Assessment of completeness of reporting of oro-facial cleft cases in Arizona using the capture-recapture method.

The Arizona Birth Defects Monitoring Program (ABDMP) is mandated to collect data on several birth defects including oro-facial clefts such as cleft lip with and without cleft palate (CL) and cleft palate without cleft lip (CP). Completeness of reporting and ascertainment validations had never been conducted on the birth defects monitored by the program. Two-source capture-recapture modeling was used to evaluate the completeness of incidence reporting of CL and CP. The Lincoln-Petersen model and the log-linear model were used to estimate the completeness of reporting of CL and CP cases based on hospital discharge and disease index data. On average, 14 cases of CP per year and 24 cases of CL per year were identified as underreported. Validations of model estimates of CL and CP were done using confirmed cases of CL and CP from the Arizona Birth Defects Registry for the years of 2004 through 2007. The agreement between the observed and estimated number of cases of CP and CL by both models was adequate as demonstrated by the observed coefficient of determination R2. Hospital discharge data had a better percent of completeness than disease index data. Stronger agreements between estimated and observed cases were found using both models for CP cases as compared to CL cases. Neither data source was adequate for complete reporting of CP and CL cases in Arizona. The estimated average case underreporting suggests that appropriate policy provisions and actions are needed to achieve better completeness of case reporting.