Perinatal & Pediatric
Who do we leave out when we exclude discordant birth certificate records? Ruby Barnard-Mayers* Ruby Barnard-Mayers Martha Werler
Background: Many studies on maternal and child health use birth certificate data. However, this data is not always consistent, and individuals with discrepancies are often excluded from analyses, as their information can’t be assumed to be reliable.
Objective: The purpose of this study is to understand who we exclude when we exclude discrepant records from birth certificate data.
Methods: Birth certificate data from the Pregnancy and Early Life Longitudinal (PELL) data system, which links birth certificate and hospital discharge records in Massachusetts from 2011 to 2018. For 4 conditions (pregnancy risk factor, delivery procedure, test or screening procedure, and labor and delivery complication) the birth certificate has a check box to indicate ‘No.” After each, the birth certificate has a list of specific conditions for checking off when present (
Results: Of the 552,806 births, 9,737 (1.8%) were discrepant. The most common discrepancy was for labor and delivery complications (67%), followed by tests and procedures (24%), parity (11%), pregnancy risk factors (6%), and finally delivery procedures (0.2%). Compared to birthing people with non-discrepant record, those with discrepant records were, on average, about one year older, were more likely to have private insurance (63.4% vs. 49.1%), were more likely to have a college degree (58.3% vs. 46.0%), were less likely to have a vaginal delivery (53.3% vs. 66.4%), and were less likely to have prenatal care (87.3% vs. 95.2%). Birthing people with discrepant records also appeared more likely to be non-Hispanic Black (18.4 vs. 9.9%) Distributions of BMI, parity, and country of birth were similar for the two groups.
Conclusion: Birthing people with discrepant birth certificate records differ from those without such discrepancies on socio-economic and reproductive indicators, raising the possibility of bias arising from their exclusion. Further research should focus on the potential impact excluding these people from analyses may have on results.
