Medical misinformation is a topic of great concern to a lot of the most powerful voices in medicine.
But what is medical misinformation?
Increasingly, I don’t really think that I know what that means because of the subjective nature of data interpretation.
Does medical misinformation only refer to disagreement with the consensus about Covid risk? Or masks? Or vaccines?
Or does medical misinformation extend to an over-extrapolation of preliminary data to suggest a higher level of confidence than the evidence deserves?
If we use the latter definition, then a lot of the people most concerned about medical misinformation need to look in the mirror.
The new study on “weekend warrior” exercise habits helps illustrate an important point
JAMA, the Journal of the American Medical Association, is one of the most important medical journals in the world.
So when it publishes a new study on exercise and heart health, you’d suspect that there’s some definitive or important new information there.
The study in question was picked up by a lot of media outlets and promoted quite a bit on social media.
Here’s Eric Topol suggesting that we have some brand new information from this study telling us that “Weekend Warriors” don’t need to be worried about their heart health:
You’ll see articles on this paper referenced all over the internet, basically all suggesting that this paper provides new and important information for us.
The author of the paper suggests that the work provides illuminating new information on cardiac risk based on exercise pattern:
This study did not show what the quotes and articles are telling you that it did
This study is actually pretty innovative. They looked at data on people who wore accelerometers for a week and then had their cardiovascular status monitored for several years.
They divided the participants into different groups:
People who did exercise clustered in a couple of days
People who did more regular physical activity spread throughout the week
People who weren’t active at all
You won’t be surprised what they found: people who do exercise are healthier than people who don’t do exercise, and it doesn’t matter if that exercise comes regularly throughout the week or clustered in a couple of days.
But there are some problems here.
First, they only measured activity levels for a week, and then assumed that reflected someone’s exercise pattern indefinitely.
Have you ever had a week where you exercised either more or less than you do regularly? I certainly have. Evaluating activity for a week is a really limited way of doing it.
Second, this isn’t an intervention trial. They didn’t randomized people to different types of exercise or inactivity. So any data you get is just correlation.
Third, and most importantly, the people who exercise are really different than the people who don’t.
Take a look:
The regular exercisers in this study have less diabetes, less of them are on high blood pressure medication, they smoke less, and they have a higher perception of their own health
What if I told you that I had 3 groups of people and that two of the groups was different than the third group in a few ways:
They had less diabetes
They had less high blood pressure
They smoked less
Their perception of their own health was better
Would you be surprised that those two groups had lower risk of heart problems than the third group?
That’s what this study showed: if you take a group of people who are healthier and follow them for a few years, they have less heart problems than a group of people who are less healthy.
We get no information about the impact of exercise here.
We just see that healthier people are healthier.
Drawing a causal conclusion here is completely incorrect
No scientist who is being truly honest about the usefulness of this data would tell you that we’ve answered any important questions with this study.
It’s certainly interesting - a novel way of measuring exercise with accelerometer data and an objective way of evaluating exercise patterns - but it’s not *actually* answering any questions about cause and effect.
I learned zero new information about whether exercising only on the weekends is better than getting more regular exercise throughout the week with this study.
I don’t really know why anyone would represent otherwise.
And while the comments on this study seem to me to be generally made in good faith, they’re still not accurate.
Does this classify as misinformation?
I think the answer to that is yes.
If people with a platform and credibility in medicine exaggerate the conclusions that you can draw from flawed observational research on exercise, it means that they can exaggerate the conclusions on lots of other topics too.
Take Covid boosters. A lot of the data that’s been used to justify their use is observational, just like this exercise study.
When you don’t actually study an intervention in a high quality way, your data is confounded and you can’t actually determine cause and effect.
With Covid boosters, there is real concern that the survival benefit seen in booster studies isn’t actually related to the booster itself, but rather the underlying health of the people who took one:
It’s not always easy to measure every single thing that might influence an outcome. Observational research - the vast majority of research that is done, as well as almost everything that you seem commented on or shared across social media - is inherently limited.
Often the data literally do not say what you’re being told that it says.
And so when it comes to the fight against medical misinformation, many of our most important voices should think harder about how they’re interpreting data.