More Primary Care Doctors or Just More Doctors?

In a Health Affairs “Web exclusive”, Dr. Richard Cooper of the Wharton School at the University of Pennsylvania has two studies, “States with More Physicians Have Better-Quality Health Care,”[1] and “States With More Health Care Spending Have Better-Quality Health Care: Lessons About Medicare”[2], which have generated a great deal of discussion; in this article I will address the first of these.

Cooper’s study purports to refute a previous study published by Katherine Baicker and Amitabh Chandra previously cited on this blog, “Medicare Spending, The Physician Workforce, And Beneficiaries’ Quality Of Care”[3], a widely quoted study that demonstrated that when states were ranked by either quality of care indicators or by Medicare spending, there was a large difference between those with higher numbers of specialists (costs up, quality down) and generalists (quality up, cost down). Cooper, a long-time advocate of increasing physician production -- but not with an emphasis on primary care or generalists (believing that market demand is an effective way of determining physician workforce) -- argues that increasing numbers of physicians increases quality of care, regardless of whether they are family physician/general practitioners (FP/GP) or specialists, and identifies what he sees as flaws with the statistical assumptions and analysis in the original Baicker and Chandra article. Cooper looks at regions, and notes that the South has low numbers of both specialists and generalists and low quality, the Northeast has high numbers of specialist and low numbers of generalists and pretty high quality, but that the Upper Midwest, Northwest, and Northern New England states, with high quality and high levels of FP/GPs are what creates this result (seeming as if more FP/GPs improve quality). He believes that this is because these states are relatively wealthy, have relatively low numbers of poor and minorities, and perhaps more progressive social policies demonstrated, for example, by the lower rates of incarceration. Some of these are good points; places that have a lot of resources, both in terms of money and social capital, do well; it does not address the issues of the incredibly higher cost of medical care in the Southern New England and Mid-Atlantic states compared to the three areas above, which are very probably the result of the high level of specialists compared to generalists. Cooper does not address cost at all.

Baicker and Chandra respond to this article in the same web issue of Health Affairs,[4] particularly addressing both Cooper’s misinterpretation of their data, and (in their assessment) poor use of statistics in his own study. (I will not try to summarize these issues, which involve the use of correlation, weighting and regression analysis; those of you who are expert enough in statistics can read the article.) They conclude that, properly analyzed, Cooper’s data (as opposed to his conclusions) supports their conclusions in the original 2004 study that higher numbers of FP/GP doctors improve quality measures and higher numbers of specialists do not. They take issue with his title and main theme, ““Quality is better in states with more physicians, both specialists and family physicians”, saying “A more careful statement would be, ‘Quality is better in states with more family physicians, but no significant association was found for specialists.’” They state, in pointing out that correlation shows direction but not in magnitude:

“The numbers of specialists and generalists per capita may have identical correlations with quality, but they have very different size effects on quality. Cooper’s own exhibits suggest that this is the case—and that generalists have a dramatically bigger effect on quality than specialists do....you would have to add roughly ten specialists per capita to move up ten spots in the quality ranking, but you would only have to add one generalist per capita to move up the same ten spots. And, apparently, even the small effect of additional specialists on quality is statistically insignificant.”

Philip Musgrove, deputy editor of Health Affairs, who wrote the Introduction to Cooper’s article, makes this point in the most recent issue of Health Affairs, which contains a series of letters relating to these articles (including letters by both Cooper and Baicker and Chandra). In responding to a letter by John Frey (see below) that asserts that Cooper’s viewpoint is supported by Musgrove, the latter endorses the analysis of Baicker and Chandra, writing “Cooper’s analysis actually agrees with theirs [Baicker and Chandra’s], since his own results show that the presence of more specialists has a much smaller (about a tenth as large) effect on quality than the presence of GPs has.” [5] Of note, Cooper’s response contains no such acknowledgement.

The letter from Dr. Frey, a leading family physician and academic who is Professor and Chair Emeritus in the Department of Family Medicine at the University of Wisconsin, suggests that:

“To test Richard Cooper’s hypothesis that it is simply more doctors, not the mix of specialty/generalists, that makes a difference in access, quality, and cost, why not close down all generalist training programs (which are well on their way toward that goal anyway, with the choices made by U.S. medical students) and see what happens? Managing complex multiple comorbidities, managing urgent and unorganized health complaints, or providing primary and secondary preventive care to large populations of chronically ill patients would be done by an increasing cadre of subspecialty providers.”[6]

Implicit in Frey’s suggestion is that they (the subspecialists, sometimes, to distinguish them from generalists, called “partialists”) would not, and perhaps could not, do so. I absolutely agree. While I do not urge the adoption of such a plan (and I’m sure Dr. Frey really does not either) any more than I advocate closing volunteer safety net clinics to stop “bailing out” government and society from the consequences of their failure to act to meet the basic health needs of all our people, I have no question but that the results would be exactly what Dr. Frey implies. And I have no doubt that the subspecialists would overwhelmingly agree. The “Joint Principles of the Patient Centered Medical Home” developed collaboratively by the American Academy of Family Physicians (AAFP), the American College of Physicians (ACP, representing internal medicine doctors), the American Academy of Pediatrics (AAP), and the American Osteopathic Association (AOA) and endorsed by the Patient Centered Primary Care Collaborative (PCPCC)[7], an industry-led coalition, never specify that these services must be provided by an identified primary care doctor, but it is very unlikely that there are many subspecialists who would choose to have to provide all these services.

However, Frey is correct about the fact that medical students are voting with their feet and running from primary care. Unquestionably, while there are many reasons that medical students might choose one specialty or another, the major issue in the dramatic change away from primary care in recent years ere is expected income combined with medical debt. As I have previously noted, much of this income differential is not “market” driven but simply reimbursement driven. If nothing is done to change the circumstances that have produced this movement, Frey’s ironic suggestion might become de facto true, and we will all suffer for it. Unfortunately, the work of scholars and policy people such as Cooper, advocating simply increase in the number of physicians without attention to the composition of the physicians workforce, may hasten rather than slow, this disastrous outcome.

[1] Cooper RA, “States with More Physicians Have Better-Quality Health Care,” Health Affairs 28, no. 1 (2009): w91–w102
[2] Cooper RA, “States With More Health Care Spending Have Better-Quality Health Care: Lessons About Medicare”, Health Affairs 28, no. 1 (2009): w103–w115
[3] Baicker K & Chandra A, “Medicare Spending, The Physician Workforce, And Beneficiaries’ Quality Of Care”, Health Affairs, 7 Apr 2004;W4.184
[4] Baicker K & Chandra A, “Cooper’s analysis is incorrect”, Health Affairs 28, no. 1 (2009): w116–w118
[5] Musgrove P, “Primary/specialty care: an author responds”, Health Affairs Mar-Apr 2009;28(2):594-5.
[6] Frey JJ, “Test the primary/specialty care hypothesis”, letter, Health Affairs Mar-Apr 2009;28(2):594.
[7] “Joint Principles of the Patient Centered Medical Home”, February 2007, http://www.pcpcc.net/node/14