From Health by Peter Gott, M.D., appearing in the Rock Hill Herald of July 20, 1997:

Dear Dr. Gott:

I have been eating gin-soaked raisins for four years and have no arthritis pain. I believe that this works, and I object to your saying that it doesn’t.

Dear Reader:

Enough of this gin-and-raisins obsession.

While I appreciate the testimonials I have received, there is no scientific evidence that this home remedy relieves the pain of arthritis.

Bear with me for a moment. I enjoy a glass of orange juice every morning with breakfast. I don’t have arthritis. Is it fair to say that the juice prevents arthritis pain? I don’t think so. But I can name a hundred people who will testify that they drink orange juice and don’t have arthritis, so there must be something to this. Right?

Wrong. The two factors may be totally unrelated.

The only way to test this hypothesis is to take 200 people with arthritis, give half of them orange juice and the other half plain, flavored water, and then see if the juice group has arthritis pain.

Here’s the outcome I predict. Ten people in the water group will experience improvement (a placebo effect), while 20 people in the juice group will feel better.

OK. Subtracting the 10 percent placebo cures from each group leaves 10 juicers who may have been helped. Here is where the problems begin. Is 10 percent a significant number, or could it be the result of chance selection alone? In short, is the difference between the groups statistically valid? Maybe.

Or maybe not. Perhaps the juice group was younger, or used low-dose aspirin to prevent stroke, or were more (or less) active—or also ate gin-soaked raisins.

This kind of analysis is what drives scientists crazy and makes some medical studies so difficult to interpret. Suppose the 10 juicers who improved were later found, on re-analysis of the data, to be overweight. They loved their large breakfasts with orange juice and decided to sit at home and not exercise regularly during the study; hence, their pains diminished.

Goodbye orange juice as a cure for arthritis.

This scenario is silly, isn’t it? But it does document the tremendous pitfalls inherent in any study, and also shows that mere testimonials are not valid reasons to use or not to use a particular product.

Enjoy your gin-soaked raisins, but—as far as I’m concerned—the issue is closed… until someone does a controlled, scientific study of 200 arthritis patients who…