Commentary
Paying Physicians and Protecting the Poor
By David J. Satin, M.D.
Pay-for-performance systems should protect the health of poor patients and level the playing field for the physicians who treat them.
It is said that whereas morality reflects the way we would like the world to work, economics reflects how the world actually does work.1 As an experiment in physician reimbursement, pay for performance (P4P) has been criticized and justified on both accounts.2 Nevertheless, P4P compensation systems appear to be here to stay for at least the near future, both in the United States and abroad. Pay-for-performance systems were developed to close the chasm between the quality of health care patients can receive and that which they actually do receive. But as these systems gain popularity, we ought to ask ourselves what effect they have on one vulnerable segment of the population: the poor.
At WONCA Europe’s 2005 meeting of general practice and family medicine physicians in Kos, Greece, I had the opportunity to sit down with primary care physicians from Great Britain and New Zealand to compare notes about our countries’ similar-but-distinct P4P systems. We concluded that P4P programs in the United States lack the safeguards necessary to protect the quality of health care for patients with a low socioeconomic status. I believe there are two ideas we could borrow from these countries to make our P4P systems more equitable both to physicians and poor patients: adjust performance goals to account for the socioeconomic status of patients and allow for limited exceptions to the program.
Adjusting P4P Goals
A hallmark of P4P programs in the United States is that they do not discriminate between clinics serving rich and poor patients. Based on my investigation of Minnesota’s major insurers, interviews, and internal documents from the University of Minnesota Department of Family Medicine and Community Health, medical practices in wealthy neighborhoods typically have the same P4P goals as practices in poor neighborhoods. At first blush, that might appear fair. And it would seem morally suspect to suggest lowering the bar for health care quality in poor neighborhoods. After all, don’t principles of justice and fairness require equality?
In fact, they do not. Any accepted theory of justice requires that like cases be treated alike. Discrimination is not inherently bad; it is the process by which we determine which cases are and are not alike. For example, outlawing blind bus drivers is a kind of justifiable discrimination. Conversely, outlawing bus drivers who are black is arbitrary and thus an unjust kind of discrimination.3 So the question of discrimination as it applies to P4P is: Does lowering P4P thresholds in poor neighborhoods constitute arbitrary discrimination? To answer this question, we need to examine some empirical data.
There is a strong, well-known correlation between low socioeconomic status and poor health outcomes.4 In my own investigation of Minnesota’s largest health insurers, I’ve found that most, if not all, of their private commercial plans consistently have healthier patients with better health outcomes than their government-assisted Prepaid Medical Assistance Programs (PMAPs).
Why is this? The phrase “social determinants of health” describes the factors that influence how healthy a person is likely to be.5 Both in the United States and abroad, health care as a formal entity is generally considered to have far less of an impact on overall health than socioeconomic factors such as food, clean water, adequate housing, transportation, education, gainful employment, and the support of family and friends.6,7 A recent Wisconsin study suggested that 40 percent of health outcomes are related to socioeconomic factors, whereas only 10 percent are related to the provision of health care.8 So why does low socioeconomic status correlate so strongly with poor health? In short, because being poor typically means having poor social determinants of health.9,10
Failure to take into account the fact that many among us lack basics such as food or a home when designing a system of health care reimbursement puts clinics, physicians, and patients in poor neighborhoods at a disadvantage. Imagine yourself as a poor patient with diabetes. Checking your LDL cholesterol annually is unlikely to take priority over going to work, feeding your children, or finding stable housing, especially if the upshot is that you’d benefit from a medication you can’t afford in the first place. Asking clinics and physicians who work primarily with poor patient populations to achieve the same results as those working with wealthier populations is effectively asking for more, and in some cases, impossibly more from these providers. The results of such unrealistic demands may be fewer and fewer providers willing to serve the already underserved.11
Both Britain and New Zealand acknowledge that it is unfair to demand equal results when starting from such unequal positions. I learned from talks with a physician from South London Family Practice that Britain adjusts its P4P goals in relation to the economic status of the clinic’s postal code. In an interview with a family physician from New Zealand, I learned that that country’s most salient health care disparities run along ethnic rather than economic lines. In a bold acknowledgement of this disparity in the determinants of health between its aboriginal and nonaboriginal peoples, New Zealand has lower P4P goals for practices with higher proportions of aboriginal patients.
What might P4P safeguards look like in the United States? For starters, P4P strategies ought to reflect the fact that private insurance plans consistently have healthier patients with better health outcomes as compared with government-assisted PMAPs. Given the precisely documented correlation between socioeconomic status and health outcomes in this country, we ought to adjust our P4P goals according to patients’ enrollment in private or government-assisted insurance.
Allow for Exceptions
All P4P programs acknowledge that achieving 100 percent on any quality marker is unrealistic. We appreciate, for example, that a 90 percent childhood vaccination rate may actually amount to perfect physician performance, given the many factors that may influence the rate. We know that despite the best educational efforts on the part of the clinic, some parents refuse to vaccinate their children. Others simply do not bring their children in despite reminder letters and phone calls. And some children, albeit a very few, have contraindications to particular vaccines. Thus, a goal of a 90 percent success rate provides a plausible margin of error, given that all physicians may have some of these patients in their practices. But what if 30 percent of a physician’s patient population happened to be of a religious faith that prohibited vaccination? That physician would automatically lose his chance at a 90 percent vaccination rate and a P4P bonus.
My own clinic provides a real-life example of how clinicians who serve disadvantaged populations are penalized in P4P systems. One of our 2005 P4P goals was a 68.5 percent child and teen checkup rate.12 As a clinic serving an inner-city Minneapolis population, one of our biggest challenges is determining the denominator of this P4P statistic: identifying our patients. Once we have a child or teen in the office, we are typically good at providing a thorough checkup. But many of the patients assigned to our clinic are transient. One need not be homeless to be impossible to contact; one need only move once or twice or lose phone service and forget to update the clinic. Such a patient may come in once every two years, or never, as in the case of the so-called “invisible patient” assigned to the clinic by his insurance carrier.13 In any case, each counts as one more inadequately treated patient, and each counts against the physician’s P4P bonus, thereby penalizing clinicians serving transient populations.
The complete loss of or inability to contact a patient may seem unusual to many clinicians. But it is common for inner-city clinics both in the United States and abroad. Pay-for-performance models in Britain and New Zealand take this into account. For example, some British physicians receive P4P bonuses for performing an annual medication review of their patients with psychiatric conditions. British physicians who send two registered letters to a patient’s last-known address and make a minimum of one telephone call to the last-known number in an attempt to get the patient to the clinic have exercised due diligence.
In my interviews, I learned that physicians who show such efforts may count those patients toward their P4P bonus. Physicians in New Zealand have similar requirements to demonstrate due diligence. Although their appropriately documented “lost-to-follow-up” patients do not count in their favor, they are removed from the denominator and do not count against the physicians’ P4P bonus.
Critics might argue that no one is losing anything, as P4P programs pay bonuses over and above existing fee-for-service or capitation arrangements. But this is increasingly not the case. Initially, P4P programs paid physicians new moneys exclusively. But some insurers have started withholding as much as 15 percent to 20 percent from their reimbursement of providers.14 In my investigation, I found that some payers in Minnesota currently withhold 3 percent of all fee-for-service reimbursements, distributing the withheld payments as a “bonus” to the clinics successfully meeting P4P criteria. This anti-Robin Hood strategy often amounts to taking a percentage of revenue from the poorest clinics and giving it to the richest.
Establishing P4P exceptions will likely be subject to fierce debate. But provided the debate maintains an eye on safeguarding our nation’s most vulnerable populations, we will be better off for it. Therefore, I propose we begin by copying the exceptions of our colleagues from overseas.
Conclusion
None of what I have written should detract from the goal of working toward a system of physician accountability—perhaps by way of an equitable P4P system in which the goals are the same for all clinics and all patient populations. But a precondition of such a system must be that all patients have reasonably similar social determinants of health. Until that happens, we should follow the paths of Britain and New Zealand, which have leveled the playing field for clinicians who treat poor patients enabling them to compete fairly for economic incentives. Only after we adjust P4P goals according to socioeconomic status and introduce appropriate exceptions can we realistically claim to be paying physicians for their performance. MM
David Satin is an assistant professor in the department of family medicine and community health, and a postdoctoral fellow in the Center for Bioethics at the University of Minnesota. He also practices family medicine at Smiley’s Clinic in Minneapolis.
References
1. Levitt S, Dubner S. Freakonomics: A Rogue Economist Explores the Hidden Side of Everything. New York, NY: HarperCollins Publishers Inc.; 2005. Paraphrased from p.13.
2. Rosenthal M. Testimony to house subcommittee on employer-employee relations hearing on examining pay-for-performance measures and other trends in employer-sponsored health care. May 17, 2005.
3. Rachels J. The Elements of Moral Philosophy. 2nd ed. New York, NY: McGraw-Hill Companies; 1993.
4. Zaslavsky AM, Hochheimer JN, Schneider EC, et al. Impact of sociodemographic case mix on the HEDIS measures of health plan quality. Med Care. 2000;38(10):981-92.
5. Wilkinson R, Marmot M (eds). Social determinants of health: The solid facts. 2nd ed. World Health Organization. 2003.
6. Peppard PE, Kindig D, Jovaag A, et al. An initial attempt at ranking population health outcomes and determinants. WMJ. 2004; 103(3):52-6.
7. Arah AO, Westert GP. Correlates of health and healthcare performance: applying the Canadian health indicators framework at the provincial-territorial level. BMC Health Services Research. 2005;5:76.
8. Kempf A, Remington PL, Peppard PE, et al. Measuring population health disparities: the Wisconsin county health rankings. Prev Chronic Dis. 2005 Apr. Available at: www.cdc.gov/pcd/issues/2005/apr/04_0142d.htm. Accessed Feb. 24, 2006.
9. European Health Report 2002. Part 2. The major determinants of health. Available at: www.euro.who.int/europeanhealthreport/determinants/20020903_1 Accessed Feb. 24, 2006.
10. Public Health Agency of Canada. What Makes Canadians Healthy or Unhealthy? Available at: www.phac-aspc.gc.ca/ph-sp/phdd/determinants/determinants.html. Accessed Feb. 24, 2006.
11. Rosenblatt R, Andrilla CHA, Curtin T, Hart LG. Shortages of medical personnel at community health centers: implications for planned expansion. JAMA. 2006; 295(9):1042-9.
12. UCare child and teen checkup pay-for-performance guidelines. Internal documents, Department of Family Medicine and Community Health, University of Minnesota. 2005.
13. Internal document, Department of Family Medicine and Community Health, University of Minnesota.
14. Terry K. Pay for performance: a double-edged sword. Med. Econ. 2005 Jan 21; 82(2):64-6, 68, 71.