Vital statistics and other administrative data are becoming an increasingly important source for epidemiologic research and surveillance. This study, the first in Canada, evaluated the usefulness of birth registry data on congenital anomalies in Alberta. We compared the number of birth defects recorded in the birth registry with the number collected through the Alberta Congenital Anomalies Surveillance System (ACASS) between 1985 and 1996. In addition, records of 3,881 (99.9 %) babies with (a) birth defect(s) from the ACASS during 1994--1996 were matched to the birth registry by deterministic linkage. Of these, 2,969 babies had single anomalies that were used for validity analysis. The anomalies were grouped by those within International Classification of Disease (ICD) ICD-9 Section XIV (ICD-9=740.0-759.9) and those outside the Section XIV. For those within Section XIV, 24 summary diagnostic categories were examined. As shown, the total case count from the birth registry was on average about 3 % lower than that from the ACASS between 1985 and 1996. The validity of diagnostic categories is high for the 24 categories examined, with an overall agreement of between 80 % and 100 %. The sensitivity, positive predictive value, and kappa are also high for all these anomalies combined during 1994 and 1996, showing 95.7 %, 99.8, and 0.81 respectively.