Validation of Birth and Birth Outcome Information in the German Pharmacoepidemiological Research Database (GePaRD)

Background: Any use of data not primarily collected for research purposes requires their prior validation. A large pharmacoepidemiological database was created in Germany, which contains records of ˜14 mill insurants from four statutory health insurances covering whole Germany (17% of the German population). The aim of the current analysis was to validate the birth and birth outcome information in GePaRD by comparing it to official statistics. Methods: We estimated the crude birth rate (number of newborns per 100,000 population), the fertility rate (number of newborns per 1,000 women in reproductive age range (15–44)), the percentage of twins and higher order births and percentage of stillbirths stratified by federal state in 2004 to 2006. We compared our results with the data from birth registration from federal health monitoring. Results: The crude birth rate agreed with the official statistics within 2% (2004: 873 vs. 855, 2005: 834 vs. 832 and 2006: 806 vs. 817). The fertility rate showed an average underestimation of 9% (2004: 39 vs. 42, 2005: 38 vs. 41 and 2006: 38 vs. 41), in line with the slight underrepresentation of younger women in reproductive age in the GePaRD compared to the German population. As expected, regional variation in fertility indicators from official statistics in Germany was better reproduced by the database estimates for the fertility rate than for crude birth rate. Twins and higher order births were recorded in 1.5% of all births in our database and in 1.6% in the official statistics for the years 2004–2006. During the same time period, 3.3% of births in the database were stillbirths, in agreement with the official statistics where 3.6% of the deliveries were stillbirths. Conclusion: Our results show a good agreement of the database with official statistics for Germany for birth and birth outcomes providing a basis of further evaluation.