Creating a CareVue-Derived Cohort from MIMIC-III

[AL0UNE]

Hi everyone,

I’m working on a project involving both the MIMIC-III and MIMIC-IV databases. Specifically, I am using MIMIC-IV as a temporal validation dataset. My analyses are based on data extracted using the script sapsii.sql from the MIMIC III and IV concepts repository, directly in BigQuery.

However, I would like to ensure a non-overlapping cohort between the two datasets, and thus I'm particularly interested in focusing on the CareVue subset of MIMIC-III.

Ideally, I would like to create a "derived" version of the MIMIC-III CareVue data using the scripts from the concepts/ folder.

My questions:

• What would be the best way to accomplish this?
• I managed to connect a colab notebook to my project in BigQuery. Is it possible to execute the scripts from the concepts folder directly within Colab (by changing the name in the make-concepts.sh)?
• If so, would it be sufficient to download the scripts, upload them to my Google Drive, and run them from there via colab?

Complete newbie, any guidance or best practices would be greatly appreciated!

Thank you in advance!