Related Experiment Video
Updated: Jun 19, 2026
14:43
A Novel Method for Involving Women of Color at High Risk for Preterm Birth in Research Priority Setting
Published on: January 12, 2018
The identification of pregnancies within the general practice research database
Scott Devine1, Suzanne West, Elizabeth Andrews
1University of NC School of Public Health, Chapel Hill, NC, USA. sdevine@unc.edu
Pharmacoepidemiology and Drug Safety
|October 14, 2009
Summary
We developed an algorithm to identify pregnancies in electronic health records, enabling better drug safety studies. This method accurately captures pregnancy outcomes for large-scale teratogenicity research.
Area of Science:
- Pharmacovigilance and Pharmacoepidemiology
- Reproductive Health and Outcomes Research
- Health Informatics and Data Science
Background:
- Electronic health records (EHRs) are increasingly used for drug safety surveillance in the United States.
- Identifying pregnancies within EHRs is a significant challenge for teratogenicity studies.
- The General Practice Research Database (GPRD) requires robust methods for pregnancy identification.
Purpose of the Study:
- To develop and validate an algorithm for identifying pregnancies within the GPRD.
- To enable the use of EHR data for studying pregnancy outcomes and teratogenicity.
- To improve the accuracy of pregnancy identification in large administrative healthcare databases.
Main Methods:
- Developed an algorithm to identify pregnancies in women aged 15-45 between 1987 and 2006.
- Included identification of live births, stillbirths, spontaneous, and elective terminations.
- Validated the algorithm using the Additional Clinical Details Maternity (ACDM) file and free-text records.
Main Results:
- Analyzed over 16 million records, identifying 580,356 pregnancies in 383,184 women.
- Detailed pregnancy outcomes included live births, preterm births, multiple deliveries, abortions, terminations, and fetal deaths.
- Pregnancy care markers were identified for 86.3% of pregnancies, with consistent results from internal validation.
Conclusions:
- Successfully developed a validated algorithm for identifying a large number of pregnancies in the GPRD.
- The hierarchical approach minimizes misclassification of pregnancy outcomes.
- This algorithm facilitates large-scale pharmacoepidemiological research on drug safety during pregnancy.
