Account for severity and risk factors when reviewing peers

When physicians are given feedback data regarding outcomes, a common response is “My outcomes are worse because my patients are sicker.” This concern is legitimate and should be addressed by using data adjusted for the severity of the patients’ illnesses or other risk factors such as smoking and obesity.

There are two primary ways to account for severity and risk factors. The first is to use electronic patient information data sets that have already been collected for other purposes. These databases use statistical methods to adjust the outcomes for factors that were captured electronically, such as age, diagnoses, procedure codes, and admission sources. Although easy to obtain and less expensive, this information is limited to the factors that were captured electronically and often does not include relevant clinical information. This data still depends on the accuracy of medical records coding and might not, for example, attribute the outcome to the correct physician.

The other way is to decide which clinical factors (e.g., blood pressure, prior episodes of heart failure) might predict outcomes based on clinical studies or expert opinion. This information is then entered into an electronic database and submitted to a central database for statistical risk adjustment. The Society of Thoracic Surgeons’ database for cardiovascular surgery is an example of this approach. This method produces more accurate adjustments and better physician attribution but is more labor intensive.