Metrics to assess extracorporeal membrane oxygenation utilization in pediatric cardiac surgery programs

Objectives: Only a small fraction of pediatric cardiac surgical patients are supported with extracorporeal membrane oxygenation following cardiac surgery, but extracorporeal membrane oxygenation use is more common among those undergoing higher complexity surgery. We evaluated extracorporeal membrane oxygenation metrics indexed to annual cardiac surgical volume to better understand extracorporeal membrane oxygenation use among U.S. cardiac surgical programs. Design: Retrospective analysis Setting: Forty-three U.S. Children's Hospitals in the Pediatric Health Information System that performed cardiac surgery and used extracorporeal membrane oxygenation. Patients: All patients (< 19 yr) undergoing cardiac surgery during January 2003 to July 2014. Interventions: None. Measurements and Main Results: Both extracorporeal membrane oxygenation use and surgical mortality were risk adjusted using Risk Adjustment for Congenital Heart Surgery 1. Extracorporeal membrane oxygenation metrics indexed to annual cardiac surgery cases were calculated for each hospital and the metric values divided into quintiles for comparison across hospitals. Among 131,786 cardiac surgical patients, 3,782 (2.9%) received extracorporeal membrane oxygenation. Median case mix adjusted rate of extracorporeal membrane oxygenation use was 2.8% (interquartile range, 1.6-3.4%). Median pediatric cardiac case mix adjusted surgical mortality was 3.5%. Extracorporeal membrane oxygenation-associated surgical mortality was 1.3% (interquartile range, 0.7-1.6%); without extracorporeal membrane oxygenation, median case mix adjusted surgical mortality would increase from 3.5% to 5.0%. Among patients who died, 36.7% (median) were supported with extracorporeal membrane oxygenation. The median reduction in case mix adjusted surgical mortality from extracorporeal membrane oxygenation surgical survival was 30.1%. The median extracorporeal membrane oxygenation free surgical survival was 95% (interquartile range, 94-96%). Centers with less than 150 annual surgical cases had significantly lower median extracorporeal membrane oxygenation use (0.78%) than centers with greater than 275 cases (≥ 2.8% extracorporeal membrane oxygenation use). Extracorporeal membrane oxygenation use and mortality varied within quintiles and across quintiles of center annual surgical case volume. Conclusions: Risk adjusted extracorporeal membrane oxygenation metrics indexed to annual surgical volume provide potential for benchmarking as well as a greater understanding of extracorporeal membrane oxygenation utilization, efficacy, and impact on cardiac surgery mortality.