It is easy and tempting to think about medical progress as a series of incremental advances. Many lay-people and even skilled practitioners think about medicine in this manner. Though incremental progress does occur, it isn't the only path in the story of medicine. In the wake of the recent U.S. Preventive Services Task Force (USPSTF) flip flop concerning aspirin therapy for primary prevention, we highlight three occurrences in the story of medicine: replacement, reversal, and reassessment.
Replacement
Replacement is stepwise advancement. This is how most of us imagine medicine moves forward. Supportive care is replaced by a good therapy, which is, in turn, replaced by another approach proven superior in robust clinical trials.
A few examples of replacement include therapies for peptic ulcer disease and Hodgkin's lymphoma, and the management of heart attacks. We once treated peptic ulcer disease with surgery, then with a series of progressively better acid suppressant medications, and now with proton pump inhibitors and antibiotics. The treatment of Hodgkin's lymphoma went from nothing to radiation to combination chemotherapy, and transformed the condition from an incurable one to a disease cured at high rates with modest toxicity. The management of acute myocardial infarction has progressed with a series of incremental advances that have saved hundreds of thousands of lives.
When history books are written, replacement is portrayed as the dominant mode of progress, in part, because of a human desire to tell a logical, linear story. But the true history of medicine is full of many missteps and false starts.
Reversal
Reversal is another story of medical progress; one which is less frequently discussed. Reversal occurs when a new medical practice is adopted without sufficient evidence and is later found to be no better, or even worse, than the previous standard of care. Examples include: the (activated; Xigris); , an arthroscopic surgery for degenerative meniscal tears; and the for babies at high risk of peanut allergies (a practice that actually increases the rate of peanut allergy).
Reversals occur when clinical practice jumps ahead of the evidence -- when recommendations are made or interventions adopted without knowing that they help. Reversals are avoidable. The acquisition of better evidence that a medical practice is effective prior to its promulgation would drastically reduce the frequency of reversal.
How do we end up with reversals? One reason is we hang our hat on a phase 2 trial that is underpowered or uses a surrogate endpoint. This can lead to false negatives, or in other cases, false positives and exaggerated effect sizes. At other times, we adopt a quality and safety initiative, validated in a single center, and deploy it widely without a multicenter trial. This led to the fate of tight glycemic control in the ICU. And sometimes we base our practice on a randomized trial, the gold standard, but a flawed one -- a trial with a poor control, a bad design, improper blinding, too generous non-inferiority margin, or bias.
But the most important reason for medical reversal, the one that underlies all the rest is: the adoption of a practice because of its bio-plausibility -- some mixture of pathophysiologic models or reasoning and retrospective observational studies. This dominates the landscape of reversal. Enthusiasm and hype lead to adoption, while well done randomized trials follow years later (if ever).
Reassessment
Reassessment is rarely documented in biomedicine, but it does happen. Reassessment occurs when the on-the-ground reality changes over time, when an intervention that once offered a benefit no longer offers one, or one that was once found to be ineffective now offers benefit. We that aspirin for primary prevention fits this category. Initial trials of aspirin occurred in people who were often smokers, had higher BMIs than those in modern day studies, and were seldom on statins. Studies from decades ago demonstrated that the benefits of aspirin therapy exceeded its harm. More recent trials, those which examined aspirin in today's world, demonstrate that it is no longer effective as preventive medication.
We believe there are other medical practices that were once favorable, but no longer are, due to evolving demographics, exposures, and risk factors. There may also be interventions, once found lacking, that may now be effective. This might especially be the case where surgical technology has advanced. Interest in testing previously adopted or dismissed therapies is low. That is why we medical evidence to come with an expiration date.
Avoiding Flip Flops In Medicine
There are several distinct patterns by which medicine progresses. Some flip flops are forgivable -- these reassessments are due to evolving populations and ancillary medical care -- but others, those due to medical reversal, occur more frequently than need be. A streamlined system for running robust clinical trials will prevent many missteps in medicine, sparing patients ineffective, costly, and even harmful treatments. For a deeper dive of these topics, we encourage you to read our book .
is a hematologist-oncologist and associate professor of medicine at the University of California San Francisco, and author of . is a general internist, and Professor of Medicine at the University of Chicago.