This folder contains code for reproducing a study on indwelling arterial catheters:
Hsu DJ, Feng M, Kothari R, Zhou H, Chen KP, Celi LA. The association between indwelling arterial catheters and mortality in hemodynamically stable patients with respiratory failure: a propensity score analysis. CHEST Journal. 2015 Dec 1;148(6):1470-6.
The study showed, in the MIMIC-II database, that after adjustment for various confounders, indwelling arterial catheters were not associated with a mortality benefit.
The code here reproduces this study in the MIMIC-III database. This involved many technical changes due to schema differences and data differences between MIMIC-II and MIMIC-III. As MIMIC-III covers four additional years, the cohort extracted here is larger than that reported in the study.
This version of the aline study has been modified to work with the MIMIC-III dataset in the AWS OpenData program and query it using AWS Athena instead of PostgreSQL. You can learn more about the details of this modification and see a performance and cost comparison in this blog post:
The study can be reproduced by:
- Use the below Launch Stack button to deploy access to the MIMIC-III dataset into your AWS account. This will give you real-time access to the MIMIC-III data in your AWS account without having to download a copy of the MIMIC-III dataset. It will also deploy a Jupyter Notebook with access to the content of this GitHub repository in your AWS account. Prior to launching this, please login to the MIMIC PhysioNet website, input your AWS account number, and request access to the MIMIC-III Clinical Database on AWS.
To start this deployment, click the Launch Stack button. On the first screen, the template link has already been specified, so just click next. On the second screen, provide a Stack name (letters and numbers) and click next, on the third screen, just click next. On the forth screen, at the bottom, there is a box that says I acknowledge that AWS CloudFormation might create IAM resources.. Check that box, and then click Create. Once the Stack has complete deploying, look at the Outputs tab of the AWS CloudFormation console for links to your Juypter Notebooks instance.
-
Running the
aline-awsathena.ipynbfile - this notebook generates the data (using the SQL files in this directory) and saves a single dataframe to CSV -
Running the
aline_propensity_score.ipynbfile - this notebook uses R and the above CSV to create a propensity score and calculate the mortality difference between matched pairs
There are a few minor differences between our reproduction and the original study.
- the original study subselected variables using a genetic algorithm, whereas we simply use the final set of variables they report
- we did not include PO2 and PCO2 in the propensity score
- we removed patients based on hospital service, not ICU service