Purpose: Since 2012, BORN Ontario, a maternal-newborn registry, has collected data on every birth in Ontario. To ensure data quality, we assessed the reliability of key elements collected in BORN by comparing these with like data elements in the Canadian Institute for Health Information-Discharge Abstract Database (CIHI-DAD).
Methods: We used provincial health card numbers to deterministically link live or stillbirth records and their corresponding mothers’ records in the BORN database to the CIHI-DAD in the fiscal years 2012-2013 to 2014-2015. Percentage agreement and Cohen Kappa statistics were used to assess agreement on main elements in both databases.
Results: The percentage agreement and Kappa coefficients were 99.98% and 0.740 (95% CI: 0.677-0.803) on live/stillbirth, respectively. The Kappa coefficients for infant sex, gestational age at birth, induction of labour, and caesarean birth were 0.989 (95% CI: 0.988-0.989), 0.920 (95% CI: 0.919-0.920), 0.782 (95% CI: 0.780-0.785), and 0.995 (95% CI: 0.995-0.996), respectively. Kappa agreement for the number of fetuses in a pregnancy was 0.979 (95% CI: 0.977-0.981). Percentage agreement was very high for infants’ birthdates (99.9%), infant postal codes (91.8%), infants’ birth weight in grams (95.5%), and mothers’ dates of birth (99.1%).
Conclusions: Overall, the BORN and CIHI-DAD databases had concordance on key birth and maternal data elements; however, additional work is needed to understand discrepancies identified.
Scientist, CHEO Research Institute