The influence of misclassification bias on the reported rates of congenital anomalies on the birth certificates for West Virginia-A consequence of an open-ended query

BACKGROUND: Passive surveillance for congenital anomalies using birth certificates are generally considered to have biased reporting, though the sources of those biases are not well-known nor controlled for. We have analyzed the congenital anomaly reporting data for 418,385 live births in West Virginia (1990-2009) from the 1989 US standard birth certificate and have newly identified a particular source of bias. METHODS: Congenital anomaly prevalence rates per 100 live births have been determined for both specified birth defects and for other congenital anomalies by county, by hospital, and by year. Extreme outliers were identified by z score. Text strings for "other congenital anomaly" reports recorded for 1998-2009 were assessed for information on congenital anomalies. RESULTS: While rates for specified birth defects reported in checked-box format showed little variation, rates for "other congenital anomaly" collected in open-ended format showed much variation. Nearly half of the "other congenital anomaly" reports were for neonatal conditions rather than for major structural congenital anomalies. This misclassification alone had elevated the state-wide congenital anomaly reporting rate from 1.1 to 1.8% of live births. Geographic clustering and a temporal bulge in congenital anomaly reports disappeared after misclassified data were removed. CONCLUSIONS: Data collected in checked-box format on specified birth defects showed consistent patterns over time and space, while data collected in open-ended format on "other congenital anomalies" showed an epidemiological pattern reflecting neonatal conditions rather than birth defects. The 2003 US standard birth certificate wisely limits data collection to specified birth defects using the checked-box format.