80 Percent Of All Medical Studies Are Lies

” 'Doctors?' said Ron startled. 'Those Muggle nutters that cut people up?' ” —J. K. Rowling, Harry Potter and the Order of the Phoenix

In A Nutshell

Somewhere around 80 percent of all medical research—the studies that determine what’s good for us and what new drugs are awesome—are flawed or outright lies made up by the people conducting them. The reason for this is a simple, sometimes-innocent conflict of interest (though it’s usually not innocent). Drug companies desire certain results from a new product for obvious, financial reasons. Researchers desire funding and often only get it if they can deliver on an outcome. And sometimes, just wanting something to happen makes a scientist more likely to see things that aren’t there.

The Whole Bushel

The problem of flawed research in the field of medicine begins with researchers who have a strong desire to have their work published in respected journals. The rejection rate for papers is high, often above 90 percent, and the ones that are accepted tend to have (as stupid as this might sound) more exciting titles that promise something new. However, coming up with surprising and exciting results is really difficult. There’s a tendency, according to Dr. John Ioannidis, the leading expert on the credibility of medical studies, for researchers to basically arrange their studies, knowingly and unknowingly sometimes, so that they get the results they anticipate. Otherwise, they don’t get published.

Even worse, the media picks up on these studies and often reports them as fact, even when, for example, there are already other studies that contradict the new and interesting one. This is the origin of shifting health claims: One day, something is good for you; the next day, it’s bad. Beer drinking, for example, may make you smarter or more stupid, depending on which study you go with.

Of course, one of these claim may in fact be true, but telling the difference is difficult, even for doctors. Ioannidis’s first study on the credibility of medical research discovered that “80 percent of non-randomized studies (by far the most common type) turn out to be wrong, as do 25 percent of supposedly gold-standard randomized trials, and as much as 10 percent of the platinum-standard large randomized trials.” So an easy fix would be to mostly rely on information gathered from studies requiring a higher standard, right?

Article Continued Below

Too bad Ioannidis’s second study showed that a lot of the most widely accepted truths of medicine, things that came from the gold and platinum standard, were also incorrect. Out of 34 widely accepted research studies, 14 were wrong or the results exaggerated dramatically. These aren’t quirky studies about beer drinking either, but actual treatments being used by patients right now.

How this is allowed to happen is well understood. The 34 studies mentioned above came from a larger body of 45 accepted pieces of research. Ioannidis could only really examine 34 because the remaining 11 had never been re-tested or looked at again, which is odd considering that actual treatment follows some of these studies. But even after further studies found problems, the findings in the original study stuck around in the medical community, still accepted as the truth, sometimes for decades.

So what should we do when it comes to dealing with all the people telling us how to improve our health with vitamins and fish oil and beer? Ioannidis says “Ignore them all” for the most part, because a lot of the research deals with marginal gains in health, none of which are based on research reliable enough to take seriously.

[NOTE: Paul K Pickett wants you to know that he is not a doctor. Doctors are doctors. And while a lot of the research in medicine has been called into question, your doctor is still the best person to give you advice concerning your health.]

Show Me The Proof

Lies, Damned Lies, and Medical Science
Drinking three pints of beer a week permanently dulls your brain 20%, study shows
STUDY: Drinking Beer Makes You Smarter