“But My Patients Are Different!”: Risk Adjustment in 2012 and Beyond

Feedback of surgical site infection (SSI) rates to surgeons improves patient outcomes and should be considered a cornerstone of any infection control program. For as long as feedback of SSI data has occurred, those in infection control have often heard a searing retort from indignant surgeons: “But my patients are different!”Fortunately, epidemiologists have several tools to use in response. One of the most commonly used approaches involves risk adjustment for differences in case mix between the group of interest (eg, a surgeon's patients) and a comparator. In other words, risk adjustment levels the playing field.Formal risk adjustment for rates of SSI has existed for almost 50 years but is still an imperfect science. In fact, risk adjustment for different variables can lead to different conclusions. Over the past 2 decades, the National Healthcare Safety Network (NHSN) risk index has been used by many hospitals to perform risk adjustment for rates of SSI. The NHSN risk index is simple and effective but has undergone considerable scrutiny. Numerous investigators have described scenarios and/or procedures for which the risk index performed poorly and have offered suggestions for improvement. Indeed, Robert Gaynes summarized some of the shortcomings of the NHSN risk index in 2 editorials 10 years ago, stating, “A composite risk index that captures the joint influence of [intrinsic patient risk] and other risk factors is required before meaningful comparisons of SSI rates can be made by surgeons, among institutions, or across time.”