Racial Disparities in Care: Looking Beyond the Clinical Encounter

In the February issue of this year, Jose Escarce contributed a thoughtful editorial column titled “How Does Race Matter, Anyway?” In it he discussed the many ways in which miscommunication between white physicians and minority patients can occur, potentially leading to disparities in care. The clinical encounter is, of course, a key component in the delivery of care, and understanding what factors influence that encounter is essential to mapping the sources and dynamics of racial disparities in care.

Escarce (2005) provides a conceptual bridge between economic and psychosocial understandings of how race matters in the clinical encounter, building upon the 2003 IOM report Unequal Treatment (Institute of Medicine, 2003) . The IOM report provided insight on three mechanisms that can influence providers' cognitive processing within the clinical encounter: stereotyping, prejudice, and clinical uncertainty. Although these three mechanisms can apply to a wide variety of social characteristics, the focus here is on racial and ethnic status. Stereotyping is defined as the use of social categories (such as race or gender) to acquire, organize (and infer), and process information about an individual. Prejudice is defined as the holding of a negative attitude (or stereotype) about members of a different social group such as a racial or ethnic group. And uncertainty in the clinical arena refers to the types of problems a clinician encounters in the process of cognitively organizing information and/or expectations about a patient presenting with particular problems and in a particular setting. Two types of uncertainty related to race and ethnicity are particularly germane: (1) physicians may only have “noisy” or imprecise indicators of the patient's clinical condition, perhaps because of imperfect diagnostic tests and (2) white physicians may routinely have noisy indicators of clinical conditions for nonwhite patients, because of miscommunication or miscues because of cultural and/or language differences. In each case, the clinician is forced to rely upon previously established (or assumed) probabilities (e.g., disease prevalence) or poorly communicated preferences or symptoms in forming their diagnostic and treatment decisions. Under both uncertainty conditions, it is likely that care recommendations would be less than maximally optimal for nonwhite patients, which may in turn lead to nonwhites being less inclined to seek follow-up care from their physician or to comply with the treatment plan. Racially contingent decision making resulting from either type of uncertainty is referred to as statistical discrimination (Balsa and McGuire 2003, Balsa, McGuire, and Meredith 2005).

These different mechanisms affecting disparities in care resulting from the clinical encounter are brought into either direct or indirect focus in two papers in this issue of Health Services Research. We can see them more clearly in the analysis of “Patient and Provider Assessments of Adherence and the Sources of Disparities: Evidence from Diabetes Care” by Karen Lutfey and Jonathan Ketcham (2005). These authors examine patient perceptions of their adherence to diabetes management plans, compared with their provider's perceptions of their adherence, using data on 156 patient–provider dyads from two endocrinology clinics in the same university-based medical center. Measures focused on the difference between patient self-ratings and providers' ratings (whether provider assessments are systematically above or below patients' assessments), and the absolute value of that difference (distance between assessments). Two findings stand out. First, patients' self-evaluations of diabetes adherence vary little by race, gender, or age; however, providers' assessments of patient adherence do vary by race (provider assessments of black patients are nearly 1.2 points on a 0–10 point scale below their average assessment of white patients). Second, the absolute value of the discrepancy between patient and provider assessments is significantly greater for black patients than white patients. These findings support the notion that a clinician's cognitive processing (at least in terms of perceptions of patient adherence to treatment regimens) might involve prejudice toward minority patients, and that providers might have greater uncertainty about the adherence of their minority patients to recommended treatment. One possible explanation for that uncertainty is the presumed miscommunication or discordance in expectations between minority patients and their providers from a different racial or ethnic group. The providers' race or ethnicity was not specifically addressed in this study (or in the following study), so we are left to speculate whether minority patients were more likely to be treated by a provider from a different race/ethnic group, thereby at least increasing their exposure to miscommunication.

The research reported by Stefanie Mollborn, Irena Stepanikova, and Karen Cook (2005) looks closely at the underlying relationship between patients and physicians, focusing on one important dimension: fiduciary trust (the patient's belief that his/her physician will act in the patient's best interests). Fiduciary trust has been linked to better communication, higher patient satisfaction, and high compliance rates. Mollborn et al. hypothesized that higher levels of fiduciary trust should be associated with fewer delays in seeking care and fewer unmet care needs, but they also examined whether and how race, lower income, and lack of private insurance might intervene to weaken the relationship between fiduciary trust and seeking care. Their analyses used a sample of patients who all have a regular source of care (N =29,994 adults from the Community Tracking Study Household Survey). Among this sample of adults with regular physicians, “… there is no evident racial/ethnic disadvantage in getting prompt care … Our findings imply that once blacks and Hispanics gain access to the health care system and become regular users, they are able to get prompt care when needed” (2005, p. 1911). This is an important distinction: it does not say that there are no disparities in access to care; in fact, their finding that there were no significant differences in care delays or levels of unmet care needs within disadvantaged groups (blacks and Hispanics) among adults with established provider relationships, serves to underscore the importance of understanding the obstacles encountered in establishing regular sources of care for nonwhites. To do that, we need to look beyond the physician–patient relationship and the single clinical encounter: we need to consider the larger systems within which access to care is structured, and the pathways through which racial/ethnic disparities have been institutionalized, both in the past and currently.

Fortunately, this year we have seen a number of important contributions to the literature on racial disparities in care, including the Special Issue of Health Affairs (March/April 2005), a social history of segregated hospitals in Mississippi from David Barton Smith (“The Politics of Racial Disparities,”The Milbank Quarterly, Vol. 83, No. 2), and growing interest in disparities in access to long-term care, disparities in the quality of care received by whites and nonwhites within long-term care settings, and the connections between racial residential segregation, racial differences in disease, and segregated access to high quality health care organizations (Mor et al. 2004; Williams and Jackson, 2005). It is worthwhile to briefly review that work.

Racial and ethnic disparities in the medical treatment of ambulatory, hospitalized and long-term care patients have been frequently documented (Smith 1999; Epstein and Ayanian 2001; Bach et al. 2004). Hospital care in the U.S. remained largely “separate and unequal” (Smith 2005a, b) through the mid-1960s. Smith recounts the history of hospital desegregation efforts in Mississippi (Smith 2005b), which were strongly tied to Civil Rights Era efforts to desegregate Veterans Administration hospitals, medical school hospitals, and private nonprofit facilities, using the threat of lost federal dollars. He has carefully distinguished among efforts for (1) ending Jim Crow era practices of segregating patients by race, (2) “… eliminating more subtle forms of segregation shaped by physician referral practices, insurance status, and residential location; and (3) assuring that once racial and ethnic minorities had access to the same health care settings as whites, they would receive nondiscriminatory treatment” (2005a, p. 317).

Outright segregation via explicit Jim Crow policies has disappeared, thanks to the federal government's efforts to tie the end of discriminatory practices to the flow of federal dollars through the implementation of Medicare. However, it is not clear how successful we have been in eliminating more subtle forms of segregation and in assuring equal quality care. Smith reminds us that there has never been a federal effort to collect the type of data needed to monitor the extent of discriminatory treatment on a parallel with such monitoring efforts in housing and employment. Further, current efforts to reduce payments for Medicare and Medicaid patients can readily translate into hospital action to increase their proportion of privately insured patients and reduce their admission of Medicaid and uninsured patients, which then contributes to the expansion of services in more affluent (and predominantly white) suburban communities, and the contraction of services in poorer areas.

The case of long-term care, especially nursing home care, is also instructive. In the past 20 years, long-term care facilities, particularly nursing homes, have become more central in providing a broad range of both acute and rehabilitation care functions, as elderly patients are discharged more quickly from hospitals and nursing home acuity levels have increased.

Nursing home care, or any institutionalized form of nonacute care, is significantly different from the typical clinical encounter in either a doctor's office or an ambulatory care clinic. The care received within nursing homes is managed primarily by nurses and nursing aids, with only infrequent input from physicians. Many primary care physicians discontinue their active care of elderly patients once the nursing home is entered, and it is unclear how frequently any physician presence is felt within nursing homes. Using Online Survey Certification and Reporting data (OSCAR) from 2000, 22 percent of all surveyed nursing homes reported they did not have a medical director, and the rate of medical director turnover appears to be high: between 1992 and 2001, over 25 percent of nursing homes reported at least three instance of medical director vacancies (personal communication with Orna Intrator; see also Feng et al. 2005).

Nursing homes are similar to total institutions (Goffman 1961), in that the organization controls most service inputs and some environmental inputs, as well as discharge processes. But unlike the ideal type of a total institution, it is possible for family members to be actively involved in the placement decision, in the monitoring of ongoing nursing and medical care of the elder family member, and at times be the primary source of social, psychological, and emotional support during the nursing home stay. Unfortunately, both access to nursing homes and the quality of care received by elderly patients in nursing homes can still be described as “separate…” (i.e., nursing homes tend to be either well financed and located in resource-rich areas, or poorly equipped and located in poor communities) “and unequal” (i.e., poorer homes tend to have lower quality of care processes and outcomes; Mor et al. 2004).

The meaning of “unequal” nursing home care goes beyond variation in resource levels; minority patients tend to be overrepresented in poorer quality homes. Recent research suggests that African Americans residing in nursing homes were nearly four times as likely to reside in a home with limited resources and historically poor performance than were white patients (Mor et al. 2004).

As with the acute and primary care sectors, there is reason to believe that nursing home segregation is related to the degree of residential segregation within the community, and such patterns would be exacerbated by disparities in insurance coverage. Both African American and Latino nursing home residents in the Chicago area tended to replicate the residential housing patterns of their communities and used nursing facilities within their communities. Latino residents were particularly disadvantaged in that Latino communities in Chicago had the fewest nursing home beds (Reed and Andes 2001). Nonetheless, in many other regions a nursing home's racial mix is often substantially different from the racial mix of the community in which it resides; we have found that in communities with small proportions of African Americans there were still nursing homes with high proportions of African American residents; conversely, in communities with large proportions of African Americans, nursing homes could be found with no African American residents (Fennell et al. 2000). Additionally, research examining hospital discharges to nursing homes found that nonwhite and poorly educated patients (who had not been in a nursing home in the past year) were much more likely to be placed in the worst quartile of nursing home in the local hospital market (Angelelli et al. 2005). Minority status (black and other minority) and poor education each increased by about 30 percent the risk of placement in a poor quality home, possibly because minorities lived in the poor neighborhoods in which poor quality providers were located.

At the same time, a dramatic and largely unexplored shift has taken place in patterns of racial and ethnic use of nursing homes. Using National Nursing Home Survey estimates, analysts compared age adjusted nursing home rates for those 65 years and older, between 1974 and 1999; rates for whites declined 32 percent during that time period, while African American resident rates increased 50 percent (National Center for Health Statistics, 2003). In 1974, the black elderly used nursing homes at less than half the rate of whites, while in 1999 their rates were 33 percent higher than whites. That expanded use has tended to concentrate African Americans in “lower tier” facilities that have lower staffing levels and are more likely to face Medicaid program termination.

Further, the historic pattern of Hispanic nonuse of nursing homes and other types of institutional LTC could be changing in the near future. Demographic shifts in the elderly population, and among immigrant Hispanic elders in particular, suggest that the role of formal long-term care may become more important in the near future, given increased labor force participation of adult Latina daughters (Angel et al. 2003). Angel and Hogan (2004) also projected (using Census projections) that by 2050 the number (and percent) of the elderly population who are minority (black, Hispanic, or Asian) will rise dramatically from 12.3 percent in 2000 to 18.3 percent in 2050. The implications of this important compositional shift (including the gender composition and expected health status of future minority elders) for long-term care policy and the provision of long term care services are substantial.

Beyond the clinical encounter then, the patterns of disparity in care access and care quality found in larger organizational settings stem from many disparate, and poorly understood, processes. Many of these relate to larger community-wide patterns of residential segregation, or to differences in organizational resources and staffing levels and their impact on care quality, or to physician referral practices that subtly differentiate between wealthy and poor patients (and between majority and minority patients), or to family preferences for care location and treatments that may be intertwined with cultural preferences.

At present, there is growing evidence about in- and outpatient care, but the extent to which these diverse factors affect pathways to long-term care for the individual, and how they might interact, is simply unknown. Beyond the boundaries of any particular health care organization, however, are a myriad of environmental factors that can also shape access and influence the health care organizations' pursuit of market resources and clients, such as the nature of the local market for acute or long-term care, variation in state regulatory policies, and the extent to which any particular organizational provider is embedded within complex ownership or management relationships with either corporate or diversified health care systems.

This editorial began by examining individual processes of stereotyping and prejudice that could affect the outcomes of any particular clinical encounter. Then attention was shifted away from the traditional dyad of the physician–patient clinical encounter, to focus on larger organizational and community or market structures that can influence disparities in care access and care quality.

There is yet another concept worth examining, the concept of cultural competence, for at least two reasons.

First, there is a growing consensus that it can influence care delivery at multiple levels, from the dyadic clinical encounter through the market-based business imperative to appeal to an increasingly diverse workforce. As described by Betancourt et al. (2005), the goal of cultural competence is “… to create a health care system and workforce that are capable of delivering the highest quality care to every patient regardless of race, ethnicity, culture, or language proficiency” (p. 499).

The concept covers a variety of strategies that affect communication between providers and patients, as well as possibly influencing levels of quality of care at the organization and health care system levels. The training of medical and health professionals has been influenced by the promulgation of cultural competence standards by the Accreditation Council of Graduate Medical Education (ACGME 2001). Managed care plans show signs of adopting initiatives that include the development of culturally competent disease management programs, internal employee diversity training, and linguistic training and/or interpreter services (Brach et al. 2005).

The second reason for taking a much closer look at cultural competence in the health care organizational setting and its increasingly widespread adoption in various forms is that its growing appeal is not currently linked to “hard evidence” that such training or special programming actually will increase the bottom line, or produce more culturally sensitive clinicians, or eliminate disparities in access or quality of care. In fact, as noted by Betancourt et al. in their exploratory interviews with academics, managed care, and government experts, there is recognition that “disparities are the result of many factors and that cultural competence alone could not address the problem.” Nonetheless, it would appear that cultural competence is being linked with quality improvement models in many managed care plans. One explanation is that it may be attractive to various care plans as a way to signal to potential clients and workers that the large, impersonal diversified healthcare system is sensitive and responsive to the needs of an increasingly diverse patient population.

In the words of institutional theorists, the symbolic value of cultural competence programs may be far-outstripping their actual effectiveness. We should probably be skeptical of any given remedy for combating disparities in health care, which can be applied to the clinical encounter, the organizational setting and the community within which segregated sources of care are found. Even more important, however, is the need to develop the databases that would allow for a careful untangling of the various mechanisms that lead to unequal access to care, to variations in the quality of care received by majority and minority patients, and to disparate tiers of high and low quality care settings. And finally, we need more information about how to redress inappropriate disparities that arise from prejudice and stereotyping while recognizing that cultural differences may occasionally lead to informed differences in care.