Regis Duvignau / Reuters

Maybe, just maybe, the secret to saving money in the U.S. health-care system is to give patients more information.

At least, that’s the hope in certain corners of the health-policy world. Armed with more knowledge about nutrition, the thinking goes, Americans might choose broccoli over burritos. For surgeries, everyone from the government to insurers is releasing cost-comparison tools so that patients can browse procedures much like they would used cars—even though some studies suggest that these tools don’t actually save any money.

A new working paper challenges the notion that better-informed patients make better decisions. It found that an especially well-informed group of patients—physicians—don’t always make the best choices when they are patients, either.

For the study, which is out this week from the National Bureau of Economic Research and is still being peer-reviewed, three economists—Michael Frakes of Duke, Anupam Jena of Harvard, and Jonathan Gruber of MIT—used the military’s health system to compare everyday patients with patients who are themselves physicians. (The military system is one of the few databases that disclose a patient’s profession, Frakes told me.)

The study authors looked at several types of treatments in which people elect to get too much medical care, against scientific evidence—albeit sometimes under the guidance of their physician. They also examined situations in which people don’t get enough medical care, even when their health conditions necessitate it.

First, they considered C-section rates, which the World Health Organization recommends should be done at about half the rate they’re currently performed in North America. Among the pregnant female physicians, 29 percent got C-sections, compared with 31 percent of other pregnant female patients. But the pregnant doctors still didn’t meet the WHO guidelines, which suggest that only 10 to 15 percent of women should have C-sections. In other words, on C-sections, the doctors did appear to make more scientifically approved medical decisions, but not by much.

On the other measures the study authors looked at, the patients who were doctors did not differ from the regular patients. The study authors examined the likelihood that patients would get unnecessary screenings, such as chest X-rays, before low-risk surgeries, such as cataract removal. “The medical community says these surgeries are low risk, and the best medical evidence suggests this is just unnecessary money being spent on these tests,” Frakes told me. Still, 35 percent of both physicians and lay patients got the tests.

The study authors also looked at treatments that patients should be getting at higher rates than they currently are. Among people who have diabetes, both doctors and regular patients failed to get medically recommended tests and exams at roughly the same rates. Drugs called statins are recommended for people with heart disease, but only 75 percent of both the physician patients and regular patients took them. Finally, the physician group wasn’t more likely to fill prescriptions for high blood pressure or high cholesterol.

In all, physicians were only slightly more diligent than other patients in following medical guidelines for their own health care. “[These patients] went through internships, residencies, fellowships. They’re super informed,” Frakes said. “And even then, they’re not doing that much better.”

Why do physicians fail to follow their own medical advice? The study wasn’t set up to answer this, but a few things could be at play. When it comes to not taking their medications or getting necessary tests done, the physicians might be too busy to stay on top of their health, just like many of their own patients. Or, when a test is theoretically skippable but the physicians opt for it anyway, they might be prioritizing their own health worries over the potential savings for the health-care system, like many other patients probably do.

Frakes put forward another interesting theory for why some of the doctors received unnecessary treatment: Patients tend to be super deferential to their doctors, even when they themselves are doctors. There’s something about being told what to do by someone in a white coat that makes you want to follow directions, even when you know the directions don’t align with the latest in scientific evidence. One recent study, for instance, found that doctors’ referrals were a much stronger determinant of where a patient got an MRI than how much that MRI provider cost. People were willing to pass by lower-priced locations in favor of the one their doctor picked.

These findings, if strengthened by further research, would put a damper on the idea that the U.S. can resolve inefficiencies in its health-care system by simply giving patients more information. For example, the concept behind high-deductible health plans—in which patients pay higher costs before insurance kicks in—is that patients will research the prices of different procedures and avoid unnecessary care. But in this study, even the best-informed patients didn’t always do that.

True, online tools such as Amazon have made comparison shoppers of us all. But getting medical care isn’t like buying a toaster. It’s personal and often scary. You might ignore certain advice, or you might know, intellectually, that a test or procedure is unnecessary. But when it’s your own body on the line, the medical guidelines might not mean much.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.