Abstract 13473: Refinement of a Risk-Standardization Model for Adverse Outcomes Following Congenital Cardiac Catheterization: Insights From from the NCDR ® IMPACT TM Registry

Abstract

Listen

Translate article

Background: Risk-standardization for adverse events following congenital cardiac catheterization is needed to equitably compare patient outcomes among different hospitals as a foundation for quality improvement. We sought to refine and update the risk-standardization methodology for the NCDR ® IMPACT TM (Improving Pediatric and Adult Congenital Treatment) Registry. Methods: We identified 39,725 consecutive patients within IMPACT undergoing cardiac catheterization between January 2011 and March 2014. New procedure type risk categories were derived by empiric data and expert opinion and markers of hemodynamic vulnerability were empirically explored and identified. A multivariable hierarchical logistic regression model to identify patient and procedural characteristics predictive of a major adverse event (MAE) or death following cardiac catheterization was derived in 70% of the cohort and validated in the remaining 30%. Results: The rate of MAE or death was 7.1% and 7.2% in the derivation and validation cohorts, respectively. Six independent indicators of hemodynamic vulnerability and six procedure-type risk groups were identified. The final risk adjustment model included procedure type risk group, hemodynamic vulnerability, renal insufficiency, single-ventricle physiology, and coagulation disorder (See Figure). The model had good discrimination with a C-statistic of 0.76 and 0.75 in the derivation and validation cohorts, respectively. Model calibration was excellent with a slope of 0.97 (standard error [SE] 0.04; p-value [for difference from 1]= 0.53) and an intercept of 0.007 (SE 0.12; p-value [for difference from 0]= 0.95). Conclusion: We updated and refined a risk-standardization model for outcomes following congenital cardiac catheterization using the large, multicenter IMPACT Registry. Our model can be used for quality reporting back to sites participating in IMPACT to guide quality improvement initiatives.