Validation of case definition algorithms for the ascertainment of congenital anomalies

Area of Science:

• Public Health Surveillance
• Health Informatics
• Pediatric Epidemiology

Background:

• Congenital anomalies (CA) are a major cause of infant mortality and disability.
• Health administrative data and case definition algorithms are crucial for CA monitoring.
• Validation of these algorithms is essential to assess data reliability and surveillance system limitations.

Purpose of the Study:

• To validate the performance of case definition algorithms used in a congenital anomaly surveillance system in British Columbia, Canada.
• To assess the accuracy of algorithms for identifying congenital anomalies in administrative health data.

Main Methods:

• Linked a cohort of births (March 2000-April 2002) to the Health Status Registry (HSR) and BC Congenital Anomalies Surveillance System (BCCASS).
• Calculated algorithm performance measures (sensitivity, specificity, positive predictive value, negative predictive value) using the HSR as the reference standard.
• Assessed agreement between databases using the kappa coefficient and followed modified Standards for Reporting Diagnostic Accuracy guidelines.

Main Results:

• Algorithm performance varied by specific congenital anomaly.
• Positive predictive values ranged from approximately 73% to 100%.
• Specificity and negative predictive values consistently exceeded 99%, while sensitivity was lower, particularly for internal anomalies or those not evident at birth.

Conclusions:

• Validated case definition algorithms are valuable tools for population-level congenital anomaly surveillance.
• Algorithm accuracy is higher for anomalies readily identifiable at birth.
• Utilizing validated case definitions facilitates robust CA monitoring and cross-jurisdictional comparisons for public health initiatives.