Glucosamine Chondroitin and the Statistics Thereof, from Charles Pennington

April 15, 2015 |

I'll just throw this out.

Intuitively, I suspect that if a fraction X gets better on a placebo, and if a fraction Y (which could overlap with X) gets real physiological benefit (as determined the by the deities), then the fraction that will REPORT being better would be something like sqrt(X^2 + Y^2). (The "reasoning" is that the real effect and the placebo effect are probably uncorrelated and therefore "add" in an orthogonal way, like the Pythagorean theorem.)

So if X is 0.6 and Y is 0.4 then 72% of people in the study would say they were better.

Of course this won't be valid if X^2+Y^2 gets close to or exceeds one.

Anyway, if that formula is right, and if 40% of people really do benefit as determined by the deities, then we'd see 72% reporting that they're better, which is not much more than the percent that "respond" to the placebo, 60%. So it's probably hard to smoke out an effect, even if it's kind of big.

anonymous writes:

Before any marathon or ultra, you hang around in the corral of runners waiting to go, (towards the back. towards the WaaaaaAAAaay back, with the jockeys, fat ladies, kids dribbling basketballs) and ask practically ANY old guy if they take it, they will tell you affirmatively. I've done that at least dozens of times. Then look around at who has had a knee replacement and is in that category. No one.

Now that does not mean that the prevalence of old guys running marathons now (whereas two or three decades ago you didn't see that, may be a function of fad, but I remember old guys who ran two or three decades ago stopped running– almost all of them because "their knees couldn't take it anymore," or they "wore out their knees.") is a result of G&C consumption, or the fact that there are so many more older people running now, the fad effect.

There is a tendency to mock anecdotal evidence such as this– but our entire lives are spent accumulating anecdotal evidence and attempting to draw conclusions, from what we consume, what the "best" route to get to a certain destination is, what time we ought to wake up, to how we trade, etc. Everything we do in life is an attempt to solve an optimization problem based most often on a statistically insignificant number of data points.

David Lillienfeld writes:

First, I'm a physician and among my areas of expertise is the evaluation of drugs (pharm, not abuse). If you want to use anecdote, then you must have little use for regression to the mean. Anecdotes are subject to publication bias, small numbers, inadequate control of bias, among others. It is human nature to work off of anecdotes. It is also misleading.

Based on anecdote, radical mastectomy would still be the standard of care for breast cancer. Based on anecdote, rehab after a heart attack would consist of sitting on one's butt for six months "for healing." Based on anecdote, there are any number of medications one might use for treating pulmonary fibrosis. They actually don't do much. None of them. Based on anecdote, laetrile would be the nectar for cancer. Guess what—it isn't. So if you want to run on anecdote, go right ahead. But don't be surprised if your results are random, because that's what's happened in medicine based on anecdote. It's the reason why evidence-based medicine has emerged from the shadows. And don't forget that regression to the mean. Relying on anecdote goes right up there with physician self-treatment of disease. BTW, my uncle treated himself for a heart attack. Wrote the orders for morphine (it was 1960). Managed to kill himself with an overdose. In the hospital.

Second, vitamin C has been looked at for any number of diseases. For the common cold, there's lots of hedging by the Cochrane Collaboration, but I'd hardly call it something where they see compelling evidence—at least for the common cold. Linus Pauling may have thought he was onto something. He was brilliant, some would say he was a genius. That doesn't give him a pass on evidence. Ronald A. Fisher believed cigarette smoking wasn't—couldn't be—a cause of lung cancer, and he was mystified by the increasing mortality rates from it. The same was true for Jacob Yerushalmy. There's a fellow in San Francisco, generally acknowledged as brilliant (he may even have a Nobel) who maintained that HIV wasn't the cause of AIDS. Genius isn't immunity from being wrong. Conjectures in science, even from geniuses, need evidence to be considered worthy of incorporation into the corpus of scientific knowledge.

I had two good friends, Bill Cochran (he of Cochran's Theorem and Abel Wolman talking at a symposium on the history of epidemiology. Cochran observed that "Evidence is a bitch." Wolman replied, "At least evidence is visible. It's the non-visible things that will get you every time." Wolman made his reputation in sanitary engineering (as it was then known) on figuring out how to get sufficient chlorine into tap water as to kill the cell present in it while maintaining that water's potability. Threats that weren't visible was his stock in trade, so to speak. But these were philosophies of science, not specific research questions.

Third, the pharmacokinetics of vit C do not suggest that more is better, ie, always gives a higher serum concentration.

Sorry about the length of this message, but it's worth noting that saying, "Guessing is a capital crime, and if you engage in it, you will lose your capital and become a criminal." I wish I could remember who said it. Can't though.

Ralph Vince writes:

I don't disagree with you (more specifically, I'm not qualified to disagree with you on this even if I were inclined to), however, as infants we learn to speak, and before that even, in our earliest life hours, we learn to learn by optimization based solely on the sparse data set of anecdotal evidence.

It's a platform that has certainly served us well, should not be disparaged, but rather ought to be acknowledged as perhaps not always best when other determination making platforms are available.

Properly done full squats are excellent for strengthening knees (assuming no preexisting damage, only weakness). One of the surprising things I've found since starting powerlifting 4 years ago, is that a lot of 'knee pain' can be corrected through better mobility (ie, stretching). New power lifters of all ages typically have to work on hip and ankle mobility before they can successfully squat. Once you have the mobility issues corrected, building strength is a matter of patience and diligence.

My wife, a RPh, thinks it MAY help, because it does seem to increase the lubricant on the joints.

However, firstly, this effect takes 2-3 months of use to develop this effect, The placebo effect is much more immediate. And most users think it works much quicker than the measurable effect to the body.

Secondly, it may simply be self selection, since as Jim and others suggest. Those willing to stick to taking 3-5 large pills a day are usually the ones willing to exercise. Diet also effect it.

Thirdly, many drugs help cause the desired response to the body, but create other problems to produce that effect. For example lowering cholesterol, but also side effect of lower calcium/electrolyte for the heart. (this is why I avoid supplements in general)

Fourth, it is not a "cure" but a MAY prevention future flare-ups, it MAY mask the symptoms. And people with arthritis have various rate of deterioration. Hence, needing a large group to determine if it helps.

With this said, many doctors and pharmacists do recommend using it.

`SELECT * FROM wp_comments WHERE comment_post_ID = '10207' AND comment_approved = '1' ORDER BY comment_date`