11 February 2024
- Now requires minimum
AMR
v2.0.0. - Updated function calls and documentation following deprecation of
AMR::as.rsi()
.
19 November 2023
- Now requires minimum
dbplyr
v2.4.0. - Unit test maintenance required by
dbplyr
v2.4.0.
11 April 2023
- Now requires minimum R v4.0.0,
dplyr
v1.1.0,dbplyr
v2.3.2 andtidyselect
v1.2.0. - Internal code maintenance in line with new tidy programming practice.
- Tolerance built into the timeline for handling of no growth results. Ramses does not
currently prescribe whether no growth should be loaded into
microbiology_isolates
or not (#108). The timeline will now show specimens without an isolate record at the time of collection.
23 October 2022
- Changed
validate_inpatient_diagnoses()
to accept two optional fieldsdiagnosis_start
anddiagnosis_end
and take advantage of clinical systems such as problem lists - Changed
therapy_timeline()
to take advantage ofdiagnosis_start
anddiagnosis_end
variables if present - Changed the
inpatient_diagnoses
mock dataset to removed episode start and end fields #103 which duplicated information ininpatient_episodes
- Changed
bridge_*()
functions to add a fieldantiinfective_type
to bridge tables #76 and prevent unrelated generation issues #104 - General code maintenance and adaptation to ongoing changes to the
dplyr
data masking approach #105.
13 September 2022
validate_investigations()
now expectsobservation_code
values to be associated with uniqueobservation_name
andobservation_display
values- improvements to namespace handling in
validate_microbiology()
.
30 August 2022
- Fix bug affecting
TherapyEpisode
instances created for more than 3 therapy episodes (id
with length > 3) #100
27 August 2022
- Local databases (
connect_local_database()
andcreate_mock_database()
) are now powered by DuckDB rather than SQLite, bringing support for datetime variable types and increased performance. Ramses now depends on theduckdb
library - Added new S4 class
Encounter
#92 to manipulate hospitalisations (admissions) with associated methods forshow()
,compute()
,collect()
,Patient()
,longitudinal_table()
,clinical_feature_*()
,therapy_timeline()
TherapyEpisode()
andEncounter()
have a new optionalextend_table_start
input controlling their longitudinal table's start. This allows the creation of longitudinal tables starting before thetherapy_start
oradmission_date
, respectively. More detail is available fromvignette("therapy-episodes")
- Added
Patient()
function to create aPatient
object from any other object of classRamsesObject
- Improved documentation and export of S4 methods
compute()
andcollect()
.
therapy_table()
is now deprecated: uselongitudinal_table()
instead #86bridge_spell_therapy_overlap()
is now deprecated : usebridge_encounter_therapy_overlap()
instead #86- Database table
inpatient_episodes
and documentation now refer to 'encounters' rather than 'spells'. Variablespell_id
is now known asencounter_id
#86. Databases built with previous versions of Ramses will not work with Ramses 0.5.0 - Function signatures for all methods are standardised: function arguments
object
are now changed tox
, except forshow(object)
. - Support for SQLite databases is withdrawn.
22 August 2022
clinical_feature_*()
functions trigger a warning rather than an error if no value matchingobservation_code
is found in theinpatient_investigations
table.
2 August 2022
- upgrade dependency on packages
dplyr
version >= 1.0.1 anddbplyr
version >= 2.1.0 - improve
validate_inpatient_episodes()
#85
- fix bug in
map_charlson_comorbidities()
withtbl_df
objects #84 - fix bug in
compute()
withTherapyEpisode
objects #83 - fix uncommon bug in
clinical_feature_*()
functions with PostgreSQL #88
28 January 2022
- upgrade dependency on
AMR
package version >= 1.8.0 - better error messages in
clinical_feature_*()
functions #80 spell_id
is no longer required byvalidate_investigations()
#77- bridge tables will be generated by
bridge_tables()
even in the absence of defined daily doses (DDDs) in drug prescription and administration records - improve
vignette("Ramses")
5 August 2021
- clarifications on
vignette("load-data")
- corrections to
antibiotic_icd_indications
7 July 2021
- faster implementation of prescription linkage using
igraph
dependency TherapyEpisode
class now supporting multiple therapy episodes- support for integer
*_id
variables - updated schema of drug prescriptions and administrations:
drug_id
refactored intodrug_code
;ATC_group
retired in favour ofdrug_group
- updated schema of microbiology isolates and susceptibilities: isolates are now identified by
isolate_id
rather thanorganism_id
- updated schema of microbiology susceptibilities:
drug_id
,drug_name
anddrug_display_name
refactored intoagent_code
,agent_name
andagent_display_name
- faster execution of
clinical_feature_*()
functions on PostgreSQL - uniform handling of
prescription_status
: therapy episodes and bridge table records will not be generated for prescriptions with status"draft"
,"entered-in-error"
,"cancelled"
, or"unknown"
. - update
reference_loinc
to version 2.70 and expand to 10 new concepts
- fix bug in
validate_prescriptions()
checks ondaily_frequency
- fix bug in
load_*()
handling ofdata.table
objects
17 March 2021
- fix bug in handling
data.table
objects inarrange_variables()
andvalidate_*()
26 February 2021
- introduce S4 classes
Patient
,MedicationRequest
,TherapyEpisode
- introduce therapy methods
therapy_table()
,parenteral_changes()
,clinical_feature_last()
,clinical_feature_ols_trend()
,clinical_feature_interval()
,clinical_feature_mean()
- remove dependencies
- improve handling of time data types by SQLite
- upgrade to
units
package version >=0.7-0
- new
ramses-objects
vignette - new
therapy-episodes
vignette
8 February 2021
- Corrected incorrect transitive closure settings
- Corrected prescription edge classification
- Complete verification of prescription edge classification
therapy_id
now always set to the id of the first prescription in the episode
25 January 2021
- Support PostgreSQL
- Introduce standalone
patients
table and dataset
8 January 2021
- First minor version