People fortunate enough to raise their families in Lake Wobegon have the reassurance that their children will always be above average. If you have diabetes and get admitted to my hospital, the admitting resident invariably will do well at locating the dorsalis pedis and posterior tibial pulses, as at the time of admission they are always 2+. If you are an elderly widow with the misfortune to end up on a ventilator in our ICU, you will at least avoid the ignominy of having suffered from dyspareunia prior to the illness: the electronic record specifically says "not present" in every review of symptoms.
Richard M. Plotzker, MD
Ah, we have entered the age of fake news and alternate facts—or, perhaps, as Professor Tom Nichols entitled his wonderful recent book, The Death of Expertise.
From the time of Hippocrates, medicine has valued some form of evidence, even if anecdotal, to drive medical advice. In the 12th century, Maimonides wrote about protecting one's body with a combination of prudent diet, physical activity, avoiding hazardous situations, and paying attention to sanitation. Without any microscopes or randomized controlled clinical trials at the time, he relied on either observation or belief, although some might say that he simply made up that advice and happened to get it right.
The sages of infection control got it wrong in proposing that malaria was related to bad air, but once the pathogenesis of malaria and other epidemics was elucidated, the medical and public health communities were pretty quick to abandon their erroneous practices.
The history of medicine is replete with those adaptations to evidence. While prefrontal lobotomy was probably more effective at treating rage behavior than is an anger management class, sometimes "good enough" or even a little inferiority is the best path forward. Depression is no longer treated with insulin infusions, with even the most ardent adherents acknowledging that there were safer and more efficacious ways to help those people.
In my own professional lifetime, there was divided opinion on whether to treat diabetic ketoacidosis (DKA) with 25 units or 5 units an hour of regular insulin. Physicians were pretty open to study results that led to the low-dose infusion becoming the norm, although historical vignettes of the high dose are still useful for quizzing residents about what happens if you inject 25 units of insulin into the gluteus of somebody with DKA. No, they do not get hypoglycemic and seize, as DKA is a form of insulin resistance. Abandoned treatments have a way of creating their own misconceptions.
So, while doctors adopt real evidence pretty well, not all evidence is incontrovertible, leaving us with a certain amount of belief and tradition to act upon when there are choices to be made: Which parathyroid adenomas need to go to the surgeon and which can be left alone? When is methimazole better and when is radioiodine better? Studies on communal preferences have shown that they vary widely from one location to another, even though the same body of evidence guides the communal standards.
In what sequence does an experienced clinician select and modify common hypoglycemic agents? We have a consensus on starting with metformin for most people, but beyond that, decisions are based on beliefs about what happens to islets when squeezed to hormonal oblivion with glipizide, whether insulin will make patients even more obese, will they remember to inject a weekly GLP-1 receptor agonist, and can I really get the A1c to go from 12% to 7.5% with a DPP4?
Some physicians will never prescribe a thiazolidinedione (TZD) because of weight gain, and others use reports of bladder carcinoma as their confirmation bias in the validity of their judgment, only to have the next consultant do exactly what they avoided and reduce the patient's A1c enough to enable the patient to discontinue U500 insulin.
The New England Journal of Medicine (NEJM) juxtaposed two articles in a recent issue that highlighted the challenges of treating low-grade thyroid disease. The first was a rather elaborate study of 700-plus individuals with slightly high thyroid-stimulating hormone (TSH) values who were randomly assigned to receive low-dose levothyroxine with periodic adjustment or placebo with sham adjustment.[1]
This rather common condition engenders a wide variety of treatment approaches, from "treat everyone" to "repeat the TSH in 3 months" to "treat if symptomatic" even though symptoms are highly nonspecific, or "treat to a normal TSH" to keep everyone from pointing a finger at a trivial lab variant which becomes a distraction to figuring out how to properly improve the patient's complaints.
Unfortunately, nonspecific symptoms may be the forerunner of contemporary alternate facts: They must be related to the thyroid if you believe they are. But you might find that you can correct the TSH without changing the symptoms; this may be a bit of confirmation bias for me, as I like to remove the distraction of the lab abnormality knowing that the symptoms will remain and can then be separated from the abnormality.
I'm not sure that I would reason it the same way if removing the distraction required an intervention with a more hazardous or expensive outcome. Moreover, the same study could be used to support the confirmation bias of those who usually opt not to offer levothyroxine for trivial TSH elevations.
In the same issue of NEJM, though, appears a more comprehensive summary of subclinical hypothyroidism presented as a common clinical management situation.[2]
As the author retraces the literature about whether adding levothyroxine improves symptoms or offers some protection from cardiovascular disease, the conclusions of who to treat and who to reassure remain open-ended. Our decisions are partly driven by how we were imprinted during training and by how thoroughly we explore a patient's history or risks. The latter alone is highly variable among practitioners as histories move from direct conversations with the patient to "copying and pasting" the last five histories that pop up on the screen.
When records have misinformation and studies are inconclusive, it becomes difficult for us as physicians to remain objective. So we rely on meta-analyses that reinforce our own preferences without really addressing the dilemmas of medical judgment in a more conclusive way.
Medscape Diabetes © 2017 WebMD, LLC
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this: The Data Show What I Expected… - Medscape - Oct 16, 2017.
Comments