Reducing Health Disparities or Improving Minority Health? The End Determines the Means

Improving health outcomes for racial and ethnic minorities in the United States has been a national priority since the Institute of Medicine identified equity as one of the six areas of priority focus for improving health care quality in the United States.¹Finding Answers: Disparities Research for Change is a national program funded by the Robert Wood Johnson Foundation that seeks to facilitate efforts to improve health care equity in part by conducting and disseminating systematic reviews of the literature on effective strategies for reducing disparities.² This issue of JGIM features five systematic reviews, conducted by the Findings Answers program, which summarize the published literature on interventions aimed at reducing disparities in asthma, human immunodeficiency virus (HIV), colorectal cancer, prostate cancer and cervical cancer. Reading these reviews, one is heartened by the progress that has been made in the decade since the Institute of Medicine report was released. Yet, it is striking that, while the Finding Answers initiative is explicitly aimed at reducing racial and ethnic disparities in care, only a few of the 178 intervention studies included in the five systematic reviews published in this issue actually examined reduction in disparities as an outcome measure. The vast majority of interventions focused on improvement in processes or outcomes of care for members of minority groups, rather than reductions in disparities; a point noted by the authors of the systematic reviews, and consistent with the findings of other systematic reviews of disparities interventions.³ The distinction between the goals of “reducing health disparities” and “improving minority health” is rarely explicitly addressed in published studies, but has important implications for healthcare organizations as they choose and implement interventions, as we describe below.

There are certain cases where an organization’s focus should clearly be on reducing disparities. The most obvious example is when language barriers have been identified as a contributing factor in poor health care processes and/or outcomes for non-English speaking populations. An effective intervention that targets those barriers should have the intended effects of improving care for non-English speakers without affecting English-speaking patients, thereby reducing the disparity. An explicit focus on reducing disparities is also warranted when observed health disparities can be attributed to provider- or system- factors, such as process-of-care measures (e.g., testing of HbA1c and LDL cholesterol level) or treatment decisions (e.g., decision to prescribe anti-retroviral therapy for HIV). In such cases, system-level interventions aimed at providing uniform, guideline-concordant care have been shown to reduce disparities.^4–6

The appropriate focus is less obvious when the disparity can be attributed, in part, to patient behaviors (e.g., control of HbA1c and LDL cholesterol level, adherence to anti-retroviral therapy), and when social determinants of health associated with patient race/ethnicity (e.g., income, education, health literacy, neighborhood), in addition to healthcare factors, contribute to the disparity by affecting those patient behaviors. In those cases, intervening on aspects of the healthcare organization (such as improving patient-centered care) could conceivably improve patient-level outcomes for both disadvantaged and advantaged groups, improving minority health but not reducing the disparity. Indeed, an analysis of racial disparities observed in the Veterans Health Administration between 2000 and 2009 (when the organization underwent a transformation that improved quality of care) found a reduction in disparities for process-of-care measures, but no reduction in disparities in outcomes such as blood pressure, glucose, and cholesterol control.⁴

Should an outcome of improved minority health but stable or worsening disparities following the implementation of an intervention be considered a failure? From an organizational perspective, the answer to this question may depend in part on the broader context that led to the decision to implement the intervention, as well as whether factors contributing to the disparities found are within the organization’s control to modify. Our objective here is not to advocate for a stance on the answer to this question, but rather to impress upon health care organizations that it is an important question to ask before implementing interventions and program improvements because some intervention strategies might be better suited to reducing disparities, whereas others might be better suited to improving minority health. To illustrate our point, we describe some common intervention strategies below, and explain how both the context of observed disparities (i.e., the factors contributing to them) and the relative priority assigned to the distinct but related goals of “reducing health disparities” and “improving minority health” can and should influence the choice of intervention strategy.

One common intervention strategy that might be well-suited to the goal of reducing disparities involves targeting an effective program to minorities. As mentioned above, interventions to address language barriers are a classic example of such a targeted, disparity reduction strategy. Another example of a targeted, disparity reduction strategy would be to offer an effective intervention to those patients with the greatest room for improvement. If minorities are disproportionately represented among those with the worst outcomes before the intervention begins, and the intervention is equally or more effective among minorities relative to others, the program should reduce disparities. However, a targeted approach may not be appropriate for all circumstances. For instance, it may not be feasible or acceptable to restrict exposure to interventions designed to modify processes at levels other than the patient to minority groups exclusively. Although it is possible to target organizational and provider directed interventions to healthcare facilities or systems with a large minority population, it is likely that such interventions would also confer benefits to white patients in those facilities, particularly those from vulnerable patient groups.

An alternative strategy for reducing disparities might involve tailoring intervention content to enhance engagement and appeal for a particular cultural group, or to address barriers specific to this group, under the assumption that such tailoring would increase the effectiveness of the intervention for this group, thereby potentially reducing disparities. This approach could be an effective approach to reducing disparities if the primary source of variation resides at the patient level, and there is evidence of unique barriers in different racial and ethnic subgroups. However, it may have little impact on disparities if community, system, or provider level factors dominate.

Strategies that improve care quality are likely to benefit minorities but may not necessarily reduce disparities. Such strategies include implementing standardized interventions designed to reduce variation in how care is provided, or redesigning care in patient-centered ways in order to better adapt care processes to individualized patient needs.⁷ One study of medical students, which found a positive association between patient-centered attitudes and clinical performance scores among African American standardized patients but not white standardized patients, concluded that “patient-centered attitudes may be more important in improving physician behaviors with African American patients than with white patients and may, therefore, play a role in reducing disparities.”⁸ However, organizations should be aware that a possible result of implementing standardized and individualized strategies is improved outcomes for minorities but either no impact on, or widening disparities. Indeed, one health insurance plan found, after implementing a standardized diabetes intervention, that both African American and white patients showed improvement in glycemic control, but that racial disparities apparent at baseline remained unchanged a year after the program was implemented.⁹

How to prioritize candidate intervention strategies might also depend on the relative priority placed on two potentially competing perspectives. From a utilitarian perspective, an intervention that benefits white and non-white patients would be preferable to an intervention that benefits only a single group. From a moral perspective, however, the concern for equity and social justice would favor an intervention that reduced racial disparities, even if it conferred a lesser net benefit to all patients. Given the shameful history of racial injustice in the United States, there is a strong, moral imperative to reduce racial inequality in health and healthcare, which motivates many of us in this area to do the research we do. This moral imperative to reduce racial inequality perhaps underlies the decision to frame this research agenda in terms of “reducing health disparities” rather than “improving minority health,” despite the fact that the latter might encompass a broader range of outcomes that our interventions and quality improvement initiatives are designed to address.

The systematic reviews in this special symposium point to both the progress made toward healthcare equity, marked by a spate of interventions that have been effective at improving health outcomes for minority patients, as well as to the gaps that remain in the literature. We agree with Chin and colleagues that the time for action is now, and that resources need to be targeted toward helping healthcare organizations develop and successfully implement interventions to improve minority health outcomes. A difficult question, but one that we believe is important to address, is whether and under what circumstances healthcare organizations should focus on “reducing health disparities” versus “improving minority health,” particularly when there is pressure for organizations to improve the quality of care for all patients. Because the best intervention strategy will often vary by these foci, we urge healthcare organizations to explicitly address this question as they consider new improvement efforts.