Incorporating Comorbidity Within Risk Adjustment for UK Pediatric Cardiac Surgery

Background When considering early survival rates after pediatric cardiac surgery it is essential to adjust for risk linked to case complexity. An important but previously less well understood component of case mix complexity is comorbidity. Methods The National Congenital Heart Disease Audit data representing all pediatric cardiac surgery procedures undertaken in the United Kingdom and Ireland between 2009 and 2014 was used to develop and test groupings for comorbidity and additional non–procedure-based risk factors within a risk adjustment model for 30-day mortality. A mixture of expert consensus based opinion and empiric statistical analyses were used to define and test the new comorbidity groups. Results The study dataset consisted of 21,838 pediatric cardiac surgical procedure episodes in 18,834 patients with 539 deaths (raw 30-day mortality rate, 2.5%). In addition to surgical procedure type, primary cardiac diagnosis, univentricular status, age, weight, procedure type (bypass, nonbypass, or hybrid), and era, the new risk factor groups of non-Down congenital anomalies, acquired comorbidities, increased severity of illness indicators (eg, preoperative mechanical ventilation or circulatory support) and additional cardiac risk factors (eg, heart muscle conditions and raised pulmonary arterial pressure) all independently increased the risk of operative mortality. Conclusions In an era of low mortality rates across a wide range of operations, non–procedure-based risk factors form a vital element of risk adjustment and their presence leads to wide variations in the predicted risk of a given operation.