It’s always interesting to read the medical research studies that come by my daily feed. Some are groundbreaking, most are so-so, and a few are downright silly. This morning we’ll look at a silly study about shoulder imaging that seems to have started with a conclusion and worked backward.
My Shoulder Hurts, so I Need an MRI
MRIs are a two-edged sword. On the one hand, they have revolutionized what we can see inside the body. On the other hand, they have likely launched millions of unnecessary surgeries. One of the more startling research studies of the last few years that threw our understanding of what’s important on its head involved rotator cuff tears. For example, instead of showing a direct correlation between patients who had their tears operated on, an improvement on the MRIs, and reports of better function, it demonstrated no correlations All that was important was the level of an inflammatory chemical inside the shoulder joint.
The Bizarrely Dumb Study
The research was conducted at an urban teaching hospital on all shoulder MRIs ordered. The authors then created an ordered list of the percentage of patients who then went on to get surgery by the department that ordered the study. Since the percentage of patients who got shoulder surgery was highest when an orthopedic surgeon ordered the study, the authors concluded, “Detailed analysis suggests that providers who do not have specific training in shoulder pathology should consider consultation with an orthopedic surgeon before ordering shoulder MRI for patients who may need additional imaging after radiography.” Huh?
Argumentum ex silentio
While the conclusion of the above study likely has a few logical fallacies that apply (circulus in demonstrando and petitio principii to name a pair), an argument from silence (argumentum ex silentio) where the conclusion is based on the absence of evidence rather than the existence of evidence is one. Basically, the authors assume that orthopedic surgeons are infallible in their judgment to operate on a shoulder based on MRI findings. Which, you know if you read this blog, is a nutty assumption. What’s thrown out the window is the opposite possibility that surgeons want to operate; hence, they would have a higher number of surgeries resulting from fewer images than say family doctors. In fact, in this study, the family doctors did have a lower number patients going to surgery, which of course could mean that the patients who never had surgery never needed it in the first place.
For example, if we look at the research above on rotator cuff healing, we’re not even sure that a rotator cuff tear seen on MRI that would lead to a surgery is even all that important for why the patient hurts or recovers. As another example, in this recent study, more than half of asymptomatic men had shoulder labral tears, suggesting that some of these MRI findings leading to surgery do little to explain why a patient hurts.
The upshot? Never put the fox in charge of the henhouse is a truism from my mother’s time growing up in rural Pennsylvania. One that the authors of this study clearly forgot that along with a number of logical fallacies!