Background: To examine the consistency and likely degree of bias in a study of cardiovascular health, linked with reproductive data over 40 years.
Methods: Linkage of vital statistics data of births to female Bogalusa Heart Study participants was compared to interviewing of female participants. The characteristics of participants, the agreement, and demographic, study-related, and medical predictors of discrepancy were analyzed, using kappa statistics, mean and median differences, and logistic regression.
Results: Overall, 3944 (66.7%) of participants were located by one or both sources. The strongest predictor of either linkage or interview was recent and/or frequent participation in the parent study. Agreement between the two sources was generally good (kappa > 0.9 for birthweight and 0.8 for gestational age). Black race, older age, and time since pregnancy were associated with greater discrepancy in reporting of outcomes, but cardiovascular risk factors generally were not.
Conclusions: Combining information from multiple sources to increase sample size and outcome ascertainment may be valid, which will increase population health sciences' ability to leverage the many existing, large-scale sources to answer previously unexplored questions, even those that the data were not initially collected to answer.
Keywords: Bias; Cardiovascular disease; Common data elements; Reproductive history; Vital statistics.