Can physician characteristics predict overuse?

Many researchers have theorized about when, where, and in what situations physicians are more likely to provide low-value health care services. We know that certain types of practices are more likely to provide low-value services; for example, hospital-based primary care clinics provide more low-value care than community-based clinics. But what about differences in overuse within institutions? Are there physician characteristics that make them more or less likely to provide low-value care?

Predicting low-value care

Previous research on the topic suggests that some physician characteristics are associated with overuse. For example, one study found that male doctors order more low-value tests than female doctors, and another found a link between financial conflicts of interest and overuse. More research is needed, however, as the topic is far from settled.

That’s where this study from Dr. Aaron Schwartz and colleagues comes in. Schwartz and other researchers at Harvard Medical School used Medicare claims data to analyze differences in provision of low-value services by physician characteristics, including: age, sex, academic degree, professorship, publication record, trial investigation, grant receipt, industry payment, and panel size (the number of patients seen regularly). The researchers only included generalist physicians, so differences in provision of low-value care between family medicine doctors and specialists were not measured. 

Surprisingly, these characteristics predicted very little about physicians’ tendencies toward overuse. There was wide variation in rates of low-value services between physicians in the same organization, but differences in observable characteristics explained only 1.4% of variation in physicians’ provision of low-value care. Because the study compared physicians within organizations, differences in organizational culture were largely accounted for. So what could be causing such big differences between physicians if not factors such as age, number of patients, and financial conflicts of interest? 

The importance of individual behavior

The answer is individual physician behavior, writes Dr. Deborah Korenstein, in an accompanying editorial in JAMA Internal Medicine. Korenstein writes that “overuse drivers go beyond incentives and culture,” so reducing overuse will require deeper understanding of physician behavior.” 

Korenstein explains some of the common cognitive biases that can lead physicians to overuse, including: Availability bias, when we give more importance to examples we more readily remember; Confirmation bias, when we give more importance to information that confirms our existing beliefs; and Loss Aversion, when we fear the outcome of losing something more than anticipate the value of gaining something.

For example, when the ORBITA trial showed that stents were no more effective than placebo in relieving stable angina, many doctors refused to believe the results. Some doctors saw heartfelt letters from patients as proof that stents helped their patients feel better. “Getting letters from patients who feel better is an emotional lift doctors don’t want to lose,” said health services expert Doug McKell in a previous Lown Institute interview. “They did something and it was the right thing, and there was improvement. You feel the loss of that greater than the win.”

Can we solve behavior problems?

The bad news is that individual behavior is hard to measure; the good news is that behavior is more easy to change than characteristics like age, gender, and professorship. Korenstein suggests putting more emphasis on clinical reasoning and shared decision making in medical school and residency, to give physicians a foundation for thinking about the complexity involved in clinical decisions. 

Korenstein also notes that we could use nudges to take advantage of our cognitive biases, such as setting EMR defaults to “less” rather than “more”; and giving physicians data on how they compare to peers to pressure high-utilizers of low-value care to conform to their colleagues.

Additionally, “simple education about cognitive biases may foster awareness and facilitate more reflective [thoughtful] decisions,” writes Korenstein. We all have biases of some sort, whether conscious or unconscious. Knowing and recognizing these biases is a good first step to overcoming them.

Judith Garber is a Senior Policy Analyst at the Lown Institute. She holds a masters degree in public policy from the Heller School of Social Policy and Management.


View our curated list of news articles about Low-value Care