Bridgewater B et al. (2004) Improving mortality of coronary surgery over first four years of independent practice: retrospective examination of prospectively collected data from 15 surgeons. BMJ 329: 421–425

In coronary artery surgery, mortality has been associated with the number of operations that surgeons have performed. This observation prompted Ben Bridgewater and colleagues to examine the 'learning curve' effect, by comparing mortality in patients operated on by newly appointed and established surgeons.

Data were collected on 18,913 patients in northwest England undergoing isolated coronary artery surgery for the first time. In 5,678 cases, the surgery was carried out by surgeons who had been appointed as consultants within the previous 4 years. The remainder of the operations were performed by more experienced colleagues. Observed mortality and predicted (EuroSCORE) mortality were recorded.

There was no significant difference in observed mortality for patients treated by newly appointed or experienced surgeons (1.9% and 2.0% mortality, respectively; P = 0.71). Mortality in patients treated during the first year of a surgeon's appointment, however, was significantly higher than in the fourth year (2.2% and 1.2%, respectively; P = 0.049). Adjusting for time and case mix, mortality decreased from 2.3% to 1.0% from the first to the fourth year (P = 0.019).

While a learning curve effect has been clearly demonstrated, the overall mortality in patients treated by newly appointed or experienced surgeons was similar. The authors stress the importance of managing the learning curve effect by supporting, monitoring and further training newly appointed surgeons, and offering experienced clinical support in nontechnical areas such as case selection and perioperative management.