Validation of birth certificate data in New York State.

Vital statistics birth certificate data are an important source of information for researchers, policy makers, and state officials to evaluate the quality of care delivered to pregnant women. The purpose of this study was to assess the validity of data elements being reported by the hospitals on the birth certificate record when compared to the medical record. This study used a random sample of birth certificates from two upstate and two downstate counties, in New York State, comprising a total of 100 records per county. The review assessed data elements from seven major categories: prenatal care, maternal medical risk factors, risk factors related to pregnancy, lifestyle risk factors, method of delivery, complications of labor and delivery, and infant information. Sensitivity, specificity, the positive predictive value and the negative predictive value assessed level of agreement between the birth certificate and the medical record. Overall, the birth certificate data reflected high specificity, because most conditions are rare. The sensitivity of the data was more varied, ranging from 0 to 100%, reflecting that if a rare condition was present it often was not documented on the birth certificate. Many of the data elements are reported accurately. However, caution should be used for data elements that are poorly reported.