On Patient Safety: Differential Standards for Medical Evidence Risks Patient Safety

To paraphrase Upton Sinclair, it is difficult to get a man to understand something when his situation depends upon his not understanding it [10]. That is the position that doctors are in today with the practice of evidence-based medicine. All physicians can cherry-pick medical evidence to find data that supports our viewpoint, while ignoring or pointing out the flaws of studies that undermine our positions. As surgeons, we typically agree with medical research that validates the benefit of those procedures we perform while looking askance at data that calls the value of our surgical practices into question. This sort of differential approach to reading leads to important patient-safety concerns.

I was reminded of this when reading a recent review [8] of a randomized clinical trial [7] led by physical therapists that compared arthroscopic surgery to physical therapy for femoroacetabular impingement. The study found no difference in Hip Outcome Score between those patients treated with surgery and those treated by therapy alone. I expect that this result likely pleased the therapy community. However, a review by orthopaedic surgeons [4] noted the high rate of patient crossover, among other limitations, to question the trial’s results. Contrast this response in which reviewers critically analyzed researchers’ methods and results to the quiet affirmation of a recent study’s conclusion that overlapping surgery caused no increased risks for surgical patients [2]. In my opinion, this study result has been largely accepted among the orthopaedic community, as I could not find any serious discussion of the study’s limitations, despite the results of another large study that identified this practice as potentially risky [9].

Like all people, our decision making is strongly affected by our natural biases. This phenomenon is well documented in Michael Lewis’s book, The Undoing Project, which details the powerful effect that our human biases have on our ability to make good decisions [6]. He describes how all people have a tendency to trust our gut instincts or anecdotal evidence in order to believe what we hope is true—even when we know we are being led astray. Several of his examples of poor decision-making involve physicians; these include clinical misjudgments in trauma care, radiologists’ diagnoses of stomach cancer, and psychologists’ evaluations of suicide risk. He shows how a great deal of research, some of it by Nobel Prize winners, demonstrates that we are not nearly as rational as we imagine ourselves to be, and that we are particularly poor with making decisions involving risk—the exact situation we find ourselves in when we discuss surgical options with patients.

In agreement with the evidence from Mr. Lewis’ book, studies have shown that clinicians typically underestimate the risk involved in procedures that they are contemplating, while overestimating the potential benefits of those treatments [5]. We must realize that even seemingly low-risk procedures are quite unsafe if there is evidence to suggest that they will not help the patient in question; there is no upside to the procedure to justify even slight risk. For example, we know that a percentage of those undergoing vertebroplasty will develop a complication, and yet, this procedure continues to be commonly performed despite a good deal of high-level evidence [1, 3] showing its ineffectiveness. When we perform ineffective procedures, every complication that results from them is a complication that should never have occurred.

This brings us back to Upton Sinclair and how we evaluate medical evidence. Mr. Sinclair was both a Pulitzer Prize-winning novelist and journalist, and he looked at the world with a discriminating eye. Orthopaedic surgeons should take the same approach to reading medical research. In light of our natural biases, we should be most critical of those results we are happiest to see.