The New Health Dialogue

Matt Yglesias of the Center for American Progress made a major departure from his usually well-reasoned analysis of politics and policy last week in his post, “America Needs More Doctors.” We felt it was important to set the record straight.

The mistake of thinking we’re going to run into a crisis-level physician shortage is an understandable one, given the dominance of that narrative among health policy observers. Most of the arguments are justified along the lines Yglesias uses: we spend a lot of money on medical care, and we pay higher prices for most of it than the rest of the world. We also have fewer doctors per capita than many other countries, especially in primary care. Add to that the limited number of residency slots that determines how many new physicians can enter the workforce each year and the rapid growth in our elderly population, and the problem seems obvious: Low supply equals high prices and therefore high health care costs.

Unfortunately, this analysis misses a couple of fundamental points. Most importantly, new doctors don’t go where they’re needed! When doctors leave residency, they tend to stay in areas that already have high levels of physician staffing. Part of this is due to the simple fact that many of the places where there are lots of doctors are attractive places to live (i.e. New York City). However, the wage adjustment that would normally happen when professionals over-concentrate in one location (i.e. wages drop in that area, and some people decide to move elsewhere to make more money) doesn’t happen much in medicine. Instead, as the Dartmouth Atlas has demonstrated, areas with lots of doctors see the volume of medical services delivered increase, allowing physicians to continue making high salaries even in places that are overendowed with doctors. One demonstration of the disconnect between need and physician distribution is in the graph to the right: there is practically no relationship between the number of very low birth weight infants and the number of neonatologists in an area. (Image from this paper.)

While it’s true that some parts of the US (particularly rural areas and inner cities) may be legitimately understaffed, it’s also important to note that not all low-supply areas have too few doctors. Rather, there appears to be some “threshold” level of physician capacity, below which people’s health suffers, but above which there are few additional health benefits.* That means in many parts of the country, we really don’t need all of the doctors that we have right now. We especially don’t need as many interventional cardiologists and other specialists as we have. (For reference, we have about two specialists per primary care doc, while other developed countries with as good or better health have about two primary care docs per specialist.) We should concentrate on getting underserved areas to have sufficient primary care capacity, but expanding capacity everywhere else is, at best, useless; at worst, it’s harmful because it will drive up spending without improving health.

Finally, Yglesias is right that we need to find more ways to get by without doctors—or, put another way, to make our current supply of doctors sufficient for more people by making sure that they’re using their time as effectively and efficiently as possible. That means making sure doctors aren’t doing unnecessary procedures, as well as re-engineering workflow so doctors spend less time filling in redundant charts and more time treating and communicating with patients. The reason we need better ways to get by with fewer doctor isn’t an impending physician crisis—it’s just because that’s the best way to get as much health as we possibly can for our dollars. We’re certainly spending enough of them.

*Dartmouth professor David Goodman, et al. demonstrated the threshhold effect in a 2002 paper on neonatal care. The paper is here; we've excerpted the critical chart, which shows a small difference between "very low supply" of neonatologists and other categories, but no improvement in survival after that point.