Who Is Accountable for Racial Equity in Health Care?

Racial disparities are a ubiquitous feature of the US medical landscape, with health care delivery substantially segregated by race/ethnicity. Recent evidence from hospitals,^{1 - 3} nursing homes,^{4 - 5} and physicians' offices⁶ suggests that those caring for minority patients do not perform as well as those who care for nonminority patients, on average. This evidence is troubling but hardly surprising because the limited resources of those who care for the poor have helped to create and sustain racial disparities. As the United States enters an era of accountability in health care, it is time to consider these familiar circumstances from a new perspective.

De facto racial segregation is a remarkably prevalent feature of US health care delivery. A national study of Medicare beneficiaries hospitalized with acute myocardial infarction in 1994 and 1995 showed that the majority of facilities in the United States admitted no black patients during that 2-year period; 85% of all black acute myocardial infarction patients were admitted to only 1000 of 4690 acute care hospitals nationwide.² Nursing homes are even more segregated, with two-thirds of black residents residing in 10% of the homes nationwide.⁴ Physicians' offices also appear to be sites of de facto segregation: 80% of all primary care visits by black Medicare beneficiaries are made to only 22% of physicians, and physicians serving black patients serve few whites.⁶

These statistics are consistent with patterns of segregation found in other social settings such as public schools and churches. Some of this racial separation is driven by residential patterns: blacks and whites live in different neighborhoods and even in different parts of the country. But in health care, there are forces at play beyond sheer residential segregation. These involve the confluence of patient race, patient income, and health care financing.

Under the US system of health care financing, revenues from patient care depend on patient insurance status. Minority patients are substantially more likely to be uninsured or hold Medicaid insurance than white patients. This “racial payer gap” means that those offering care in concentrated minority settings do so at a significant financial disadvantage, on average. For example, the typical physician who provides primary care to a black Medicare beneficiary is 40% more reliant on Medicaid for practice revenues than physicians providing care to the typical white Medicare beneficiary.⁶ At the hospital level, National Association of Public Hospital and Health Systems (NAPH) member institutions (with inpatient discharges >50% nonwhite) had average operating margins of 1.2% in 2004 compared with average operating margins of 5.2% for hospitals nationwide.⁷ Among US nursing homes, 5% are “highly dependent” on Medicaid revenues. Forty percent of black nursing home residents reside in one of the highly dependent facilities compared with 9% of white residents.⁵

Because those providing care are generally free to consider ability to pay in choosing their patients, the racial payer gap compounds the effect of sheer residential segregation. For uninsured patients, many doors are closed. For Medicaid beneficiaries, options are limited, particularly for physician services and nursing home placements. Selective clinician and organization participation likely concentrates nonwhite patients within physician practices and institutions in areas in which the racial payer gap is substantial.

Chronically depleted finances can be expected to blunt organizational capacity by reducing the ability to attract qualified staff and diminishing the potential to purchase and maintain infrastructure. In turn, this takes a toll on performance and on organizational capacity to improve performance.

Revenues and the Ability to Attract and Retain Skilled Personnel. Dedication to the community and institutional mission attract some outstanding clinicians to underresourced settings. But recruitment of specialist physicians,⁸ nurses,⁷ and pharmacists⁷ disproportionately affects underfunded hospitals. Federally qualified community health centers (CHCs) face a shortage of willing clinicians generally, but the problem is particularly acute for obstetrician/gynecologists, psychiatrists, dentists, and pharmacists. Noncompetitive compensation is the primary barrier to recruitment.⁹

Geographic isolation is a further barrier to recruitment and retention of qualified clinicians in rural areas. Lack of spousal employment opportunities, a paucity of cultural activities, and poor-quality schools compound the problem of low pay.⁹ Although much attention has been given to difficulties in attracting clinicians, there are likely analogous problems in attracting seasoned health care managers. These rural workforce difficulties take a disproportionate toll on health care of black Americans, who are 3 times as likely to live in poverty as their white rural counterparts and disproportionately reside in the “persistent poverty counties” in the rural South.¹⁰

Revenues and Investment in Infrastructure. Thin operating margins preclude large capital outlays. Purchasing and maintaining information technology (IT) systems is now receiving considerable attention, in view of the role of these technologies in performance monitoring and reporting. NAPH member institutions lag substantially behind similarly sized academic medical centers: 97% of academic medical centers have electronic inpatient medical records vs 64% of NAPH hospitals.¹¹ A substantial lag in IT capacity also has been documented in CHCs,¹² with the greatest IT deficits in centers serving high volumes of uninsured and very poor patients.¹³

Revenues and Performance. In hospitals, chronic financial pressures have been linked to decreased patient safety and declining compliance with Joint Commission standards.¹⁴ A recent study showed modest but significant increases in 30-day mortality and lower adherence to Centers for Medicare & Medicaid Services process measures for patients with acute myocardial infarction in highly Medicaid-dependent urban hospitals compared with less Medicaid-dependent urban institutions.¹⁵ However, there was much heterogeneity in performance among hospitals that were highly Medicaid dependent. Consistent with this result, another analysis showed an association between high Medicaid dependence and performance on the Centers for Medicare & Medicaid Services process measures for several conditions.¹⁶ However, that association was found only in nonteaching hospitals. In teaching hospitals, performance was unrelated to Medicaid dependence.¹⁶ Performance is undoubtedly determined by many factors beyond finances.

Nursing homes with high rates of Medicaid dependence perform worse on a range of Minimum Data Set−based quality measures, including the incidence of pressure ulcers, use of restraints, and use of antipsychotic medications.⁵ Although there are few data on the performance of office-based physicians who have limited revenues, studies of clinicians with heavily Medicaid-dependent practices show that these physicians are less likely to have mainstream credentials (ie, be US medical school graduates, be board certified).¹⁷

Revenues and the Capacity to Improve Performance. Research suggests that health care organizations with the capacity to improve performance have strong stable leadership (committed chief executive officers and board members), collaborative physician-staff relations, organizational cultures that foster give and take, and reasonably sound infrastructure.^{18 - 19} It is unknown how these assets and capacities are distributed in the health care delivery system. However, economic theory and common sense suggest that in a market-driven system, poorly funded clinicians and health care organizations are less likely to have these organizational advantages. To the extent that there is a cumulative effect of underfunding, disadvantage may be compounded. To the degree that performance is linked to reimbursement in the future, the problem may become worse.

All of this suggests that performance incentives must be must be mindfully arranged if incentives are to dovetail with the goal of enhancing racial equity. A survey of current practices suggests that incentive programs are not crafted with racial equity as a consideration.²⁰ Prescriptions for doing so are difficult because little is known about how performance is produced or how it can be improved.²¹ Nevertheless, the following seem promising to address the issues raised here.

Set achievable performance targets. Performance assessment should be based on improvement in performance rather than the achievement of a particular level of performance. Alternatively, different rates of improvement or different standards could be expected of clinicians who are willing to care for uninsured patients and those covered by Medicaid. Although some might argue that this strategy would enforce a “soft bigotry of low expectations,” it seems equally unjust to ignore the consequences of chronic underfunding. This is admittedly a difficult balance.

“Jump start” improvement in underresourced environments by providing targeted infusions of expertise or technology. Although the transformation of chronically underresourced organizations will take time, temporary assistance may help catalyze improvement. For example, a recent evaluation of pay for performance in California Medicaid examined the ability of office-based pediatric care clinicians to improve well-child visit rates. To improve those rates, clinicians and practices needed to contact parents of children who were due for routine care. However, most practices lacked the requisite in-house technical capacity to generate the appropriate contact lists. Clinicians and office practices that received computerized lists from third parties with technical capacity—in this case, the health plans that covered the children—found this assistance crucial in being able to meet performance targets.²²

Provide funding for investment in needed infrastructure. Several initiatives are already under way to make IT investment affordable to those serving low-revenue patients. The NAPH has proposed a “Hill Burton Act for IT,” whereby health care organizations and practices would receive federal grants and low-interest loans for IT investment in exchange for providing free or low-cost medical services to those in need.²³ In CHCs, the California Clinics Initiative has provided a model for successful investment, and a national program of support for IT development in CHCs has recently been proposed.¹²

Offer supplemental funding to those caring for high volumes of low-revenue patients. The current system of subsidy to safety net clinicians and organizations is a frayed patchwork.²⁴ Hospitals may recover some of the shortfall from Medicare and Medicaid Disproportionate Share Hospital payments, local governmental transfers, and state risk pools; those supports vary substantially by local circumstances. For nursing homes, there is even more local variation. In the realm of ambulatory care, CHCs are reimbursed on a rather generous cost basis; however, non−clinic-based ambulatory care providers receive little or no supplement for caring for low-revenue patients. Medicaid professional fees are notoriously inadequate, posing a significant barrier to equity in outpatient care. In short, cross-subsidies vary widely. As long as those caring for uninsured and Medicaid patients have substantially depressed resources, it is hard to understand how they can (or should) be expected to perform at the level of their better-reimbursed peers.

This Commentary has considered the resources and capacities of clinicians and organizations, rather than the characteristics of patients. As others have noted, the era of accountability must also grapple with patient characteristics to avoid exacerbating racial disparities.²⁵ Inadequate adjustment for patient clinical, social, and behavioral characteristics; poor choices of measures; and “cherry picking” less ill patients all threaten to widen racial disparities, in addition to the factors discussed here.²⁵

The observation that money matters does not mean that other factors are irrelevant, and the efforts of those who provide care under financially difficult conditions should not be disparaged. Some assert that those currently providing care to minority patients may ultimately be positioned to offer them the best care because of cumulative individual and organizational cultural competence. This claim not only bears further investigation but also underscores the need to ensure that “performance” reflects ability to care for patients of all kinds.

The era of accountability affords the opportunity to identify, champion, and learn from clinicians and institutions that perform well, including those that do so against the odds. But the long, complex, and lamentable history that has led to racial disparities in health and health care in America today cannot be ignored. To a great extent, that is a story of the systematic disadvantaging of minority patients, and one with crucial implications for clinician and health care organization accountability.²⁵ The focus here has been on a parallel history. It would be unfortunate if the accountability era were to proceed without attending to the ways in which segregation and health care financing have shaped the organizations and individuals that are now to be held accountable.