Pre-surgical judgement and decision making in paediatric cardiac surgery

This thesis investigates pre-surgical risk assessment and the choice between surgical procedures for children with congenital heart disease. The psychology of judgement and decision making provided the theoretical and methodological starting point for this investigation. A series of studies examined the estimation of the likelihood of early mortality following a modified Fontan operation for repair of univentricular heart. The participants were cardiologists and surgeons from a single UK centre. Doctors differed widely in their estimates, but agreed on a common core of factors that influenced whether a patient was more, or less, likely to survive. This agreed core of factors was consistent with the findings of published retrospective risk factor analyses. Doctors and statistical models taken from the published literature failed to discriminate appropriately between patients more, or less, likely to die. Models derived from analysis of the operations performed at the institution where patients underwent surgery identified higher and lower risk patient groups on the basis of only one or two predictive factors. Models taken from the published literature were unrealistically optimistic, whereas other methods were unbiased. These results suggest that prediction using only the best single predictor of an outcome can be as accurate as more complex methods. Two studies investigated the choice between different surgical strategies for the management of univentricular heart. Results indicate that the preference for procedures is more closely related to the perceived likelihood of good heart function than the likelihood of survival, and more closely related to long-term than shorter-term outcomes. It appears that doctors seek to maximise the likelihood of the best outcome in preference to minimising the likelihood of the worst outcome. The subjective likelihood judgements most closely related to option preference, are those for which there is greatest variation among doctors