Abstract
BACKGROUND: Health administrative datasets have the potential to provide valuable insights into pediatric solid organ transplantation; however, validation is necessary to ensure their accuracy. This study aimed to assess the validity of administrative data by comparing it to direct transplant records from a major pediatric transplant center in Ontario, Canada (1991-2011). METHODS: Using linked administrative healthcare databases, we conducted a retrospective analysis to evaluate the validity of physician billing claims and hospital diagnostic and procedural codes in identifying pediatric solid organ transplants. Sensitivity and positive predictive value (PPV) were calculated for various algorithms. RESULTS: During the study period, a total of 347 kidney, 250 liver, 200 heart, and 28 lung transplants were performed. The best algorithm for identifying these transplants utilized hospital procedural codes from the Canadian Institute for Health Information Discharge Abstract Database. Compared to transplant center records, these codes demonstrated a sensitivity of 91% (95% CI: 89-93) and PPV of 93% (95% CI: 91-95) when including all organ types, and performed similarly well when evaluating individual organ types. CONCLUSION: This study is the first to validate administrative data for identifying pediatric solid organ transplant recipients, demonstrating the reliability of procedural codes for population-level health research in this domain.