It has been said that a surgeon’s skill and judgment account for between 80% and 90% of a patient’s outcome. (I believe this is true for both surgical and nonsurgical treatments.) Throw in a physician’s ability to listen and clearly communicate with patients, and I am sure we are approaching that 90% mark. That means that when we conduct randomized trials comparing two types of knee prostheses or fracture-fixation constructs, we are, in essence, scrutinizing only about 10% of the patient-outcome equation.
So how do we best evaluate the 90% of the outcome equation that is physician-dependent? With the advent of “bundled” episodes of care, the orthopaedic community has emphasized the need for risk-adjustment in evaluating surgeon performance. Clearly, there are certain patients who are at higher risk for worse outcomes than others, such as those with diabetes, nicotine abuse, advanced age, and less social support.
In the December 19. 2018 issue of The Journal, Thigpen et al. report on patient outcomes 6 months after arthroscopic rotator cuff repair in 995 patients treated by 34 surgeons. The authors evaluated patient-reported outcomes from all surgeons using both unadjusted and adjusted ASES change scores. The adjusted scores took into account about a dozen baseline patient characteristics, including symptom severity, functional and mental scores, medical comorbidities, and Workers’ Compensation status. Relative to performance rankings based on unadjusted data, risk adjustment significantly altered the rankings for 91% of the surgeons. According to the authors, these findings “underpin the importance of risk-adjustment approaches to accurately report surgeon performance.”
But what is of even greater interest to me is that risk adjustment led to positive increases in patient outcomes for some surgeons, while decreasing outcomes for other surgeons. Some of these outcome differences likely reflect each surgeon’s patient-selection biases, but in the words of the authors, the numbers strongly suggest “that there is a meaningful, distinguishable difference in patient outcomes between surgeons.”
What should we do with this data? In my opinion, surgeons in the lower 80% of the list, at least, ought to be engaging with the surgeons who demonstrated the highest adjusted performance scores to understand what is helping them obtain outcomes that are superior to everyone else’s. We owe it to our patients to understand what our personal outcomes are for at least the most common conditions we treat. I believe it borders on unethical behavior to quote patients outcome data of a procedure from the peer-reviewed literature when we have no idea how our personal results compare. Orthopaedic surgeons need to be more active in lobbying our groups and health systems to support best practices for clinical outcome data collection and reporting so we can, in turn, improve our care by adopting the best practices of the surgeons with the best outcomes.
Marc Swiontkowski, MD