Development and Validation of Revised Risk Adjustment for Congenital Heart Surgery Method

  • Dr Kathy Jenkins, Children's Hospital Boston, United States
  • Dr Jeffrey Jacobs, University of South Florida, United States
  • Dr Marshall Jacobs, Drexel University College of Medicine, United States
  • Dr Emile Bacha, Children's Hospital Boston, United States
  • Dr Bohdan Maruszewski, Children's Memorial Health Institute, Poland
  • Dr James Moller, University of Minnesota, United States
  • Dr Sean O'Brien, Duke University Medical Center, United States
  • Dr Kimberlee Gauvreau, Children's Hospital Boston, United States

    • Background: The Risk Adjustment for Congenital Heart Surgery (RACHS-1) method has been used extensively for comparative analyses of risk of death for children undergoing congenital heart surgery. Ongoing risk reduction, emergence of standardized nomenclature and new procedures necessitate reconsideration of the original methodology.

    Methods: A panel of experts used clinical judgement, and then examined mortality rates from multiple data sources (Society of Thoracic Surgeons, European Association for Cardiothoracic Surgery), Pediatric Cardiac Care Consortium, Pediatric Health Information System, and Kids Inpatient Database) to group procedures, using International Pediatric and Congenital Cardiac Codes, into candidate risk groups (1 to 5) with similar within-group risk of in-hospital death.

    Results: Using input from 30 pediatric cardiologists and cardiac surgeons, procedure-based risks of death for 180 procedures were considered. Procedural mortality rates were calculated using data from nearly 100,000 cases. Median mortality rates for preliminary risk categories are: 1 - 0.4%, 2 - 1.7%, 3 - 4.1%, 4 - 6.4%, 5 - 20.7%. Final risk category development and decisions about incorporation of original RACHS-1 clinical variables (age, prematurity, noncardiac anomaly) or other clinical factors will be determined, based on statistical significance, contribution to ROC area, and consistency of findings, at the next meeting of the panel in January 2009. The final model will be validated in subsequent databases.

    Conclusion: Similar to RACHS-1, this updated version of the Risk Adjustment for Congenital Heart Surgery method should be useful to adjust for current, baseline risk of death after congenital heart surgery, using various clinical data sources.