A replicable, open-source, data integration method to support national practice-based research & quality improvement...

Area of Science:

• Neurology
• Health Informatics
• Data Science

Background:

• The Epilepsy Learning Healthcare System (ELHS) aims to improve epilepsy outcomes, but data system fragmentation hinders progress.
• Integrating real-world clinical data into national research and quality improvement (QI) systems is crucial for advancing epilepsy care.

Purpose of the Study:

• To test the feasibility of an open-source Data Integration (DI) method for connecting epilepsy clinical data to national systems.
• To overcome barriers in data fragmentation for improved reporting and participation in the ELHS.

Main Methods:

• Developed and validated a DI method to extract, preprocess (using text mining), and structure epilepsy data from the Electronic Health Record (EHR) Enterprise Data Warehouse (EDW).
• Utilized EPIC SmartPhrases/Forms to collect standardized epilepsy data, including diagnosis, seizure type, frequency, and medication adherence.
• Matched extracted data fields using the ELHS data dictionary to create a structured dataset for uploading to the national ELHS registry.

Main Results:

• The DI method automated data mining, preprocessing, and exporting of structured data for weekly access by clinicians and quality improvers.
• During the study period (12/2018-12/2021), 1806 patients were included, with 5168 clinic visits documented.
• The structured dataset, containing seizure control information, was successfully uploaded to the national ELHS registry.

Conclusions:

• Structured epilepsy patient data can be feasibly extracted from clinical text notes.
• This DI method enables weekly reporting to a national learning healthcare system, facilitating quality improvement and research in epilepsy care.