Statistical drugs: Depression and cholesterol


It’s summer, higher education slows a bit, and I get to consider other things. So, we have another depressing headline in the news:

CDC: Teen suicide rate up 70% from 2006 to 2016

This is a pretty hefty increase over a mere decade. Some news sites are trying to spin this as inadequate medical coverage, but I’m not buying it. Americans are the most medicated population on the planet, though we pay very dearly for those drugs, even when insured. It’s not just teens in any event, suicide rates in general are increasing in this country.

But wait just a minute here. Anti-depressant drugs are very easily prescribed, doctors pass them out all the time—when I was diagnosed with cancer, the doc had a bottle in my hand very quickly (I refused to take them, almost certainly a great decision on my part).

These things really are being passed out like candy on Halloween:

Between 2011 and 2014, approximately one in nine Americans of all ages reported taking at least one antidepressant medication in the past month, according to national survey data released by the Centers for Disease Control and Prevention (CDC). Three decades ago, less than one in 50 people did.

Now, the ol’ “conspiracy theory” sites say these drugs are useless, and doctors who claim such things (there are a few) generally have their careers smashed but…the bottom line is we have over 30 million Americans taking drugs to offset depression, the most debilitating symptom of which is death by suicide.

And yet the suicide rates are increasing.

These drugs supposedly help based on statistical studies.

Now, I know a thing or two about statistics, and I know how trivial it is to manipulate statistics in a study to say whatever you want. To get a drug approved, a drug company need only bribe convince our government into saying the drug works, and all that takes is money a statistical study.

These studies typically have only a few dozen people, maybe a few hundred, and this study might be performed over a few weeks, perhaps a year. On the basis of such studies we’ve now inflicted these drugs on more than 30,000,000 human beings, for decades. Most importantly, we’re giving these drugs to human beings which our doctors believe are most vulnerable to the most debilitating symptom of depression, suicide. If these drugs really helped, we’d be seeing a reduction in suicide in general statistics for the population.

In short, I have a study with 30,000,000 subjects on a drug which is supposed to reduce the suicide rate over the course of years…and the rate has gone up, dramatically. So…how exactly are these drugs helping? There are various studies saying these drugs are addictive, these drugs have very harmful side effects (including, amazingly enough, suicide), and so on, but challenging these studies is completely irrelevant, because we now have unbiased data more powerful than any study performed by the drug companies, which shows the drugs aren’t helping (it’s possible society is just so horrible now that the rates would be far worse without the drugs…but I just don’t buy that one at all, not without extraordinary evidence).

I could toss in how there’s a strong link between these types of psychotropic drugs and mass shootings but, again, it’s beside the point. It would be up to the philosophers to decide if the occasional mass murder was worth it in exchange for helping millions of people. We need not consider this, however, because the evidence doesn’t show these drugs helping the general populace at all.

I’m no doctor, and you must of course follow your heart regarding taking these kinds of drugs. I totally understand how one might take these drugs in desperation, and the placebo effect is legitimate enough that even if the drugs are worthless medically, they might still have a benefit in the short run for an individual.

But, bottom line, in terms of statistics? I deeply dispute studies performed by the drug companies or a corrupt government taking money from the drug companies (i.e., with a conflict of interest, direct or indirect) which say these drugs do much against the worst side effect of depression, because the general evidence says, if anything, these drugs are increasing the rate of suicide.

I was confronted with another statistical drug recently. I’m easily in the last half of my life, so I went to the doctor for my first checkup in five years.

Doc: “Your cholesterol level is high, I’m going to put you on some statins.”
Me: “How high? What are my risks here? How safe are these statins?”
Doc: “You’re at a 15% elevated risk of a heart attack. You need to take these drugs for life.”
Me: “Hmm. Okeedoke.”

–pretty much the conversation I had with the doctor, a young man who struck me as fairly bright. It really was surprising that concepts such as “diet” and “exercise” didn’t come up.

We’re really at the stage of “big data” for our drugs, where benefits are not determined by direct observation, but by statistical manipulation. Now, I grant I’m no athlete, but I’m fairly active, go up and down 4 flights of stairs without any issues (I teach in an old building, built before elevators, but does have nice views of the modern administrative palaces on campus), go to the gym and lift weights weekly, and no specific health problems. Checking the ol’ stats, I see I have around a 1% chance of having a heart issue/heart disease, based on my age, gender, and race. That’s an average, and I’m pretty sure my fitness level, minimal drinking, no smoking, good diet, and having no health issues will put me in that area (no family history of heart disease, either), give or take a few tenths of a percent.

So the doc says I have a 15% elevated risk. Now that sounds like a lot, and I imagine most people get suckered by hearing it expressed that way. But I know the math, and I know that now my chance of a heart attack (again focusing on the worst effect of what the drug should prevent) soars from 0.01 all way up to 0.0115.

With effects this small, this is another statistical drug, you really need to look at a sizeable part of the population to have any hope of determining any benefit. The gentle reader should compare these drugs to say, aspirin, where the fever-reducing affects are immediately noticeable on any lone individual who takes a few pills—you don’t have to study a few hundred people over the course of a year to calculate a few tenths’ of a percent improvement with aspirin.

So, the doc says if I take the drugs (and they work perfectly) we’ll have a reduction in heart attack/death here, changing my percentage from a low 1.15% all the way down to 1%. That’s not so easy to visualize.

Hmm. Let’s put that in terms most people can more easily understand. Under the assumption these drugs were perfectly effective, then a group of 700 males like me who used these drugs would see 7 heart attacks in a year (i.e., 700 times 1%), while a group of 700 who didn’t use the drugs would see 8 heart attacks in a year (i.e., 700 times 1.15%, or 8.05 heart attacks—I’m rounding a bit).

8 heart attacks without the drugs, but 7 for the group using the drugs for this group of 700, and you could scale that for any group size. So, if I don’t take the drugs, the chance of that decision hurting me is 1/700. Hey, heart attacks are bad things, and if these drugs were completely harmless, I’d have no problem doing something to give even a piffling 1/700 chance of not having a heart attack.

But these drugs do have side effects. They increase the rate of diabetes (I’m citing an official medical site, not a conspiracy by any means), and diabetes dramatically increases the chance of all other health issues, including heart attacks, by far more than 1/700. I’ve a family history of diabetes so I’m pretty motivated not to get it. The drugs also might well cause dementia and memory problems…I like having a functional brain, so another strike against taking these drugs.

Do these drugs even work? Well, the doc says a 15% increased rate, so they must have been studied pretty hard. To quote me that number, he must have been given a margin of error of 0.0005. A quick check of the sample size for this margin of error shows a study of around 80,000 people much like me, for a year (assuming 95% confidence, with an estimated 1% chance) must have been done at some point.

I would have liked to have seen that study, as that’s basically a very large proportion of the population with stats somewhat like me…I don’t know anyone who was in it (and I know plenty of people in my age group), nor could I find evidence of such a study being performed. There are at least a dozen age/sex/race categories, so assuming they did this for every group indicated, we’re talking a clinical trial of over a few million people for over a year. I don’t think it actually happened despite the doctor’s confidence. Hmm, just how many numbers are simply being made up, here?

On the other hand, we have been using the population of this country to experiment with this drug. 25,000,000 Americans take statins now, and, much like with anti-depressants, these drugs are given to people who, supposedly, are most vulnerable to heart disease (if I can use the formal description for a range of heart-related issues, though like the general public I’ll simply express this a as a heart attack).

With such a massive prescribing of drugs to reduce heart disease, we totally should see a reduction in the general population, right? Statins were first introduced in the late 80s. Again, I focus on the worst effect of a bad heart: death. So, are deaths from heart disease decreasing?

It turns out, they are. More for males than females, even, and I’ll just consider males for statistics here.

From 1990 (when these drugs started being given to nigh everyone) to 2014 (most recent stats), heart disease deaths dropped almost 10%. That’s pretty impressive. Is all of this due to the minor benefits (at best) claimed by the drugs? That drop doesn’t even cover what the doctor told me, 15%, for my particular numbers, and these drugs are supposedly more effective for more “at risk” people. Have we truly made absolutely no other medical advances in the last 24 years? It really seems like there have been improvements in medicine in the last generation.

My suspicions aroused, I look harder at the data. I note that from 1969 (earliest stats provided above) to 1990, deaths from heart disease dropped 14% Seriously, the rate was dropping faster without the drugs. So…again I’m pretty wary of using these drugs, as unbiased statistical evidence points in the exact opposite direction which the drug company claims. (Scaling for population changes over this time period, incidentally, only makes these drugs look more irrelevant.)

The gentle reader must, of course, follow his heart. I do encourage you to make an informed opinion, however. (I again consider the possibility that perhaps we’re damaging our hearts so much that death rate would be worse without the drugs, but again I would need to see some extraordinary evidence to accept this as a possibility.)

Personally, I filled the prescription, but I’ll not be taking the drugs. The doctor is only claiming a 1/700 chance that I’d be helped by them, after all, and that’s the best case scenario of assuming no negative side effects, assuming there have been no heart-related medical advances in the last 24 years, and assuming the empirical evidence of the 21 years before that is falsified. I’m willing to bet 1/700th of my life against at least one of those assumptions being invalid.

I personally find more realistic options, such as cutting down sugar, losing weight, and fast walking a mile a day on the new treadmill, are more likely to help me…admittedly, “more likely to help” is a pretty low bar than what the doc offered. I’ll let you know how well my plan works out after a few months.

If I suddenly stop posting, well…1/700 is greater than zero, and disregarding doctor’s orders always has risks. Like I said, follow your heart.