Association Between Extracorporeal Membrane Oxygenation Center Volume and Mortality Among Children With Heart Disease: Propensity and Risk Modeling

Objectives: To evaluate the relationship between extracorporeal membrane oxygenation center volume and mortality in children undergoing heart operations using propensity score matching in a multiinstitutional cohort. Design: Post hoc analysis of data from an existing national database, Pediatric Health Information System. Propensity score matching was performed to 1-1-1 match patients in low-volume (0-30 cases per year), medium-volume (31-50 cases per year), and high-volume (>50 cases per year) categories. We tested the sensitivity of our findings by repeating the primary analyses using traditional statistical techniques (traditional regression-based methods and covariate adjustment using propensity score). Setting: Forty-two children's hospitals across the Unites States. Patients: Patients 18 years old or younger receiving extracorporeal membrane oxygenation before or after pediatric heart operation at a Pediatric Health Information System participating hospital (2004-2013) were included. Interventions: None. Measurements and Main Results: A total of 3,502 from 42 hospitals qualified for inclusion. Using propensity score matching, 1,962 patients were matched 1-1-1 to compare the three volume categories (654 patients in each category). Overall mortality was 1,493 patients (43%). Before matching and adjustment, low-and medium-volume centers were associated with higher mortality (low versus high volume: unadjusted odds ratio, 1.99; 95% CI, 1.68-2.36; p < 0.001). After matching, there was no significant association between center volume and mortality in unadjusted and adjusted analyses (low versus high volume: unadjusted odds ratio, 1.06; 95% CI, 0.85-1.32; p = 0.62 and adjusted odds ratio, 0.97; 95% CI, 0.63-1.50; p = 0.90). This relationship remained similar for analyses using traditional statistical techniques (regression adjustment, low versus high volume: adjusted odds ratio, 1.23; 95% CI, 0.80-1.89; p = 0.35 and covariate adjustment using propensity score, low versus high volume: adjusted odds ratio, 1.16; 95% CI, 0.77-1.74; p = 0.49). Conclusions: We demonstrated no relationship between extracorporeal membrane oxygenation center volume and mortality. Further analyses are needed to evaluate this relationship.