic model. Typically the output of +tpmatrix().

        • +

      • utility: The utility values to attach to states and used to simulate +quality-adjusted life-years in the economic model. Either a vector (in +which case utility is the same in each health state) or a +data.table/data.frame/matrix with a column for each (non-death) +health state.

        • +

      • costs: A named list of costs for each category used to simulate +costs in the economic model. Each element of the +list must be in the same format as utility.

        • +
      times

      Distinct times denoting the stopping time of time intervals.

      ...

      Additional arguments to pass to the environment used to evaluate +expr.

      x

      An object of class tparams_def.

      input_data

      An object of class expanded_hesim_data (as +in eval_model()) expanded by the distinct times in times.

      rng_params

      Random samples of the parameters returned by eval_rng().

      + +

      Value

      + +

      define_tparams() returns an object of class tparams_def, +which is a list containing the unevaluated "transformation" expressions +passed to expr, times, and any additional arguments passed to +... . eval_tparams() evaluates the tparams_def object +and should return a list of transformed parameter objects.

      +

      Details

      + +

      define_tparams() is evaluated when creating economic models as a +function of model_def objects defined with define_model(). Operations +are "vectorized" in the sense that they are performed for each unique combination +of input_data and params. expr is evaluated in an environment including +each variable from input_data, all elements of rng_params, and a variable +time containing the values from times. The time variable can be used +to create models where parameters vary as a function of time. +eval_tparams() is not exported and is only meant for use within eval_model().

      +

      See also

      + +

      define_model(), define_rng()

      + +
      + +
      + + +
      +
      +

      Developed by Devin Incerti, Jeroen P. Jansen.

      +
      + +
      +

      Site built with pkgdown 1.5.1.

      +
      + +
      +
      + + + + + + + + diff --git a/docs/reference/expand.hesim_data.html b/docs/reference/expand.hesim_data.html index 90bfd09a..f760277a 100644 --- a/docs/reference/expand.hesim_data.html +++ b/docs/reference/expand.hesim_data.html @@ -8,21 +8,29 @@ Expand hesim_data — expand.hesim_data • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,15 +38,12 @@ - - + + - - @@ -52,9 +57,10 @@ + - +
      @@ -85,28 +91,46 @@
    • News
    -
    • - - + +
      • @@ -117,101 +141,119 @@ +

      Expand hesim_data

      - Source: R/input-data.R + Source: R/hesim_data.R
      - -

      Expand the data tables from an object of class hesim_data into a data table in -long format (one row for each combination of observations as specified with the -ID variables from the tables specified with the by argument). - See "Details" for an explanation of how the expansion is done.

      - +

      Create a data table in long format from all combinations of specified tables +from an object of class hesim_data and optionally time intervals. See "Details" for +an explanation of how the expansion is done.

      # S3 method for hesim_data
-expand(object, by = c("strategies", "patients"))
      - +expand(object, by = c("strategies", "patients"), times = NULL) +

      Arguments

      - + - + + + + +
      object

      An object of class hesim_data.

      An object of class hesim_data.
      by

      A character vector of the names of the data tables in hesim_data to expand by.

      A character vector of the names of the data tables in hesim_data to expand by.

      times

      Either a numeric vector of distinct times denoting the start of time intervals or +q time_intervals object.
      - +

      Value

      -

      An object of class "expanded_hesim_data", which is a data.table with an "id_vars" -attribute containing the names of the ID variables in the data table.

      - +

      An object of class expanded_hesim_data, which is a data.table with an "id_vars" +attribute containing the names of the ID variables in the data table and, if times is +not NULL, a time_intervals object derived from times.

      Details

      -

      This function is similar to expand.grid, but works for data frames or data tables. -Specifically, it creates a data.table from all combinations of the supplied tables in object. -The supplied tables are determined using the by argument. The resulting dataset is sorted by -prioritizing ID variables as follows: (i) strategy_id, (ii) line, (iii) patient_id, -(iv) the health-related ID variable (either state_id or transition_id), and -(v) the time interval (i.e., time_id).

      - +

      This function is similar to expand.grid(), but works for data frames or data tables. +Specifically, it creates a data.table from all combinations of the supplied tables in object +and optionally the start of times intervals in times. +The supplied tables are determined using the by argument. The resulting dataset is sorted by +prioritizing ID variables as follows: (i) strategy_id, (ii) patient_id, +(iii) the health-related ID variable (either state_id or transition_id, and +(iv) the time intervals from times.

      Examples

      - 
      strategies <- data.frame(strategy_id = c(1, 2))
-patients <- data.frame(patient_id = seq(1, 3), age = c(65, 50, 75),
-                          gender = c("Female", "Female", "Male"))
-states <- data.frame(state_id =  seq(1, 3),
-                     state_var = c(2, 1, 9))
+    
      strategies <- data.frame(strategy_id = c(1, 2))
+patients <- data.frame(patient_id = seq(1, 3), age = c(65, 50, 75),
+                          gender = c("Female", "Female", "Male"))
+states <- data.frame(state_id =  seq(1, 3),
+                     state_var = c(2, 1, 9))
 hesim_dat <- hesim_data(strategies = strategies,
                         patients = patients,
                         states = states)
-expand(hesim_dat, by = c("strategies", "patients"))
      #>    strategy_id patient_id age gender
+expand(hesim_dat, by = c("strategies", "patients"))
      #>    strategy_id patient_id age gender
 #> 1:           1          1  65 Female
 #> 2:           1          2  50 Female
 #> 3:           1          3  75   Male
 #> 4:           2          1  65 Female
 #> 5:           2          2  50 Female
-#> 6:           2          3  75   Male
      
+#> 6:           2          3  75   Male
      expand(hesim_dat, by = c("strategies", "patients"),
+       times = c(0, 2, 10))
      #>     strategy_id patient_id time_id age gender time_start time_stop
+#>  1:           1          1       1  65 Female          0         2
+#>  2:           1          1       2  65 Female          2        10
+#>  3:           1          1       3  65 Female         10       Inf
+#>  4:           1          2       1  50 Female          0         2
+#>  5:           1          2       2  50 Female          2        10
+#>  6:           1          2       3  50 Female         10       Inf
+#>  7:           1          3       1  75   Male          0         2
+#>  8:           1          3       2  75   Male          2        10
+#>  9:           1          3       3  75   Male         10       Inf
+#> 10:           2          1       1  65 Female          0         2
+#> 11:           2          1       2  65 Female          2        10
+#> 12:           2          1       3  65 Female         10       Inf
+#> 13:           2          2       1  50 Female          0         2
+#> 14:           2          2       2  50 Female          2        10
+#> 15:           2          2       3  50 Female         10       Inf
+#> 16:           2          3       1  75   Male          0         2
+#> 17:           2          3       2  75   Male          2        10
+#> 18:           2          3       3  75   Male         10       Inf
      - +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/expand.html b/docs/reference/expand.html index 5c94caa4..0cfc6ed1 100644 --- a/docs/reference/expand.html +++ b/docs/reference/expand.html @@ -8,21 +8,29 @@ Expand object — expand • hesim + - + - + + + + - + + + - + + - + - - + + + @@ -30,13 +38,11 @@ - - + + - - @@ -50,9 +56,10 @@ + - +
      @@ -83,28 +90,46 @@
    • News
    - -
    • - - + +
      • @@ -116,6 +141,7 @@ +
      @@ -127,15 +153,13 @@

      Random generation for generalized gamma distribution

      - -

      Draw random samples from a generalized gamma distribution using the +

      Draw random samples from a generalized gamma distribution using the parameterization from flexsurv. Written in C++ -for speed. Equivalent to flexsurv::rgengamma.

      - +for speed. Equivalent to flexsurv::rgengamma.

      fast_rgengamma(n, mu = 0, sigma = 1, Q)
      - +

      Arguments

      @@ -145,7 +169,7 @@

      Arg

      - @@ -157,55 +181,52 @@

      Arg

      mu

      Vector of location parameters. +

      Vector of location parameters. and columns correspond to rates during specified time intervals.

      Vector of shape parameters.
      - +

      Value

      -

      A vector of random samples from the generalized gamma distribution. The length of the sample is -determined by n. The numerical arguments other than n are recycled so that the number of samples is +

      A vector of random samples from the generalized gamma distribution. The length of the sample is +determined by n. The numerical arguments other than n are recycled so that the number of samples is equal to n.

      -

      Examples

      n <- 1000
 m <- 2 ; s <- 1.7; q <- 1
-ptm <- proc.time()
+ptm <- proc.time()
 r1 <- fast_rgengamma(n, mu = m, sigma = s, Q = q)
-proc.time() - ptm
      #>    user  system elapsed 
-#>   0.001   0.000   0.001 
      ptm <- proc.time()
-library("flexsurv")
-r2 <- flexsurv::rgengamma(n, mu = m, sigma = s, Q = q)
-proc.time() - ptm
      #>    user  system elapsed 
-#>   0.009   0.000   0.004 
      summary(r1)
      #>     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
-#>   0.0000   0.6979   3.5662  10.0344  11.4067 139.7566 
      summary(r2)
      #>     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
-#>   0.0000   0.8895   3.7453  11.0040  11.9801 296.0166 
      
+proc.time() - ptm
      #>    user  system elapsed 
+#>   0.001   0.000   0.001 
      ptm <- proc.time()
+library("flexsurv")
+r2 <- flexsurv::rgengamma(n, mu = m, sigma = s, Q = q)
+proc.time() - ptm
      #>    user  system elapsed 
+#>   0.002   0.000   0.002 
      summary(r1)
      #>     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
+#>   0.0000   0.7394   3.7148   9.9329  11.9836 139.7566 
      summary(r2)
      #>     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
+#>   0.0000   0.8895   3.6439  11.2069  12.5807 193.2178
      - +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/flexsurvreg_list.html b/docs/reference/flexsurvreg_list.html index 1e778f38..52e7a235 100644 --- a/docs/reference/flexsurvreg_list.html +++ b/docs/reference/flexsurvreg_list.html @@ -8,21 +8,29 @@ List of <code>flexsurvreg</code> objects — flexsurvreg_list • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,13 +38,10 @@ - - + + - - @@ -50,9 +55,10 @@ + - +
      @@ -83,28 +89,46 @@
    • News
    - - - -
    • - - + +
      • @@ -116,56 +141,45 @@ +

      Data for health-economic simulation modeling

      - Source: R/input-data.R + Source: R/hesim_data.R
      -

      A list of tables required for health-economic simulation modeling. -Each table must either be a data.frame or data.table. All ID variables within +Each table must either be a data.frame or data.table. All ID variables within each table must be numeric vectors of integers.

      -
      - 
      hesim_data(
-  strategies,
-  patients,
-  lines = NULL,
-  states = NULL,
-  transitions = NULL,
-  times = NULL
-)
      - + 
      hesim_data(strategies, patients, states = NULL, transitions = NULL)
      +

      Arguments

      - +describing the characteristics of a treatment strategy.

      - - - - - @@ -180,70 +194,51 @@

      Arg transition_id, which denotes a unique transition; from, which denotes the starting health state; and to which denotes the state that will be transitioned to.

      -

      - - -
      strategies

      A table of treatment strategies. +

      A table of treatment strategies. Must contain the column strategy_id denoting a unique strategy. Other columns are variables - describing the characteristics of a treatment strategy.
      patients

      A table of patient observations. -Must contain the column patient_id denoting a unique patient. The -number of rows should be equal to the number of patients in the model. +

      A table of patients. +Must contain the column patient_id denoting a unique patient. The +number of rows should be equal to the number of patients in the model. The table +may also include columns for grp_id for subgroups and patient_wt specifying +the weight to apply to each patient (within a subgroup). If grp_id is +NULL, then it is assumed that there is only 1 subgroup. If patient_wt is +NULL. then each patient is given the same weight. Weights +within subgroups are normalized to sum to one. Other columns are variables describing the characteristics of a patient.

      lines

      A table of treatment lines used for each treatment strategy. Must contain the columns -strategy_id, denoting a treatment strategy, and line, denoting a treatment line. Other -columns are variables describing the characteristics of a treatment line for a given treatment -strategy. A column denoting the treatment used for a given strategy and line would often -be specified. Not currently supported.
      states
      times

      A table of time intervals. Must contain the column time_start, -which denotes the starting time of times intervals defined between time_start - and time_stop. hesim_data() automatically creates the columns - time_stop and time_id (a vector of integers sorted by time_start). - Time intervals are closed on the left and open on the right and the final time - interval is defined from time_start to infinity.
      - +

      Value

      -

      Returns an object of class "hesim_data", which is a list of data tables for +

      Returns an object of class hesim_data, which is a list of data tables for health economic simulation modeling.

      -

      See also

      -

      expand.hesim_data

      - +

      expand.hesim_data()

      Examples

      - 
      strategies <- data.frame(strategy_id = c(1, 2))
-patients <- data.frame(patient_id = seq(1, 3), age = c(65, 50, 75),
-                          gender = c("Female", "Female", "Male"))
-states <- data.frame(state_id =  seq(1, 3),
-                        state_var = c(2, 1, 9))
-times <- data.frame(time_id = c(1, 2, 3),
-                    time_start = c(0, 4, 9),
-                    time_stop = c(4, 9, Inf))
+    
      strategies <- data.frame(strategy_id = c(1, 2))
+patients <- data.frame(patient_id = seq(1, 3), age = c(65, 50, 75),
+                       gender = c("Female", "Female", "Male"))
+states <- data.frame(state_id =  seq(1, 3),
+                        state_var = c(2, 1, 9))
 hesim_dat <- hesim_data(strategies = strategies,
                          patients = patients,
-                         states = states,
-                         times = times)
      
+                         states = states)
      - +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/hesim_survdists.html b/docs/reference/hesim_survdists.html index 643b80cd..7d34e8c3 100644 --- a/docs/reference/hesim_survdists.html +++ b/docs/reference/hesim_survdists.html @@ -8,21 +8,29 @@ List of survival distributions — hesim_survdists • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,14 +38,12 @@ - - - - + - + @@ -51,9 +57,10 @@ + - +
      @@ -84,28 +91,46 @@
    • News
    -
    • - - + +
      • @@ -116,6 +141,7 @@ +
      @@ -127,61 +153,62 @@

      List of survival distributions

      - -

      List of additional distributions for parametric survival analysis that are -not contained in flexsurv. Can be used to fit models with +

      List of additional distributions for parametric survival analysis that are +not contained in flexsurv. Can be used to fit models with flexsurvreg. Same format as flexsurv.dists in flexsurv.

      - 
      hesim_survdists
      - + +

      Format

      -

      A list with the following elements:

      -
      name

      Name of the probability distribution.

      -
      pars

      Vector of strings naming the parameters of the distribution. - These must be the same names as the arguments of the density and probability functions.

      -
      location

      Name of the location parameter.

      -
      transforms

      List of R functions which transform the range of values - taken by each parameter onto the real line. For example, - c(log, log) for a distribution with two positive parameters.

      -
      inv.transforms

      List of R functions defining the corresponding inverse - transformations. Note these must be lists, even for single parameter - distributions they should be supplied as, e.g. c(exp) or list(exp).

      -
      inits

      A function of the observed survival times t (including - right-censoring times, and using the halfway point for interval-censored times) - which returns a vector of reasonable initial values for maximum likelihood - estimation of each parameter. For example, function(t){ c(1, mean(t)) } - will always initialize the first of two parameters at 1, and the second - (a scale parameter, for instance) at the mean of t.

      +

      A list with the following elements:

      +
      name

      Name of the probability distribution.

      +
      pars

      Vector of strings naming the parameters of the distribution. +These must be the same names as the arguments of the density and probability functions.

      +
      location

      Name of the location parameter.

      +
      transforms

      List of R functions which transform the range of values +taken by each parameter onto the real line. For example, +c(log, log) for a distribution with two positive parameters.

      +
      inv.transforms

      List of R functions defining the corresponding inverse +transformations. Note these must be lists, even for single parameter +distributions they should be supplied as, e.g. c(exp) or list(exp).

      +
      inits

      A function of the observed survival times t (including +right-censoring times, and using the halfway point for interval-censored times) +which returns a vector of reasonable initial values for maximum likelihood +estimation of each parameter. For example, function(t){ c(1, mean(t)) } +will always initialize the first of two parameters at 1, and the second +(a scale parameter, for instance) at the mean of t.

      +
      - - - + +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/icea.html b/docs/reference/icea.html index cc7a02f3..46da2110 100644 --- a/docs/reference/icea.html +++ b/docs/reference/icea.html @@ -8,21 +8,29 @@ Individualized cost-effectiveness analysis — icea • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,21 +38,20 @@ - - - + - + @@ -58,9 +65,10 @@ + - +
      @@ -91,28 +99,46 @@
    • News
    -
    • - - + +
      • @@ -123,6 +149,7 @@ +
      @@ -134,18 +161,17 @@

      Individualized cost-effectiveness analysis

      -

      Conduct individualized cost-effectiveness analysis (ICEA) given output of an economic -model. That is, summarize a probabilistic sensitivity analysis (PSA) by subgroup.

        +model; that is, summarize a probabilistic sensitivity analysis (PSA) by subgroup.

        • icea() computes the probability that each treatment is most cost-effective, output for a cost-effectiveness acceptability frontier, the expected value of perfect information, and the net monetary benefit for each treatment.

        • icea_pw() conducts pairwise ICEA by comparing strategies to a comparator. Computed -quantities include the incremental cost-effectiveness ratio, the +quantities include the incremental cost-effectiveness ratio, the incremental net monetary benefit, output for a cost-effectiveness plane, and output for a cost-effectiveness acceptability curve.

        - + 
      icea(x, ...)
@@ -153,12 +179,12 @@ 
      Individualized cost-effectiveness analysis
      
 icea_pw(x, ...)
 
 # S3 method for default
-icea(x, k = seq(0, 2e+05, 500), sample, strategy, grp = NULL, e, c, ...)
+icea(x, k = seq(0, 2e+05, 500), sample, strategy, grp = NULL, e, c, ...)
 
 # S3 method for default
 icea_pw(
   x,
-  k = seq(0, 2e+05, 500),
+  k = seq(0, 2e+05, 500),
   comparator,
   sample,
   strategy,
@@ -169,11 +195,11 @@ 
      Individualized cost-effectiveness analysis
      
 )
 
 # S3 method for ce
-icea(x, k = seq(0, 2e+05, 500), dr_qalys, dr_costs, ...)
+icea(x, k = seq(0, 2e+05, 500), dr_qalys, dr_costs, ...)
 
 # S3 method for ce
-icea_pw(x, k = seq(0, 2e+05, 500), comparator, dr_qalys, dr_costs, ...)
      - +icea_pw(x, k = seq(0, 2e+05, 500), comparator, dr_qalys, dr_costs, ...) +

      Arguments

      @@ -216,7 +242,7 @@

      Arg

      - + @@ -227,120 +253,119 @@

      Arg

      comparator

      Name of the comparator strategy in x.

      Name of the comparator strategy in x.
      dr_qalys

      Discount rate for costs.
      - +

      Value

      -

      icea returns a list containing four data.tables:

      -
      -
      summary

      A data.table of the mean, 2.5% quantile, and 97.5% - quantile by strategy and group for clinical effectiveness and costs.

      -
      mce

      The probability that each strategy is the most effective treatment - for each group for the range of specified willingness to pay values. In addition, - the column best denotes the optimal strategy (i.e., the strategy with the - highest expected net monetary benefit), which can be used to plot the - cost-effectiveness acceptability frontier (CEAF).

      -
      evpi

      The expected value of perfect information (EVPI) by group for the range - of specified willingness to pay values. The EVPI is computed by subtracting the expected net - monetary benefit given current information (i.e., the strategy with the highest - expected net monetary benefit) from the expected net monetary benefit given - perfect information.

      -
      nmb

      The mean, 2.5% quantile, and 97.5% quantile of net monetary benefits - for the range of specified willingness to pay values.

      +

      icea() returns a list of four data.table elements.

      + +
      summary

      A data.table of the mean, 2.5% quantile, and 97.5% +quantile by strategy and group for clinical effectiveness and costs.

      +
      mce

      The probability that each strategy is the most effective treatment +for each group for the range of specified willingness to pay values. In addition, +the column best denotes the optimal strategy (i.e., the strategy with the +highest expected net monetary benefit), which can be used to plot the +cost-effectiveness acceptability frontier (CEAF).

      +
      evpi

      The expected value of perfect information (EVPI) by group for the range +of specified willingness to pay values. The EVPI is computed by subtracting the expected net +monetary benefit given current information (i.e., the strategy with the highest +expected net monetary benefit) from the expected net monetary benefit given +perfect information.

      +
      nmb

      The mean, 2.5% quantile, and 97.5% quantile of net monetary benefits +for the range of specified willingness to pay values.

      +
      -

      icea_pw also returns a list containing four data.tables:

      -
      summary

      A data.table of the mean, 2.5% quantile, and 97.5% - quantile by strategy and group for clinical effectiveness and costs.

      -
      delta

      Incremental effectiveness and incremental cost for each simulated - parameter set by strategy and group. Can be used to plot a cost-effectiveness plane.

      -
      ceac

      Values needed to plot a cost-effectiveness acceptability curve by - group. The CEAC plots the probability that each strategy is more cost-effective than - the comparator for the specified willingness to pay values.

      -
      inmb

      The mean, 2.5% quantile, and 97.5% quantile of - incremental net monetary benefits for the range of specified willingness to pay values.

      +

      icea_pw also returns a list of four data.table elements:

      +
      summary

      A data.table of the mean, 2.5% quantile, and 97.5% +quantile by strategy and group for clinical effectiveness and costs.

      +
      delta

      Incremental effectiveness and incremental cost for each simulated +parameter set by strategy and group. Can be used to plot a cost-effectiveness plane.

      +
      ceac

      Values needed to plot a cost-effectiveness acceptability curve by +group. The CEAC plots the probability that each strategy is more cost-effective than +the comparator for the specified willingness to pay values.

      +
      inmb

      The mean, 2.5% quantile, and 97.5% quantile of +incremental net monetary benefits for the range of specified willingness to pay values.

      +
      -

      Examples

      # simulation output
 n_samples <- 100
-sim <- data.frame(sample = rep(seq(n_samples), 4),
-                  c = c(rlnorm(n_samples, 5, .1), rlnorm(n_samples, 5, .1),
-                        rlnorm(n_samples, 11, .1), rlnorm(n_samples, 11, .1)),
-                  e = c(rnorm(n_samples, 8, .2), rnorm(n_samples, 8.5, .1),
-                        rnorm(n_samples, 11, .6), rnorm(n_samples, 11.5, .6)),
-                  strategy = rep(paste0("Strategy ", seq(1, 2)),
+sim <- data.frame(sample = rep(seq(n_samples), 4),
+                  c = c(rlnorm(n_samples, 5, .1), rlnorm(n_samples, 5, .1),
+                        rlnorm(n_samples, 11, .1), rlnorm(n_samples, 11, .1)),
+                  e = c(rnorm(n_samples, 8, .2), rnorm(n_samples, 8.5, .1),
+                        rnorm(n_samples, 11, .6), rnorm(n_samples, 11.5, .6)),
+                  strategy = rep(paste0("Strategy ", seq(1, 2)),
                                  each = n_samples * 2),
-                  grp = rep(rep(c("Group 1", "Group 2"),
+                  grp = rep(rep(c("Group 1", "Group 2"),
                             each = n_samples), 2)
 )
 
 # icea
-icea <- icea(sim, k = seq(0, 200000, 500), sample = "sample", strategy = "strategy",
+icea <- icea(sim, k = seq(0, 200000, 500), sample = "sample", strategy = "strategy",
              grp = "grp", e = "e", c = "c")
-names(icea)
      #> [1] "summary" "mce"     "evpi"    "nmb"    
      # The probability that each strategy is the most cost-effective 
+names(icea)
      #> [1] "summary" "mce"     "evpi"    "nmb"    
      # The probability that each strategy is the most cost-effective 
 # in each group with a willingness to pay of 20,000
-library("data.table")
+library("data.table")
 icea$mce[k == 20000]
      #>        k   strategy     grp best prob
 #> 1: 20000 Strategy 1 Group 1    0 0.47
 #> 2: 20000 Strategy 2 Group 1    1 0.53
-#> 3: 20000 Strategy 1 Group 2    0 0.43
-#> 4: 20000 Strategy 2 Group 2    1 0.57
      
+#> 3: 20000 Strategy 1 Group 2    1 0.55
+#> 4: 20000 Strategy 2 Group 2    0 0.45
      
 # icea_pw
-icea_pw <-  icea_pw(sim,  k = seq(0, 200000, 500), comparator = "Strategy 1",
+icea_pw <-  icea_pw(sim,  k = seq(0, 200000, 500), comparator = "Strategy 1",
                     sample = "sample", strategy = "strategy", grp = "grp",
                      e = "e", c = "c")
-names(icea_pw)
      #> [1] "summary" "delta"   "ceac"    "inmb"   
      # cost-effectiveness acceptability curve
-head(icea_pw$ceac[k >= 20000])
      #>        k   strategy     grp prob
+names(icea_pw)
      #> [1] "summary" "delta"   "ceac"    "inmb"   
      # cost-effectiveness acceptability curve
+head(icea_pw$ceac[k >= 20000])
      #>        k   strategy     grp prob
 #> 1: 20000 Strategy 2 Group 1 0.53
-#> 2: 20000 Strategy 2 Group 2 0.57
-#> 3: 20500 Strategy 2 Group 1 0.57
-#> 4: 20500 Strategy 2 Group 2 0.60
-#> 5: 21000 Strategy 2 Group 1 0.60
-#> 6: 21000 Strategy 2 Group 2 0.62
      icer_tbl(icea_pw)
      #> $`Group 1`
+#> 2: 20000 Strategy 2 Group 2 0.45
+#> 3: 20500 Strategy 2 Group 1 0.63
+#> 4: 20500 Strategy 2 Group 2 0.49
+#> 5: 21000 Strategy 2 Group 1 0.68
+#> 6: 21000 Strategy 2 Group 2 0.57
      icer_tbl(icea_pw)
      #> $`Group 1`
 #>                   Strategy 1 Strategy 2                
-#> Incremental QALYs "-"        "2.98 (1.83, 4.20)"       
-#> Incremental costs "-"        "58,928 (47,414, 71,155)" 
-#> Incremental NMB   "-"        "89,869 (32,459, 154,012)"
-#> ICER              "-"        "19,802"                  
+#> Incremental QALYs "-"        "3.10 (1.59, 4.23)"       
+#> Incremental costs "-"        "60,071 (49,644, 72,671)" 
+#> Incremental NMB   "-"        "94,708 (26,763, 152,056)"
+#> ICER              "-"        "19,405"                  
 #> Conclusion        "-"        "Cost-effective"          
 #> 
 #> $`Group 2`
 #>                   Strategy 1 Strategy 2                
-#> Incremental QALYs "-"        "3.06 (2.14, 4.08)"       
-#> Incremental costs "-"        "59,870 (50,922, 71,718)" 
-#> Incremental NMB   "-"        "93,158 (39,806, 143,108)"
-#> ICER              "-"        "19,562"                  
+#> Incremental QALYs "-"        "2.96 (1.82, 3.86)"       
+#> Incremental costs "-"        "60,563 (50,026, 71,975)" 
+#> Incremental NMB   "-"        "87,279 (31,241, 135,557)"
+#> ICER              "-"        "20,482"                  
 #> Conclusion        "-"        "Cost-effective"          
 #>
      - +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/icer_tbl.html b/docs/reference/icer_tbl.html index 19c018ea..1760e082 100644 --- a/docs/reference/icer_tbl.html +++ b/docs/reference/icer_tbl.html @@ -8,21 +8,29 @@ ICER table — icer_tbl • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,13 +38,11 @@ - - + + - - @@ -50,9 +56,10 @@ + - +
      @@ -83,28 +90,46 @@
    • News
    -
    • - - + +
      • @@ -115,6 +140,7 @@ +
      @@ -126,10 +152,8 @@

      ICER table

      - -

      Generate a table of incremental cost-effectiveness ratios given output from -icea_pw.

      - +

      Generate a table of incremental cost-effectiveness ratios given output from +icea_pw().

      icer_tbl(
@@ -139,18 +163,18 @@ 
      ICER table
      
   prob = 0.95,
   digits_qalys = 2,
   digits_costs = 0,
-  output = c("matrix", "data.table"),
+  output = c("matrix", "data.table"),
   rownames = NULL,
   colnames = NULL,
   drop = TRUE
 )
      - +

      Arguments

      - + @@ -158,7 +182,7 @@

      Arg

      - @@ -189,58 +213,54 @@

      Arg

      -
      x

      An object of class "icea_pw" returned by icea_pw.

      An object of class icea_pw returned by icea_pw().
      k
      cri

      If TRUE, credible intervals are computed; otherwise +

      If TRUE, credible intervals are computed; otherwise they are not.
      drop

      If TRUE, then the result is coerced to the lowest possible dimension. +

      If TRUE, then the result is coerced to the lowest possible dimension. Relevant if output = "matrix" and there is one group, in which case a single matrix will be returned if drop = TRUE and a list of length 1 will be returned if drop = FALSE.
      - +

      Value

      If output = "matrix", then a list of matrices (or a matrix if drop = TRUE) reporting incremental cost-effectiveness ratios (ICERs) -by group. Specifically, each matrix contains five rows for: (i) +by group. Specifically, each matrix contains five rows for: (i) incremental quality-adjusted life-years (QALYs), (ii) incremental costs, -(iii) the incremental net monetary benefit (NMB), (iv) the ICER, +(iii) the incremental net monetary benefit (NMB), (iv) the ICER, and (v) a conclusion stating whether each strategy is cost-effective relative - to a comparator. The number of columns is equal to the +to a comparator. The number of columns is equal to the number of strategies (including the comparator).

      If output = "data.table", then the results are reported as a data.table, with one row for each strategy and group combination.

      -

      See also

      -

      icea_pw

      - +

      icea_pw()

      - +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/id_attributes.html b/docs/reference/id_attributes.html new file mode 100644 index 00000000..115b2b7a --- /dev/null +++ b/docs/reference/id_attributes.html @@ -0,0 +1,293 @@ + + + + + + + + +Attributes for ID variables — id_attributes • hesim + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
      +
      + + + + +
      + +
      +
      +
      +

      Attributes for ID variables

      + Source: R/hesim_data.R + +
      + +
      +

      Stores metadata related to the ID variables used to index input_mats +and transformed parameter objects already predicted from covariates.

      +
      + + 
      id_attributes(
+  strategy_id,
+  n_strategies,
+  patient_id,
+  n_patients,
+  state_id = NULL,
+  n_states = NULL,
+  transition_id = NULL,
+  n_transitions = NULL,
+  time_id = NULL,
+  time_intervals = NULL,
+  n_times = NULL,
+  sample = NULL,
+  n_samples = NULL,
+  grp_id = NULL,
+  patient_wt = NULL
+)
      + +

      Arguments

      + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
      strategy_id

      A numeric vector of integers denoting the treatment strategy.

      n_strategies

      A scalar denoting the number of unique treatment strategies.

      patient_id

      A numeric vector of integers denoting the patient.

      n_patients

      A scalar denoting the number of unique patients.

      state_id

      A numeric vector of integers denoting the health state.

      n_states

      A scalar denoting the number of unique health states.

      transition_id

      A numeric vector denoting the +health state transition. This is only used for state transition models.

      n_transitions

      A scalar denoting the number of unique transitions.

      time_id

      A numeric vector of integers denoting a unique time interval.

      time_intervals

      A data.table denoting unique time intervals. Must +contain the columns time_id, time_start, and time_stop. +time_start is the starting time of an interval and time_stop is +the stopping time of an interval. Following the survival package, +time intervals are closed on the right and +open on the left (except in the final interval where time_stop is equal to +infinity).

      n_times

      A scalar denoting the number of time intervals. Equal to the +number of rows in time_intervals.

      sample

      A numeric vector of integer denoting the sample from the posterior +distribution of the parameters.

      n_samples

      A scalar denoting the number of samples.

      grp_id

      An optional numeric vector of integers denoting the subgroup.

      patient_wt

      An optional numeric vector denoting the weight to apply to each patient +within a subgroup.

      + +

      Details

      + +

      When using the ID variables to index input_mats, sorting order should be +the same as specified in expand.hesim_data(); that is, +observations must be sorted by: (i) strategy_id, (ii) patient_id, +and (iii) the health-related ID variable (either state_id or +transition_id). When using ID variables to index transformed parameter +objects and sample is used for indexing, then observations must be sorted by: +(i) sample, (ii) strategy_id, (iii) patient_id, and +(iv) the health-related ID variable.

      +

      See also

      + +

      hesim_data(),expand.hesim_data(), input_mats

      + +
      + +
      + + +
      +
      +

      Developed by Devin Incerti, Jeroen P. Jansen.

      +
      + +
      +

      Site built with pkgdown 1.5.1.

      +
      + +
      +
      + + + + + + + + diff --git a/docs/reference/incr_effect.html b/docs/reference/incr_effect.html index dca37d19..fd14fd86 100644 --- a/docs/reference/incr_effect.html +++ b/docs/reference/incr_effect.html @@ -8,21 +8,29 @@ Incremental treatment effect — incr_effect • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,13 +38,11 @@ - - - - + - + @@ -50,9 +56,10 @@ + - +
      @@ -83,28 +90,46 @@
    • News
    -
    • - - + +
      • @@ -115,6 +140,7 @@ +
      @@ -126,14 +152,12 @@

      Incremental treatment effect

      - -

      Computes incremental effect for all treatment strategies +

      Computes incremental effect for all treatment strategies on outcome variables from a probabilistic sensitivity analysis relative to a comparator.

      - 
      incr_effect(x, comparator, sample, strategy, grp = NULL, outcomes)
      - +

      Arguments

      @@ -168,64 +192,61 @@

      Arg

      Name of columns to compute incremental changes for.
      - +

      Value

      -

      A data.table containing the differences in the values of each variable -specified in outcomes between each treatment strategy and the +

      A data.table containing the differences in the values of each variable +specified in outcomes between each treatment strategy and the comparator.

      -

      Examples

      # simulation output
 n_samples <- 100
-sim <- data.frame(sample = rep(seq(n_samples), 4),
-             c = c(rlnorm(n_samples, 5, .1), rlnorm(n_samples, 5, .1),
-                    rlnorm(n_samples, 11, .1), rlnorm(n_samples, 11, .1)),
-             e = c(rnorm(n_samples, 8, .2), rnorm(n_samples, 8.5, .1),
-                   rnorm(n_samples, 11, .6), rnorm(n_samples, 11.5, .6)),
-             strategy = rep(paste0("Strategy ", seq(1, 2)),
+sim <- data.frame(sample = rep(seq(n_samples), 4),
+             c = c(rlnorm(n_samples, 5, .1), rlnorm(n_samples, 5, .1),
+                    rlnorm(n_samples, 11, .1), rlnorm(n_samples, 11, .1)),
+             e = c(rnorm(n_samples, 8, .2), rnorm(n_samples, 8.5, .1),
+                   rnorm(n_samples, 11, .6), rnorm(n_samples, 11.5, .6)),
+             strategy = rep(paste0("Strategy ", seq(1, 2)),
                            each = n_samples * 2),
-             grp = rep(rep(c("Group 1", "Group 2"),
+             grp = rep(rep(c("Group 1", "Group 2"),
                            each = n_samples), 2)
 )
 # calculate incremental effect of Strategy 2 relative to Strategy 1 by group
 ie <- incr_effect(sim, comparator = "Strategy 1", sample = "sample",
-                        strategy = "strategy", grp = "grp", outcomes = c("c", "e"))
-head(ie)
      #>    sample   strategy     grp       ic       ie
-#> 1:      1 Strategy 2 Group 1 52301.34 3.664019
-#> 2:      1 Strategy 2 Group 2 61294.32 2.372158
-#> 3:      2 Strategy 2 Group 1 60549.64 2.576371
-#> 4:      2 Strategy 2 Group 2 63767.20 3.557785
-#> 5:      3 Strategy 2 Group 1 55596.69 2.578505
-#> 6:      3 Strategy 2 Group 2 64515.01 3.662150
      + strategy = "strategy", grp = "grp", outcomes = c("c", "e")) +head(ie)
      #> sample strategy grp ic ie +#> 1: 1 Strategy 2 Group 1 77063.21 2.767131 +#> 2: 1 Strategy 2 Group 2 60556.68 2.098251 +#> 3: 2 Strategy 2 Group 1 63041.84 2.265125 +#> 4: 2 Strategy 2 Group 2 51160.36 3.114136 +#> 5: 3 Strategy 2 Group 1 46096.08 1.887154 +#> 6: 3 Strategy 2 Group 2 64185.74 3.286395
      - +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/index.html b/docs/reference/index.html index acda4020..9b1d78c6 100644 --- a/docs/reference/index.html +++ b/docs/reference/index.html @@ -8,21 +8,29 @@ Function reference • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,12 +38,12 @@ - - + + @@ -46,9 +54,10 @@ + - +
      @@ -79,28 +88,46 @@
    • News
    -     -
    • - - + +
      • @@ -118,175 +142,71 @@ +

      Input matrices for a statistical model

      - Source: R/input-data.R + Source: R/input_mats.R
      - -

      Create an object of class "input_mats", which contains inputs matrices -for simulating a statistical model. Consists of (i) input matrices, X, -(ii) ID variables indexing the rows of each matrix in X, and (iii) the dimensions of -the X matrices. More details are provided under "Details" below. Note that an "input_mats" -object should be created using create_input_mats.

      - +

      Create an object of class input_mats, which contains inputs matrices +for simulating a statistical model. Consists of (i) input matrices, X, and +(ii) metadata used to index each matrix in X. +More details are provided under "Details" below.

      - 
      input_mats(
-  X,
-  strategy_id,
-  n_strategies,
-  patient_id,
-  n_patients,
-  line = NULL,
-  n_lines = NULL,
-  state_id = NULL,
-  n_states = NULL,
-  transition_id = NULL,
-  n_transitions = NULL,
-  time_id = NULL,
-  time_intervals = NULL,
-  n_times = NULL,
-  time_reset = NULL,
-  time_fun = NULL
-)
      - + 
      input_mats(X, ...)
      +

      Arguments

      - - - - - - - - - + - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - + +
      X

      A list of input matrices for predicting the values of each parameter in a statistical model. May also be -a list of lists of input matrices when a list of separate models is fit (e.g., with flexsurvreg_list).

      strategy_id

      A numeric vector of integers denoting the treatment strategy represented by each row -in X.

      n_strategies

      A scalar denoting the number of unique treatment strategies.

      A list of input matrices for predicting the values of each parameter +in a statistical model. May also be a list of lists of input matrices when a +list of separate models is fit (e.g., with flexsurvreg_list()).
      patient_id

      A numeric vector of integers denoting the patient represented by each row -in X.

      n_patients

      A scalar denoting the number of unique patients.

      line

      A numeric vector of integers denoting the treatment line represented by each row -in X. Not supported by currently available models.

      n_lines

      A data.table denoting the number of treatment lines associated -with each treatment strategy. Should contain a column, "strategy_id", and a column, -"N". Not supported by currently available models.

      state_id

      A numeric vector of integers denoting the health state represented by each row -in X.

      n_states

      A scalar denoting the number of unique health states.

      transition_id

      A numeric vector denoting the -health state transition represented by each row in X. This must only be specified when -estimating the health state transitions with a joint likelihood function. If independent -models are fit for each transition, then separate X matrices must be specified -for each transition. Note that this is not currently supported but -will be supported once hesim provides support for state transition modeling.

      n_transitions

      A scalar denoting the number of unique transitions. -Not supported by currently available models.

      time_id

      A numeric vector of integers denoting a unique time interval.

      time_intervals

      A data.table denotes unique time intervals. Must -contain the columns time_id, time_start, and time_stop. -time_start is the starting time of an interval and time_stop is -the stopping time of an interval. Time intervals are closed on the left and -open on the right, and in the final interval, time_stop is equal to -infinity.

      n_times

      A scalar denoting the number of time intervals. Equal to the -number of rows in time_intervals.

      time_reset

      If TRUE, then time intervals reset each time a patient enters a new health -state. This is relevant if, for example, costs vary over time within health states. -If FALSE then time intervals are based on time since the start of the simulation.

      time_fun

      A pointer to a C++ functor that can be used to update X as a function -of time in a simulation model. Not currently supported.

      		...

      Arguments to pass to id_attributes().
      - +

      Details

      -

      Each row of each matrix X is an input vector, \(x_{hijk}\), where \(h\) denotes -a health-related index, \(i\) indexes a patient, \(j\) indexes a treatment line, -and \(k\) is a treatment strategy. A health-related index is either a health state -(e.g., state_id) or a transition between health states (e.g., transition_id). +

      Each row of each matrix X is an input vector, \(x_{hik}\), where \(h\) denotes +a health-related index, \(i\) indexes a patient, and \(k\) is a treatment strategy. +A health-related index is either a health state +(e.g., state_id or a transition between health states (e.g., transition_id). In some cases, the health-related index \(h\) can be suppressed and separate models -can be fit for each health index. This is, for instance, the case in a partitioned survival -model where separate models are fit for each survival endpoint. Likewise, models -can be fit without multiple treatment lines as would, again, be the case in a -partitioned survival analysis where sequential treatment would be incorporated by -adding additional health states rather than by using the index \(j\).

      -

      The rows of the matrices in X must be sorted in a manner consistent with the ID variables. -The sorting order should be the same as specified in expand.hesim_data; that is, -the rows of X must be sorted by: (i) strategy_id, (ii) line, -(iii) patient_id, and (iv) the health-related ID variable (either state_id or - transition_id).

      - +can be fit for each health index. This is, for instance, the case in a partitioned survival model where separate models are fit for each survival endpoint.

      +

      The rows of the matrices in X must be sorted in a manner consistent with the ID variables as +described in id_attributes().

      See also

      -

      create_input_mats

      - +

      create_input_mats()

      Examples

      - 
      strategies <- data.frame(strategy_id = c(1, 2))
-patients <- data.frame(patient_id = seq(1, 3),
-                          age = c(45, 47, 60),
-                          female = c(1, 0, 0),
-                          group = factor(c("Good", "Medium", "Poor")))
+    
      strategies <- data.frame(strategy_id = c(1, 2))
+patients <- data.frame(patient_id = seq(1, 3),
+                          age = c(45, 47, 60),
+                          female = c(1, 0, 0),
+                          group = factor(c("Good", "Medium", "Poor")))
 hesim_dat <- hesim_data(strategies = strategies,
                         patients = patients)
 
-dat <- expand(hesim_dat, by = c("strategies", "patients"))
-input_mats <- input_mats(X = list(mu = model.matrix(~ age, dat)),
+dat <- expand(hesim_dat, by = c("strategies", "patients"))
+input_mats <- input_mats(X = list(mu = model.matrix(~ age, dat)),
                          strategy_id = dat$strategy_id,
-                         n_strategies = length(unique(dat$strategy_id)),
+                         n_strategies = length(unique(dat$strategy_id)),
                          patient_id = dat$patient_id,
-                        n_patients = length(unique(dat$patient_id)))
-print(input_mats)
      #> $X
+                        n_patients = length(unique(dat$patient_id)))
+print(input_mats)
      #> $X
 #> $X$mu
 #>   (Intercept) age
 #> 1           1  45
@@ -311,37 +231,36 @@ 
      Examp
 #> $n_patients
 #> [1] 3
 #> 
+#> $grp_id
+#> [1] 1 1 1 1 1 1
+#> 
 #> attr(,"class")
 #> [1] "input_mats"
      
   
      
-  
 
+
       
      
       
      
   
      Developed by Devin Incerti, Jeroen P. Jansen.
      
 
      
 
 
      
-  
      Site built with pkgdown 1.3.0.
      
+  
      Site built with pkgdown 1.5.1.
      
 
      
+
      + + diff --git a/docs/reference/joined.html b/docs/reference/joined.html index 5b65de61..600c65c1 100644 --- a/docs/reference/joined.html +++ b/docs/reference/joined.html @@ -8,21 +8,29 @@ Join statistical models at specified times — joined • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,17 +38,15 @@ - - - + - + @@ -54,9 +60,10 @@ + - +
      @@ -87,28 +94,46 @@
    • News
    -
    • - - + +
      • @@ -119,6 +144,7 @@ +
      @@ -130,14 +156,12 @@

      Join statistical models at specified times

      -

      Construct classes which are used to join statistical models at specified time points. The first model is used prior to the first time point, the second model is used prior to the second time point, and so on. There should be N + 1 models and N time points. Can be used to join a single set of models (e.g., for survival analysis or for making predictions from a jointly -estimated multi-state model) or to join multiple sets of models (e.g., predicting curves in a +estimated multi-state model) or to join multiple sets of models (e.g., predicting curves in a partitioned survival analysis).

      - 
      joined_formula(..., times)
@@ -147,14 +171,14 @@ 
      Join statistical models at specified times
      
 joined_flexsurvreg(..., times)
 
 joined_flexsurvreg_list(..., times)
      - +

      Arguments

      @@ -164,34 +188,34 @@

      Arg with the length of each list element equal to the number of sets of models.

      ...

      Model objects to join. Currently supports objects of class -formula, formula_list, flexsurvreg, and +formula, formula_list, flexsurvreg, and flexsurvreg_list.
      - +

      Value

      -

      Returns a class of the same name as the function that is a -list with two variables:

      +

      Returns a class of the same name as the function that is a +list with two variables:

      models

      List of joined models.

      times

      Time points to join models.

      +
      -

      Examples

      # joined_formula
-joined_fm <- joined_formula(f1 = formula(~ age), f2 = formula(~ 1),
-                                 f3 = formula(~age + age2),
-                                 times = c(2, 3))
-print(joined_fm)
      #> $models
+joined_fm <- joined_formula(f1 = formula(~ age), f2 = formula(~ 1),
+                                 f3 = formula(~age + age2),
+                                 times = c(2, 3))
+print(joined_fm)
      #> $models
 #> $models$f1
 #> ~age
-#> <environment: 0x7fa755a2a120>
+#> <environment: 0x7fbab3a5f9b0>
 #> 
 #> $models$f2
 #> ~1
-#> <environment: 0x7fa755a2a120>
+#> <environment: 0x7fbab3a5f9b0>
 #> 
 #> $models$f3
 #> ~age + age2
-#> <environment: 0x7fa755a2a120>
+#> <environment: 0x7fbab3a5f9b0>
 #> 
 #> 
 #> $times
@@ -200,46 +224,44 @@ 
      Examp
 #> attr(,"class")
 #> [1] "joined_formula"
      
 # joined_formula_list
-fm_list1 <- formula_list(f1 = formula(~ age), f2 = formula(~ 1))
-fm_list2 <- formula_list(f1 = formula(~ age), f2 = formula(~ age2), f3 = formula(~age3))
+fm_list1 <- formula_list(f1 = formula(~ age), f2 = formula(~ 1))
+fm_list2 <- formula_list(f1 = formula(~ age), f2 = formula(~ age2), f3 = formula(~age3))
 joined_fm_list <- joined_formula_list(model1 = fm_list1, model2 = fm_list2,
-                                     times = list(5, c(2, 4)))
-class(joined_fm_list)
      #> [1] "joined_formula_list"
      
+                                     times = list(5, c(2, 4)))
+class(joined_fm_list)
      #> [1] "joined_formula_list"
      
 # joined_flexsurvreg
-library("flexsurv")
-fit_wei <- flexsurv::flexsurvreg(formula = Surv(futime, fustat) ~ 1,
+library("flexsurv")
+fit_wei <- flexsurv::flexsurvreg(formula = Surv(futime, fustat) ~ 1,
                                 data = ovarian, dist = "weibull")
-fit_exp <- flexsurv::flexsurvreg(formula = Surv(futime, fustat) ~ 1,
+fit_exp <- flexsurv::flexsurvreg(formula = Surv(futime, fustat) ~ 1,
                                  data = ovarian, dist = "exp")
 joined_fsr <- joined_flexsurvreg(fit_wei, fit_exp, times = 5)
-class(joined_fsr)
      #> [1] "joined_flexsurvreg"
      +class(joined_fsr)
      #> [1] "joined_flexsurvreg"
      - +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/lm_list.html b/docs/reference/lm_list.html index 1d776c24..057afa11 100644 --- a/docs/reference/lm_list.html +++ b/docs/reference/lm_list.html @@ -8,21 +8,29 @@ List of <code>lm</code> objects — lm_list • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,13 +38,10 @@ - - + + - - @@ -50,9 +55,10 @@ + - +
      @@ -83,28 +89,46 @@
    • News
    - -
    • - - + +
      • @@ -116,6 +141,7 @@ +
      @@ -127,15 +153,13 @@

      Method of moments for beta distribution

      -

      Compute the parameters shape1 and shape2 of the beta distribution -using method of moments given the mean and standard +using method of moments given the mean and standard deviation of the random variable of interest.

      - 
      mom_beta(mean, sd)
      - +

      Arguments

      @@ -148,21 +172,19 @@

      Arg

      Standard deviation of the random variable.
      - +

      Value

      A list containing the parameters shape1 and shape2.

      -

      Details

      -

      If \(\mu\) is the mean and +

      If \(\mu\) is the mean and \(\sigma\) is the standard deviation of the random variable, then the method of moments estimates of the parameters shape1 = \(\alpha > 0\) and shape2 = \(\beta > 0\) are: $$\alpha = \mu \left(\frac{\mu(1-\mu)}{\sigma^2}-1 \right)$$ and $$\beta = (1 - \mu) \left(\frac{\mu(1-\mu)}{\sigma^2}-1 \right)$$

      -

      Examples

      mom_beta(mean = .8, sd = .1)
      #> $shape1
@@ -171,7 +193,7 @@ 
      Examp
 #> $shape2
 #> [1] 3
 #> 
      # The function is vectorized.
-mom_beta(mean = c(.6, .8), sd = c(.08, .1))
      #> $shape1
+mom_beta(mean = c(.6, .8), sd = c(.08, .1))
      #> $shape1
 #> [1] 21.9 12.0
 #> 
 #> $shape2
@@ -179,34 +201,30 @@ 
      Examp
 #>
      - +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/mom_gamma.html b/docs/reference/mom_gamma.html index 24b375e0..eb44ac52 100644 --- a/docs/reference/mom_gamma.html +++ b/docs/reference/mom_gamma.html @@ -8,21 +8,29 @@ Method of moments for gamma distribution — mom_gamma • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,13 +38,11 @@ - - - + - + @@ -50,9 +56,10 @@ + - +
      @@ -83,28 +90,46 @@
    • News
    -
    • - - + +
      • @@ -115,6 +140,7 @@ +
      @@ -126,14 +152,12 @@

      Method of moments for gamma distribution

      -

      Compute the shape and scale (or rate) parameters of the gamma distribution using method of moments for the random variable of interest.

      - 
      mom_gamma(mean, sd, scale = TRUE)
      - +

      Arguments

      @@ -151,15 +175,14 @@

      Arg the rate parameter is returned.

      - +

      Value

      If scale = TRUE, then a list containing the parameters shape and scale; otherwise, - if scale = FALSE, then a list containing the parameters shape and rate.

      - +if scale = FALSE, then a list containing the parameters shape and rate.

      Details

      -

      If \(\mu\) is the mean and +

      If \(\mu\) is the mean and \(\sigma\) is the standard deviation of the random variable, then the method of moments estimates of the parameters shape = \(\alpha > 0\) and scale = \(\theta > 0\) are: @@ -167,7 +190,6 @@

      Details and $$\alpha = \frac{\mu}{\theta}$$

      The inverse of the scale parameter, \(\beta = 1/\theta\), is the rate parameter.

      -

      Examples

      mom_gamma(mean = 10000, sd = 2000)
      #> $shape
@@ -176,7 +198,7 @@ 
      Examp
 #> $scale
 #> [1] 400
 #> 
      # The function is vectorized.
-mom_gamma(mean = c(8000, 10000), sd = c(1500, 2000))
      #> $shape
+mom_gamma(mean = c(8000, 10000), sd = c(1500, 2000))
      #> $shape
 #> [1] 28.44444 25.00000
 #> 
 #> $scale
@@ -184,34 +206,30 @@ 
      Examp
 #>
      - +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/mstate3_exdata.html b/docs/reference/mstate3_exdata.html new file mode 100644 index 00000000..5e2d0b45 --- /dev/null +++ b/docs/reference/mstate3_exdata.html @@ -0,0 +1,251 @@ + + + + + + + + +Example data for a 3-state multi-state model — mstate3_exdata • hesim + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
      +
      + + + + +
      + +
      +
      +
      +

      Example data for a 3-state multi-state model

      + Source: R/data.R + +
      + +
      +

      Example multi-state data for parameterizing a continuous time state +transition model. Costs and utility +are also included to facilitate cost-effectiveness analysis.

      +
      + + 
      mstate3_exdata
      + + +

      Format

      + +

      A list containing the following elements:

        +

      • transitions A data frame containing the times at which patient transitions between health states based +on the prothr dataset from the mstate package.

        • +

      • costs A list of data frames. The first data frame contains summary medical cost estimates and the +second data frame contains drug cost data.

        • +

      • utility A data frame of summary utility estimates.

        • +
      + +

      Transitions data

      + + + +

      The data frame has the following columns:

      +
      strategy_id

      Treatment strategy identification number.

      +
      patient_id

      Patient identification number.

      +
      age

      Patient age (in years).

      +
      female

      1 if a patient is female; 0 if male.

      +
      from

      Starting state.

      +
      to

      Receiving state.

      +
      trans

      Transition number.

      +
      Tstart

      Starting time.

      +
      Tstop

      Transition time.

      +
      years

      Elapsed years between Tstart and Tstop.

      +
      status

      Status variable; 1=transition, 0=censored.

      + + + +

      Cost data

      + + + +

      The cost list contains two data frames. The first data frame contains +data on the drug costs associated with each treatment strategy.

      +
      strategy_id

      The treatment strategy identification number.

      +
      costs

      Annualized drug costs.

      + + + +

      The second data frame contains summary data on +medical costs by health state, and contains the following columns:

      +
      state_id

      The health state identification number.

      +
      mean

      Mean costs.

      +
      se

      Standard error of medical costs.

      + + + +

      Utility data

      + + + +

      The data frame has the following columns:

      +
      state_id

      The health state identification number.

      +
      mean

      Mean utility

      +
      se

      Standard error of utility

      + + + + +
      + +
      + + +
      +
      +

      Developed by Devin Incerti, Jeroen P. Jansen.

      +
      + +
      +

      Site built with pkgdown 1.5.1.

      +
      + +
      +
      + + + + + + + + diff --git a/docs/reference/multinom3_exdata.html b/docs/reference/multinom3_exdata.html new file mode 100644 index 00000000..ba07fa1b --- /dev/null +++ b/docs/reference/multinom3_exdata.html @@ -0,0 +1,255 @@ + + + + + + + + +Example data for a 3-state multinomial model — multinom3_exdata • hesim + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
      +
      + + + + +
      + +
      +
      +
      +

      Example data for a 3-state multinomial model

      + Source: R/data.R + +
      + +
      +

      Example discrete time health state transitions data simulated using +multinomial logistic regression. Costs and utility +are also included to facilitate cost-effectiveness analysis.

      +
      + + 
      multinom3_exdata
      + + +

      Format

      + +

      A list containing the following elements:

        +

      • transitions A data frame containing patient transitions between health +states at discrete time intervals (i.e., on a yearly basis).

        • +

      • costs A list of data frames. The first data frame contains +drug cost data and the second contains summary medical cost estimates.

        • +

      • utility A data frame of summary utility estimates.

        • +
      + +

      Transitions data

      + + + +

      The data frame has the following columns:

      +
      patient_id

      Patient identification number.

      +
      strategy_id

      Treatment strategy identification number.

      +
      strategy_name

      Treatment strategy name.

      +
      age

      Patient age (in years).

      +
      age_cat

      A factor variable with 3 age groups: (i) age less than 40, +(ii) age at least 40 and less than 60, and (iii) age at least 60.

      +
      female

      1 if a patient is female; 0 if male.

      +
      year

      The year since the start of data collection with the first year equal to 1.

      +
      state_from

      State making a transition from.

      +
      state_to

      State making a transition to.

      +
      year_cat

      Factor variable for year with 3 categories: (i) year 3 and below, (ii) +year between 3 and 6, and (iii) year 7 and above.

      + + + +

      Cost data

      + + + +

      The cost list contains two data frames. The first data frame contains +data on the drug costs associated with each treatment strategy.

      +
      strategy_id

      The treatment strategy identification number.

      +
      strategy_name

      The treatment strategy name.

      +
      costs

      Annualized drug costs.

      + + + +

      The second data frame contains summary data on +medical costs by health state, and contains the following columns:

      +
      state_id

      The health state identification number.

      +
      state_name

      The name of the health state.

      +
      mean

      Mean medical costs.

      +
      se

      Standard error of medical costs.

      + + + +

      Utility data

      + + + +

      The data frame has the following columns:

      +
      state_id

      The health state identification number.

      +
      state_name

      The name of the health state.

      +
      mean

      Mean utility

      +
      se

      Standard error of utility.

      + + + + +
      + +
      + + +
      +
      +

      Developed by Devin Incerti, Jeroen P. Jansen.

      +
      + +
      +

      Site built with pkgdown 1.5.1.

      +
      + +
      +
      + + + + + + + + diff --git a/docs/reference/multinom_list.html b/docs/reference/multinom_list.html new file mode 100644 index 00000000..7bc8457b --- /dev/null +++ b/docs/reference/multinom_list.html @@ -0,0 +1,220 @@ + + + + + + + + +List of <code>multinom</code> objects — multinom_list • hesim + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
      +
      + + + + +
      + +
      +
      +
      +

      List of multinom objects

      + Source: R/model-fits.R + +
      + +
      +

      Combine multinom objects into a list.

      +
      + + 
      multinom_list(...)
      + +

      Arguments

      + + + + + + +
      ...

      Objects of class multinom, which can be named.

      + +

      Value

      + +

      An object of class multinom_list.

      + +

      Examples

      + 
       library("nnet")
+ library("data.table")
+ trans_data <- data.table(multinom3_exdata$transitions)
+ dat_healthy <- trans_data[state_from == "Healthy"]
+ fit_healthy <- multinom(state_to ~ strategy_name + female + age_cat + year_cat,
+                          data = dat_healthy)
      #> # weights:  24 (14 variable)
+#> initial  value 18461.080899 
+#> iter  10 value 18166.573671
+#> final  value 18139.640776 
+#> converged
       dat_sick <- trans_data[state_from == "Sick"]
+ dat_sick$state_to <- droplevels(dat_sick$state_to)
+ fit_sick <- multinom(state_to ~ strategy_name + female + age_cat + year_cat,
+                      data = dat_sick)
      #> # weights:  8 (7 variable)
+#> initial  value 6800.466988 
+#> iter  10 value 6733.594702
+#> iter  10 value 6733.594663
+#> iter  10 value 6733.594663
+#> final  value 6733.594663 
+#> converged
       fits <- multinom_list(healthy = fit_healthy, sick = fit_sick)
+ class(fits)
      #> [1] "multinom_list"
      +
      + +
      + + +
      +
      +

      Developed by Devin Incerti, Jeroen P. Jansen.

      +
      + +
      +

      Site built with pkgdown 1.5.1.

      +
      + +
      +
      + + + + + + + + diff --git a/docs/reference/params.html b/docs/reference/params.html new file mode 100644 index 00000000..c457883a --- /dev/null +++ b/docs/reference/params.html @@ -0,0 +1,194 @@ + + + + + + + + +Parameter object — params • hesim + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
      +
      + + + + +
      + +
      +
      +
      +

      Parameter object

      + Source: R/params.R + +
      + +
      +

      Objects prefixed by "params_" are lists containing the parameters of a statistical model +used for simulation modeling. The parameters are used to simulate outcomes +as a function of covariates contained in input matrices (input_mats).

      +
      + + + +

      See also

      + +

      tparams

      + +
      + +
      + + +
      +
      +

      Developed by Devin Incerti, Jeroen P. Jansen.

      +
      + +
      +

      Site built with pkgdown 1.5.1.

      +
      + +
      +
      + + + + + + + + diff --git a/docs/reference/params_joined_surv.html b/docs/reference/params_joined_surv.html index c1ad520b..8dd1ffdb 100644 --- a/docs/reference/params_joined_surv.html +++ b/docs/reference/params_joined_surv.html @@ -8,21 +8,29 @@ Parameters of joined survival models — params_joined_surv • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,13 +38,11 @@ - - - + - + @@ -50,9 +56,10 @@ + - +
      @@ -83,28 +90,46 @@
    • News
    -
    • - - + +
      • @@ -115,6 +140,7 @@ +
      @@ -126,14 +152,12 @@

      Parameters of joined survival models

      -

      Create a list containing the parameters of survival models joined at specified time points. See joined for more details.

      - 
      params_joined_surv(..., times)
      - +

      Arguments

      @@ -146,35 +170,35 @@

      Arg

      A numeric vector of times at which to join models.
      - +

      Value

      -

      An object of class "params_joined_surv", which is a list containing two elements:

      +

      An object of class "params_joined_surv", which is a list containing two elements:

      models

      A list of params_surv objects from each statistical model to be joined.

      times

      Equivalent to the argument times.

      +
      -

      Examples

      - 
      library("flexsurv")
-fit_exp <- flexsurv::flexsurvreg(formula = Surv(futime, fustat) ~ 1,
+    
      library("flexsurv")
+fit_exp <- flexsurv::flexsurvreg(formula = Surv(futime, fustat) ~ 1,
                                  data = ovarian, dist = "exp")
-fit_wei <- flexsurv::flexsurvreg(formula = Surv(futime, fustat) ~ 1,
+fit_wei <- flexsurv::flexsurvreg(formula = Surv(futime, fustat) ~ 1,
                                  data = ovarian, dist = "weibull")
 params_surv_exp <- create_params(fit_exp, n = 2)
 params_surv_wei <- create_params(fit_wei, n = 2)
 params_joined_surv <- params_joined_surv(exp = params_surv_exp,
                                          wei = params_surv_wei,
                                          times = 3)
-print(params_joined_surv)
      #> $models
+print(params_joined_surv)
      #> $models
 #> $models$exp
 #> $coefs
 #> $coefs$rate
 #>           rate
-#> [1,] -7.407699
-#> [2,] -7.186788
+#> [1,] -7.312052
+#> [2,] -7.260642
 #> 
 #> 
 #> $dist
@@ -190,13 +214,13 @@ 
      Examp
 #> $coefs
 #> $coefs$shape
 #>          shape
-#> [1,] 0.5908195
-#> [2,] 0.1636856
+#> [1,] 0.4708211
+#> [2,] 0.5861358
 #> 
 #> $coefs$scale
 #>         scale
-#> [1,] 6.790387
-#> [2,] 6.592881
+#> [1,] 6.331590
+#> [2,] 6.886001
 #> 
 #> 
 #> $dist
@@ -215,32 +239,30 @@ 
      Examp
 #> attr(,"class")
 #> [1] "params_joined_surv"
      
   
      
-  
 
+
       
      
       
      
   
      Developed by Devin Incerti, Jeroen P. Jansen.
      
 
      
 
 
      
-  
      Site built with pkgdown 1.3.0.
      
+  
      Site built with pkgdown 1.5.1.
      
 
      
+
       
      
    
 
   
 
+
   
 
 
+
diff --git a/docs/reference/params_joined_surv_list.html b/docs/reference/params_joined_surv_list.html
index 15adae8a..376ccb5e 100644
--- a/docs/reference/params_joined_surv_list.html
+++ b/docs/reference/params_joined_surv_list.html
@@ -8,21 +8,29 @@
 
 Parameters of joined lists of survival models — params_joined_surv_list • hesim
 
+
 
-
+
 
-
+
+
 
-
+
+
+
+
+
 
 
-
+
+
 
 
-
+
 
-
-
+
+
+
 
 
 
@@ -30,13 +38,11 @@
 
 
 
-  
-  
-
 
+
 
-
+
 
 
 
@@ -50,9 +56,10 @@
 
 
 
+
   
 
-  
+  
     
      
       
      
       
 
@@ -83,28 +90,46 @@
   
   
 
 
    • 
   News
 
      •
    -
    • - - + +
      • @@ -115,6 +140,7 @@ +
      @@ -126,14 +152,12 @@

      Parameters of joined lists of survival models

      -

      Create a list containing the parameters of multiple sets of survival models, each joined at specified time points. See joined for more details.

      - 
      params_joined_surv_list(..., times)
      - +

      Arguments

      @@ -147,24 +171,24 @@

      Arg to the number of sets of models.

      - +

      Value

      -

      An object of class "params_joined_surv_list", which is a list containing two elements:

      -
      models

      A list of params_surv_list, each containing codeparams_surv +

      An object of class "params_joined_surv_list", which is a list containing two elements:

      +
      models

      A list of params_surv_list, each containing codeparams_surv objects to be joined.

      times

      Equivalent to the argument times.

      +
      -

      Examples

      - 
      library("flexsurv")
-fit_exp <- flexsurv::flexsurvreg(formula = Surv(futime, fustat) ~ 1,
+    
      library("flexsurv")
+fit_exp <- flexsurv::flexsurvreg(formula = Surv(futime, fustat) ~ 1,
                                  data = ovarian, dist = "exp")
-fit_wei <- flexsurv::flexsurvreg(formula = Surv(futime, fustat) ~ 1,
+fit_wei <- flexsurv::flexsurvreg(formula = Surv(futime, fustat) ~ 1,
                                  data = ovarian, dist = "weibull")
-fit_lnorm <- flexsurv::flexsurvreg(formula = Surv(futime, fustat) ~ 1,
+fit_lnorm <- flexsurv::flexsurvreg(formula = Surv(futime, fustat) ~ 1,
                                   data = ovarian, dist = "lognormal")
 
 params_exp <- create_params(fit_exp, n = 2)
@@ -175,15 +199,15 @@ 
      Examp
 params_list2 <- params_surv_list(params_exp, params_lnorm)
 params_joined <- params_joined_surv_list(model1 = params_list1,
                                          model2 = params_list2,
-                                         times = list(3, 5))
-print(params_joined)
      #> $models
+                                         times = list(3, 5))
+print(params_joined)
      #> $models
 #> $models$model1
 #> [[1]]
 #> $coefs
 #> $coefs$rate
 #>           rate
-#> [1,] -6.937318
-#> [2,] -7.494966
+#> [1,] -7.255898
+#> [2,] -7.188072
 #> 
 #> 
 #> $dist
@@ -198,14 +222,14 @@ 
      Examp
 #> [[2]]
 #> $coefs
 #> $coefs$shape
-#>           shape
-#> [1,] -0.3444579
-#> [2,]  0.2861824
+#>          shape
+#> [1,] 0.1864200
+#> [2,] 0.2972981
 #> 
 #> $coefs$scale
 #>         scale
-#> [1,] 7.266289
-#> [2,] 6.928428
+#> [1,] 7.590157
+#> [2,] 7.268442
 #> 
 #> 
 #> $dist
@@ -225,8 +249,8 @@ 
      Examp
 #> $coefs
 #> $coefs$rate
 #>           rate
-#> [1,] -6.937318
-#> [2,] -7.494966
+#> [1,] -7.255898
+#> [2,] -7.188072
 #> 
 #> 
 #> $dist
@@ -242,13 +266,13 @@ 
      Examp
 #> $coefs
 #> $coefs$meanlog
 #>       meanlog
-#> [1,] 6.791645
-#> [2,] 7.431594
+#> [1,] 6.550419
+#> [2,] 6.825682
 #> 
 #> $coefs$sdlog
-#>          sdlog
-#> [1,] 0.2705199
-#> [2,] 0.5021063
+#>           sdlog
+#> [1,]  0.1856125
+#> [2,] -0.1294268
 #> 
 #> 
 #> $dist
@@ -275,32 +299,30 @@ 
      Examp
 #> attr(,"class")
 #> [1] "params_joined_surv_list"
      
   
      
-  
 
+
       
      
       
      
   
      Developed by Devin Incerti, Jeroen P. Jansen.
      
 
      
 
 
      
-  
      Site built with pkgdown 1.3.0.
      
+  
      Site built with pkgdown 1.5.1.
      
 
      
+
       
      
    
 
   
 
+
   
 
 
+
diff --git a/docs/reference/params_lm.html b/docs/reference/params_lm.html
index 5be8ac5e..a079fe57 100644
--- a/docs/reference/params_lm.html
+++ b/docs/reference/params_lm.html
@@ -8,21 +8,29 @@
 
 Parameters of a linear model — params_lm • hesim
 
+
 
-
+
 
-
+
+
 
-
+
+
+
+
+
 
 
-
+
+
 
 
-
+
 
-
-
+
+
+
 
 
 
@@ -30,12 +38,10 @@
 
 
 
-  
-  
-
 
+
 
-
+
 
 
 
@@ -49,9 +55,10 @@
 
 
 
+
   
 
-  
+  
     
      
       
      
       
 
@@ -82,28 +89,46 @@
   
   
 
 
    • 
   News
 
      •
    -
    • - - + +
      • @@ -114,6 +139,7 @@ +
      @@ -125,59 +151,54 @@

      Parameters of a linear model

      -

      Create a list containing the parameters of a fitted linear regression model.

      - 
      params_lm(coefs, sigma = NULL)
      - +

      Arguments

      - + -
      coefs

      Matrix of samples from the posterior distribution of the -regression coefficients.

      Matrix of samples of the coefficients under sampling uncertainty.
      sigma

      A vector of samples of the standard error of the regression model. -Must only be specified if the model is used to randomly simulate values +

      A vector of samples of the standard error of the regression model. +Must only be specified if the model is used to randomly simulate values (rather than to predict means).
      - +

      Value

      -

      An object of class "params_lm", which is a list containing coefs, +

      An object of class params_lm, which is a list containing coefs, sigma, and n_samples. n_samples is equal to the number of rows in coefs.

      -

      Details

      Fitted linear models are used to predict values, \(y\), - as a function of covariates, \(x\), +as a function of covariates, \(x\), $$y = x^T\beta + \epsilon.$$ Predicted means are given by \(x^T\hat{\beta}\) where \(\hat{\beta}\) -is the vector of estimated regression coefficients. Random samples are obtained by -sampling the error term from a normal distribution, +is the vector of estimated regression coefficients. Random samples are obtained by +sampling the error term from a normal distribution, \(\epsilon \sim N(0, \hat{\sigma}^2)\).

      -

      Examples

      - 
      library("MASS")
+    
      library("MASS")
 n <- 2
-params <- params_lm(coefs = MASS::mvrnorm(n, mu = c(.5,.6),
-                                            Sigma = matrix(c(.05, .01, .01, .05), nrow = 2)),
-                      sigma <- rgamma(n, shape = .5, rate = 4))
-print(params)
      #> $coefs
+params <- params_lm(coefs = MASS::mvrnorm(n, mu = c(.5,.6),
+                                            Sigma = matrix(c(.05, .01, .01, .05), nrow = 2)),
+                      sigma <- rgamma(n, shape = .5, rate = 4))
+print(params)
      #> $coefs
 #>           [,1]      [,2]
-#> [1,] 0.3731079 0.5771039
-#> [2,] 0.6574268 0.9524617
+#> [1,] 0.4120082 0.9566176
+#> [2,] 0.7203465 0.6891463
 #> 
 #> $sigma
-#> [1] 0.10306939 0.00953951
+#> [1] 0.04613469 0.02744829
 #> 
 #> $n_samples
 #> [1] 2
@@ -186,34 +207,30 @@ 
      Examp
 #> [1] "params_lm"
      
 
      
   
      
-  
 
+
       
      
       
      
   
      Developed by Devin Incerti, Jeroen P. Jansen.
      
 
      
 
 
      
-  
      Site built with pkgdown 1.3.0.
      
+  
      Site built with pkgdown 1.5.1.
      
 
      
+
       
      
    
 
   
 
+
   
 
 
+
diff --git a/docs/reference/params_lm_list.html b/docs/reference/params_lm_list.html
index 44427cef..78eb388f 100644
--- a/docs/reference/params_lm_list.html
+++ b/docs/reference/params_lm_list.html
@@ -8,21 +8,29 @@
 
 Parameters of a list of linear models — params_lm_list • hesim
 
+
 
-
+
 
-
+
+
 
-
+
+
+
+
+
 
 
-
+
+
 
 
-
+
 
-
-
+
+
+
 
 
 
@@ -30,12 +38,10 @@
 
 
 
-  
-  
-
 
+
 
-
+
 
 
 
@@ -49,9 +55,10 @@
 
 
 
+
   
 
-  
+  
     
      
       
      
       
 
@@ -82,28 +89,46 @@
   
   
 
 
    • 
   News
 
      •
    -
    • - - + +
      • @@ -114,6 +139,7 @@ +
      @@ -125,47 +151,44 @@

      Parameters of a list of linear models

      -

      Create a list containing the parameters of a list of fitted linear regression models.

      - 
      params_lm_list(...)
      - +

      Arguments

      - +
      ...

      Objects of class params_lm, which can be named.

      Objects of class params_lm, which can be named.
      - +

      Value

      -

      An object of class "params_lm_list", which is a list containing params_lm +

      An object of class params_lm_list, which is a list containing params_lm objects.

      -

      Examples

      n <- 2
-coefs1 <- MASS::mvrnorm(n, mu = c(.5,.6),
-                        Sigma = matrix(c(.05, .01, .01, .05), nrow = 2))
-sigma1 <- rgamma(n, shape = .5, rate = 4)
+coefs1 <- MASS::mvrnorm(n, mu = c(.5,.6),
+                        Sigma = matrix(c(.05, .01, .01, .05), nrow = 2))
+sigma1 <- rgamma(n, shape = .5, rate = 4)
 params1 <- params_lm(coefs = coefs1, sigma = sigma1)
-coefs2 <- MASS::mvrnorm(n, mu = c(.2,.9),
-                        Sigma = matrix(c(.08, .02, .02, .08), nrow = 2))
-sigma2 <- rgamma(n, shape = .9, rate = 4)
+coefs2 <- MASS::mvrnorm(n, mu = c(.2,.9),
+                        Sigma = matrix(c(.08, .02, .02, .08), nrow = 2))
+sigma2 <- rgamma(n, shape = .9, rate = 4)
 params2 <- params_lm(coefs = coefs2, sigma = sigma2)
 params_list <- params_lm_list(params1, params2)
-print(params_list)
      #> [[1]]
+print(params_list)
      #> [[1]]
 #> $coefs
-#>            [,1]      [,2]
-#> [1,] -0.0291177 0.3268880
-#> [2,]  0.6672634 0.7249889
+#>           [,1]      [,2]
+#> [1,] 0.2349742 0.4368422
+#> [2,] 0.2165152 0.4914059
 #> 
 #> $sigma
-#> [1] 0.3332290 0.1076738
+#> [1] 0.3484404 0.1225498
 #> 
 #> $n_samples
 #> [1] 2
@@ -176,11 +199,11 @@ 
      Examp
 #> [[2]]
 #> $coefs
 #>            [,1]     [,2]
-#> [1,] -0.2793113 0.829587
-#> [2,]  0.7367600 1.100981
+#> [1,] 0.39052881 0.956252
+#> [2,] 0.03119033 1.300492
 #> 
 #> $sigma
-#> [1] 0.2479514 0.5455200
+#> [1] 0.04985075 0.06382896
 #> 
 #> $n_samples
 #> [1] 2
@@ -191,32 +214,30 @@ 
      Examp
 #> attr(,"class")
 #> [1] "params_lm_list"
      - +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/params_mlogit.html b/docs/reference/params_mlogit.html new file mode 100644 index 00000000..90ed627a --- /dev/null +++ b/docs/reference/params_mlogit.html @@ -0,0 +1,220 @@ + + + + + + + + +Parameters of a multinomial logit model — params_mlogit • hesim + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
      +
      + + + + +
      + +
      +
      +
      +

      Parameters of a multinomial logit model

      + Source: R/params.R + +
      + +
      +

      Store the parameters of a fitted multinomial logistic +regression model. The model is used to predict probabilities of \(K\) +classes.

      +
      + + 
      params_mlogit(coefs)
      + +

      Arguments

      + + + + + + +
      coefs

      A 3D array of stacked matrices. The number of matrices (i.e., +the number of slices in the cube) should be equal to \(K-1\). Each +matrix is contains samples of the regression coefficients under sampling uncertainty +corresponding to a particular class. Rows index parameter samples and +columns index coefficients.

      + +

      Value

      + +

      An object of class params_mlogit, which is a list containing coefs +and n_samples, where n_samples is equal to the number of rows +in each element of coefs.

      +

      Details

      + +

      Multinomial logit models are used to predict the probability of +membership for subject \(i\) in each of \(K\) classes as a function of covariates: +$$Pr(y_i = c) = \frac{e^{\beta_c x_i}}{\sum_{k=1}^K e^{\beta_k x_i}}$$

      + +

      Examples

      + 
      params <- params_mlogit(coefs = array(
+  c(matrix(c(intercept = 0, treatment = log(.75)), nrow = 1),
+    matrix(c(intercept = 0, treatment = log(.8)), nrow = 1)),
+  dim = c(1, 2, 2)
+))
      +
      + +
      + + +
      +
      +

      Developed by Devin Incerti, Jeroen P. Jansen.

      +
      + +
      +

      Site built with pkgdown 1.5.1.

      +
      + +
      +
      + + + + + + + + diff --git a/docs/reference/params_mlogit_list.html b/docs/reference/params_mlogit_list.html new file mode 100644 index 00000000..295004d5 --- /dev/null +++ b/docs/reference/params_mlogit_list.html @@ -0,0 +1,200 @@ + + + + + + + + +Parameters of a list of multinomial logit models — params_mlogit_list • hesim + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
      +
      + + + + +
      + +
      +
      +
      +

      Parameters of a list of multinomial logit models

      + Source: R/params.R + +
      + +
      +

      Create a list containing the parameters of multiple fitted multinomial logit models.

      +
      + + 
      params_mlogit_list(...)
      + +

      Arguments

      + + + + + + +
      ...

      Objects of class params_mlogit, which can be named.

      + +

      Value

      + +

      An object of class params_mlogit_list, which is a list containing +params_mlogit objects.

      + +
      + +
      + + +
      +
      +

      Developed by Devin Incerti, Jeroen P. Jansen.

      +
      + +
      +

      Site built with pkgdown 1.5.1.

      +
      + +
      +
      + + + + + + + + diff --git a/docs/reference/params_surv.html b/docs/reference/params_surv.html index 8b2d37af..68784cdb 100644 --- a/docs/reference/params_surv.html +++ b/docs/reference/params_surv.html @@ -8,21 +8,29 @@ Parameters of a survival model — params_surv • hesim + - + - + + + + - + + + - + + - + - - + + + @@ -30,13 +38,11 @@ - - - - + - + @@ -50,9 +56,10 @@ + - +
      @@ -83,28 +90,46 @@
    • News
    -
    • - - + +
      • @@ -115,6 +140,7 @@ +
      @@ -126,22 +152,20 @@

      Parameters of a survival model

      - -

      Create a list containing the parameters of a single fitted parametric or +

      Create a list containing the parameters of a single fitted parametric or flexibly parametric survival model.

      - 
      params_surv(coefs, dist, aux = NULL)
      - +

      Arguments

      - @@ -153,108 +177,115 @@

      Arg

      coefs

      A list of length equal to the number of parameters in the -survival distribution. Each element of the list is a matrix of samples from -the posterior distribution of the regression coefficients used to predict +

      A list of length equal to the number of parameters in the +survival distribution. Each element of the list is a matrix of samples +of the regression coefficients under sampling uncertainty used to predict a given parameter.

      Auxiliary arguments used with splines or fractional polynomials. See "Details".
      - +

      Value

      -

      An object of class "params_surv", which is a list containing coefs, -dist, and n_samples. n_samples is equal to the number of rows -in each element of coefs, which must be the same. The list may also contain aux if +

      An object of class params_surv, which is a list containing coefs, +dist, and n_samples. n_samplesis equal to the number of rows in each element ofcoefs, which must be the same. The list may also contain aux` if a spline or fractional polynomial model is fit.

      -

      Details

      -

      Survival is modeled as a function of \(L\) parameters \(\alpha_l\). +

      Survival is modeled as a function of \(L\) parameters \(\alpha_l\). Letting \(F(t)\) be the cumulative distribution function, survival at time \(t\) is given by $$1 - F(t | \alpha_1(x_{1}), \ldots, \alpha_L(x_{L})).$$ -The parameters are modeled as a function of covariates, \(x_l\), with an +The parameters are modeled as a function of covariates, \(x_l\), with an inverse transformation function \(g^{-1}()\), $$\alpha_l = g^{-1}(x_{l}^T \beta_l).$$ \(g^{-1}()\) is typically \(exp()\) if a parameter is strictly positive and the identity function if the parameter space is unrestricted.

      The types of distributions that can be specified are:

      • exponential or exp Exponential distribution. coef -must contain the rate parameter on the log scale and the same parameterization as in -Exponential.

        • -

      • weibull or weibull.quiet Weibull distribution. The first +must contain the rate parameter on the log scale and the same parameterization as in +stats::Exponential.

        • +

      • weibull or weibull.quiet Weibull distribution. The first element of coef is the shape parameter (on the log scale) and the second element is the scale parameter (also on the log scale). The parameterization is -that same as in Weibull.

        • -

      • gamma Gamma distribution. The first +that same as in stats::Weibull.

        • +

      • gamma Gamma distribution. The first element of coef is the shape parameter (on the log scale) and the second element is the rate parameter (also on the log scale). The parameterization is -that same as in GammaDist.

        • -

      • lnorm Lognormal distribution. The first +that same as in stats::GammaDist.

        • +

      • lnorm Lognormal distribution. The first element of coef is the meanlog parameter (i.e., the mean on the log scale) and the second element is the sdlog parameter (i.e., the standard deviation on the log scale). The parameterization is -that same as in Lognormal.

        • -

      • gompertz Gompertz distribution. The first +that same as in stats::Lognormal.

        • +

      • gompertz Gompertz distribution. The first element of coef is the shape parameter and the second element is the rate parameter (on the log scale). The parameterization is -that same as in Gompertz.

        • -

      • llogis Log-logistic distribution. The first +that same as in flexsurv::Gompertz.

        • +

      • llogis Log-logistic distribution. The first element of coef is the shape parameter (on the log scale) and the second element is the scale parameter (also on the log scale). The parameterization is -that same as in Llogis.

        • -

      • gengamma Generalized gamma distribution. The first +that same as in flexsurv::Llogis.

        • +

      • gengamma Generalized gamma distribution. The first element of coef is the location parameter mu, the second element is the scale parameter sigma (on the log scale), and the third element is the shape parameter Q. The parameterization is -that same as in GenGamma.

        • +that same as in flexsurv::GenGamma.

      • survspline Survival splines. Each element of coef is a parameter of the spline model (i.e. gamma_0, gamma_1, \(\ldots\)) with length equal to the number of knots (including the boundary knots). See below for details on the -auxiliary arguments. The parameterization is that same as in Survspline.

        • +auxiliary arguments. The parameterization is that same as in flexsurv::Survspline.

      • fracpoly Fractional polynomials. Each element of coef is a parameter of the fractional polynomial model (i.e. gamma_0, gamma_1, \(\ldots\)) with length equal -to the number of powers minus 1. See below for details on the auxiliary arguments +to the number of powers minus 1. See below for details on the auxiliary arguments (i.e., powers).

      -

      Auxiliary arguments for spline models should be specified as a list containing the elements:

      + +

      Auxiliary arguments for spline models should be specified as a list containing the elements:

      knots

      A numeric vector of knots.

      scale

      The survival outcome to be modeled -as a spline function. Options are "log_cumhazard" for the log cumulative hazard; -"log_hazard" for the log hazard rate; "log_cumodds" for the log cumulative odds; -and "inv_normal" for the inverse normal distribution function.

      -
      timescale

      If "log" (the default), then survival is modeled as a spline function -of log time; if "identity", then it is modeled as a spline function of time.

      +as a spline function. Options are "log_cumhazard" for the log cumulative hazard; +"log_hazard" for the log hazard rate; "log_cumodds" for the log cumulative odds; +and "inv_normal" for the inverse normal distribution function.

      +
      timescale

      If "log" (the default), then survival is modeled as a spline function +of log time; if "identity", then it is modeled as a spline function of time.

      +
      -

      Auxiliary arguments for fractional polynomial models should be specified as a list containing the elements:

      + +

      Auxiliary arguments for fractional polynomial models should be specified as a list containing the elements:

      powers

      A vector of the powers of the fractional polynomial with each element - chosen from the following set: -2. -1, -0.5, 0, 0.5, 1, 2, 3.

      +chosen from the following set: -2. -1, -0.5, 0, 0.5, 1, 2, 3.

      +
      -

      Furthermore, when splines (with scale = "log_hazard") or fractional -polynomials are used, numerical methods must be used to compute the cumulative + +

      Furthermore, when splines (with scale = "log_hazard") or fractional +polynomials are used, numerical methods must be used to compute the cumulative hazard and for random number generation. The following additional auxiliary arguments -can therefore be specified:

      -
      cumhaz_method

      Numerical method used to compute cumulative hazard +can therefore be specified:

      +
      cumhaz_method

      Numerical method used to compute cumulative hazard (i.e., to integrate the hazard function). Always used for fractional polynomials but only used for splines if scale = "log_hazard". -Options are "quad" for adaptive quadrature and "riemann" for Riemann sum.

      -
      random_method

      Method used to randomly draw from - an arbitrary survival function. Options are "invcdf" for the inverse CDF and - "sample" for randomly sampling from discrete survival probabilities.

      -
      step

      Step size for computation of cumulative hazard with -numerical integration. Only required when using "riemann" to compute the -cumulative hazard or using "sample" for random number generation.

      +Options are "quad" for adaptive quadrature and "riemann" for Riemann sum.

      +
      random_method

      Method used to randomly draw from +an arbitrary survival function. Options are "invcdf" for the inverse CDF and +"discrete" for a discrete time approximation that randomly samples +events from a bernoulli distribution at discrete times.

      +
      step

      Step size for computation of cumulative hazard with +numerical integration. Only required when using "riemann" to compute the +cumulative hazard or using "discrete" for random number generation.

      +
      - +

      Examples

      - 
      library("flexsurv")
-fit <- flexsurvreg(Surv(futime, fustat) ~ 1, data = ovarian, dist = "weibull")
-params <- params_surv(coefs = list(shape = fit$res.t["shape", "est", drop = FALSE],
+    
      library("flexsurv")
+fit <- flexsurvreg(Surv(futime, fustat) ~ 1, data = ovarian, dist = "weibull")
+params <- params_surv(coefs = list(shape = fit$res.t["shape", "est", drop = FALSE],
                                    scale = fit$res.t["scale", "est", drop = FALSE]),
                      dist = fit$dlist$name)
-print(params)
      #> $coefs
-#> $coefs$shape
+print(params)
      #> $coefs
+#> $shape
 #>             est
 #> shape 0.1026105
 #> 
-#> $coefs$scale
+#> $scale
 #>            est
 #> scale 7.111038
 #> 
+#> attr(,"class")
+#> [1] "matlist"
 #> 
 #> $dist
 #> [1] "weibull.quiet"
@@ -265,34 +296,30 @@ 
      Examp
 #> attr(,"class")
 #> [1] "params_surv"
      
   
      
-  
 
+
       
      
       
      
   
      Developed by Devin Incerti, Jeroen P. Jansen.
      
 
      
 
 
      
-  
      Site built with pkgdown 1.3.0.
      
+  
      Site built with pkgdown 1.5.1.
      
 
      
+
       
      
    
 
   
 
+
   
 
 
+
diff --git a/docs/reference/params_surv_list.html b/docs/reference/params_surv_list.html
index 155815eb..54fad8fa 100644
--- a/docs/reference/params_surv_list.html
+++ b/docs/reference/params_surv_list.html
@@ -8,21 +8,29 @@
 
 Parameters of a list of survival models — params_surv_list • hesim
 
+
 
-
+
 
-
+
+
 
-
+
+
+
+
+
 
 
-
+
+
 
 
-
+
 
-
-
+
+
+
 
 
 
@@ -30,12 +38,10 @@
 
 
 
-  
-  
-
 
+
 
-
+
 
 
 
@@ -49,9 +55,10 @@
 
 
 
+
   
 
-  
+  
     
      
       
      
       
 
@@ -82,28 +89,46 @@
   
   
 
 
    • 
   News
 
      •
    - -
    • - - + +
      • @@ -117,6 +142,7 @@ +
      @@ -128,16 +154,14 @@

      Partitioned survival regression object

      -

      Create a partitioned survival regression object of class "partsurvfit". The object contains a list of fitted survival models fit using either flexsurvreg or flexsurvspline (i.e., -an object of class flexsurvreg_list) and the data frame used to perform the fit of each model. +an object of class flexsurvreg_list) and the data frame used to perform the fit of each model. The same data frame must have been used for each fit.

      - 
      partsurvfit(object, data)
      - +

      Arguments

      @@ -151,52 +175,49 @@

      Arg flexsurvreg.

      - +

      Value

      -

      Returns an object of class "partsurvfit", which is a list containing two elements. -The first element, "models", contains the survival models passed to object, and the second +

      Returns an object of class "partsurvfit", which is a list containing two elements. +The first element, "models", contains the survival models passed to object, and the second element, "data" contains the data frame passed to data.

      -

      Examples

      - 
      library("flexsurv")
-fit1 <- flexsurv::flexsurvreg(formula = Surv(endpoint1_time, endpoint1_status) ~ age,
+    
      library("flexsurv")
+fit1 <- flexsurv::flexsurvreg(formula = Surv(endpoint1_time, endpoint1_status) ~ age,
                               data = psm4_exdata$survival,
                               dist = "weibull")
-fit2 <- flexsurv::flexsurvreg(formula = Surv(endpoint2_time, endpoint2_status) ~ age,
+fit2 <- flexsurv::flexsurvreg(formula = Surv(endpoint2_time, endpoint2_status) ~ age,
                               data = psm4_exdata$survival,
                               dist = "weibull")
 fsreg_list <- flexsurvreg_list(endpoint1 = fit1, endpoint2 = fit2)
 fits <- partsurvfit(fsreg_list, data = psm4_exdata$survival)
-class(fits)
      #> [1] "partsurvfit"
      
+class(fits)
      #> [1] "partsurvfit"
      - +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/psm4_exdata.html b/docs/reference/psm4_exdata.html index 3f7e9777..46c1c857 100644 --- a/docs/reference/psm4_exdata.html +++ b/docs/reference/psm4_exdata.html @@ -8,21 +8,29 @@ Example data for a 4-state partitioned survival model — psm4_exdata • hesim + - + - + + + + - + + + - + + - + - - + + + @@ -30,13 +38,11 @@ - - - + - + @@ -50,9 +56,10 @@ + - +
      @@ -83,28 +90,46 @@
    • News
    -
    • - - + +
      • @@ -115,6 +140,7 @@ +
      @@ -126,14 +152,13 @@

      Example data for a 4-state partitioned survival model

      -

      A collection of example datasets containing simulated survival, costs, and utility data for a 4-state partitioned survival model.

      - 
      psm4_exdata
      - + +

      Format

      A list containing the following elements:

        @@ -141,11 +166,12 @@

        Format

        CostsA list of data frames. The first data frame contains medical cost data and the second data frame contains drug cost data.

      - +

      Survival data

      -

      The survival data frame contains a list of 3 survival curves, each containing the following columns.

      + +

      The survival data frame contains a list of 3 survival curves, each containing the following columns.

      female

      An indicator variable equal to 1 if the patient is female and 0 otherwise.

      age

      The age of the patient in years.

      strategy_id

      The id of the treatment strategy used.

      @@ -155,52 +181,54 @@

      endpoint2_status

      A status indicator for survival endpoint 2 equal to 0 if alive and 1 if dead.

      endpoint3_time

      Follow up time with right censored data to survival endpoint 3.

      endpoint3_status

      A status indicator for survival endpoint 3 equal to 0 if alive and 1 if dead.

      +

      - +

      Cost data

      -

      The cost list contains two data frames.The first data frame contains data on the -medical costs by patient and health state, and contains the following columns:

      -
      patient_id

      An integer denoting the id of the patient.

      -
      female

      An indicator variable equal to 1 if the patient is female and 0 otherwise.

      -
      state_name

      A categorical variable denoting the three possible health states.

      -
      costs

      Annualized medical costs.

      -
      -

      The second data frame contains data on the drug costs associated with each treatment strategy.

      -
      strategy_id

      The id of each treatment strategy.

      -
      costs

      Annualized drug costs.

      -
      - - - + +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/qalys.html b/docs/reference/qalys.html new file mode 100644 index 00000000..8cd3e8e8 --- /dev/null +++ b/docs/reference/qalys.html @@ -0,0 +1,208 @@ + + + + + + + + +Quality-adjusted life-years object — qalys • hesim + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
      +
      + + + + +
      + +
      +
      +
      +

      Quality-adjusted life-years object

      + Source: R/statevals.R + +
      + +
      +

      An object of class qalys returned from methods +$sim_qalys() in model classes that store simulated +quality-adjusted life-years (QALYs).

      +
      + + + +

      Components

      + + + +

      A qalys object inherits from data.table and contains +the following columns:

      + +
      sample

      A random sample from the PSA.

      +
      strategy_id

      The treatment strategy ID.

      +
      patient_id

      The patient ID.

      +
      state_id

      The health state ID.

      +
      dr

      The rate used to discount QALYs.

      +
      category

      A single category always equal to "qalys".

      +
      qalys

      The simulated values of QALYs.

      + +

      If the argument lys = TRUE, then the data.table also contains a column +lys containing simulated life-years.

      + +
      + +
      + + +
      +
      +

      Developed by Devin Incerti, Jeroen P. Jansen.

      +
      + +
      +

      Site built with pkgdown 1.5.1.

      +
      + +
      +
      + + + + + + + + diff --git a/docs/reference/rcat.html b/docs/reference/rcat.html index 02db0bd9..5ee04bcb 100644 --- a/docs/reference/rcat.html +++ b/docs/reference/rcat.html @@ -8,21 +8,29 @@ Random generation for categorical distribution — rcat • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,13 +38,11 @@ - - - + - +rcat is vectorized and written in C++ for speed." /> + @@ -50,9 +56,10 @@ + - +
      @@ -83,28 +90,46 @@
    • News
    - -
    • - - + +
      • @@ -115,6 +140,7 @@ +
      @@ -126,14 +152,16 @@

      Random generation for multiple Dirichlet distributions

      - -

      Draw random samples from multiple Dirichlet distributions. -rdirichlet_mat is vectorized and written in C++ for speed.

      - +

      Draw random samples from multiple Dirichlet distributions for use in +transition probability matrices.

      - 
      rdirichlet_mat(n, alpha)
      - + 
      rdirichlet_mat(
+  n,
+  alpha,
+  output = c("array", "matrix", "data.frame", "data.table")
+)
      +

      Arguments

      @@ -145,22 +173,33 @@

      Arg

      + + + +
      alpha

      A matrix where each row is a separate vector of shape parameters.
      output

      The class of the object returned by the function. Either an +array, matrix, data.frame, or data.table.

      - +

      Value

      -

      An array of matrices where each row of each matrix is a sample from the Dirichlet distribution.

      - +

      If output = "array", then an array of matrices is returned +where each row of each matrix is a sample from the Dirichlet distribution. +If output results in a two dimensional object (i.e., a matrix, +data.frame, or data.table, then each row contains +all elements of the sampled matrix from the Dirichlet distribution +ordered rowwise; that is, each matrix is flattened. In these cases, +the number of rows must be less than or equal to the number of columns.

      Details

      -

      This function is particularly useful for representing the distribution of -transition probabilities in a transition matrix.

      - +

      This function is meant for representing the distribution of +transition probabilities in a transition matrix. The (i,j) element of +alpha is a transition from state i to state j. It is vectorized and +written in C++ for speed.

      Examples

      - 
      alpha <- matrix(c(100, 200, 500, 50, 70, 75), ncol = 3, nrow = 2, byrow = TRUE)
+    
      alpha <- matrix(c(100, 200, 500, 50, 70, 75), ncol = 3, nrow = 2, byrow = TRUE)
 samp <- rdirichlet_mat(100, alpha)
-print(samp[, , 1:2])
      #> , , 1
+print(samp[, , 1:2])
      #> , , 1
 #> 
 #>           [,1]      [,2]      [,3]
 #> [1,] 0.1131831 0.2561972 0.6306197
@@ -173,34 +212,30 @@ 
      Examp
 #> [2,] 0.2976925 0.3570395 0.3452680
 #> 
      
   
      
-  
 
+
       
      
       
      
   
      Developed by Devin Incerti, Jeroen P. Jansen.
      
 
      
 
 
      
-  
      Site built with pkgdown 1.3.0.
      
+  
      Site built with pkgdown 1.5.1.
      
 
      
+
       
      
    
 
   
 
+
   
 
 
+
diff --git a/docs/reference/rng_distributions.html b/docs/reference/rng_distributions.html
new file mode 100644
index 00000000..f8149543
--- /dev/null
+++ b/docs/reference/rng_distributions.html
@@ -0,0 +1,322 @@
+
+
+
+  
+  
+
+
+
+Random number generation distributions — rng_distributions • hesim
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  
+
+  
+    
      
+      
      
+      
+
+      
+
+      
      
+
+
      
+  
      
+    
      
+    
      Random number generation distributions
      
+    Source: R/model_def.R
+    
+    
      
+
+    
      
+    
      A collection of functions for randomly generating deviates from probability
+distributions with define_rng().
      
+    
      
+
+    
      beta_rng(shape1 = 1, shape2 = 1, mean = NULL, sd = NULL, names = NULL)
+
+dirichlet_rng(alpha, names = NULL)
+
+fixed(est, names = NULL)
+
+custom(x, names = NULL)
+
+gamma_rng(mean, sd, names = NULL)
+
+lognormal_rng(meanlog, sdlog, names = NULL)
+
+multi_normal_rng(mu, Sigma, names = NULL, ...)
+
+normal_rng(mean, sd, names = NULL)
+
+uniform_rng(min, max, names = NULL)
      
+
+    
      Arguments
      
+    
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
      shape1, shape2
      Non-negative parameters of the Beta distribution.
      mean, sd
      Mean and standard deviation of the random variable.
      names
      Names for columns if an object with multiple columns is returned
+by the function.
      alpha
      A matrix where each row is a separate vector of shape parameters.
      est
      A vector of estimates of the variable of interest.
      x
      A numeric vector, matrix, data.frame, or data.table containing
+random samples of the variable of interest from a suitable probability distribution. This would
+typically be a posterior distribution from a Bayesian analysis.
      meanlog, sdlog
      Mean and standard deviation of the distribution on the
+log scale.
      mu, Sigma
      mu is a vector giving the means of the variables and
+Sigma is a positive-definite symmetric matrix specifying the
+covariance matrix of the variables.
      ...
      Additional arguments to pass to underlying random number generation
+functions. See "details".
      min, max
      Lower and upper limits of the distribution. Must be finite.
      
+
+    
      Value
      
+
+    
      Functions either return a vector of length n or an n by k data.table.
+Multivariate distributions always return a data.table. If a
+univariate distribution is used, then a data.table is returned if each
+parameter is specified as a vector with length greater than 1; otherwise, if
+parameters are scalars, then a vector is returned. In the data.table case,
+k is equal to the length of the parameter vectors
+entered as  arguments. For example, if the probability distribution contained
+mean as an argument and mean were
+of length 3, then an n by 3 matrix would be returned. The length of all
+parameter vectors must be the same. For instance, if the vector mean
+were of length 3 then all additional parameters (e.g., sd)
+must also be of length 3.
      
+
      If a data.table is returned by a distribution, then its column names are set
+according to the following hierarchy:
        
+
      1. With the names argument if it is not NULL
        2. 
+
      2. With the names of the parameter vectors if they are named vectors. If there
+are multiple parameter vector arguments, then the names of the first parameter
+vector with non NULL names is used. For instance, if mean and sd are
+both arguments to a random number generation function and mean is a
+named vector, then the names from the vector mean are used.
        3. 
+
      3. As v1, ..., vk if the names argument is NULL and there are no
+named parameter vectors.
        4. 
+
      
+
+    
      Details
      
+
+    
      These functions are not exported and are meant for use with
+define_rng(). They consequently assume that the number of samples to draw, n,
+is defined in the parent environment. Convenience random number generation
+functions include:
      
+
      beta_rng()
      If mean and sd are both not NULL, then
+parameters of the beta distribution are derived using
+the methods of moments with mom_beta(). Beta variates are generated with
+stats::rbeta().
      
+
      custom()
      Use previously sampled values from a custom probability distribution.
+There are three possibilities: (i) if n is equal to the number previously
+sampled values (say n_samples), then x is returned as is; (ii) if
+n < n_samples, then samples from x are sampled without replacement;
+and (iii) if n > n_samples, then samples from x are sampled with replacement
+and a warning is provided.
      
+
      dirichlet_rng()
      Dirichlet variates for each row in the matrix are
+generated with rdirichlet_mat(). The sampled values are stored in a data.table
+where there is a column for each element of alpha
+(with elements ordered rowwise).
      
+
      fixed()
      This function should be used when values of the variable
+of interest are fixed (i.e., they are known with certainty). If length(est) > 1,
+an n by length(est) data.table is returned meaning that each element of est
+is repeated n times; otherwise (if length(est) == 1), a vector is returned
+where est is repeated n times.
      
+
      gamma_rng()
      The parameters of the gamma distribution are derived using
+the methods of moments with mom_gamma() and gamma variates are generated
+with stats::rgamma().
      
+
      lognormal_rng()
      Lognormal variates are generated with stats::rlnorm().
      
+
      multi_normal_rng()
      Multivariate normal variates are generated with MASS::mvrnorm().
      
+
      normal_rng()
      Normal variates are generated with stats::rnorm().
      
+
      uniform_rng()
      Uniform variates are generated with stats::runif().
      
+
+
+
+
+    
      See also
      
+
+    
      define_rng()
      
+
+  
      
+  
+
      
+
+
+      
      
+      
      
+  
      Developed by Devin Incerti, Jeroen P. Jansen.
      
+
      
+
+
      
+  
      Site built with pkgdown 1.5.1.
      
+
      
+
+      
      
+   
      
+
+  
+
+
+  
+
+
+
diff --git a/docs/reference/rpwexp.html b/docs/reference/rpwexp.html
index 0e41073e..4240690a 100644
--- a/docs/reference/rpwexp.html
+++ b/docs/reference/rpwexp.html
@@ -8,21 +8,29 @@
 
 Random generation for piecewise exponential distribution — rpwexp • hesim
 
+
 
-
+
 
-
+
+
 
-
+
+
+
+
+
 
 
-
+
+
 
 
-
+
 
-
-
+
+
+
 
 
 
@@ -30,13 +38,11 @@
 
 
 
-  
-  
-
 
+
 
-
+rpwexp is vectorized and written in C++ for speed." />
+
 
 
 
@@ -50,9 +56,10 @@
 
 
 
+
   
 
-  
+  
     
      
       
      
       
 
@@ -83,28 +90,46 @@
   
   
 
 
    • 
   News
 
      •
    -
    • - - + +
      • @@ -115,6 +140,7 @@ +
      @@ -126,14 +152,12 @@

      Random generation for piecewise exponential distribution

      -

      Draw random samples from an exponential distribution with piecewise rates. - rpwexp is vectorized and written in C++ for speed.

      - +rpwexp is vectorized and written in C++ for speed.

      rpwexp(n, rate = 1, time = 0)
      - +

      Arguments

      @@ -143,7 +167,7 @@

      Arg

      - @@ -152,52 +176,49 @@

      Arg times at which the rate changes

      rate

      A matrix of rates where rows correspond to observations +

      A matrix of rates where rows correspond to observations and columns correspond to rates during specified time intervals.
      - +

      Value

      -

      A vector of random samples from the piecewise exponential distribution. The length of the sample is -determined by n. The numerical arguments other than n are recycled so that the number of samples is +

      A vector of random samples from the piecewise exponential distribution. The length of the sample is +determined by n. The numerical arguments other than n are recycled so that the number of samples is equal to n.

      -

      Examples

      - 
      rate <- c(.6, 1.2, 1.3)
+    
      rate <- c(.6, 1.2, 1.3)
 n <- 100000
-ratemat <- matrix(rep(rate, n/2), nrow = n,
+ratemat <- matrix(rep(rate, n/2), nrow = n,
                   ncol = 3, byrow = TRUE)
-t <- c(0, 10, 15)
-ptm <- proc.time()
+t <- c(0, 10, 15)
+ptm <- proc.time()
 samp <- rpwexp(n, ratemat, t)
-proc.time() - ptm
      #>    user  system elapsed 
-#>   0.009   0.001   0.011 
      summary(samp)
      #>    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
+proc.time() - ptm
      #>    user  system elapsed 
+#>   0.007   0.001   0.008 
      summary(samp)
      #>    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
 #>  0.0000  0.4793  1.1517  1.6597  2.3021 14.4929 
      
   
      
-  
 
+
       
      
       
      
   
      Developed by Devin Incerti, Jeroen P. Jansen.
      
 
      
 
 
      
-  
      Site built with pkgdown 1.3.0.
      
+  
      Site built with pkgdown 1.5.1.
      
 
      
+
       
      
    
 
   
 
+
   
 
 
+
diff --git a/docs/reference/sim_ev.html b/docs/reference/sim_ev.html
new file mode 100644
index 00000000..a2a93637
--- /dev/null
+++ b/docs/reference/sim_ev.html
@@ -0,0 +1,233 @@
+
+
+
+  
+  
+
+
+
+Expected values — sim_ev • hesim
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  
+
+  
+    
      
+      
      
+      
+
+      
+
+      
      
+
+
      
+  
      
+    
      
+    
      Expected values
      
+    Source: R/statevals.R
+    
+    
      
+
+    
      
+    
      Simulate costs and quality-adjusted life-years (QALYs) as a function of
+simulated state occupancy probabilities.
      
+    
      
+
+    
      sim_qalys(object, utility_model, dr, method, lys)
+
+sim_costs(object, cost_models, dr, method)
      
+
+    
      Arguments
      
+    
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
      object
      A stateprobs object.
      utility_model
      A single object of class StateVals used
+to simulate utility.
      dr
      Discount rate.
      lys
      If TRUE, then life-years are simulated in addition to
+QALYs.
      cost_models
      A list of objects of class StateVals used
+to simulate costs.
      integrate_method
      Method used to integrate state values when computing
+weighted length of stay. Options are trapz for the trapezoid rule,
+riemann_left left for a left Riemann sum, and
+riemann_right right for a right Riemann sum.
      
+
+    
      Value
      
+
+    
      sim_costs() and sim_qalys() return objects of class
+costs and qalys, respectively.
      
+    
      Details
      
+
+    
      See vignette("expected-values") for details.
      
+
+  
      
+  
+
      
+
+
+      
      
+      
      
+  
      Developed by Devin Incerti, Jeroen P. Jansen.
      
+
      
+
+
      
+  
      Site built with pkgdown 1.5.1.
      
+
      
+
+      
      
+   
      
+
+  
+
+
+  
+
+
+
diff --git a/docs/reference/stateprobs.html b/docs/reference/stateprobs.html
new file mode 100644
index 00000000..b0883f3a
--- /dev/null
+++ b/docs/reference/stateprobs.html
@@ -0,0 +1,207 @@
+
+
+
+  
+  
+
+
+
+State probability object — stateprobs • hesim
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  
+
+  
+    
      
+      
      
+      
+
+      
+
+      
      
+
+
      
+  
      
+    
      
+    
      State probability object
      
+    Source: R/statevals.R
+    
+    
      
+
+    
      
+    
      An object of class stateprobs returned from methods
+$sim_stateprobs() in model classes.
      
+    
      
+
+
+
+    
      Components
      
+
+    
+
+
      A stateprobs object inherits from data.table and contains
+the following columns:
      
+
+
      sample
      A random sample from the PSA.
      
+
      strategy_id
      The treatment strategy ID.
      
+
      patient_id
      The patient ID.
      
+
      state_id
      The health state ID.
      
+
      t
      The time at which a state probability is computed.
      
+
      prob
      The probability of being in a given health state.
      
+
+
+
+
      When simulating individual-level models, the patient_id column is
+not included as state probabilities are computed by averaging across patients.
      
+
+  
      
+  
+
      
+
+
+      
      
+      
      
+  
      Developed by Devin Incerti, Jeroen P. Jansen.
      
+
      
+
+
      
+  
      Site built with pkgdown 1.5.1.
      
+
      
+
+      
      
+   
      
+
+  
+
+
+  
+
+
+
diff --git a/docs/reference/stateval_tbl.html b/docs/reference/stateval_tbl.html
index 087a3761..78b6aaef 100644
--- a/docs/reference/stateval_tbl.html
+++ b/docs/reference/stateval_tbl.html
@@ -8,21 +8,29 @@
 
 Table to store state value parameters — stateval_tbl • hesim
 
+
 
-
+
 
-
+
+
+
+
 
-
+
+
+
 
 
-
+
+
 
 
-
+
 
-
-
+
+
+
 
 
 
@@ -30,14 +38,12 @@
 
 
 
-  
-  
-
 
-
+
-
+
 
 
 
@@ -51,9 +57,10 @@
 
 
 
+
   
 
-  
+  
     
      
       
      
       
 
@@ -84,28 +91,46 @@
   
   
 
 
    • 
   News
 
      •
    -
    • - - + +
      • @@ -116,6 +141,7 @@ +
      @@ -127,162 +153,160 @@

      Table to store state value parameters

      - -

      Create a table for storing parameter estimates used to simulate costs or +

      Create a table for storing parameter estimates used to simulate costs or utility in an economic model by treatment strategy, patient, health state, and (optionally) time interval.

      - 
      stateval_tbl(
   tbl,
-  dist = c("norm", "beta", "gamma", "lnorm", "unif", "fixed", "custom"),
-  hesim_data = NULL
+  dist = c("norm", "beta", "gamma", "lnorm", "unif", "fixed", "custom"),
+  hesim_data = NULL,
+  grp_var = NULL
 )
      - +

      Arguments

      - - - +Not required if tbl includes one row for each treatment strategy, patient, and +health state combination. Patients are matched to groups by specifying both a +patient_id and a grp_var column in the patients table.

      + + + +
      tbl

      A data.frame or data.table for storing parameter +

      A data.frame or data.table for storing parameter values. See "Details" for specifics.
      dist

      Probability distribution used to sample parameters for a +

      Probability distribution used to sample parameters for a probabilistic sensitivity analysis (PSA).
      hesim_data

      A hesim_data object. Required to specify +

      A hesim_data object. Required to specify treatment strategies , patients, - and/or health states not included as columns +and/or health states not included as columns in tbl, or, to match patients in tbl to groups. - Not required if tbl includes one row for each treatment strategy, patient, and -health state combination. Patients are matched to groups by specifying both a -patient_id and a grp_id column in the patients table; -if hesim_data = NULL but grp_id is included as a column -in tbl, then each group is assumed to be a unique patient.

      grp_var

      The name of the variable used to group patients.
      - +

      Value

      An object of class "stateval_tbl", which is a data.table of -parameter values with attributes for dist and optionally -strategy_id, patients, and state_id. tbl -is in the same format as described in "Details". patients is a -data.table with one column containing patient_id and -optionally a second column containing grp_id.

      - +parameter values with attributes for dist and optionally +strategy_id, patients, state_id, and grp_var. tbl +is in the same format as described in "Details". patients is a +data.table with one column containing patient_id and +optionally a second column containing grp_var.

      Details

      -

      tbl is a data.table containing columns for treatment -strategies (strategy_id), patient subgroups (grp_id), -health states (state_id), and/or the start of time intervals +

      tbl is a data.table containing columns for treatment +strategies (strategy_id), patients (patient_id), +health states (state_id), and/or the start of time intervals (time_start). The table must contain at least one column -named strategy_id, grp_id, or state_id, -but does not need to contain all of them. Each row denotes a unique -treatment strategy, patient subgroup, health state, and/or time interval pair.

      -

      tbl must also contain columns summarizing the state values for each -row, which depend on the probability distribution selected with dist. +named strategy_id, patient_id, or state_id, +but does not need to contain all of them. Each row denotes a unique +treatment strategy, patient, health state, and/or time interval pair. +tbl may also contain a column with the name specified in grp_var +(rather than patient_id) so that state values are assigned to +groups of patients.

      +

      tbl must also contain columns summarizing the state values for each +row, which depend on the probability distribution selected with dist. Available distributions include the normal (norm), beta (beta), gamma (gamma), lognormal (lnorm), and uniform (unif) - distributions. In addition, the option fixed can be used if estimates - are known with certainty and custom can be used if - parameter values for a PSA have been previously +distributions. In addition, the option fixed can be used if estimates +are known with certainty and custom can be used if +parameter values for a PSA have been previously sampled from an arbitrary probability distribution. - The columns in tbl that must be included, - by distribution, are:

      -
      +The columns in tbl that must be included, +by distribution, are:

      +
      norm

      mean and sd

      beta

      mean and se or shape1 and shape2

      -
      gamma

      mean and se, shape and rate, -or shape and scale

      +
      gamma

      mean and se, shape and rate, +or shape and scale

      lnorm

      meanlog or sdlog

      unif

      min and max

      fixed

      est

      custom

      sample and value

      +
      -

      Note that if dist = "custom", then tbl must include a column + +

      Note that if dist = "custom", then tbl must include a column named sample (an integer vector denoting a unique random draw) and - value (denoting the value of the randomly sampled parameter). In this case, there is a unique -row in tbl for each random draw (sample) and - each combination of strategies, patients, health states, and/or time intervals. +value (denoting the value of the randomly sampled parameter). In this case, +there is a unique row in tbl for each random draw (sample) and +each combination of strategies, patients, health states, and/or time intervals. Again, tbl must contain at least one column -named strategy_id, grp_id, or state_id, - but does not need to contain them all.

      - +named strategy_id, patient_id (or grp_var), or state_id, +but does not need to contain them all.

      See also

      -

      create_StateVals

      - +

      create_StateVals, StateVals

      Examples

      - 
      strategies <- data.frame(strategy_id = c(1, 2))
-patients <- data.frame(patient_id = seq(1, 3),
-                       grp_id = c(1, 1, 2),
-                       age = c(45, 50, 60),
-                       female = c(0, 0, 1))
-states <- data.frame(state_id = c(1, 2))
+    
      strategies <- data.frame(strategy_id = c(1, 2))
+patients <- data.frame(patient_id = seq(1, 3),
+                       grp = c(1, 1, 2),
+                       age = c(45, 50, 60),
+                       female = c(0, 0, 1))
+states <- data.frame(state_id = c(1, 2))
 hesim_dat <- hesim_data(strategies = strategies,
                         patients = patients,
                         states = states)
 
 # Utility varies by health state and patient group
-utility_tbl <- stateval_tbl(data.frame(state_id = rep(states$state_id, 2),
-                                       grp_id = rep(rep(c(1, 2)), each = nrow(states)),
-                                       mean = c(.8, .7, .75, .55),
-                                       se = c(.18, .12, .10, .06)),
+utility_tbl <- stateval_tbl(data.frame(state_id = rep(states$state_id, 2),
+                                       grp = rep(rep(c(1, 2)), each = nrow(states)),
+                                       mean = c(.8, .7, .75, .55),
+                                       se = c(.18, .12, .10, .06)),
                             dist = "beta",
-                            hesim_data = hesim_dat)
-print(utility_tbl)
      #>    grp_id state_id mean   se
-#> 1:      1        1 0.80 0.18
-#> 2:      1        2 0.70 0.12
-#> 3:      2        1 0.75 0.10
-#> 4:      2        2 0.55 0.06
      utilmod <- create_StateVals(utility_tbl, n = 2)
+                            hesim_data = hesim_dat,
+                            grp_var = "grp")
+print(utility_tbl)
      #>    grp state_id mean   se
+#> 1:   1        1 0.80 0.18
+#> 2:   1        2 0.70 0.12
+#> 3:   2        1 0.75 0.10
+#> 4:   2        2 0.55 0.06
      utilmod <- create_StateVals(utility_tbl, n = 2)
 
 # Costs vary by treatment strategy
-cost_tbl <- stateval_tbl(data.frame(strategy_id = strategies$strategy_id,
-                                    mean = c(5000, 3000),
-                                    se = c(200, 100)),
+cost_tbl <- stateval_tbl(data.frame(strategy_id = strategies$strategy_id,
+                                    mean = c(5000, 3000),
+                                    se = c(200, 100)),
                          dist = "gamma",
                          hesim_data = hesim_dat)
-print(cost_tbl)
      #>    strategy_id mean  se
+print(cost_tbl)
      #>    strategy_id mean  se
 #> 1:           1 5000 200
 #> 2:           2 3000 100
      costmod <- create_StateVals(cost_tbl, n = 2)
      
   
      
-  
 
+
       
      
       
      
   
      Developed by Devin Incerti, Jeroen P. Jansen.
      
 
      
 
 
      
-  
      Site built with pkgdown 1.3.0.
      
+  
      Site built with pkgdown 1.5.1.
      
 
      
+
       
      
    
 
   
 
+
   
 
 
+
diff --git a/docs/reference/summarize_ce.html b/docs/reference/summarize_ce.html
index f09bdd8d..e55149bf 100644
--- a/docs/reference/summarize_ce.html
+++ b/docs/reference/summarize_ce.html
@@ -8,21 +8,29 @@
 
 Summarize costs and effectiveness — summarize_ce • hesim
 
+
 
-
+
 
-
+
+
 
-
+
+
+
+
+
 
 
-
+
+
 
 
-
+
 
-
-
+
+
+
 
 
 
@@ -30,14 +38,12 @@
 
 
 
-  
-  
-
 
+
 
-
+from an economic model. The summary output is used to perform
+cost-effectiveness analysis with icea() and icea_pw()." />
+
 
 
 
@@ -51,9 +57,10 @@
 
 
 
+
   
 
-  
+  
     
      
       
      
       
 
@@ -84,28 +91,46 @@
   
   
 
 
    • 
   News
 
      •
    -
    • - - + +
      • @@ -116,6 +141,7 @@ +
      @@ -127,70 +153,71 @@

      Summarize costs and effectiveness

      -

      Summarize costs and quality-adjusted life-years (QALYs) given output simulated -from an economic model. The summary output is used to perform -cost-effectiveness analysis with icea and icea_pw.

      - +from an economic model. The summary output is used to perform +cost-effectiveness analysis with icea() and icea_pw().

      - 
      summarize_ce(costs, qalys)
      - + 
      summarize_ce(costs, qalys, by_grp = FALSE)
      +

      Arguments

      - + - + + + + +
      costs

      Simulated costs by category.

      Simulated costs by category (objects of class costs).
      qalys

      Simulated QALYs.

      Simulated QALYs (objects of class qalys).

      by_grp

      If TRUE, then costs and QALYs are computed by subgroup. If +FALSE, then costs and QALYs are aggregated across all patients (and subgroups).
      - +

      Value

      -

      An object of class ce.

      - +

      An object of class ce.

      Details

      -

      If mean costs and/or QALYs have already been computed -(i.e., an average within a population), then there -must be one observation for each discount rate (dr), -PSA sample (sample), treatment strategy (strategy_id), -and health state (state_id). Alternatively, there can be column +

      If mean costs and/or QALYs have already been computed +(i.e., an average within a population), then there +must be one observation for each discount rate (dr), +PSA sample (sample), treatment strategy (strategy_id), +and health state (state_id). Alternatively, there can be a column denoting a patient (patient_id), in which case outcomes will first be -averaged across patients.

      - +averaged across patients. A grp_id column can also be used so that +outcomes are computed for each subgroup (if by_grp = TRUE); otherwise it is assumed that +there is only one subgroup.

      - +

      Developed by Devin Incerti, Jeroen P. Jansen.

      -

      Site built with pkgdown 1.3.0.

      +

      Site built with pkgdown 1.5.1.

      +
      + + diff --git a/docs/reference/surv_quantile.html b/docs/reference/surv_quantile.html index cd8fe8c3..4a6a88ec 100644 --- a/docs/reference/surv_quantile.html +++ b/docs/reference/surv_quantile.html @@ -8,21 +8,29 @@ Survival quantiles — surv_quantile • hesim + - + - + + - + + + + + - + + - + - - + + + @@ -30,12 +38,10 @@ - - - + - + @@ -49,9 +55,10 @@ + - +
      @@ -82,28 +89,46 @@
    • News
    -
    • - - + +
      • @@ -114,6 +139,7 @@ +
      @@ -125,13 +151,11 @@

      Survival quantiles

      -

      Compute quantiles from survival curves.

      - 
      surv_quantile(x, probs = 0.5, t, surv_cols, by)
      - +

      Arguments

      @@ -149,7 +173,7 @@

      Arg

      - @@ -157,55 +181,52 @@

      Arg

      surv_cols

      A character vector of the names of columns containing +

      A character vector of the names of columns containing survival curves.

      A character vector of the names of columns to group by.
      - +

      Value

      -

      A data.table of quantiles of each survival curve in +

      A data.table of quantiles of each survival curve in surv_cols by each group in by.

      -

      Examples

      - 
      library("data.table")
-t <- seq(0, 10, by = .01)
-surv1 <- seq(1, .3, length.out = length(t))
-surv2 <- seq(1, .2, length.out = length(t))
-strategies <- c("Strategy 1", "Strategy 2")
-surv <- data.table(strategy = rep(strategies, each = length(t)),
-                   t = rep(t, 2),
-                   surv = c(surv1, surv2))
-surv_quantile(surv, probs = c(.4, .5), t = "t",
+    
      library("data.table")
+t <- seq(0, 10, by = .01)
+surv1 <- seq(1, .3, length.out = length(t))
+surv2 <- seq(1, .2, length.out = length(t))
+strategies <- c("Strategy 1", "Strategy 2")
+surv <- data.table(strategy = rep(strategies, each = length(t)),
+                   t = rep(t, 2),
+                   surv = c(surv1, surv2))
+surv_quantile(surv, probs = c(.4, .5), t = "t",
               surv_cols = "surv", by = "strategy")
      #>      strategy prob quantile_surv
 #> 1: Strategy 1  0.4          5.72
 #> 2: Strategy 2  0.4          5.00
 #> 3: Strategy 1  0.5          7.15
 #> 4: Strategy 2  0.5          6.25
      
   
      
-  
 
+
       
      
       
      
   
      Developed by Devin Incerti, Jeroen P. Jansen.
      
 
      
 
 
      
-  
      Site built with pkgdown 1.3.0.
      
+  
      Site built with pkgdown 1.5.1.
      
 
      
+
       
      
    
 
   
 
+
   
 
 
+
diff --git a/docs/reference/time_intervals.html b/docs/reference/time_intervals.html
new file mode 100644
index 00000000..26ed100f
--- /dev/null
+++ b/docs/reference/time_intervals.html
@@ -0,0 +1,217 @@
+
+
+
+  
+  
+
+
+
+Time intervals — time_intervals • hesim
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  
+
+  
+    
      
+      
      
+      
+
+      
+
+      
      
+
+
      
+  
      
+    
      
+    
      Time intervals
      
+    Source: R/hesim_data.R
+    
+    
      
+
+    
      
+    
      Create a table of time intervals given a vector or data frame of unique times.
+This would typically be passed to id_attributes.
      
+    
      
+
+    
      time_intervals(times)
      
+
+    
      Arguments
      
+    
+    
+    
+      
+      
+    
+    
      times
      Either a vector of times for each interval or a
+data.frame with at least one column named time_start.
      
+
+    
      Value
      
+
+    
      An object of class time_intervals that inherits from
+data.table in the same format as time_intervals as
+described in id_attributes.
      
+    
      See also
      
+
+    
      id_attributes
      
+
+    
      Examples
      
+    
      time_intervals(c(0, 3, 5))
      #>    time_id time_start time_stop
+#> 1:       1          0         3
+#> 2:       2          3         5
+#> 3:       3          5       Inf
      time_intervals(data.frame(time_start = c(0, 3, 5),
+                          time_cat = c("Time <= 3", "3 < Time <= 5",
+                                       "Time > 5")))
      #>    time_id time_start      time_cat time_stop
+#> 1:       1          0     Time <= 3         3
+#> 2:       2          3 3 < Time <= 5         5
+#> 3:       3          5      Time > 5       Inf
      
+  
      
+  
+
      
+
+
+      
      
+      
      
+  
      Developed by Devin Incerti, Jeroen P. Jansen.
      
+
      
+
+
      
+  
      Site built with pkgdown 1.5.1.
      
+
      
+
+      
      
+   
      
+
+  
+
+
+  
+
+
+
diff --git a/docs/reference/tparams.html b/docs/reference/tparams.html
new file mode 100644
index 00000000..b7ed01c9
--- /dev/null
+++ b/docs/reference/tparams.html
@@ -0,0 +1,198 @@
+
+
+
+  
+  
+
+
+
+Transformed parameter object — tparams • hesim
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  
+
+  
+    
      
+      
      
+      
+
+      
+
+      
      
+
+
      
+  
      
+    
      
+    
      Transformed parameter object
      
+    Source: R/params.R
+    
+    
      
+
+    
      
+    
      Objects prefixed by "tparams_" are lists containing transformed parameters used
+to simulate outcomes. The parameters have presumably already been transformed as a
+function of input data and consequently do not need to be used alongside
+input matrices. In other words, transformed parameters are parameters that have
+already been predicted as a function of covariates.
      
+    
      
+
+
+
+    
      See also
      
+
+    
      params
      
+
+  
      
+  
+
      
+
+
+      
      
+      
      
+  
      Developed by Devin Incerti, Jeroen P. Jansen.
      
+
      
+
+
      
+  
      Site built with pkgdown 1.5.1.
      
+
      
+
+      
      
+   
      
+
+  
+
+
+  
+
+
+
diff --git a/docs/reference/tparams_mean.html b/docs/reference/tparams_mean.html
new file mode 100644
index 00000000..26823e9d
--- /dev/null
+++ b/docs/reference/tparams_mean.html
@@ -0,0 +1,251 @@
+
+
+
+  
+  
+
+
+
+Predicted means — tparams_mean • hesim
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  
+
+  
+    
      
+      
      
+      
+
+      
+
+      
      
+
+
      
+  
      
+    
      
+    
      Predicted means
      
+    Source: R/params.R
+    
+    
      
+
+    
      
+    
      Create a list containing means predicted from a statistical model.
      
+    
      
+
+    
      tparams_mean(value, ...)
      
+
+    
      Arguments
      
+    
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
      value
      Matrix of samples from the distribution of the
+mean. Columns denote random samples and rows denote means for different observations.
      ...
      Arguments to pass to id_attributes. Each row in
+value must be a prediction for a strategy_id,
+patient_id, state_id, and optionally time_id combination.
      
+
+    
      Value
      
+
+    
      An object of class tparams_mean, which is a list containing value,
+n_samples, and the ID attributes passed to id_attributes.
      
+    
      See also
      
+
+    
      tparams
      
+
+    
      Examples
      
+    
      tparams_mean(value = matrix(1:8, nrow = 4),
+             strategy_id = rep(1:2, each = 2),
+             n_strategies = 2,
+             patient_id = rep(1, 4),
+             n_patients = 1,
+             state_id = rep(1:2, times = 2),
+             n_states = 2)
      #> $value
+#>      [,1] [,2]
+#> [1,]    1    5
+#> [2,]    2    6
+#> [3,]    3    7
+#> [4,]    4    8
+#> 
+#> $n_samples
+#> [1] 2
+#> 
+#> $strategy_id
+#> [1] 1 1 2 2
+#> 
+#> $n_strategies
+#> [1] 2
+#> 
+#> $patient_id
+#> [1] 1 1 1 1
+#> 
+#> $n_patients
+#> [1] 1
+#> 
+#> $state_id
+#> [1] 1 2 1 2
+#> 
+#> $n_states
+#> [1] 2
+#> 
+#> $grp_id
+#> [1] 1 1 1 1
+#> 
+#> attr(,"class")
+#> [1] "tparams_mean"
      
+
      
+  
      
+  
+
      
+
+
+      
      
+      
      
+  
      Developed by Devin Incerti, Jeroen P. Jansen.
      
+
      
+
+
      
+  
      Site built with pkgdown 1.5.1.
      
+
      
+
+      
      
+   
      
+
+  
+
+
+  
+
+
+
diff --git a/docs/reference/tparams_transprobs.html b/docs/reference/tparams_transprobs.html
new file mode 100644
index 00000000..43776d14
--- /dev/null
+++ b/docs/reference/tparams_transprobs.html
@@ -0,0 +1,279 @@
+
+
+
+  
+  
+
+
+
+Transition probabilities — tparams_transprobs • hesim
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  
+
+  
+    
      
+      
      
+      
+
+      
+
+      
      
+
+
      
+  
      
+    
      
+    
      Transition probabilities
      
+    Source: R/params.R
+    
+    
      
+
+    
      
+    
      Create a list containing predicted transition probabilities at discrete times.
+Since the transition probabilities have presumably
+already been predicted based on covariate values, no input data is required for
+simulation. The class can be instantiated from either an array, a data table, or
+a data frame.
      
+    
      
+
+    
      tparams_transprobs(object, ...)
+
+# S3 method for array
+tparams_transprobs(object, times = NULL, grp_id = NULL, patient_wt = NULL)
+
+# S3 method for data.table
+tparams_transprobs(object)
+
+# S3 method for data.frame
+tparams_transprobs(object)
      
+
+    
      Arguments
      
+    
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
      object
      An object of the appropriate class.
      ...
      Further arguments passed to or from other methods. Currently unused.
      times
      An optional numeric vector of distinct times to pass to
+time_intervals representing time intervals indexed by the 4th dimension of
+the array. May either be the start or the end of intervals.
+This argument is not required if there is only one time interval.
      grp_id
      An optional numeric vector of integers denoting the subgroups. Must
+be the same length as the 3rd dimension of the array.
      patient_wt
      An optional numer vector denoting the weight to apply to each
+patient within a subgroup. Must be the same length as the 3rd dimension of the array.
      
+
+    
      Value
      
+
+    
      An object of class tparams_transprobs,
+which is a list containing value and relevant ID attributes. The element value is an
+array of predicted transition probability matrices from the probability
+distribution of the underlying statistical model. Each matrix in
+value is a prediction for a sample, strategy_id,
+patient_id, and optionally time_id combination.
      
+    
      Details
      
+
+    
      The format of object depends on its class:
      
+
      array
      Must be a 4D array of matrices (i.e.,
+a 6D array). The dimensions of the array should be indexed as follows:
+1st (sample), 2nd (strategy_id), 3rd (patient_id),
+4th (time_id), 5th (rows of transition matrix), and
+6th (columns of transition matrix). In other words, an index of
+[s, k, i, t] represents the transition matrix for the sth
+sample, kth treatment strategy, ith patient, and tth
+time interval.
      
+
      data.table
      Must contain the following:
+
        
+
      • ID columns for the parameter sample (sample),
+treatment strategy (strategy_id), and patient (patient_id).
+If the number of time intervals is greater than 1 it must also contain the
+column time_start denoting the starting time of a time interval. A column
+patient_wt may also be used to denote the weight to apply to each
+patient.
        • 
+
      • Columns for each element of the transition probability matrix.
+They should be prefixed with "probs_" and ordered rowwise.
+For example, the following columns would be used for a 2x2 transition
+probability matrix:
+probs_1 (1st row, 1st column),
+probs_2 (1st row, 2nd column),
+probs_3 (2nd row, 1st column), and
+probs_4 (2nd row, 2nd column).
        • 
+
      
+
      data.frame
      Same as data.table.
      
+
+
+
+
      A tparams_transprobs object is also instantiated when creating a
+cohort discrete time state transition model using define_model().
      
+    
      See also
      
+
+    
      define_model(), create_CohortDtstm()
      
+
+  
      
+  
+
      
+
+
+      
      
+      
      
+  
      Developed by Devin Incerti, Jeroen P. Jansen.
      
+
      
+
+
      
+  
      Site built with pkgdown 1.5.1.
      
+
      
+
+      
      
+   
      
+
+  
+
+
+  
+
+
+
diff --git a/docs/reference/tpmatrix.html b/docs/reference/tpmatrix.html
new file mode 100644
index 00000000..9a17bffd
--- /dev/null
+++ b/docs/reference/tpmatrix.html
@@ -0,0 +1,224 @@
+
+
+
+  
+  
+
+
+
+Transition probability matrix — tpmatrix • hesim
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  
+
+  
+    
      
+      
      
+      
+
+      
+
+      
      
+
+
      
+  
      
+    
      
+    
      Transition probability matrix
      
+    Source: R/model_def.R
+    
+    
      
+
+    
      
+    
      tpmatrix() both defines and evaluates a transition probability matrix in which
+elements are expressions. This function is used within define_tparams() to
+create a transition probability matrix used for simulation modeling.
      
+    
      
+
+    
      tpmatrix(...)
      
+
+    
      Arguments
      
+    
+    
+    
+      
+      
+    
+    
      ...
      Named values of expressions defining elements of the matrix. The parameter
+values of the matrix elements should refer to parameters defined using
+define_rng() or define_tparams().
      
+
+    
      Value
      
+
+    
      Returns a data.table where each column is an element of the
+transition probability matrix with elements ordered rowwise.
      
+    
      Details
      
+
+    
      The matrix is filled rowwise, meaning that each row should sum to 1. It is evaluated
+in the environment used by eval_tparams() so that any objects available
+within define_tparams() can be used by tpmatrix(). The complementary probability
+equal to 1 minus the sum of the probabilities of all other rows can be
+conveniently referred to as C.
      
+    
      See also
      
+
+    
      define_model(), define_tparams()
      
+
+    
      Examples
      
+    
      p <- c(.7, .6)
+tpmatrix(
+  C, p,
+  0, 1
+)
      #>    s1_s1 s1_s2 s2_s1 s2_s2
+#> 1:   0.3   0.7     0     1
+#> 2:   0.4   0.6     0     1
      
+  
      
+  
+
      
+
+
+      
      
+      
      
+  
      Developed by Devin Incerti, Jeroen P. Jansen.
      
+
      
+
+
      
+  
      Site built with pkgdown 1.5.1.
      
+
      
+
+      
      
+   
      
+
+  
+
+
+  
+
+
+
diff --git a/docs/reference/tpmatrix_names.html b/docs/reference/tpmatrix_names.html
new file mode 100644
index 00000000..7406eeeb
--- /dev/null
+++ b/docs/reference/tpmatrix_names.html
@@ -0,0 +1,226 @@
+
+
+
+  
+  
+
+
+
+Names for elements of a transition probability matrix — tpmatrix_names • hesim
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  
+
+  
+    
      
+      
      
+      
+
+      
+
+      
      
+
+
      
+  
      
+    
      
+    
      Names for elements of a transition probability matrix
      
+    Source: R/model_def.R
+    
+    
      
+
+    
      
+    
      Create names for all elements of a transition probability matrix given
+names for the health states. This is useful for flattening a transition
+probability matrix (rowwise) into a vector and naming the resulting vector.
+The name of an element of the flattened vector representing a transition from
+the ith state to the jth state is of the form
+paste0(prefix, state_i, sep, state_j).
      
+    
      
+
+    
      tpmatrix_names(states, prefix = "p_", sep = "_")
      
+
+    
      Arguments
      
+    
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
+      
+      
+    
+    
      states
      A character vector of the names of health states in the
+transition matrix.
      prefix
      A prefix that precedes the described transitions between states.
      sep
      A character string to separate the terms representing
+state i and state j.
      
+
+    
      Value
      
+
+    
      A character vector containing a name for each element of the transition
+probability matrix encompassing all possible transitions.
      
+
+    
      Examples
      
+    
      tpmatrix_names(LETTERS[1:4])
      #>  [1] "p_A_A" "p_A_B" "p_A_C" "p_A_D" "p_B_A" "p_B_B" "p_B_C" "p_B_D" "p_C_A"
+#> [10] "p_C_B" "p_C_C" "p_C_D" "p_D_A" "p_D_B" "p_D_C" "p_D_D"
      tpmatrix_names(LETTERS[1:4], prefix = "")
      #>  [1] "A_A" "A_B" "A_C" "A_D" "B_A" "B_B" "B_C" "B_D" "C_A" "C_B" "C_C" "C_D"
+#> [13] "D_A" "D_B" "D_C" "D_D"
      tpmatrix_names(LETTERS[1:4], prefix = "", sep = ".")
      #>  [1] "A.A" "A.B" "A.C" "A.D" "B.A" "B.B" "B.C" "B.D" "C.A" "C.B" "C.C" "C.D"
+#> [13] "D.A" "D.B" "D.C" "D.D"
      
+
      
+  
      
+  
+
      
+
+
+      
      
+      
      
+  
      Developed by Devin Incerti, Jeroen P. Jansen.
      
+
      
+
+
      
+  
      Site built with pkgdown 1.5.1.
      
+
      
+
+      
      
+   
      
+
+  
+
+
+  
+
+
+
diff --git a/docs/reference/uv_rng.html b/docs/reference/uv_rng.html
new file mode 100644
index 00000000..95d54ad8
--- /dev/null
+++ b/docs/reference/uv_rng.html
@@ -0,0 +1,188 @@
+
+
+
+  
+  
+
+
+
+Generate variates for univariate distributions — uv_rng • hesim
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+  
+
+  
+    
      
+      
      
+      
+
+      
+
+      
      
+
+
      
+  
      
+    
      
+    
      Generate variates for univariate distributions
      
+    Source: R/model_def.R
+    
+    
      
+
+    
      
+    
      Generate variates for univariate distributions
      
+    
      
+
+    
      uv_rng(n, params, rng_fun, mom_fun = NULL, names = NULL)
      
+
+
+
+  
      
+  
+
      
+
+
+      
      
+      
      
+  
      Developed by Devin Incerti, Jeroen P. Jansen.
      
+
      
+
+
      
+  
      Site built with pkgdown 1.5.1.
      
+
      
+
+      
      
+   
      
+
+  
+
+
+  
+
+
+
diff --git a/docs/reference/weibullNMA.html b/docs/reference/weibullNMA.html
index 2d98e6c2..e42ce676 100644
--- a/docs/reference/weibullNMA.html
+++ b/docs/reference/weibullNMA.html
@@ -8,21 +8,29 @@
 
 Parameterization of the Weibull distribution for network meta-analysis — weibullNMA • hesim
 
+
 
-
+
 
-
+
+
 
-
+
+
+
+
+
 
 
-
+
+
 
 
-
+
 
-
-
+
+
+
 
 
 
@@ -30,13 +38,11 @@
 
 
 
-  
-  
-
 
+
 
-
+
 
 
 
@@ -50,9 +56,10 @@
 
 
 
+
   
 
-  
+  
     
      
       
      
       
 
@@ -83,28 +90,46 @@
   
   
 
 
    • 
   News
 
      •
    -
    • - - + +
      • @@ -115,6 +140,7 @@ +
      @@ -126,10 +152,8 @@

      Parameterization of the Weibull distribution for network meta-analysis

      -

      Density, distribution function, hazards, quantile function and random generation for the Weibull distribution when parameterized for network meta-analysis.

      - 
      dweibullNMA(x, a0, a1 = FALSE, log = FALSE)
@@ -147,7 +171,7 @@ 
      Parameterization of the Weibull distribution for network meta-analysis
      
 rmst_weibullNMA(t, a0, a1, start = 0)
 
 mean_weibullNMA(a0, a1)
      - +

      Arguments

      @@ -191,46 +215,42 @@

      Arg mean survival will be conditional on survival up to this time.

      - +

      Value

      dweibullNMA gives the density, pweibullNMA gives the distribution function, qweibullNMA gives the quantile function, rweibullNMA generates random deviates, HweibullNMA returns the cumulative hazard and hweibullNMA the hazard.

      -

      See also

      -

      dweibull

      - +

      dweibull

      - + + +