Covid-19 and Evidence-Based Medicine

I recently read an article in New York Time Magazine about the controversy and acrimony surrounding the use of convalescent plasma in the treatment of Covid-19 patients. This brought to mind a conversation I had last year with a couple of economists during an overseas sojourn my wife and I took. One of the economists told me she was puzzled by all the current buzz about “evidence-based medicine.” She wondered what doctors have been using all this time instead of evidence to decide what procedures to perform and medicines to prescribe.

My response to the economist was that doctors have been using evidence all along, but now an increasing range of medical recommendations and decisions are guided by the results of the best evidence, placebo-controlled clinical trials, rather than by tradition, anecdote, and plausibility. The New York Time Magazine article focused on the relative merits of these two approaches, but missed some important considerations.

In placebo-controlled clinical trials, participants are typically randomly assigned to two groups. Members of one group get medicine and members of the other get a placebo. Neither the participants nor the physicians know who is getting which. Outcomes among members of the two groups are compared to see if the medication is safe and effective.

This procedure produces the best evidence, but some patients don’t want to risk getting a placebo. They want, instead, to use therapies that have not proven safe and effective for their condition, but that have been FDA approved as safe and effective for other conditions, and that would plausibly be helpful to them. This is called “off-label use.” An example is the use of beta blockers, FDA approved to treat certain heart conditions, to treat migraines.

Generally speaking, off-label use has some real benefits. It allows physicians more flexibility to tailor treatments to their patients’ particular needs, as managing human health isn’t one-size-fits all. People differ from one another in so many ways that there are more variables than can be tested scientifically one-by-one in randomized trials.

Most of the discussion about using of convalescent plasma in the treatment of Covid-19 has centered on the conflict between patients and their doctors eager to use a promising therapy, on the one hand, and medical researchers wanting clinical trials first, to know if that therapy really works, on the other.

Researchers point out that for decades people thought that egg yolks increased our bad cholesterol (LDL) and should therefore be avoided by people with heart disease. Recent studies have shown, however, that eating egg yolks doesn’t appreciably increase cholesterol in our blood. Even better in my view, a recent study showed that flossing teeth doesn’t improve oral health. Hydroxychloroquine was for a while all the rage for treating Covid-19. But when the treatment was subject to strict clinical trials, it was shown to be neither safe nor effective.

A second reason against using hydroxychloroquine to treat Covid-19 before clinical trials were completed, a feature not sufficiently covered in the debate, is that the drug is approved for and needed by patients with some auto-immune diseases. Its speculative, off-label use to treat Covid-19, combined with limited supply and herd anxiety during a pandemic, threatened to deny these patients a drug they need to maintain their health. This side effect of off-label use doesn’t appear in the use of beta blockers to treat migraines.

A third consideration is the impact of off-label use on the conduct of clinical trials, which can’t take place unless two conditions are met. Besides a sufficient supply of the drug to be studied, there must be enough people willing to take part. The successful completion of clinical trials of new migraine drugs, such as Ubrelvy and Lasmiditan, shows that off-label use of beta blockers didn’t impair clinical trials of new drugs to treat migraines. The use of hydroxychloroquine to treat Covid-19, by contrast, not only threatened the supply for those who use it on-label, but threatened to deprive researchers of needed participants. Herd optimism discouraged people from taking a chance on receiving a placebo.

This is why using convalescent plasma to treat Covid-19 before clinical trials are completed is so controversial. First, it may not be both safe and effective. Second, the supply of convalescent plasma is so limited that there may not be enough to run clinical trials if herd enthusiasm induces physicians to prescribe it often for Covid-19 patients. Finally, optimism may discourage patients from enrolling in the kinds of studies that alone can produce secure evidence about safety and efficacy, because those patients assume without the best evidence that convalescent plasma will be helpful.

Both sides are right. There is a place for off-label use of pharmaceuticals, but not during a pandemic when supplies are low and people need the best evidence-based medicine.