Why We Need More Medical Studies

Megan McArdle is a Bloomberg View columnist. She wrote for the Daily Beast, Newsweek, the Atlantic and the Economist and founded the blog Asymmetrical Information. She is the author of "“The Up Side of Down: Why Failing Well Is the Key to Success.”
Read More.
a | A

A few years back, I wrote about the tyranny of blood levels in medicine, a problem that thyroid patients have been complaining about for years:

Thyroid deficiency is measured by checking the levels of something called Thyroid Stimulating Hormone (TSH). When your thyroid levels are too low, your body releases TSH to get things moving again. Only if your thyroid is conking out, the TSH doesn't do much, so your body releases more. Most labs consider the "normal" range to be somewhere between 0.5 and 5.

The American Association of Clinical Endocrinologists now considers this to be too conservative; they've revised their guidelines to between 0.3 and 3. My TSH levels were right at the edge of the new, narrower range.

There are real risks to taking too much thyroid hormone -- it can cause heart palpitations and increase your risk of fractures. Unfortunately, too little thyroid hormone can leave you fat, bald, constipated, and depressed.

Doctors used to treat mostly based on symptoms; they kept giving you thyroid supplement until you lost weight, regained your hair, and perked the hell up. If you developed heart palpitations, sweating, insomnia, or anxiety, they cut back your dose.

But then it became easier to test for TSH. Guidelines were developed -- conservative guidelines that erred on the side of hypothyroidism, since being hyperthyroid can kill you, while being hypothyroid just, well, makes you fat, bald, constipated, and depressed.

As a result, as thyroid patients have now been complaining for years, doctors stopped paying attention to the symptoms. They treated the number instead of the disease.

This problem has eased in recent years, but it has not gone away. I know of a number of thyroid patients who complain that they still obviously have symptoms, but their doctor insists that they're fine -- or, worse, accuses them of trying to use thyroid hormone as a weight-loss drug and tells them to get their fat butt to the gym. Just to illustrate how ridiculous this is, when my normally slim mother's thyroid gave out, she packed on more than 20 pounds in a month. She responded by cutting her diet back to nothing but sashimi, and she worked out so hard that she repeatedly injured herself. But even on this extreme regime, she struggled to take off a few pounds. When they finally diagnosed her and put her on thyroid hormone, she dropped 20 pounds in a month. Accusing thyroid patients of being lazy gluttons may not be medical malpractice, but it's astonishingly stupid for a group of people who are supposed to be smart and knowledgeable about how the body works. Yet it's still shockingly common.

This is not just a problem in treatment; it's also a problem in drug development. Drugs are often developed -- and prescribed -- not because they've been shown to reduce heart attacks or strokes, but because they've been shown to lower blood levels that are associated with heart attacks or strokes. More than occasionally, as John LaMattina notes in Forbes, it turns out that your treatment does indeed affect those blood levels but doesn't actually improve the health outcomes we'd really like to target. Or, worse, that your treatment is not merely useless, but actively harmful:

One of the major complications of diabetes is end-stage renal disease (ESRD). One way of slowing its progression is to treat these patients with blood pressure lowering agents like lisinopril (an ACE inhibitor) or losartan (an ARB), which help protect the kidney not just by lowering blood pressure but also by lowering protein levels in the urine. For the most part, physicians used either an ACE inhibitor or an ARB to treat ESRD as both drugs acted in the same biological pathway. However, under the premise that if one drug works then using two should be better, many physicians were treating their patients with both agents. Why not? They are relatively safe, have been on the market for more than 20 years, and are generic. What could be the harm?

Well, the VA NEPHRON-D study showed that combining lisinopril and losartan was associated with an increased risk of adverse events, specifically, hyperkalemia and kidney failure. Although the combination did lower blood pressure and urinary protein levels, the harmful effects of increasing blood potassium outweighed any benefits.

Unfortunately, as LaMattina observes, hard-outcomes trials are really difficult and expensive to do. It's relatively easy to ascertain whether putting someone on a statin lowers their cholesterol; it takes years to find out whether you've made that same person less likely to have a heart attack. Multiyear studies cost a lot of money, and people are more likely to disappear or just drop out, making your results harder to assess. And drug companies that have already spent a bunch of money getting their drug approved and marketed based on blood levels don't have great incentives to go back and do outcomes testing.

This is the sort of research that the government should be funding: a genuine public good that takes a lot of money and expertise and produces information that's valuable to everyone. And we do fund some of it. But if we want more of it, we're going to have to find more funding, because when bad incentives combine with lean times in the pharmaceutical industry, the free market is unlikely to deliver more hard-outcome studies. Of course, these are rather lean times for the government, too, so I doubt we'll see the studies we want anytime soon.

This column does not necessarily reflect the opinion of Bloomberg View's editorial board or Bloomberg LP, its owners and investors.

To contact the author on this story:
Megan McArdle at mmcardle3@bloomberg.net