By Aaron Bright on January 15, 2020

Can we trust our own medical research?

Wine and chocolate are great and horrible

I’ve had a personal morbid fascination with the medical literature since medical school. I’ve lived long enough to see several things be portrayed in the professional and layperson literature as miraculously beneficial treatments only to be the polar opposite in the next study or after failing to be replicable long term. It’s nuts. I’m not a statistician but I’ve been trying to become more facile with statistics (I highly recommend this amazing little book) and especially the many kinds of bias that can influence the “conclusions” of studies.

It happens in all areas of medicine, although it is more likely to be in the New York Times when the topic is chocolate or wine or how working out for 32 seconds is the same as running 10 miles. But, if you live long enough you’ll see us go change our opinions on things we espoused as gospel over and over again. Why? So many things can steer us away from the truth that it almost makes it more likely that we are wrong. I will write more about this because it is too dense a topic to cover in one post but there is one really cool article from back in 2005 that I always cite in these kinds of discussions.

Great JAMA article

Published in 2005, this article by John Ioannidis in JAMA is so interesting. He studied all of the research from major journals that had been cited >1000 times in the literature from 1990 to 2003. These are high impact studies and practice changing findings. He then asked how often these studies’ conclusions were contradicted by later studies, had been shown to have much less effect than later studies,

 

Published by Aaron Bright January 15, 2020