Research conducted using large administrative data sets has increased in recent decades, but reports on the fidelity and reliability of such data have been mixed. The goal of this project was to compare data from a large, administrative claims data set with a quality improvement registry in order to ascertain similarities and differences in content.
Data on children younger than 12 months with nonsyndromic craniosynostosis who underwent surgery in 2012 were queried in both the Kids’ Inpatient Database (KID) and the American College of Surgeons Pediatric National Surgical Quality Improvement Program (Peds NSQIP). Data from published clinical craniosynostosis surgery series are reported for comparison.
Among patients younger than 12 months of age, a total of 1765 admissions were identified in KID and 391 in Peds NSQIP in 2012. Only nonsyndromic patients were included. The mean length of stay was 3.2 days in KID and 4 days in Peds NSQIP. The rates of cardiac events (0.5% in KID, 0.3% in Peds NSQIP, and 0.4%-2.2% in the literature), stroke/intracranial bleeds (0.4% in KID, 0.5% in Peds NSQIP, and 0.3%-1.2% in the literature), infection (0.2% in KID, 0.8% in Peds NSQIP, and 0%-8% in the literature), wound disruption (0.2% in KID, 0.5% in Peds NSQIP, 0%-4% in the literature), and seizures (0.7% in KID, 0.8% in Peds NSQIP, 0%-0.8% in the literature) were low and similar between the 2 data sets. The reported rates of blood transfusion (36% in KID, 64% in Peds NSQIP, and 1.7%-100% in the literature) varied between the 2 data sets.
Both the KID and Peds NSQIP databases provide large samples of surgical patients, with more cases reported in KID. The rates of complications studied were similar between the 2 data sets, with the exception of blood transfusion events where the retrospective chart review process of Peds NSQIP captured almost double the rate reported in KID.