Why I Don’t Adhere to Evidence-Based Medical Care

My concierge doctor and I the other day were going over the results of my latest blood test, which he insists that I get with some regularity.

My cholesterol was fine. Free testosterone was perfect. But I was below where I need to be on vitamin D. Ditto for the DHEA vitamin…

My concierge doctor and I the other day were going over the results of my latest blood test, which he insists that I get with some regularity.

My cholesterol was fine. Free testosterone was perfect. But I was below where I need to be on vitamin D. Ditto for the DHEA vitamin…

Wait a minute. Testosterone? No, free testosterone. Okay, where is there any randomized controlled trial showing that free testosterone is good for your health? The ones that I know of are not much help. And there probably won’t be others any time soon. Even if they get around to another clinical trial, I may not pay any attention to it. More on that below.

On my own

Why should you care? Because we are about to enter a world in which health insurers will only pay for procedures and drugs that are strictly evidenced-based. In fact we are about to enter a world in which doctors will be encouraged to reflexively practice evidence-based medicine for all patients ― regardless of who is paying the bill.

This will have consequences. Just about everything my doctor and I were doing the other day probably violates evidence-based guidelines, including ordering a blood test on an otherwise healthy patient.

As for free testosterone, there is a book on it written by Abraham Morgentaler, a professor at Harvard Medical School. Dr Morgentaler’s bio describes him as a “medical maverick” and a “pioneer” in men’s health, which is another way of saying he doesn’t practice mainstream medicine.

But remember, every advance in medicine started with someone doing something out of the ordinary, something different, something no one else was doing.

If you insist on only employing therapies that are evidence-based, you will only get access to many useful therapies years — or perhaps decades — after other patients have befitted from them.

What about vitamin D? Actually there have been a number of trials with it and it looks like it’s good for you. Unfortunately, there is no evidence-based guideline for testing for vitamin D deficiency, which of course is what my doctor and I did. If your insurer is telling your doctor what to do, odds are you will never know whether you are vitamin D deficient.

There have also been trials with DHEA and they are positive, but probably not sufficiently so for your insurer to pay for it and certainly not enough for your insurer to approve screening for DHEA deficiency.

There have been trials on vitamins as a whole  and the results haven’t been good. This, despite the fact that Prof. Bruce Ames (University of California at Berkley) has found striking effects of vitamins in rodent studies. Like Bruce Ames, my doctor believes in vitamins.

This brings us to the whole problem with much medical research. Suppose we conduct a clinical trial on eating peanut butter — just to see what difference it makes. And low and behold we find that there is no significant effect for 99 diseases, but peanut butter eaters have significantly lower risk of Parkinson’s disease.

There are three problems with this finding. The first is called “data mining.” If the distributions are normal, statistics teaches us that about 5 percent of the time we are going to find a significant effect, even though the relationship is entirely spurious. When you hear the term “95 percent confidence interval,” that normally implies we are 95 percent confident the relationship is real. Here is the flip side of that: if we are data mining — searching for anything significant among hundreds of variables — we are going to find spurious relationships as well.

The second problem is that this is testing without theory. There is a statistically significant relationship between the cock crowing and the sun rising, but without any theory relating the two we cannot conclude that the former causes the latter.

Finally, our peanut butter test left out variables that we know are important — including the role of genes. (We didn’t ask the participants if there was a history of Parkinson’s in their family.)

As I have noted before, most medical research would be rejected if the standards of the economics journals were applied.

What should insurance pay for? I don’t have a problem with evidence-based insurance. That’s probably a good way to keep premiums low and shield us from the cost of other people trying out every cockeyed claim that comes their way.

But it is advisable for each of us to have a Health Savings Account, so that we can use our own judgment and make our own choices when it appears in our interest to do so.