The WSJ Health Blog gives us an update on the ongoing Quest to Produce More Physicians. We learn that over the last few years, there have been efforts to expand, start, and accredit more medical schools in order to produce more medical school graduates. Sounds like a plan, right?
Wrong. As the entry points out, residency is the bottleneck in physician training in the US. Increases in medical school enrollment will largely be offset by reductions in the numbers of international medical graduates (IMGs) who fill US residency slots. Without new residency slots, many of those new US medical students will simply displace IMGs, such that any increases in medical school aggregate size will only yield much smaller corresponding increases in physician numbers.
So why has so much of the past decade’s quest to produce more physicians taken the form of increasing medical school numbers when that isn’t even the limiting factor? If I had to guess, I would say that it’s the accounting. Medical student education costs are borne by states, private university endowments and donors, students themselves (or their families) in the form of tuition, and the part of the federal government that deals with student lending (the Department of Education?).
But, again, medical school isn’t the bottleneck. Who pays for virtually all residency slots, and graduate medical education (GME) expenses in general? Medicare. Yes, that Medicare. And funding for GME has been frozen-ish since the mid 1990s. To quote the thought leaders of my generation: “epic fail!”
Is this a rational policy for Medicare? If the market for physicians’ Medicare services operated like a real market, then the smaller number of physicians would lead (presumably) to a shortage, which would in turn result in fewer services being provided at a higher price. To figure out whether this would result in a net increase in Medicare expenditures, it would be time to start looking at elasticities and all sorts of other things they taught me in intermediate microeconomics. Of course, it doesn’t work that way, so it’s a moot point. There’s only so much any one physician can do (and bill Medicare for); capping the number of physicians is certainly one way to keep Part B costs from growing even faster than they are now, though I’d hesitate to ascribe that motivation to the policymakers behind the funding freeze.
This leaves us with a situation in which the traditional sources of funding for undergraduate medical education are still there, but the traditional source for GME funding is drying up, leaving residency growth to be funded by institutional endowments/revenues, or even corporate donors in some cases. Not quite the best environment for our Quest.
Of course, as with all things health policy, it gets more complicated than that. It’s not just that the US is suffering from a shortage of physicians in general, but that the US is hurting really, really badly for primary care physicians. The WSJ Health Blog also linked to a piece in the Huffington Post wherein a physician argues that the way forward is actually for Medicare to radically shift funding from specialty to primary care residency slots.
Dr. Gottfried, pointing out “Kazakistan” and “Azerbajin” as examples of countries with more physicians and worse health than the US, doesn’t think that crude physician concentration measures are a good way of looking at things [“Kazakistan?!” Is there a shortage of copy editors too?]. He prefers to look at the ratio of specialists to primary care physicians, claiming that “an ideal health care system consists of 70% primary care doctors […] and 30% specialists. Unfortunately, America has it backward-70% specialists and 30% primary care.” Though he cites no research to back up this claim, the “ideal” of 70-30 is floating around out there, and passes the smell test.
Is the “root cause” of the “top-heavy” US physician population really the alleged overabundance of specialty residency slots? It certainly stands to reason that you can’t have more specialists than primary care physicians without more training slots for specialists. That said, I don’t think that the overabundance of specialty slots is the “root cause” of the primary care shortage. More importantly, dramatic one-time reallocation of residency slots will likely have deleterious short- and long-term effects on the supply of primary care physicians and specialists alike.
As it is, there are compelling arguments out there that there are shortages of physicians in some specialties, not just primary care (general surgery comes to mind). Similarly, while there may not appear to be shortages in the national-level data, there certainly are regional shortages of PCPs and specialists alike that can’t be corrected by simply axing specialty and increasing primary care slots.
Reallocating more residency slots to primary care fields without changing the nature of clinical work in those fields might result in a short-term spike in residents. For the first few years, most of those already in the medical education pipeline will probably decide that it’s better to match to a primary care residency than to go unmatched, even if they might prefer a specialty. But unless the jobs are made more attractive — as it is, internal medicine and family practice residencies lean heavily on IMGs to fill slots in the aggregate, and even then end up a few hundred short every year — this is likely to result in fewer students choosing medical school to begin with, to the extent that the existing medical student population has aggregate revealed preference for specialty training programs that can be projected backwards onto pre-medical students. This would mean fewer specialists and primary care physicians in the long run.
Rather than a one-time reallocation, I might support an approach based on differential growth: grow the number of specialty slots more slowly than the number of primary care slots. But even with that, there is no getting around the fact that more and more medical students are shying away from primary care training programs. Forcing students into these programs by reallocating residency positions might work in the short run, but in the long run, the most effective way to increase the number of primary care physicians is to make the job something that appeals to more students than it does currently. This could take the form of increasing pay, decreasing administrative hassles and overhead, or even lowering the relative length of residency compared to specialty fields (in Canada, family practice residencies are 2 years compared to almost every other specialty at 5+).
Ultimately, you need people to want to fill the positions in order to fill the positions. When family practice residencies start to fill up more and more, we’ll know that we’re on the right track and that the time has come to increase the number of slots. This obviously isn’t as easy as decreeing “thou shalt not train specialists,” but in my view it’s the only sustainable way to get the increase in primary care supply that’s needed.