py-mimic-fhir 1.0.0

A package to help convert MIMIC-IV to FHIR

py-mimic-fhir

The py-mimic-fhir package is used to provide functions to convert MIMIC to FHIR. This package is dependent on:

• MIMIC-IV database installed
• MIMIC-FHIR SQL scripts run
• environment variables set with .env

py_mimic_fhir modes

---

There are three modes:

1. validate - Allows the user to validate N patients
   • --num_patients specifies number of patients to validate.
   • --export allows user to export resources after validation completes.
   • --export_limit specifies the limit of resources exported. If not specified will export all resources.
   • --rerun flag to rerun failed bundles
   • --cores allows the user to specify how many cores to use in parallel processing, default to single core
2. export - Allows the user to export resources to file.
   • --export_limit specifies the limit of resources exported. If not specified will export all resources.
   • --patient_bundle flag to export patient-everything bundles from GCP
   • --num_patients specifies the number of patients to export patient-everything bundles
   • --count specifies number of resources allowed per page of patient-everything request
   • --resource_types specifies resource types to be included in patient-everything bundles. Comma separated list
   • --ndjson_by_patient flag to export NDJSON by patient for Postgres
3. terminology - Generate CodeSystems and ValueSets for the latest MIMIC version
   • --version specifies the version of the terminology. This should be synced up with the MIMIC-IV version
   • --status specifies the status either as draft or complete. Defaults to draft
   • --post will post terminology stored in the terminology folder to the specified FHIR server
   • --generate_and_post will generate all the terminology and post immediately to the FHIR server

A couple example calls:

• Validation call: py_mimic_fhir validate --num_patients 5
• Validation call with export: py_mimic_fhir validate --num_patients 5 --export
• Re-validate failed bundles: py_mimic_fhir validate --rerun
• Validate in parallel: py_mimic_fhir validate --num_patients=40 --cores=4
• Export all: py_mimic_fhir export
• Export limited amount: py_mimic_fhir export --export_limit 1
• Export patient-everything: py_mimic_fhir export --patient_bundle --num_patients=100 --resource_types='Patient,Encounter,Condition'
• Export ndjson by patient: py_mimic_fhir export --ndjson_by_patient
• Terminology generation call: py_mimic_fhir terminology --version 0.4 --status complete
• Terminology post call: py_mimic_fhir terminology --post
• Terminology generate and post call: py_mimic_fhir terminology --generate_and_post --version 0.4

py_mimic_fhir environment variables

Along with the main arguments that the user enters, there are expected environment variables. The user should export all environment variables found in their .env to the bash using source .env. The variables pulled into py_mimic_fhir are:

Argument	Environment variable	Notes
--sqluser	SQLUSER	The SQL database username
--sqlpass	SQLPASS	The SQL database password
--dbname_mimic	DBNAME_MIMIC	The name of the mimic database locally
--host	DBHOST	Where the database is hosted (typically localhost)
--port	PGPORT	Which port the database is located on
--fhir_server	FHIR_SERVER	The fhir server, in our case HAPI
--output_path	MIMIC_JSON_PATH	The output path for exported resources
--log_path	MIMIC_FHIR_LOG_PATH	The path to the log files
--dbname_hapi	DBNAME_HAPI	The name of the hapi database locally
--terminology_path	MIMIC_TERMINOLOGY_PATH	The directory to output terminology to
--err_path	MIMIC_FHIR_LOG_PATH	The directory for logs to be stored
--gcp_project	GCP_PROJECT	The GCP Project name
--gcp_topic	GCP_TOPIC	The GCP topic name to submit bundles to
--gcp_location	GCP_LOCATION	The GCP location of services
--gcp_bucket	GCP_BUCKET	The Google Cloud Storage bucket where bundle errors can be logged
--gcp_dataset	GCP_DATASET	Google Healthcare API dataset
--gcp_fhirstore	GCP_FHIRSTORE	Google Healthcare API FHIR store
--gcp_export_folder	GCP_EXPORT_FOLDER	Google Cloud Storage folder to export resources to
--validator	FHIR_VALIDATOR	FHIR Validator being used. One of HAPI, GCP, or JAVA
--db_mode	BD_MODE	Database mode, either Postgres or BigQuery