Reliability of birth certificate reporting of congenital anomalies.

Birth certificates comprise an important source of data on the prevalence of genetic conditions and for monitoring possible teratogens in the population. Investigators have found wide variability (12 to 100%) in the accuracy of reporting. In a large public hospital, of those congenital anomalies detected at birth, only 5.4% were recorded on the birth certificate. This is one of the lowest rates ever reported. An underreporting correction factor may be applied if congenital anomalies are distributed randomly with respect to reporting status, and the rate of reporting is sufficient to comprise a valid sample for estimating a correction factor (that is, 20% or more reported). In this study, factors such as numbers or types of anomalies, race, infant birthweight, or estimated gestational age did not significantly influence the rate of birth certificate reporting. Thus, our data satisfied the first but not the second criterion for derivation of a correction factor. In conducting epidemiologic studies, birth certificate data should be used with: (1) great caution; (2) a system of validation with the medical record to estimate the degree of underreporting and to derive a correction factor; and (3) a priori knowledge that underreporting of congenital anomalies on this document is highly prevalent.