A Plan for Action: Key Perspectives from the Racial/Ethnic Disparities Strategy Forum

Abstract

Context

Racial and ethnic disparities in health care in the United States have been well documented, with research largely focusing on describing the problem rather than identifying the best practices or proven strategies to address it.

Methods

In 2006, the Disparities Solutions Center convened a one-and-a-half-day Strategy Forum composed of twenty experts from the fields of racial/ethnic disparities in health care, quality improvement, implementation research, and organizational excellence, with the goal of deciding on innovative action items and adoption strategies to address disparities. The forum used the Results Based Facilitation model, and several key recommendations emerged.

Findings

The forum's participants concluded that to identify and effectively address racial/ethnic disparities in health care, health care organizations should: (1) collect race and ethnicity data on patients or enrollees in a routine and standardized fashion; (2) implement tools to measure and monitor for disparities in care; (3) develop quality improvement strategies to address disparities; (4) secure the support of leadership; (5) use incentives to address disparities; and (6) create a messaging and communication strategy for these efforts. This article also discusses these recommendations in the context of both current efforts to address racial and ethnic disparities in health care and barriers to progress.

Conclusions

The Strategy Forum's participants concluded that health care organizations needed a multifaceted plan of action to address racial and ethnic disparities in health care. Although the ideas offered are not necessarily new, the discussion of their practical development and implementation should make them more useful.

Background

Over the last decade, highlighted by the release of the Institute of Medicine (IOM) Report Unequal Treatment in 2002, the issue of racial and ethnic disparities in health care in the United States has attracted significant attention (IOM 2001; Smedley, Smith, and Nelson 2003). Unequal Treatment defined disparities as “racial or ethnic differences in the quality of healthcare that are not due to access-related factors or clinical needs, preferences, and appropriateness of intervention” (Smedley, Smith, and Nelson 2003, pp. 3–4). Even with the same insurance and socioeconomic status and when comorbidities, stage of presentation, and other confounders are controlled for, minorities often receive a lower quality of health care than do their white counterparts. The IOM's report Crossing the Quality Chasm made the case that equity is a component of quality, thus signaling that racial/ethnic disparities in health care are an important quality-of-care issue (Fiscella et al. 2000; IOM 2001).

The evidence of racial/ethnic disparities in health care was carefully detailed in Unequal Treatment, and new findings have continued to emerge. For instance, disparities have been found in the utilization of cardiac diagnostic and therapeutic procedures (such as African Americans being referred less often than whites for cardiac catheterization and bypass grafting) (Ayanian and Epstein 1991; Harris, Andrews, and Elixhauser 1997; Schulman et al. 1999; Thomas et al. 2007; Woolf et al. 2004), prescriptions of analgesia for pain control (African Americans and Latinos receiving less pain medication than whites for long-bone fractures and cancer) (Bernabei et al. 1998; Green et al. 2006; Pletcher et al. 2008), and surgical treatment of lung cancer (African Americans receiving less curative surgery than whites for non-small cell lung cancer) (Bach et al. 1999). Disparities have also been found in referrals for renal transplantation (African Americans with end-stage renal disease being referred less often than whites to the transplant list) (Ayanian et al. 1999), treatment of pneumonia and congestive heart failure (African Americans receiving less optimal care than whites when hospitalized for these conditions) (Ayanian et al. 1999), and the utilization of preventive services covered by Medicare (i.e., immunizations and mammograms) (Gornick et al. 1996). Despite the increased awareness of this issue, disparities have persisted over time in both the use of major surgical procedures among the elderly and, for example, the management of myocardial infarction (Jha et al. 2007; Vaccarino et al. 2005). Additional research has shown that racial/ethnic disparities in health care are due to differences in care provided within hospitals, as well as where minority patients receive their care (i.e., specific hospitals and providers) (Bach 2005; Bach et al. 2004; Hasnain-Wynia et al. 2007; Jha et al. 2007).

Because the root causes for racial/ethnic disparities are complex, Unequal Treatment grouped them into three basic areas:

1. Health system factors (HS) related to the complexity of the health care system and how it may be disproportionately difficult for minority patients or those with a limited proficiency in English to navigate.

2. Care process variables (CP) related to the health care provider, such as stereotyping, the impact of race/ethnicity on clinical decision making, and clinical uncertainty due to poor communication.

3. Patient-level variables (PL) resulting in patients' refusal of services, poor adherence to treatment, and delays in seeking care.

Health System Factors

Even for those people who are insured and well educated and have a high degree of health literacy, navigating the health care system can be extremely complicated and confusing (IOM 2004). Some individuals, however, may be at higher risk for receiving substandard care because of their difficulty navigating the complexities of the U.S. health care system. These persons may be those from cultures unfamiliar with the Western model of health care delivery, those with limited proficiency in English, those with low health literacy, and those who mistrust the health care system (IOM 2004; Smedley, Smith, and Nelson 2003). People with such characteristics may have difficulty knowing how and where to go for a referral to a specialist, how to prepare for a procedure like a colonoscopy, or how to follow up abnormal results from a test like a mammogram. Since people of color in the United States tend to be overrepresented among the preceding groups, the inherent complexity of navigating our health care system has been seen as a root cause for racial/ethnic disparities in health care (Smedley, Smith, and Nelson 2003).

Care Process Variables

Several care process variables may contribute to racial/ethnic disparities in health care, such as the following:

Communication between Provider and Patient

Much evidence highlights the impact on health and clinical care of sociocultural factors, race, ethnicity, and limited proficiency in English (Berger 1998). Health care professionals frequently care for patient populations with different perspectives, values, beliefs, and behaviors regarding health and well-being, such as variations in the recognition of symptoms, thresholds for seeking care, comprehension of management strategies, expectations of care (including preferences for or against diagnostic and therapeutic procedures), and adherence to preventive measures and medications. In addition, patients' and providers' sociocultural differences influence communication and clinical decision making, and are pertinent to evidence linking communication between provider and patient to greater patient satisfaction, adherence, and, subsequently, better health outcomes (Eisenberg 1979; Stewart et al. 1999). Thus, when the patient's and provider's sociocultural differences are not appreciated, explored, understood, or communicated effectively in the medical encounter, the patient's dissatisfaction, poor adherence, poorer health outcomes, and racial/ethnic disparities in care may be the result (Betancourt, Carrillo, and Green 1999).

A survey of 6,722 Americans aged eighteen and older further elucidated the link between provider-patient communication and health outcomes (Commonwealth Fund 2001). White, African American, Hispanic, and Asian Americans who had had a medical visit in the last two years were asked whether they had trouble understanding their doctor, whether they felt the doctor did not listen, and whether they had medical questions they were afraid to ask. The survey found that 19 percent of all patients experienced one or more of these problems, yet whites experienced them 16 percent of the time, compared with 23 percent of the time for African Americans, 33 percent for Hispanics, and 27 percent for Asian Americans.

In addition, provider-patient communication in the presence of a language barrier but without an interpreter is a major challenge to effective health care delivery (Erzinger 1991; Perez-Stable, Napoles-Springer, and Miramontes 1997; Seijo 1991). Research has shown that Spanish-speaking patients discharged from the emergency room are less likely than their English-speaking counterparts to understand their diagnosis, prescribed medications, special instructions, and plans for follow-up care (Crane 1997). They are less likely to be satisfied with their care or willing to return if they have a problem, more likely to report problems with their care (Carrasquillo et al. 1999), and less satisfied with the patient-provider relationship (Baker, Hayes, and Fortier 1998). Communication issues related to language barriers disproportionately affect minorities and others with limited proficiency in English and may contribute to racial/ethnic disparities in health care.

Clinical Decision Making

Theory and research on clinical decision making suggest that physicians' understanding and interpretation of information obtained from patients, as well as assumptions about the patients themselves, may contribute to racial and ethnic disparities in health care (Hooper et al. 1982; McKinlay, Potter, and Feldman 1996; van Ryn and Burke 2000). Two factors are central to this process: clinical uncertainty and stereotyping (Smedley, Smith, and Nelson 2003). A doctor's decision-making process is nested in clinical uncertainty, and doctors make inferences about severity based partly on information they obtain from the patient. Therefore, if doctors have difficulty understanding their patients, their decisions may not be the same for patients who present with the same condition. Because different racial and ethnic groups may express their symptoms differently, doctors—the overwhelming majority of whom are white—may understand less well the symptoms of patients from minority groups. The consequence may be that white patients may be treated differently than minority patients are. In this situation, clinical decisions could differ even if the doctor had the same regard for each patient (no prejudice).

Stereotyping is the process by which people use social categories (e.g., race, gender) when acquiring, processing, and recalling information about others (Fiske 1998). The literature on social cognitive theory examines the ways in which natural tendencies to stereotype may influence clinical decision making. Stereotyping can result in different treatments if providers have assumptions (conscious or unconscious) related to dimensions such as race, gender, and age (Smedley, Smith, and Nelson 2003). Stereotypes also are influenced by the messages conveyed in society (Smedley, Smith, and Nelson 2003), such as the media's portrayals of minorities as being less educated, more violent, and less likely to adhere to health care recommendations. Stereotypes can affect clinical decision making and even lead to certain groups being deemed less worthy of diagnostic or therapeutic procedures or resources (Einbinder and Schulman 2000; van Ryn and Burke 2000). Stereotypes tend to be most common in environments in which persons are stressed, multitasking, and pressured for time: the hallmarks of the clinical encounter (Macrae, Milne, and Bodenhausen 1994; Smedley, Smith, and Nelson 2003).

Patient-Level Factors

Trust is crucial to the therapeutic alliance between patient and health care provider. It enables open communication and is directly related to the patient's satisfaction and adherence to the physician's recommendations (Peterson 2002). Not surprisingly, patients who mistrust their health care providers are less satisfied with the care they receive (Thom and Campbell 1997), and their mistrust of the health care system and their physicians results in poor continuity of care, doctor shopping, self-medicating, and an increased demand for referrals and diagnostic tests (Safran et al. 1998).

Based on historical factors of discrimination, segregation, and medical experimentation, African Americans may be especially mistrustful of providers (Gamble 1997). Indeed, the exploitation of African Americans by the U.S. Public Health Service during the Tuskegee syphilis study left a legacy of mistrust that persists in this population (Brandt 1978; Corbie-Smith 1999). A national survey also found significant mistrust of the health care system by other minority populations. Of the 3,884 individuals surveyed, 36 percent of Hispanics and 35 percent of African Americans (compared with 15 percent of whites) felt they had been treated unfairly in the health care system in the past based on their race and ethnicity. Perhaps even more alarming is that 65 percent of African Americans and 58 percent of Hispanics (compared with 22 percent of whites) were afraid of being treated unfairly in the future based on their race/ethnicity (KFF 1999).

This mistrust may contribute to patients' wariness in accepting or following recommendations, undergoing invasive procedures, or participating in clinical research. This in turn may lead doctors either to assume that African Americans might not want certain procedures or to stop explaining a procedure to a patient because they feel that it is a lost cause and that these patients may adhere less strictly to or be less interested in aggressive treatments.

Although the relative contribution of these various factors has not been calculated, the overall framework can provide some direction and areas of focus for possible interventions. Given the many causes of disparities, it is clear that there are no simple solutions for addressing them, just as there are no simple solutions for improving the overall quality of health care. Strategies to address disparities will likely require a multidisciplinary, multimethod, stepwise approach.

The Strategy Forum

Most of the research on disparities describes racial/ethnic disparities in health care rather than identifying best practices or proven strategies to address them. To date, few efforts have been made to identify the critical factors for success and the key concepts of change that address disparities and can be readily transferred from one organization or environment to another.

Because of this gap and with the support of the Robert Wood Johnson Foundation (RWJF), the Disparities Solutions Center (DSC) at Massachusetts General Hospital (MGH) convened the Strategy Forum to draw up a blueprint to address racial/ethnic disparities in health care nationally. In the summer of 2006, the DSC convened this one-and-a-half-day Strategy Forum in Oakland, California, composed of twenty experts (listed in the appendix) from the fields of racial/ethnic disparities in health care, quality improvement, implementation research (the translation of research findings into clinical practice), and organizational excellence (the transformation of organizational culture to maximize performance). The goal of bringing these fields into the disparities discussion for the first time was to create a set of innovative actions and adoption strategies to resolve disparities. Insights into addressing disparities come from the fields of research on racial and ethnic disparities and on quality improvement, whereas insights into bringing about change come from research on implementation science and organizational excellence (Dijkstra et al. 2006; Grimshaw et al. 2001; Runy 2002; Ryan 2004, 2006; Spath 2004).

Participants in the Strategy Forum were asked to identify (1) activities essential to eliminating disparities, (2) proven strategies for changing providers' behavior, (3) quality improvement initiatives to reduce disparities, and (4) critical factors for success in organizational excellence that would help hospitals, health plans, and physicians' organizations address this issue effectively.

The forum was designed and facilitated using the Results Based Facilitation (RBF) model. RBF is a process that supports moving from talk to action by focusing on achieving results and developing an accountability framework for action (Pillsbury 2007). The DSC staff collected, compiled, and reviewed detailed notes from the Strategy Forum and identified key themes from the meeting. These themes were placed into six major areas: (1) collecting race and ethnicity data, (2) measuring disparities, (3) developing disparities interventions, (4) securing the support of leadership, (5) finding incentives to address disparities, and (6) creating appropriate messages and communicating efforts effectively.

We hope that our article makes the following contributions: First, it brings together the perspectives of leaders from four fields (disparities, quality improvement, implementation research, and organizational excellence) on how these areas can together address disparities; second, it provides a practical look at what is going on in the field of disparities; and finally, it explores issues of implementation, including the barriers that must be overcome to achieve success. We begin by presenting the Strategy Forum's principal themes and recommendations in the context of both current efforts to address racial and ethnic disparities in health care and barriers to progress.

Principal Themes and Recommendations

Collection of Race/Ethnicity Data

The ability of hospitals, health plans, and other health care organizations to identify and address racial/ethnic disparities hinges on their collecting information about their patients' race and ethnicity. This essential step was recommended in Unequal Treatment (IOM 2001; Smedley, Smith, and Nelson 2003) and was emphasized by the Strategy Forum's participants. A significant amount of research has been conducted on how to do this effectively, including which race/ethnicity categories should be made standard, as well as the importance of collecting information about socioeconomic status (via educational level), racial/ethnic subgroups (when relevant, such as “Puerto Rican” within the Hispanic/Latino group), and primary language (Smedley, Smith, and Nelson 2003; Swift 2002; Ver Ploeg and Perrin 2004; Williams 1996).

In the past few years, two major studies examined whether hospitals routinely collect race/ethnicity data about their patients. In 2004, of 272 hospitals surveyed (from a sample of one thousand members of the American Hospital Association), 82 percent collected data on their patients' race and ethnicity, and 67 percent collected information on their patients' primary language, but the data were not collected in a systematic or standard manner. For example, the categories of race/ethnicity and the collection methods were different (such as patients' self-reports versus clerks assigning race/ethnicity to patients based only on appearance). In addition, the data were often not shared, even among different departments within the same hospital (Hasnain-Wynia, Pierce, and Pittman 2004). In 2006, of 501 hospitals (of 1,100) that were similarly surveyed, 78.4 percent collected race information, 50 percent collected data on patient ethnicity, and 50 percent collected data on primary language (Regenstein and Sickler 2006). Again, the problem is that information about race/ethnicity was not collected using standard race/ethnicity categories or collection methods. Given these studies' limited size and response rate, it is difficult to ascertain how many hospitals collected and used race/ethnicity data, but the lack of standardization is noteworthy.

Progress has been made, however, in giving hospitals a blueprint for collecting such data, primarily through the work of the American Hospital Association's Health Resource and Educational Trust. The association's “Toolkit for Collecting Race, Ethnicity, and Primary Language from Patients” is the premier resource in the field (Hasnain-Wynia and Pierce 2005), and the hope is that more hospitals will use this device to collect race/ethnicity data. At the state level, Massachusetts mandated the collection of race/ethnicity data in hospitals as part of the disparities reduction legislation in the nationally recognized health care reform initiative that the state passed in 2007 (MGL 2007).

America's Health Insurance Plans (AHIP) conducted two surveys of member and nonmember health plans—one in 2003 and the other in 2006—to determine whether information about its members' race/ethnicity was being collected. In 2003, from the 137 health plans (of 300) surveyed, 53.5 percent of enrollees were in plans that collected race/ethnicity data, and in 2006, from the 156 health plans (of 260) surveyed, 67 percent of enrollees were in plans that collected such data (AHIP 2006). This information was obtained from the enrollees, usually during their enrollment in the plan or in special programs. Some plans collected this information indirectly through geocoding and surname analysis of their members (Fiscella and Fremont 2006). Although some progress appears to have been made, collecting race/ethnicity data has yet to become an industry standard, thus limiting health plans' ability to identify and address racial/ethnic disparities in care.

Pioneering efforts in this area have included those of the National Health Plan Collaborative, funded by the Robert Wood Johnson Foundation and the Agency for Healthcare Research and Quality (AHRQ). The eleven participating health plans (which represent nearly half of all commercially insured Americans and the millions served by Medicare and Medicaid) have worked on several issues related to disparities, including developing effective strategies to collect racial, ethnic, and language data (RWJF 2007).

Despite the progress in data collection by both health plans and hospitals, obstacles remain. Leaders often express concerns about the legality of collecting this information (Betancourt and Weissman 2006; Hassett 2005), doubts about whether it is permitted under the Health Insurance Portability and Affordability Act, concerns that patients or members will feel offended if asked, and worries about the costs of retooling information systems to achieve this goal. Uncertainty also remains about which race and ethnicity categories should be used; how, where, and when these data should be collected; what other information should be collected to avoid confounding findings about racial/ethnic disparities; what information technology infrastructure adaptations are needed; what kinds of staff training and quality assurance are necessary; and what (and to whom) information should be routinely reported. Even though the cost and information systems adjustments have not yet been resolved, research has demonstrated that there are no legal barriers to collecting race/ethnicity data, that patients will generally provide the information when asked, and that there are models for what categories to collect and when and how to collect this information (Baker et al. 2005; Hasnain-Wynia and Baker 2006; Perot and Youdelman 2001; Regenstein and Sickler 2006; Rosenbaum et al. 2006). Several organizations have also developed staff training and models of quality assurance to enhance the accuracy of the process (Betancourt and Weissman 2006; MGH CRED 2003).

The participants in the Strategy Forum agreed on the need for the national adoption of standard race/ethnicity categories, guidance for health care organizations on data acquisition strategies (including how to develop information technology to collect race/ethnicity data), training materials for staff who collect this information, and techniques for communicating to patients/members the reasons for and importance of collecting this information. Ultimately, all efforts to reduce disparities rest on the effective collection of race/ethnicity data from patients and would help advance toward that goal.

Disparities Measures and Reporting Mechanisms

Collecting race and ethnicity data is not enough to address disparities. Once such data are collected, health care organizations must analyze and review them internally to monitor for disparities. It would be helpful if hospitals, for example, could recognize any racial/ethnic differences in the care their patients were receiving for their condition. Information about race and ethnicity also could be used in outpatient settings, for example, to determine whether patients were receiving the recommended testing and appropriate medications. Disparities may be amenable to quality improvement interventions. For example, if an organization found disparities in diabetics' glucose control, it could try to determine the root causes (e.g., language barriers or limited health literacy) and use them to improve the quality of initiatives to reduce disparities.

In the 2006 survey of 501 hospitals described earlier, fewer than one in five of the hospitals that collected race/ethnicity information routinely used it to assess disparities in quality of care, health care outcomes, or patient satisfaction. Building on this theme, the Strategy Forum's participants concluded that health plans and hospitals needed guidance on monitoring for disparities in quality by race/ethnicity. This guidance would include what a routine reporting mechanism should require, including which specific quality measures should be stratified by race/ethnicity, how frequently the measures should be analyzed, and to whom the report should go (such as to a hospital or health plan quality leaders).

The capacity of health plans and hospitals to routinely measure, monitor, and report disparities has been explored several times. Perhaps the earliest examination was the “Minority Health Report Card Project,” a collaborative effort of researchers and initially eight, but later thirteen, health plans (commercial, Medicaid, and Medicare) (Nerenz et al. 2002). The demonstration created report cards, and the researchers recommended that race/ethnicity data be collected and used to measure the quality (i.e., Healthcare Effectiveness Data and Information Set [HEDIS], Consumer Assessment of Health Plans Survey [CAHPS]) of health plans and also for external reporting and internal quality purposes. Despite this work, however, health plans do not routinely report quality measures by race/ethnicity.

Similarly, researchers recently investigated public hospitals to find out whether current public reporting efforts could include data by race or ethnicity (Siegel, Regenstein, and Jones 2007). In particular, they looked at the feasibility of using the Hospital Quality Alliance framework to collect quality measures by race/ethnicity and to gauge these measures' usefulness for supporting hospitals' quality improvement activities designed to reduce disparities (Siegel, Regenstein, and Jones 2007). The Hospital Quality Alliance is a public-private collaboration (American Hospital Association, Centers for Medicare and Medicaid Services, the Federation of American Hospitals, and the Association of American Medical Colleges) to improve the quality of care provided by the nation's hospitals by measuring and publicly reporting the management of patients with congestive heart failure, acute myocardial infarction, and community-acquired pneumonia (Hospital Quality Alliance 2008). A useful lesson from this work was that race/ethnicity data could be used to measure quality in order to identify those disparities that could be addressed by quality improvement initiatives.

An example of other evidence of hospital activity in this area is Massachusetts General Hospital's work beginning in 2005 on its Disparities Dashboard, which would mirror the quarterly quality dashboard distributed to hospital leaders but would focus exclusively on disparities (http://www.mghdisparities.org/sub_quality.html). Using the principles of dashboard reporting—three to five pages of easily processed information used to convey key points to the leadership—the disparities dashboard tried to better identify the hospital's patient population by race/ethnicity and by the National Hospital Quality Measures (congestive heart failure, acute myocardial infarction, and pneumonia), HEDIS outpatient measures, and patient satisfaction measures, also by race/ethnicity. The Disparities Dashboard was first released in 2006 and has served as the foundation for disparities-specific quality improvement initiatives in diabetes and colorectal cancer screening.

If disparities are to be identified and addressed, efforts like these must become routine for both hospitals and health plans. Appropriate measures must be selected; quality information technology must be developed to generate reports easily and seamlessly; and sensitive information must be communicated carefully. The Strategy Forum's participants concluded that hospitals and health plans can benefit from guidance on developing standard monitoring systems to identify racial/ethnic disparities in quality of care.

Disparities Interventions

Investigators and health care organizations are now beginning to design and implement interventions to reduce or eliminate disparities in care, particularly using quality improvement approaches. The Strategy Forum's participants underscored the need for readily available interventions to address disparities once they are identified. The participants also discussed whether standard quality improvement tools could “lift all boats” and thereby reduce disparities, or whether interventions should concentrate on specific disparities. The consensus was that general quality improvement may help but probably is not sufficient to eliminate disparities. The participants felt that interventions should be tailored to address the particular root causes leading to disparities, whether these were at the health system (HS), care process (CP), or patient level (PL) (Smedley, Smith, and Nelson 2003), rather than using a “one size fits all” approach that assumed that the same barriers applied to all patients. The number of studies examining the impact on disparities of general quality improvement efforts that are not specifically tailored to reduce disparities has been growing (Beach et al. 2006; Casalino 2005; Jha et al. 2007; Miranda et al. 2003). These efforts range from information systems–based interventions to improve the quality of diabetes treatment (Sequist et al. 2006) to performance feedback and education for clinicians to improve dialysis care (Sehgal 2003). Most of these studies show mixed results, with fewer disparities for some quality indicators but the same or more for others.

If a general improvement in quality is not enough to eliminate racial/ethnic disparities, how should we intervene? Many studies show that disease-specific interventions tailored to a particular racial/ethnic group can improve process and outcomes measures of quality (La Roche, Koinis-Mitchell, and Gualdron 2006; Whittemore 2007). Quality improvement efforts could target specific disparities by several means (whose root causes we will identify using the abbreviations HS, CP, or PL).

They could

1. Target an intervention to specific minorities who have been receiving lower-quality care, in either heavily minority communities or integrated communities (using race/ethnicity data to identify patients eligible for the intervention). For example, the Multi-Family Asthma Group Treatment (MAGT) program was designed to enhance asthma management by acknowledging, discussing, and incorporating families' cultural beliefs about asthma; targeted inner-city African American and Latino families; and significantly reduced asthma-related emergency visits one year after the intervention (La Roche, Koinis-Mitchell, and Gualdron 2006). This program addressed both the care process variables through improved communication and the patient-level variables through education and trust building (CP, PL). Another example is Aetna's diabetes disease-management pilot program directed toward its African American and Latino policyholders (HS, CP). The preliminary results show that this intervention improved the rates of LDL cholesterol screening and glycosolated hemoglobin testing for pilot enrollees (National Health Plan Collaborative 2006).

2. Target all patients but use approaches (materials, personnel, etc.) that are culturally and linguistically appropriate to minority patients and thus accommodate their needs or address/navigate those barriers that apply specifically to them. For example, a quality improvement intervention for the care of depressed patients used local experts to educate clinicians, mental health nurses, and psychotherapists, who then made modest changes in their usual practices to improve their care of minorities (Miranda et al. 2003). These changes were involving experts in minority mental health, translating patients' informational materials into Spanish, and including minority providers in videotapes for patients and information about cultural beliefs and barriers to mental health in training materials for providers. The results in a randomized trial showed better rates of appropriate care for all patients and better clinical outcomes for ethnic minorities (HS, CP).

3. Improve providers' ability to care for minority patients. Interventions by providers at the care process level (CP) include the use of physician extenders, such as having trained and culturally and linguistically competent nurses offer screening to patients. Another approach at the CP level is using training programs to educate clinicians on providing more culturally competent care. These programs could also address the effect of biases and stereotypes in clinical decision making (Green et al. 2007; van Ryn and Fu 2003). Such programs can improve the quality of care for racial/ethnic minority patients, as some studies are beginning to show (Beach et al. 2006).

Several other strategies have shown promise for reducing disparities. Health care navigators or promoters, often community members trained to serve in this capacity, can help identify minority patients and overcome barriers to their care. These strategies can address CP and PL but especially HS issues and have been helpful for cancer care, among other conditions (Frelix et al. 1999; Jandorf et al. 2005), and have improved health outcomes for all patients, not just minorities (Bradford, Coleman, and Cunningham 2007). Culturally competent disease management strategies are similar but often emphasize telephonic contact (HS, CP). Electronic medical records can help direct interventions to minority patients at risk (Fiscella and Geiger 2006). For example, evidence has shown that tracking/reminder systems for providers are effective in improving the quality of care for racial/ethnic minority patients (HS) (Beach et al. 2006). One study found that using electronic medical records to focus outreach and deliver clinical reminders improved some aspects of diabetes care (HS) (Sequist et al. 2006). The forum's participants recognized that the literature on proven strategies to address disparities is limited but growing and also that there is a clear need for more research, innovation, evaluation, and dissemination in this area. The foregoing examples offer some general guidance about how health care organizations might use quality improvement interventions to address racial and ethnic disparities in care.

Moving to Action

We have summarized the Strategy Forum participants' ideas about what hospitals and health plans can do to identify and address disparities. Next we turn to how to adopt these approaches by securing the support of leadership, developing incentives, and other possible approaches. These ideas come primarily from the fields of implementation research (incentives to change behaviors) and organizational excellence (securing support).

Securing the Support of Leadership

Both Crossing the Quality Chasm (IOM 2001) and Unequal Treatment (Smedley, Smith, and Nelson 2003) emphasize the importance of addressing racial/ethnic disparities in health care by improving quality. The Strategy Forum's participants concluded that eliminating disparities was not a priority for many health care leaders across the country and that it was necessary to develop strategies and tools to help them become interested in doing something about this. For example, an organizational assessment survey distributed to approximately 150 to 200 executives, physicians, and hospital managers at ten hospitals showed that only 3 percent of executives agreed or strongly agreed that disparities in health care were a major problem in their hospital and that 85 percent disagreed or strongly disagreed that disparities were a major problem (McAlearney, personal communication, Ohio State University, 2006).

The forum's participants thus cited the need to (1) make the “case” (business, quality, safety, etc.) for addressing disparities and (2) provide incentives to raise the issue's importance. The health care literature offers a few examples of such approaches. One that has attracted attention is Aetna's disparities agenda (Betancourt and Weissman 2006; Hassett 2005). Aetna's former chairman and CEO, Dr. John W. Rowe, showed that change is possible when leaders decide to address disparities (Hassett 2005). Under his leadership, Aetna formed its Task Force on Racial and Ethnic Disparities in Health Care and became the first major health plan to collect race/ethnicity data from its members. Aetna also required cultural competency training for medical directors and case managers and created health education tools and disease management programs for its members (Hassett 2005). Aetna was a pioneer in its field.

The city of Boston's “Disparities Project” (BPHC 2006), which was instigated by the mayor of Boston, Thomas Menino, included the City Data Report on Disparities, the Task Force Report on Addressing Disparities in Health Care, and the Hospital Working Group Report on Addressing Disparities in Health Care. In September 2005, Mayor Menino announced that $1 million in grant funds would be distributed to more than thirty organizations to carry out the report's recommendations and help eliminate racial and ethnic disparities in health and health care.

Other efforts to secure support from employers by making the business case to address disparities include the National Business Group on Health (NBGH), a national organization that represents more than 175 Fortune 500 companies, which launched its Health Disparities Initiative in 2001. The initiative cites two major financial incentives—the possibility of lower direct and indirect health care costs, as well as the retention in the workforce of racial and ethnic minority employees—as critical to making the business case for addressing disparities (Weinstock 2003). The Office of Minority Health at the U.S. Department of Health and Human Services and the National Business Group on Health recently announced a two-year, $300,000 project that seeks to reduce racial and ethnic disparities in health care for minorities enrolled in employer-sponsored health plans. The project will educate employers about disparities in health care for minorities and will recommend steps to address the issue. Such steps include the use of health plans with networks of ethnically representative providers and the establishment of on-site health clinics for minorities, many of whom tend not to visit physicians (KFF 2008).

In summary, the overarching concern expressed by the Strategy Forum's participants was that if the leaders of hospitals and health plans are neither aware of racial/ethnic disparities in health care nor concerned about addressing them, progress will be minimal. Although several opinion leaders on this topic have emerged, there remains a need both to raise the visibility of racial/ethnic disparities as a quality problem and to create the “case” for why leaders should identify and address them.

Incentives for Addressing Disparities

To eliminate racial and ethnic disparities in health care, providers and leaders must feel a sense of urgency to act. The Strategy Forum's participants recognized that this sense could be driven by either a “carrot” (incentives) or a “stick” (regulation) approach, although most favored the former. They cited several sorts of incentives and regulations that might help eliminate disparities, such as pay-for-performance (P4P) mechanisms, public reporting, and awards programs, as well as actions by payers and accreditation standards. Although several of these approaches are being tried today, it is too soon to draw any conclusions about their impact (MGL 2007; NCQA 2006; Weinick 2007; Wilson-Stronks and Galvez 2007).

During the past decade, public and private purchasers of health care have begun implementing P4P programs to encourage physicians and hospitals to invest in improving quality (Casalino and Elster 2007). A variety of concerns have been raised, however, about the possibility that pay-for-performance quality incentives could end up increasing racial and ethnic disparities in care (Weinick 2007). For example, if the incentives were based on specific quality targets (e.g., 80 percent of diabetic patients have a HbA1c level below 9.0), then physicians who care for low-income minorities with greater barriers to care are likely to be less successful and miss out on financial incentives, which in turn could lead them to avoid caring for minorities. But if designed carefully, P4P programs could also do the opposite. For example, they could reward quality improvement rather than specific quality targets, or they could adjust for minority, lower-socioeconomic-status patients. They could even specifically reward the elimination of measured disparities. Work is currently being done to better understand the complexity of such incentives (Weinick 2007).

One of the more interesting efforts in this regard is being conducted by the state of Massachusetts (MGL 2007). In 2006, health care reform legislation was enacted that included a provision to make Medicaid hospital rate increases contingent on quality measures, such as those measuring the reduction of racial and ethnic disparities in health care (Duchon and Smith 2006).

To date, we know of no other pay-for-performance programs that have measured the reduction of racial and ethnic disparities in health care (Weinick 2007). In May 2007, the Medicaid director for the Commonwealth of Massachusetts outlined the approach to the pay-for-performance program that is currently being considered. This program would base incentives on five areas, one of which is racial and ethnic disparities in care. For the first year, only measures related to the Culturally and Linguistically Appropriate Services (CLAS) standards are included. For the second year, the plan will report “clinical measures in four categories relevant to the Medicaid population for hospitals with sufficient volume of patient race/ethnic mix” (Weinick 2007, p. 5).

Awards programs have been used primarily to recognize hospitals or health plans that have successfully implemented various quality improvement interventions. The Strategy Forum's participants thought that this approach could help find new ways to address disparities. Awards could help identify potential models for replication in other settings and provide a forum to recognize high-performing organizations, thereby demonstrating the importance and feasibility of these efforts. One example is the National Committee for Quality Assurance's (NCQA) Recognizing Innovation in Multicultural Health Care award program, which recognizes health plans for exemplary efforts and demonstrates effectiveness in promoting cultural competence and addressing the health care needs of diverse members (NCQA 2006). Through these awards, NCQA hopes to advance efforts to improve multicultural health care and eliminate health care disparities. The Baldrige Awards program was discussed as a possible model for performance excellence, as its criteria are accepted widely not only in the United States but also around the world. The criteria are designed to help organizations enhance their competitiveness by focusing on two goals: delivering growing value to customers and improving overall organizational performance (Ryan 2004, 2006; Spath 2004). Disparities activities could be made part of the awards criteria, or a similar award could be created for organizations that innovate in this area.

The “stick” approach to eliminate disparities is less clearly defined. The Joint Commission has hospital accreditation standards related to Culturally and Linguistically Appropriate Services (CLAS), but these currently focus on language barriers and interpretation rather than on recognizing and addressing disparities in care (The Joint Commission 2007). In the future, these may include culturally competent care and disparities (Wilson-Stronks and Galvez 2007). Similarly, while public reporting on disparities is still a sensitive subject, it may follow the path of quality reporting and eventually become an important impetus for hospitals and other health care organizations to provide equitable care. While discrimination in health care based on race or ethnicity or otherwise is clearly a violation of the U.S. Civil Rights Act, the workshop's participants did not specifically identify legal action as a promising strategy for eliminating disparities, probably because of the difficulty in proving discrimination in a court of law and in linking an outcome to discrimination by race/ethnicity (Rosenbaum and Teitelbaum 2005). Title VI of the Civil Rights Act has, however, been used to press charges against certain health care organizations when, owing to the lack of access to an interpreter, a poor health outcome resulted (Chen, Youdelman, and Brooks 2007).

In a market-driven health care environment, incentives for organizations to reduce racial/ethnic disparities are a practical approach to the problem. Pay-for-performance mechanisms and awards programs are two promising strategies. While accrediting bodies such as the Joint Commission could play a larger role in this area, this may still be a few years away.

Messages and Communication

The final issue that emerged during the Strategy Forum was the importance of messages and communication about disparities efforts. This concept had two major components. First, the forum's participants felt that those hospitals and health plans that were interested in identifying and addressing disparities needed to communicate internally to their management, providers, staff, and patients why this was being done and what they could do to become partners in the process. For example, if a hospital made it known publicly that it was identifying and addressing disparities, might patients assume that there were disparities and thus be concerned about being treated differently? Might they become more hesitant to provide their race/ethnicity? Might providers be concerned that disparities efforts might in some way implicate them as culprits in the process? Might this initiative make patients suspicious of their health care providers? For all these reasons, the participants suggested that the goals of any disparities effort be communicated in a thoughtful way to a broad internal audience, without alienating anyone or creating unnecessary suspicions. In an attempt to address these concerns, Massachusetts General Hospital (MGH) designed several posters to communicate its disparity-reduction efforts internally (http://www.mghdisparities.org/sub_edaware.html). The posters acknowledge the existence of disparities, highlight MGH's commitment to identifying and addressing them, and provide an overview of activities in this area. Similarly, the Boston Public Health Commission distributed posters and leaflets in several languages that all hospitals could use to explain to patients why hospitals were required to collect race/ethnicity information (BPHC 2006).

Second, participants raised the idea of a media strategy that would increase the country's awareness externally of racial/ethnic disparities in health care among health care providers and the public. Because health care providers are essential to the effort to eliminate disparities, they need to be familiar with the issue, including the relevant evidence. Moreover, informing the public of the issue might create more informed and more active health care consumers. To address this need, the Henry J. Kaiser Family Foundation (KFF) and the Robert Wood Johnson Foundation (RWJF), in partnership with the nation's leading heart organizations, began using the mainstream media and the Internet to raise physicians' and the public's awareness of disparities in cardiac care (KFF 2002). To address physicians' awareness in particular, the KFF launched its Initiative to Engage Physicians in Dialogue about Racial/Ethnic Disparities in Medical Care in 2002. This initiative centered on a media campaign involving a partnership of thirteen medical organizations and had three components: an advertisement campaign in major medical publications, a review of the evidence regarding racial/ethnic disparities in cardiac care, and an outreach effort to engage physicians in dialogue.

Similarly, the National Business Group on Health, with the support of the RWJF, implemented a communications plan that featured a two-phase radio tour and media campaign for the Business Group's Health Disparities Initiative: Promoting Health for a Culturally Diverse Workforce. The goal was to give large employers information and practical solutions to address racial and ethnic disparities in health and health care.

While it is difficult to evaluate the impact of messages, communication, and public awareness efforts, there is no doubt that these strategies are important to any change process. There are few examples of how hospitals and health plans communicate why and how they are doing this work to their workforce, patients, and members, which remains a pertinent and necessary area for development.

Conclusion

Since the release of the IOM Report Unequal Treatment, the issue of racial/ethnic disparities in health care has come to be recognized as a quality-of-care issue. Research has documented that minorities have a higher mortality for conditions that are amenable to surgery (Bach et al. 1999), more missed health promotion/disease prevention opportunities (American Cancer Society 2006; Goodman et al. 2006; Walsh et al. 2004), more tests in emergency rooms when they have limited proficiency in English (Goldman, Amin, and McPherson 2006; Hampers et al. 1999; Harris, Andrews, and Elixhauser 1997; Waxman and Levitt 2000), higher rates of medical errors, and longer lengths of stay for the same clinical condition (Divi et al. 2007; John-Baptiste et al. 2004). To date, the disparities agenda has been heavy on research to define the problem but light on effective, real-world strategies to address it.

While the recommendations of Unequal Treatment provided some overarching principles for what is needed to address disparities, experts must be enlisted to provide a real-world context and a blueprint for action. This was the goal of the Strategy Forum. Efforts to address racial/ethnic disparities in health care depend on both the collection of patient race/ethnicity data by hospitals and health plans and the use of those data to identify differences in quality. There has been significant progress on the former but less progress on the latter. Research has shown that few hospitals and health plans both collect race/ethnicity information and examine it with an eye to finding disparities. Identifying areas where improvement is needed leads to the next important point in this process: developing interventions. Proven interventions to address disparities are few, and a debate continues about whether general quality improvement strategies can eliminate disparities, or whether interventions need to be tailored to the particular disadvantaged population.

The Strategy Forum's participants concluded that there was a critical need for a multifaceted plan of action to address racial and ethnic disparities in health care. They discussed a variety of approaches from the fields of implementation research and organizational excellence. Foremost among these was the need to secure the support of leadership, develop incentives to address disparities, and generate awareness and communicate disparities reduction efforts in a way that does not alienate the main actors. A review of the current context reveals some random progress, but we still do not have a strategy that takes into account the barriers to achieving equity in health care. Ultimately, the challenge lies not only in the development of strategies that would eliminate disparities but also in the difficult and often time-consuming process of persuading health care organizations across the country to adopt these strategies.

Appendix

David Baker, MD, MPH. Chief, Division of General Internal Medicine, Feinberg School of Medicine; Associate Director of Research, Institute for Healthcare Studies, Northwestern University.

Judy Bigby, MD. Massachusetts Secretary of Health and Human Services. Formerly, Director, Office for Women, Family and Community Programs, Brigham and Women's Hospital; Associate Professor of Medicine, Harvard Medical School.

John Billings, JD. Associate Professor, Robert F. Wagner Graduate School of Public Service at New York University; Director, Center for Health and Public Service Research.

Emilio Carrillo, MD, MPH. Vice President, Community Health Development of the New York–Presbyterian Hospital; President and Chief Medical Officer, New York–Presbyterian Community Health Plan; Director, Institute of Healthcare Access Research; Associate Professor of Clinical Public Health and Clinical Medicine, Weill Medical College of Cornell University.

Arthur Chen, MD. Chief Medical Officer, Alameda (Calif.) Alliance for Health.

Linda Chin, MS. Senior Vice President, Planning, Marketing, and Public Relations, Cambridge (Mass.) Health Alliance.

Marshall Chin, MD, MPH, FACP. Associate Professor of Medicine, University of Chicago School of Medicine; Co-Director, General Internal Medicine Research; Director, RWJF Finding Answers: Disparities Research for Change National Program Office; Co-Director, Chicago NIDDK Diabetes Research and Training Center; Associate Director, Robert Wood Johnson Clinical Scholars Program; Research Associate, Center for Health Administration Studies and the Center on the Demography and Economics of Aging, University of Chicago.

Crystal L. Clark, MD, MPH. Assistant Vice President of Corporate Quality Improvement, New York City Health and Hospital Corporation. Formerly, George W. Merck Fellow, Institute for Healthcare Improvement; Director of Quality Management, Elmhurst (N.Y.) Hospital Center.

Marilyn French Hubbard, PhD. Corporate Vice President of Organization and Community Partnerships, Henry Ford Health System.

Don Goldmann, MD. Senior Vice President, Institute for Healthcare Improvement.

Kathleen Goonan, MD. Executive Director, MGH Center for Performance Excellence.

Eli Korner, PharmD, MPH. Medical Liaison Virology, Scientific Field Operations, Medical Affairs, Roche Laboratories, Inc., Denver. Formerly, Research Investigator, Clinical Research Unit, Kaiser Permanente Colorado; Clinical Assistant Professor, University of Colorado School of Pharmacy.

Ann McAlearney, ScD, MS. Assistant Professor, Division of Health Services Management and Policy, Ohio State University School of Public Health.

Brian Mittman, PhD. Senior Social Scientist, VA/UCLA/RAND Center for the Study of Healthcare Provider Behavior; Formerly, Visiting Professor, Department of Health Services, UCLA School of Public Health.

David Nerenz, PhD. Director, Center for Health Services Research, Henry Ford Health System.

Cheryl Pegus, MD, MPH. Chief Medical Officer, LifeScan USA, a Johnson & Johnson Company. Formerly, Head, Clinical Products, Aetna Medical Product Business Unit.

Joan Quinlan, MPA. Director, Community Benefits, Massachusetts General Hospital.

Bruce Siegel, MD, MPH. Research Professor, Department of Health Policy, George Washington University Medical Center School of Public Health and Health Services.

Kurt C. Stange, MD, PhD. Gertrude Donnelly Hess MD Professor of Oncology Research; Professor of Family Medicine, Epidemiology and Biostatistics, Oncology and Sociology, Case Western Reserve University; Associate Director, Prevention, Control and Population Research, Case Comprehensive Cancer Center.

Michelle Tierney, MPA, SPHR. Director, Organizational Development, Southcentral Foundation, Anchorage.