[chongli]

Do these studies control for people who quit running due to an injury? It doesn't appear that they do. They look only at "healthy long distance runners." Sorry, but that is not very compelling. It's like studying 90-year old chess grandmasters and concluding that chess prevents Alzheimer's.

You have to control for all the people who don't do an activity because they're too unhealthy to continue. And there is no obvious way to do that.

[pakitan]

It says they followed runners and non-runners for 18 years. Presumably, that means they monitored the same bunch of people over the 18 years and the effect you mention would have been noticed, if it existed.

[chongli]

The question is: did they only count someone as a runner if they were still a runner at the end of 18 years? Or if someone was running for 1 year, got injured and then quit running, did they get dropped from the study or even grouped with the non-runners?

[pakitan]

Well, I don't have the time to go through the whole thing but on a first glance:

1. Seems like they identify someone as runner or walker based on a questionnaire filled. After X years those people fill another questionnaire to indicate whether they had OA or hip replacement. I would think that at the time of the second questionnaire a runner will still be assumed a runner, even if he doesn't run anymore.

2. Exercise other than running increased rates of OA and hip replacement.

[chongli]

That's not a very good study then. They should be taking a group of people, none of whom are runners, and randomly assign them to runner or non-runner groups, and then follow them for 18 years. The activity levels of all participants should be tracked at all times by GPS devices with heart rate monitors. Their health should be monitored by professionals, though they should not be given any instruction in proper running techniques. At the end of the study they should give everyone a full examination to determine the outcomes for each group.

Of course, I'm being facetious. Such a study would never be done because it's completely unethical. But the point is that you can't determine an intervention (running) is safe and effective without a proper trial like that. A survey that asks people whether they're a runner or not, before even beginning, has a problem with selection bias.

[ToFundorNot]

I skimmed the report, they break down the % of people still running. They took x-rays and measured the space between the joints amongst other tests to see how much wear/damage occurred in the 18 year period. In both the control, and the runners group there were runners (17% vs 97.8%). By the end the runners group still had 55.6% running vs 3.8%.

Unfortunately they didn't break down the km's ran to classify the individuals. Also The runners group was almost more likely to do 'vigorous exercise'.

To answer your question, they did continue to classify as runners even if they stopped (and almost half did stop).

[pakitan]

I'd suggest you read the actual study. It's way more nuanced and detailed than the 2 sentence summary I provided. And no, I don't think it provides a 100% guarantee that running is safe but that's not what I'm expecting from a study. It's just one more data point in the ever growing body of evidence that active lifestyle is better than sedentary lifestyle.

The problem is not that you're being facetious, the problem is that if you insist on a single study proving something with a 100% certainty, we shouldn't be doing any science. For example, I can find numerous holes in your supposedly perfect study design. It turns out that active runners have less joint issues indeed? How can you be sure that it's due to running and not due to increased exposure to sunlight? Why aren't you testing different exercises? What if the benefits of running are solely due to improved cardiovascular performance but running specifically is actually bad on the joints bearing the heaviest load? If you dig deep enough, we can go on and on.

[chongli]

If you go back to the beginning of the discussion, the person I was replying to was citing these studies to claim that running is perfectly safe. My reply was intended to raise doubts about that claim.

The burden of proof is on running advocates to show that running is safe and to clearly lay out the conditions for safety. As a former runner (who quit after breaking an ankle and speaking to an orthopedic surgeon who saw runners all the time), my strong suspicion is that running is only safe for those who:

a) run with perfect form

b) have perfect-fitting shoes

c) eat a well-balanced athletic diet

d) are already slim and fit before they begin running

I feel like that last one is the real key. If you're massively overweight and someone tells you to just go for a run, you're going to get hurt. To be a runner you need to get in shape first and have someone help you buy proper shoes and teach you proper form. Anything other than that is bad advice.

[dognotdog]

Thanks for mentioning that, this is something I often wonder about, as this is at least often left out of the press. I don't know how/if the underlying research accounts for it, but I'd be very happy if such statistical complications were mentioned more often to show hard it is to get the science right.