Why don't graduate medical education programs produce the doctors America needs?

In 2010, Fitzhugh Mullan and others from the George Washington University, with collaborators from the American Academy of Family Physicians' (AAFP) Robert Graham Policy Research Center, published the seminal article “The social mission of medical education: ranking the schools” in the Annals of Internal Medicine (discussed in this blog on June 10, 2010 in A New Way of Ranking Medical Schools: Social Mission).  This paper looked at the production of medical schools in the US in terms of whether their graduates were in 3 categories associated with social mission: practice in an underserved area, practicing a primary care specialty, and having a higher percentage of graduates who were members of underrepresented minority groups.  It was the first time such data had been published, and the results showed that most medical schools don’t do very well, and that, in general, those that do the worst are those most often identified as “top” schools by criteria such as National Institutes of Health (NIH) funding or rankings by US News. This is not surprising; enrolling students from high income families who have had top grades at the most elite private and suburban schools, training them in a setting in which the mix of doctors is heavily skewed to the most subspecialized and research-intensive, located in a densely populated urban area, is the precisely wrong formula for recruiting physicians from underserved backgrounds and training them to practice primary care in areas of need. Unfortunately, although there have been small programs implemented at many schools, this model has not seen any significant change in most medical schools, particularly those “most elite”.

This year, this group from the Graham Center and George Washington, with collaborators from the the American Board of Family Medicine (ABFM), take the next obvious step in examining the production of the nation’s doctors. In “Toward Graduate Medical Education (GME) Accountability: Measuring the Outcomes of GME Institutions”,[1] in Academic Medicine (not available free on-line; abstract at http://www.ncbi.nlm.nih.gov/pubmed/23752037), Candice Chen and her colleagues look at how institutions that sponsor GME (and, importantly, get very large amounts of money from Medicare and Medicaid -- $9.3 billion and $3.18 Billion, respectively, in 2009) do in producing specialists in short supply: primary care physicians (defined as family medicine, general internal medicine (GIM), general pediatrics, internal medicine–pediatrics, internal medicine geriatrics, and family medicine geriatrics), general surgeons, obstetrician - gynecologists and psychiatrists. Identifying which program produced who is quite a bit harder than when looking at medical school outcomes; while a medical school is a medical school, GME-sponsoring institutions can be consortia of a medical school and one or more hospitals, and may often sponsor more than one residency program in the same specialty.

As in all measures of primary care production, the big complication is in graduates of internal medicine programs, the majority of whose graduates go into subspecialty (e.g., cardiology, gastroenterology, pulmonary and critical care) fellowships rather than remaining in primary care/general internal medicine (GIM). Chen’s work accounts for this, but explicitly notes that they are unable to account for the percent of GIM graduates who do not go on to subspecialty training but work as hospitalists. These are clearly not in primary care, and by many accounts may represent a majority of those completing the basic internal medicine training but not going on to fellowship. The authors note that, by their calculations, the average for primary care was 25.2%,”… this overestimates primary care production, as we could not account for primary care physicians practicing as hospitalists.”

Just as with medical schools, there was a wide variation in the percent of graduates of different sponsoring institutions who entered these specialties-in-need. Of 759 sponsoring institutions, they ‘…found that 158 institutions produced no primary care graduates, and 184 institutions produced more than 80%; the latter tended to be smaller institutions.” Again no surprise; the larger, more elite and famous sponsoring institutions (most often hospitals associated with elite medical schools) did a terrible job, while the smaller sponsoring institutions -- often hospitals with one (usually family medicine) or a few residencies, in smaller cities and towns, and affiliated with Federally-Qualified Health Centers (FQHCs) or Area Health Education Centers (AHECs), based in in underserved urban and rural communities, did well. The Robert Graham Center website provides interactive tools to allow you to map the density of primary care physicians by state, county and other areas, the output of each institution producing residents in terms of location and specialty, and the footprint of graduates from each GME program.

Let me restate: 158 institutions sponsoring graduate medical education produced NO primary care graduates! Let’s add some other numbers from the study: 198 institutions (more than 25% of the total) produced NO rural physicians, while only 10 institutions had all graduates go to rural areas; the average percentage of graduates providing direct patient care in rural areas was 4.8%. 283 institutions (37%) produced NO physicians practicing in FQHCs or rural health clinics (RHCs); 479 (63%!) institutions produced no National Health Service Corps (NHSC; a program for sponsoring physicians for underserved areas) physicians. This performance can be described by a single, simple word: ABYSMAL!

Elite academic medical centers have values that lead them to attract and select students and trainees with characteristics that are the opposite of those needed for training physicians to meet the needs of the American people. They value caring for rarer and highly specialized tertiary and quarternary conditions, which both are highly reimbursed and provide the basis for research in narrow areas of disease (and almost never of health). They value receiving large sums of money from the National Institutes of Health (NIH) for research, most of which is basic laboratory research. These are not bad values; they are, in fact, necessary. We need research, including basic laboratory research and first-in-human studies, for medical science to advance (although we also need a lot more funding for population-based research into the causes of health and disease, and community-based efforts to address them). We need tertiary-care medical centers where rarer or more complex conditions can be most effectively treated by physicians and surgeons whose narrower expertise makes them more experienced and effective.

The problem is that concentrating all these subspecialists in the very places where students and residents are trained gives those learners a very skewed idea of the ratio of subspecialists to primary care doctors, and makes the teachers want to get the “best” students to go into their narrow, subspecialized areas. In addition, selection for medical school (and to a large extent for residencies that are more competitive because they have fewer slots) tend to be for students who have the characteristics that help them to do well in basic science courses, standardized tests, and possibly in laboratory research. These students tend to come from the highest income families of professionals, in the largest metropolitan areas, from elite public or suburban schools. They look a great deal like the people who have plenty of doctors to care for them, and very little like the people whose communities suffer from a dearth of physicians in inner city and rural areas. I have previously noted that, while about 20% of Americans live in rural areas, only about 9% of doctors practice in such areas; this study shows it is actually much worse, with only 4.8% of graduates in rural practice!

The situation is exacerbated by the inequity in income between primary care doctors and subspecialists, an inequity also seen (albeit at a higher level) for general surgeons vs. subspecialty surgeons, especially when hours of work are considered. Thus, students, generally selected from a population not representative of the American people, who have increasing debt (even for those from upper-middle-income, not to mention middle-income and the rare student from low income families), are attracted to specialties that pay the most; if assessments of “lifestyle” are added to the calculus, those that pay the most for the least work (or whose income/work ratios are the highest). This is a formula to continue what we have, not to make things better.

The authors of the paper say that “Primary care physician production of 25.2% and rural physician production of 4.8% will not sustain the current workforce, solve problems of maldistribution, or address acknowledged shortages. The relatively small number of physicians choosing to work in RHCs, FQHCs, HPSAs, and the NHSC will not support a doubling of the capacity of safety net services envisioned by the

Affordable Care Act.” They have that right. Medical schools and GME-sponsoring institutions have for too long been allowed to continue being self-serving, with the biggest institutions being entirely pitiful in terms of producing the doctors America needs. It needs an immediate, far-reaching, large-scale change, where the biggest training programs see themselves in the business of producing primary care doctors for underserved people.

Why would they do that? Well, there is the $9.5 billion in Medicare and $3.18 billion in Medicaid GME funding that could be withheld if they don’t, for a start…

---

[1] Chen C, et al., “Toward Graduate Medical Education (GME) Accountability: Measuring the Outcomes of GME Institutions”,  Academic Medicine, Vol. 88, No. 9 / September 2013

GME funding must be targeted to Primary Care

Much of the cost of training physicians is currently borne by Medicare (and, to a lesser extent, Medicaid). This is known as Graduate Medical Education, or GME, funding, and it pays some, all, or more than all (depending upon the hospital and based upon a complicated formula discussed on May 25, 2009, Funding Graduate Medical Education) of the cost of training residents in the various specialties that comprise medicine. For those unfamiliar with medical education, graduation from medical school, while it confers the MD (or DO, Doctor of Osteopathy) degree and the title “doctor”, no longer permits practice in any of the US states. A least one, and in some states two, years of residency (“GME”) is required for licensure, and most doctors complete an entire residency of 3 or more years to make them eligible for certification as a specialist in a field (eg, family medicine, general surgery, internal medicine, psychiatry, etc.). Fellowship training is requires addition years beyond the core residency to become a sub-specialist – for example, those who complete an internal medicine residency can then do additional years to become a cardiologist, gastroenterologist, endocrinologist, etc.

Medicare augments its payments to institutions (usually hospitals, although there are a few consortia and federally-qualified health centers) with two types of payments, Direct GME which is intended to pay residents’ salaries and cost of teaching, and Indirect ME which is for the additional costs that training hospitals bear for a variety of reasons. (In addition to the piece linked above, see also Training rural family doctors, Nov 5 2010; PPACA, The New Health Reform Law: How will it affect the public's health and primary care?, Apr 22,2010; Primary Care and Residency Expansion, Jan 7, 2010.) These payments have been the cornerstones for funding residency education. Because the amount is tied to the percent of Medicare patients in a hospital, rather than the total number of patients cared for in hospitals or outpatient settings, it could be (and has been) argued that funding GME should be done comprehensively and separately from Medicare. The most persuasive argument is that private insurers should also contribute to GME (they don’t, although Medicaid does in some, but not all, states). On the other side, many fear that uncoupling GME funds from Medicare would make it easier for a Congress looking at ways to cut the budget to cut GME than having it as part of Medicare.

Except this year, with exceptionally high pressure to cut the budget, Medicare is not even sacrosanct, although, as I have recently argued, (Medicare: A lifeline, not a Ponzi scheme, Dec 2, 2011) most of the proposals to cut it across the board by tactics such as raising the age of eligibility are poorly conceived. So there are now proposals to cut the funding from Medicare for GME. Unsurprisingly, this has created great anxiety in the community of academic health centers, and the Association of American Medical Colleges (AAMC), which has strongly supported expansion of GME residency slots, is quite alarmed (Preserve Medicare support for physician training, revised Oct 2, 2011). The Accreditation Council on Graduate Medical Education (ACGME), which accredits institutions that sponsor residency programs and, through its subsidiary Review Committees (RCs), each individual specialty and subspecialty, has done a study that shows that cuts in residency positions have already occurred and more major cuts are threatened if Medicare decreases its funding ("The Potential Impact of Reduction in Federal GME Funding in the United States: A Study of the Estimates of Designated Institutional Officials”). ACGME CEO Thomas Nasca, MD, is quoted by AAFP News Now as saying “We will actually reduce the number of physicians who are trained in the United States at a time when all workforce studies are demonstrating a mounting deficit of physicians….That will place us in a position where our physician-to-population ratio in 2020 and beyond is below (that of) most of the developed countries in the world." The study found that “With a 33 percent reduction in GME funding

• 68.3 percent of responders said they would reduce the number of core residency positions,
• 60.3 percent would reduce the number of subspecialty fellowship positions,
• 4.3 percent would close all core residency programs, and
• 7.8 percent would close all subspecialty programs.”

Because there are many more “core” residency positions than subspecialty fellowship positions, these would be disproportionately affected by across-the-board cuts. In addition, residency programs in primary care, which are not as profitable to the sponsoring institution, are even more likely to be cut despite the service that they provide to patients, especially those most in need. Perry Pugno, M.D., M.P.H., AAFP vice president for education, notes in that same article that "…any cuts to GME that go across the board are going to hurt primary care -- especially those of us who disproportionately take care of adults with chronic illnesses….In communities where primary care residency programs are present, those programs become the access point for the poor and disenfranchised of the area.” He says that it's not unusual for family medicine residency programs to see patients who live both in poverty and with numerous chronic illnesses. "The payment for taking care of those patients is so low that the local medical community often doesn't want to provide that care…But residency programs take all comers."

The key issue that Pugno is addressing is one that is very important issue and is not usually made explicit in national policy discussions: our current method of allocating Medicare GME funds to institutions (hospitals) rather than to individual residency programs tends to encourage funding the funding of positions in specialties that most profit those hospitals. The interests of the American people, in regard to the kinds of specialists they need, are not necessarily (and I would argue in fact are not) the same as the interests of the hospitals that sponsor residencies. Hospitals like to fund specialties whose trainees’ work enhances their revenue (e.g., cardiology fellows, who can increase the number of profitable procedures that are done) or at least decrease their loss (e.g., emergency medicine residents, who can fill gaps in seeing patients in emergency departments). Indeed, when hospitals can afford to, they often augment Medicare GME with their own funds to create more such positions. This is about their own financial interest, and does not take into account whether or not the US needs more cardiologists or ER docs, or more family physicians and general surgeons.

This contrast between the interests of the hospital (what kind of residency positions are most beneficial to its bottom line) and the needs of the population, is, of course, a subset of the larger tension. We train doctors in highly-specialized tertiary care academic health centers, while they will mostly practice in the community. There are a number of reasons that this is not brought up more often. For the general lay public, including most members of Congress and their staffs, it seems like a subtle difference. For experts, such as the AAMC, the issue is that they represent the interests of the medical schools, and want to have those interests seen as also representing the interests of the US population. Of course, they do not always, especially the interests of the most rural, poor, minority and other underserved portions of that population.

I think we need to use every opportunity to make this issue more clear and open. While it is probably true that it is a mistake to decrease federal funding for GME, it is absolutely necessary to increase the support for primary care and, in particular family medicine. And this will only happen if GME funding is explicitly tied to requiring it to be spent on primary care programs, and “prevents substitutions”.

Primary Care and Residency Expansion

In discussing some of the things I liked in the bill that passed the Senate (December 23, 2009, Health Reform: The good, the bad, and the bigoted ) one of them was that the proposal to expand by 15,000 the number Medicare-supported residency (or Graduate Medical Education, GME) positions did not get included. The reason was that it did not explicitly require that these positions be used for primary care, which I believe is an essential requirement. I cited the strong arguments made by Shannon Brownlee and David Goodman in their New York Times op-ed of the same date, "Doctors no one needs". Most primary care, and particularly family medicine, groups were also unenthusiastic to opposed.

The Association of American Medical Colleges (AAMC) does not agree, unsurprisingly to those who are familiar with medical education, and was the biggest advocate for that provision. As AAMC President Darrell Kirch wrote in his December 21, 2009 communication “Leader to Leader” (not publicly available on the AAMC website), the news on the Senate bill “…was a great disappointment because we viewed this as a truly historic opportunity to make a positive impact on our future workforce.” In what many in the primary care community saw as a more combative statement, he went on to say “During this process we were deeply concerned that some members of the primary care community spoke out against the amendment, and argued that it would not support the expansion of the primary care workforce. Facing an extremely tight timetable, Senate staff clearly indicated to us that such opposition would discourage the leadership from moving forward on any GME language. The AAMC expressed strong concern that the vocal opposition of the family medicine community threatened to halt progress on GME legislation that did indeed benefit all training programs.”

In a recent letter to Senator Harry Reid, staking out the organization’s positions on what needs to be included (read “benefits academic medicine”) as the Senate and House move to reconcile their health bills in conference, Dr. Kirch writes that the GME expansion is critical, and that the AAMC is “…supporting the inclusion of this workforce expansion as part of provisions to strengthen primary care.” That sounds, good, making nice. However, other parts of the letter indicated that AAMC’s attack on primary care, and particularly family medicine, groups, for not supporting the its agenda of expanding (“benefit[ing] all training programs”) has progressed.

"The AAMC recognizes that primary care is an integral part of health care delivery. Primary care, however, may be provided by many types of physicians and other practitioners. We support defining primary care by the types of services provided and not by a specialty of the physician or other provider.”

What is the problem here? Surely the assertion above is reasonable, that defining primary care by services provided rather than the specialty of the provider makes sense. And the AAMC is saying it is supportive of primary care, and even including goals for more residents in the primary care specialties in the expansion of GME slots. It is a big step for the AAMC to be so supportive, publicly, of primary care, as they have not always been. And, in addition, there are other specialty areas (e.g., general surgery) that are also in great shortage. Indeed, the movement has been to sub-specialization and sub-sub-specialization, so we are seeing fewer physicians who are even generalists in their own sub-specialties (such as cardiology). The goal should definitely not be to increase slots only for primary care, but to target those specialties in which there is a mismatch between the number of doctors being trained and the number needed by the community.

Considering primary care, however, there are several problems with the current AAMC proposal.

1. The proposed bill is about expanding residency slots, not about defining the content of a primary care practice. Yes, there are subspecialists who provide comprehensive patient-centered care for their patients. Particularly in pediatrics, but also in adult internal medicine; people who have mainly one serious chronic disease (kidney failure, cancer, heart disease) sometimes receive most of their comprehensive care from nephrologists, oncologists, or cardiologists (more often in pediatrics because having only one chronic disease is the norm in children, but much less common in adults). Many of these subspecialists do not. In identifying practices as providing primary care for, say, increased reimbursement, looking at services provided is quite reasonable. However, in looking at a strategy for creating greater primary care capacity, what makes sense is to expand the residency programs in specialties that are particularly about training physicians to practice primary care, and whose graduates actually do so – family medicine, general pediatrics, and general internal medicine. This is especially true when looking at how we can provide comprehensive primary care to communities, not simply to selected individuals. To say “let’s just train more doctors altogether, and some will probably do some primary care" (radiologists? anesthesiologists? ophthalmologists?) is a nonsense strategy.

2. The significant impact on the health of the population that is related to increased primary care capacity only occurs with more primary care doctors. It does not occur with just more doctors, some of whom might do some primary care. (This is the point of the Brownlee and Goodman piece cited above.) These results have been documented repeatedly, in a variety of geographic areas and populations. Yes, there is also a contribution made by “non-physician” primary care providers including nurse practitioners and physician’s assistants, but they are not the concern of the AAMC, and, moreover, are increasing not practicing primary care. (See “myths” 2 & 3 in Dr. Bowman’s guest blog of January 15, 2009, Ten Biggest Myths Regarding Primary Care in the Future.) I addressed the issue of specialty choice in More Primary Care Doctors or Just More Doctors? (April 3, 2009). Of note, Dr. Richard Cooper, whose positions I criticize in that piece, has more recently been advocating for the needs of poor and minority communities, a good thing. His main point is that the Dartmouth Atlas data on geographic variation do not account for socioeconomic differences (debatable, certainly); however, I have not seen any retraction of his AAMC-type support for “more doctors” rather than more primary care doctors.

3. There are not enough students currently interested in entering primary care to fill currently existing positions. Thus, even if a greater priority were given to family medicine and other primary care residency positions, the new positions would, barring a major change – that would, as discussed in many previous pieces, have to be systemic and involve large, not simply cosmetic, changes in reimbursement – also be unfilled, at least by US graduates. Then, of course, the teaching hospitals and medical schools would use them for other specialties. Indeed, a big reason even more students do not enter the “ROAD” specialties described by Pauline Chen (“Primary Care’s Image Problem”, New York Times November 12, 2009, and discussed in this blog November 17, 2009, as Primary Care’s Image: A Problem?) is the limited number of slots; increasing slots without increasing the attractiveness of primary care as a career option will just increase the mismatch between the proportion of primary care doctors needed by the society and that being produced by medical schools. To the extent that primary care residency positions are filled by international medical graduates, it continues to contribute to the “brain drain”, where third-world countries bear the cost of educating physicians to provide care to first-world citizens.

The only way an increase in the number of GME slots could be beneficial is not only if a majority are targeted at primary care, but are required to be primary care, or they don’t happen. And that only specifically primary care internal medicine residencies count, and that the slots are withdrawn if, after 5 years, more than 25% of graduates have entered subspecialty training.

In greater detail, Patrick Dowling, chair of family medicine at UCLA, comments on the AAMC letter:

“Granted these are complex issues but in the end I read this as: ‘give us more of the same—we need more doctors, more funding for academic centers and we need to get reimbursed better!’ I think the AAMC would have much more credibility if they stepped up to the plate and said:

‘The US health care delivery system is terribly flawed and we are a significant part of the problem. We have terrible geographic and specialty maldistribution of physicians, our costs continue to be way out of line compared to any other industrialized country and we have unacceptable racial and ethnic disparities in outcomes of care.

‘Moreover, because the graduates of our medical schools have overwhelming chosen to practice subspecialty medicine in green leafy suburbs we must import international physicians, to staff our inner cities and rural towns in exchange for visas. Although we are fortunate to have someone to send, these docs face overwhelming linguistic and cultural barriers, especially in the provision of care to low income minority populations. And in some instances they represent a “brain drain” in from the donor countries

‘Further, as the baby boomers begin to hit age 65 at the rate of 5,000 per day on Jan 1, 2011, the epidemic of chronic diseases linked to aging will soar. We would propose the following new innovative steps to insure that we have a geographically dispersed physician workforce that delivers cost effect, high quality care with a physician workforce that is optimally balanced by specialty.

‘If you provide $X billion in extra funding for Academic Medical Centers (AMCs), enhanced funding for NIH budgets and thousands of more Medicare funded GME spots we will insure that the number of HPSAs will be reduced by X, that the actual number of USMGs choosing bona fide primary care specialties will increase by Y number which will result in a primary care to specialty ration of A to B, a ratio which works very well in other industrialized democracies.

‘Finally, if we are funded we promise to bend the unacceptable curve of increasing costs so that average yearly increases are less than X% of CPI. If we fail to meet these objectives we agree to decreased funding over the following years of $Z billion.’

“If I was in the US Congress,” Dr. Dowling concludes, “I would tell the AAMC that rather than stuffing their pockets it is time to put some skin in the game and actually become the leaders in the science of health care delivery and solve these problems.”

Hear, hear.
.