Validation of the NVSCA registry for common oral clefts: Study design and first results
Objective: Since 1997 the Dutch Association for Cleft Palate and Craniofacial Anomalies (NVSCA) has maintained a national registry of congenital craniofacial anomalies. This study validates data on three common oral cleft categories (cleft lip/alveolus = CL/A; cleft lip/alveolus and palate = CL/AP; and cleft palate = CP) and general items. Design: Retrospective observational study. Setting: All 15 Dutch cleft palate teams registered presurgery patients with common oral clefts (n = 2553) from 1997 to 2003. Patients: A random sample of 250 cases was used; 13 cases were excluded. Main Outcome Measures: The corresponding medical data were reviewed; these medical data served to validate the NVSCA registry data. Prevalence comparisons, 2 X 2 tables and validity measures were performed. Results: The cleft categories most accurately recorded were CL/A and CP. Both categories had an observed agreement of 98%, kappa of 0.94, and a sensitivity and specificity of 97%. Cleft lip/alveolus and palate had an observed agreement of 95%, kappa of 0.89, a sensitivity of 90%, and a specificity of 99%. Regarding the general items, observed agreement and kappa were highest for adoption/foster child (99%; 0.76) and lowest for remarks about pregnancy (63%; 0.20). Sensitivity ranged from 25% (consanguinity) to 97% (white mother) and specificity was high for all items (>93%) except for white father and mother (approximately 35%). Conclusions: The NVSCA registry is a valuable tool for quality improvement and research because validity on all three common oral cleft categories is very good. Validity on the general items is reasonable to satisfying and appears to be related to the type of information.