EHR data analysis tools

This ehr-utils application provides some simple analytical capabilities for EHR data.

Instruction for use (End User)

Requirements

To use this ehr-utils application, you need to have a working Python 3.10.8+ environment.

Installation

Clone the repository:

git clone [email protected]:biostat821-2023/ehr-utils-Yer1k.git

Usage

To use the ehr-utils application, you need to import the analysis module:

import analysis

Patient data and lab data tabular files sturcture

The patient data and lab data are stored in two Tab Delimited files, either in .txt or .csv format, such as patients.txt and labs.txt, respectively.

The patients.txt file contains the following columns:

• PatientID: a unique identifier for each patient
• PatientGender: gender of the patient
• PatientDateOfBirth: date of birth of the patient in the format of YYYY-MM-DD HH:MM:SS.SSS
• PatientRace: race of the patient
• PatientMaritalStatus: marital status of the patient
• [Optional]PatientLanguage: language of the patient
• [Optional]PatientPopulationPercentageBelowPoverty: percentage of the population below poverty of the patient's residence

The labs.txt file contains the following columns:

• PatientID: a unique identifier for each patient
• AdmissionID: a unique identifier for each admission
• LabName: name of the lab test
• LabValue: value of the lab test
• LabUnits: units of the lab test
• LabDateTime: date and time of the lab test in the format of YYYY-MM-DD HH:MM:SS.SSS

Then, you can use the analysis module to perform analyses, including parsing the patient and lab data with parse_data function which returns a dictionary of Patient objects, calculating the age of the patient with property age of Patient object, checking whether the patient is sick with is_sick method of Patient object, and calculating the age of the patient when their earliest lab was recorded with first_admit method of Patient object.

Parse data

The function patient_data(patient_filename: str) -> dict[str, Patient] should take the path to the patient data and the path to the lab data and return a dictionary of Patient objects. For example,

>> parse_data(patient_file, lab_file)
"patient.db created"

Patient age property

The property age of Patient object should take the data and return the age of the patient. For example,

>> Patient("1A8791E3-A61C-455A-8DEE-763EB90C9B2C").age
49

First admission age property

The property first_admit of Patient object should take the data and return the age of the patient when their earliest lab was recorded. For example,

>> Patient("1A8791E3-A61C-455A-8DEE-763EB90C9B2C").first_admit
18

Sick patient check

The method is_sick of Lab object should take the data and return whether the patient is sick. For example,

>> Lab("1A8791E3-A61C-455A-8DEE-763EB90C9B2C").is_sick( "METABOLIC: ALBUMIN", ">", 4.0)
True

Instruction for contributors

For generalization purpose, the ehr-utils application is designed to be used by both end users and developers. The end users can use the ehr-utils application to perform simple analyses on EHR data. The developers can use the ehr-utils application as a template to develop their own EHR data analysis tools.

Development requirements

Pull requests are welcome.

For major changes, please open an issue first to discuss what you would like to change.

Before submitting a pull request, please make sure that your code passes all the tests.

Please make sure to update tests as appropriate. To contribute to this project, you need to have a working Python 3.10.8+ environment, and all source code should be formatted with black, mypy, pycodestyle, and pydocstyle.

You may view the specifics of the checks in this repository's workflow specification: .github/workflows

Tests

For testing, you would need the pytest package. To run the tests, you can use the following command:

    pytest filename.py

To see coverage report, you can use the coverage package. To run the coverage report, you can use the following command:

    coverage run -m pytest filename.py > test_report.txt
    coverage report --show-missing

For example:

You may need at least 70% coverage to pass the tests; however, if the coverage is not 100%, please make sure to explain why in the pull request.