Development and implementation of an entity relationship diagram for perinatal data.

Area of Science:

• Perinatal health research
• Health informatics
• Clinical data standardization

Background:

• Severe maternal morbidity and mortality rates in the U.S. exceed those in other high-income nations.
• Significant disparities persist in maternal health outcomes.
• Electronic Health Records (EHRs) contain valuable data for research but require standardized extraction methods.

Purpose of the Study:

• To develop a standardized approach for extracting perinatal data from EHRs.
• To facilitate research on severe maternal morbidity and mortality.
• To address disparities in perinatal outcomes.

Main Methods:

• Harmonized perinatal EHR data into a common data model.
• Developed an Entity Relationship Diagram (ERD) for aggregating data at granular levels (mothers, infants, encounters).
• Indexed observations by gestational age and time from delivery.
• Created a standard extract, transform, and load (ETL) process for pregnancy-related observations.

Main Results:

• An ERD was created to aggregate perinatal EHR data effectively.
• A standardized ETL process was developed for EHR data.
• The approach enables data inclusion in the PCORnet® Common Data Model.

Conclusions:

• The developed ERD facilitates cross-institutional perinatal research.
• This structured approach can identify at-risk populations and prompt interventions.
• Standardized EHR data extraction accelerates perinatal research and can improve outcomes.