The challenge of expanded Medicaid and the dearth of primary care physicians

.
The shortage – current and future – of primary care physicians in the US has been a recurrent theme on both this blog and many other venues in journals and the blogosphere. In addition, several posts relating to the Affordable Care Act (ACA) have noted the difficulty Massachusetts has had since it implemented an individual mandate, with many more people covered and not enough primary care doctors to see them (Solving Medicare costs and the budget deficit: primary care, cost-effectiveness, and universal health coverage, Jan 5, 2011; PPACA, The New Health Reform Law: How will it affect the public's health and primary care?, Apr 12, 2010). Apparently, there are a lot of states that will be in worse shape than Massachusetts, according to a study by Leighton Ku and colleagues in the New England Journal of Medicine.

In The States' Next Challenge — Securing Primary Care for Expanded Medicaid Populations (NEJM 10Feb2011;364(6):493-5) the authors look at the expansion of Medicaid mandated in ACA, to 133% of poverty for nonelderly adults by 2014, and note that this expansion will be far greater for states that currently have the most restrictive eligibility for Medicaid than those that have the most generous (e.g., Massachusetts). They note that this will require far more primary care doctors, and observe that many of these same states have the greatest deficit in primary care capacity. Using a creative approach, they create a “Medicaid expansion index” to identify how much a state’s Medicaid population will grow as well as a “primary care capacity index”. They standardized them so 100 was the average, and combined them to get a relative assessment of each state’s “challenge”; over 100 is worse than average, under is better. The scores ranged from 212.6 (Oklahoma) to 15.2 (Massachusetts, in fact!). Unsurprisingly, the states that face the greatest challenges are mostly in the South and Midwest, like Oklahoma, while the other states with low scores tend to be in the Northeast. Some of this is, as Ku notes, a primary care physician deficit in those states with high scores, while and much of it is a result of the fact that these high-score states have such limited current eligibility for their current Medicaid programs that they will have the largest number of newly-insured people.

The authors acknowledge that the “Access to care is determined in local service areas, not at the state level. Access problems could be more severe in rural or inner-city areas than in suburban communities, for example.” No doubt they will be; as many posts on this blog (e.g., Primary Care and Rural Areas, Apr 28, 2010) have noted, distribution of providers (not only physicians but nurse practitioners and other “midlevels”) is not even close to adequate, both for primary and subspecialty care. There are too few providers in the inner-city, but in rural areas the situation is worse -- 20% of Americans live in these areas, but well under 10% of doctors practice there. Only family physicians distribute in proportion to the population, but 20% of family physicians is not anywhere near 20% of doctors. Other primary care specialties, such as pediatrics, are very concentrated in urban areas (Primary Care, Pediatrics, and Physician Distribution, May 21, 2009).

Ku, et.al., express some guarded optimism, suggesting that expanded insurance coverage will support more primary care doctors – but note that the expansion also doesn’t begin until 2014. Given the long time frame to create physicians in any specialty, this will at best leave us with several years of shortages. And, at the current rate, “at best” is unlikely. The authors emphasize the need for training more primary care doctors, especially in the most “challenged” states, but really make no suggestions that are likely to have a significant impact, citing such things as expanded scope for “midlevels” (does not address distribution) and expectation that increased funding for Federally-Qualified Health Centers (FQHCs).

In the very next article in the same issue of the NEJM, Stephen R. Smith does make some suggestions for change that would likely produce more primary care doctors. In A Recipe for Medical Schools to Produce Primary Care Physicians[1], he starts with the admissions process, suggesting that admissions be MCAT (Medical College Admissions Test) “blind”, meaning that above a pre-defined minimum score, MCAT scores will not be considered (so that a student with a very high score is considered “more desirable” than one with just a high score). He emphasizes the need to select students “…who express a desire to serve underserved populations, who demonstrate altruism, and who are committed to social responsibility” because “they are more likely to go into primary care”. (However,expressing interest is not the same as actually having interest. See the experience of Pennsylvania’s Commonwealth Medical College. [2]) He suggests that the curriculum be based on a “patient-centered learning approach” with continuity follow-up of actual patients and teaching of “basic science” in the context of these actual patients. He urges that the entire curriculum be built around the competencies needed for a primary care physician, that students be taught in inter-professional teams, and that community-based settings be used for training.

These are all good ideas. They are consonant with recommendations I have made (of course, this makes them good :)!). While they do not look at “output variables” (mainly income/reimbursement), they do address the two areas over which medical schools have the greatest control – the students they admit (“input variables”) and the curriculum (the “process”). The suggestions that Smith makes have all been tried, and they all work to a significant degree to increase the number and percent of primary care doctors. At the University of Kansas, for example, we do have essentially “MCAT-blind” admissions, and look for the characteristics he suggests, among others, believing that such personal characteristics as caring, altruism, and communication skills are not only important for primary care, but for all physicians. The problem is absolutely not that we don’t know what works; we do. The problem is that we have, nationally, lacked the commitment to implement these strategies on a large enough scale to have a sufficient impact on the supply of physicians.

There are two big issues, though. Obviously, the first is that “output variables” – mainly the enormous differential in expected physician income – are not addressed. This is critical. As long as reimbursement policies by Medicare (see “Outing the RUC: Medicare reimbursement and Primary Care”, Feb 2, 2011) and other insurers dramatically favor subspecialists and especially proceduralists, there will not be enough primary care doctors. Indeed, the other “problem” medical students often identify with primary care – less than appealing “lifestyle” (read: “too much work”) is related to this; if you make a lot more per hour, you have to work fewer hours.

The other big issue is that Smith addresses his suggestions to the many new allopathic (“MD”) medical schools being currently created. He notes that these are (mostly, although not all) designed to increase production of primary care physicians (although, as noted in the footnote about the Commonwealth Medical College, even those may have trouble getting students who are actually interested in primary care), and he is correct that adopting his suggestions, among others, is more likely to keep them on that path. However, this is too simple; it forgives existing medical schools from fulfilling this responsibility, and they absolutely should not be so forgiven. This is particularly true for the most “elite” schools, many of them private and in the Northeast and very “selective” (indeed “selectivity” – the percent of applicants that you turn down – is a criterion for high rank by US News and World Report). Such schools are also the ones with the highest amounts of National Institutes of Health (NIH) research support, and pride themselves on producing researchers. Different schools, the refrain goes, have different mission; we produce “physician scientists”, somebody else should produce the primary care doctors (hey, like those “new schools!”).

The problem is twofold. First, these schools produce a lot of physicians, and they need to produce the kind of physicians that the community needs. Second, these schools set the standard for what most other schools want to be like – to be highly ranked by US News and get lots of NIH money. The last big expansion of medical schools, in the early 1970s, was also supposed to produce primary care doctors, but many or most of them immediately abandoned that mission and began trying to be like Harvard or Johns Hopkins. What needs to happen is that Harvard and Johns Hopkins need to look more like the University of Kansas, and produce a much higher percent of community-serving primary care doctors. In fact, so does the Warren Alpert School of Medicine at Brown University, where Dr. Smith works.

So, in case I haven’t been clear, two things need to happen:
1. Current physician reimbursement formulas need to be abandoned, and Medicare needs to adopt a reimbursement scheme that will result in primary care physicians having at least 70% of the income of subspecialists. Where Medicare leads, private insurers will follow.
2. All medical schools must adopt admissions policies that de-emphasize high exam scores and emphasize desirable personal characteristics, and lead to much greater diversity of students by socioeconomic status, geographic origins, and race/ethnicity. They need a curriculum that reinforces these skills, problem solving, independent learning, and communication. The elite private schools should take the lead; where they lead others will follow.

When? What should be the timeline? Immediately. Right now. No delays. Both should have been done yesterday.

[1] Smith SR, “A Recipe for Medical Schools to Produce Primary Care Physicians”, NEJM 10Feb2011;364(6):496-7 (online available only to subscribers)
[2] The Commonwealth Medical College in Scranton, Pennsylvania selected its students based on an expressed interest in primary care, but found that in their first class, on a pre-matriculation survey (before they even started school!) only 23% still said they wanted to be primary care doctors! (Tracy & Smego, “Discordance of Self-reported Career Goals of First-year Medical Students During Admission Interviews and Prematriculation Orientation”, Family Medicine, Jul-Aug 2010.
.

Primary Care, IMGs, and the Health of the People

.
For several years now, there has been a great deal of discussion about increasing the number of physicians in the US through increasing the number of students in US medical schools. The Association of American Medical Colleges (AAMC) has called for a very large increase, and it is in fact happening, both through the creation of new medical schools and the expansion of class size in existing schools. AAMC has also called for the expansion of post-graduate specialty training (residency) positions because medical school graduates have to do residencies before they become practicing physicians. What specialties those new positions are will thus determine the makeup of our physician workforce. If we need more primary care physicians we will need both more primary care residency positions and a greater interest on the part of medical students in entering those residencies, which I have discussed previously (Primary care specialty choice: student characteristics , July 12, 2010; Primary Care and Residency Expansion, January 7, 2010).

To recapitulate, increasing the probability that students will choose primary care requires using criteria actually associated with primary care choice, which are both demographic and based on the individual’s previous activities, mainly in volunteer service. The risk of relying on intention as expressed in an essay or interview is made clear in a recent letter to the editor in Family Medicine from the new Commonwealth Medical College in Pennsylvania. An “overwhelming majority” of the students who were accepted to the school had expressed, in their essays and interviews, a very high level of interest in primary care, and had “consistently cited a predilection for small towns,” high priorities for the school. By the time of matriculation, that is when they started school, that only 23% had any interest in any of the primary care specialties, including OB-Gyn!

If US medical schools graduate more students without comparable residency expansion, the probable outcome would be the displacement of graduates of foreign medical schools by graduates of US medical schools. This might, intuitively, sound like a good thing, given the question of whether graduates of foreign medical schools provide care of the same quality that US graduates do. This concern can be more than xenophobia; in the US accrediting bodies, the Liaison Committee for Medical Education (LCME) for allopathic medical schools and the Accreditation Council for Graduate Medical Education (ACGME) for allopathic residencies, along with the American Osteopathic Association (AOA) for osteopathic schools and residencies, provide very rigorous standards enforced by regular re-accreditation. Internationally, the thousands of schools are, in most countries, less standardized; not only may there be dramatic differences in medical education between countries but among medical schools within countries.

It is in this context that Norcini et. al. published “Evaluating the quality of care provided by graduates of international medical schools” in Health Affairs, August, 2010 (29[8]:1462-68), to significant national coverage; the article in the New York Times by Denise Grady on August 3, 2010 is called “Foreign born doctors give equal care in the US”, which seems to give us the answer. They do. That is not only reassuring, but could be raise the question “Why, then, increase the number of US medical graduates if the international graduates who come fill unfilled residency spots are just as good?” Well, for one thing there is the very important issue of “brain drain”; physicians from other countries, often underdeveloped countries with great physician shortages of their own, come to train in the US. Ostensibly, for most of them on training visas, the idea is that they will go back to their own countries with the new skills that they have acquired in the US and benefit their own people. In reality, most of them want to, and usually do, find a way to stay in the US. From an individual point of view – the ability of an individual to seek a better life for his/her family, or at least a higher income – it is consistent with the history of the US. From a societal point of view, however, this leaves their home countries with marked shortages of doctors; there are more Ghanaian trained physicians in the US and UK than in Ghana[1]. And, to the extent, which is often the case, that the medical education was paid for by the government and people of the country of origin, not the individual, it can be particularly inappropriate.

Another question is “is it true? That is, do Norcini and colleagues actually demonstrate that “foreign born doctors give equal care in the US”? The population that they studies was doctors in Pennsylvania, a big state with a lot of variety (rural/urban, rich/poor). They looked at physicians who were US-born graduates of US medical schools (USMGs), and compared them to both foreign-born graduates of foreign medical schools (IMGs) and to US-born graduates of foreign medical schools (USIMGs), most from those schools in the Caribbean, which I have previously discussed (Who will care for the underserved? The role of off-shore medical schools, June 2, 2010). They measured the “quality of care” by measuring length of stay and mortality rate of patients hospitalized for acute myocardial infarction (heart attack, MI) and congestive heart failure. They also looked at the outcomes by specialty (cardiologist, general internist, family physician). The results showed that the percentage of in-hospital deaths for these diagnoses were lower for IMGs than for USMGs, and for USMGs lower than USIMGs. These differences were small but statistically significant. For length of stay, USMGs were lower than either, and IMGs were lower than USIMGs. How much to make of these differences since there were other variables: longer time since medical school graduation, being rural vs. urban, and not being a cardiologist resulted in longer stays; interestingly being a cardiologist resulted in higher mortality.

One can think of all kinds of possible explanations, including unmeasured differences in severity of illness, and the authors, in their Discussion, identify several. The most obvious is that they looked at only two parameters (death and length of stay) in two diseases in hospitalized patients. The authors acknowledge this, although they point out that these are very common diagnoses. They virtually ignore that measuring care in the hospital is only one dimension of care, most of which takes place in the outpatient setting, and run the risk (although they do not explicitly say this) of implying that if a doctor can deliver quality care in the hospital, when people are sicker, they obviously can do it in the “simpler” outpatient setting. This is an egregious fallacy, most obviously (see "Uncomplicated" Primary Care?, October 8, 2009) because in the hospital doctors have far more control, while for outpatients they are at best advisors to their patients. The authors did a credible job given the difficulty of measuring what they want to measure – quality of care delivered by physicians – but the validity of the results suffer from another fallacy , that what is measured is what is easy/possible to measure, not necessarily what you are interested in (see Defining "Streetlight" Research, February 26, 2009). Still, it is good work.

The real problem is in identifying the cause(s), speculating on what they might be and then taking this to the next level, raising problems that might exist if the speculations on cause on correct. IMGs might perform well because they are “top performers” in their countries, and have often had prior post-graduate training in their own countries prior to coming to the US and entering residency; much of this is hospital-based. The authors worry about the pool of USIMGs, noting that “Part of this performance difference may be due to variability in the quality of the medical schools that U.S.-citizen international graduates attend, but to some degree, it may also reflect their ability. It will be important to monitor this possibility, since the pool of U.S. applicants to international schools is a potential source of students for U.S. medical schools as they expand.” There is very likely a difference in the training and education of US students at many off-shore medical schools, although, like other foreign schools, they vary a great deal. The danger is in identifying “ability”.

The most important health problem in the US is that some people do not have access, for financial or geographic reasons most commonly. Thus a study like the current one, which looks only at patients who have received hospital care for their diagnoses, are looking at a somewhat skewed sample, and can miss the total impact on population health that comes from including those people not counted because they got no care at all.

Students who get into US allopathic medical schools have higher grades and test scores than the ones who don’t. While many students choose osteopathic schools because of their interest in osteopathy, a large number choose them because they didn’t get into allopathic schools; on average their grades and standardized test scores are lower. Those who do not get into either school may choose offshore schools. Are they less able? Does lower, but still good, performance on standardized tests make a candidate less able? The data that exist show poor correlation between MCAT scores and grades and clinical performance. Moreover, are the students who attend Caribbean medical schools representative of those who do not get into US schools, and might get in if more students are accepted? Not entirely, since on average they come from even higher socioeconomic status than the already high US medical students.

Many outstanding students, measured in many ways, are not accepted in medical schools in the US every year (Medical Student Selection, December 14, 2008). Taking more students (by virtue of larger classes or more schools) may lower the mean MCAT score, but is not, in itself, likely to decrease the clinical performance of graduates. Indeed, if those new students are more likely, because of their backgrounds and/or values, to care for populations that are currently underserved, rural and urban, they will increase the health status of the American people.


[1]Hagopian A, et. al., “The flight of physicians from West Africa: Views of African physicians and implications for policy”, Social Science & Medicine, Volume 61, Issue 8, October 2005, Pages 1750-1760.
.