Health Care Disparities and Diabetes Care: Practical Considerations for Primary Care Providers

Ever since the Institute of Medicine issued its 2002 report titled, “Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care,”¹  there has been a steady increase in the awareness, measurement, and documentation of disparate health care trends across America. Health care disparities have been defined as, “[differences] in the overall rate of disease incidence, prevalence, morbidity, mortality, or survival rates in [a given subpopulation] as compared to the health status of the general population.”²  Health care disparities can be broadly classified as either differences in the quality of care received or in access to care. In general, these disparities encompass a broad spectrum of disease processes and result from a complex interplay of multiple factors.

Despite some recent successes in narrowing disparate gaps in care among select populations,³  the reduction or elimination of most health disparities has remained elusive to providers, researchers, and policy-makers. The Agency for Healthcare Research and Quality (AHRQ) has recognized several priority groups as targets for addressing disparities. These include racial and ethnic minorities, women, children, low-income groups, the elderly, residents of rural areas, and individuals with disabilities.⁴ 

Diabetes care represents an important area of national focus as efforts continue toward eliminating health care disparities and improving the overall health of all Americans. With that in mind, the goals of this article are to: 1) provide an overview of disparities specific to the care of patients with diabetes, 2) highlight some of the potential causes of these disparities while providing an overview of past and current efforts to address them, and 3) provide considerations for the primary care provider's role in helping to reduce and eliminate disparate trends in diabetes care.

The current prevalence of diabetes in the United States is startling, with nearly 24 million affected individuals (~ 8% of the U.S. population) and another 57 million individuals (~ 19% of the U.S. population) believed to be at considerable clinical risk of developing diabetes (i.e., having pre-diabetes).⁵  Racial and ethnic minorities are known to carry a disproportionate burden of diabetes, with the prevalence among African Americans at ~ 12% and that of non-white Hispanics at ~ 11%, compared to whites, whose overall prevalence is ~ 7%.⁵ 

As the population continues to age and becomes increasingly diverse, the number of individuals with diabetes is expected to increase unless the current trajectory is interrupted. This is an important consideration because diabetes has been identified as the sixth-leading cause of death nationally,⁶  and the annual costs directly attributed to the care of patients with diabetes are estimated to be $174 billion.⁷ 

Significant disparities in both the processes of care and health outcomes relevant to diabetes management persist across the country. According to the most recent National Healthcare Disparities Report (NHDR), the proportion of patients with diabetes who had all three annual services recommended by the American Diabetes Association (i.e., appropriate measurement of A1C, retinal eye exams, and foot exams) in the past year was significantly lower for poor to middle-income individuals, Hispanics, and those without at least some college education compared to their respective comparison groups. Lower-extremity amputation rates among patients with diabetes have also been consistently higher among African Americans, Hispanics, and those who live in communities with median incomes < $45,000.³ 

As a nation in general, we are doing poorly in terms of the overall quality of glycemic, blood pressure, and lipid control among patients with diabetes.^8-10  For example, during the years 1999-2004, only 48.7% of patients with diabetes met the recommended A1C goal of < 7%, and 56.6% of patients reported blood pressures < 140/80 mmHg. Within this context, nonetheless, both African Americans and Mexican Americans had significantly less likelihood of having an A1C < 7% compared to whites, and African Americans and those classified as poor to middle income had significantly fewer individuals at goal blood pressure.

It should be noted that blood pressure estimates for this period were based on previous recommendations that have since been updated (i.e., a blood pressure goal of < 135/80 mmHg),¹¹  suggesting a potential underestimate of the current disparate gap in blood pressure control.³  On a brighter note, when cholesterol control was examined for the same time period, there were no longer any statistically significant racial differences, and the previous gap between high-income and poor individuals with diabetes had been closed. Nonetheless, the overall rates of cholesterol control (i.e., total cholesterol < 200 mg/dl) also remain poor, with only 48.2% of adults meeting recommended goals.^12,13 

Most practicing clinicians know that dietary and physical activity counseling are imperative to the care of patients with diabetes. Unfortunately, the NHDR in 2007 also showed that significantly fewer Hispanics, African Americans, lower-income individuals, and individuals with less than a high school education were told by their physician that they were overweight. Similarly, obese individuals who were Hispanic, poor, had less than a high school education, or were uninsured were significantly less likely to be given instructions from their physician on physical activity.³  These results, although not specific to the population with diabetes, should be taken into consideration when thinking about disparities in diabetes care. Table 1 provides a brief list of selected studies that further highlight disparities in diabetes care that affect each of the aforementioned AHRQ priority groups.

The U.S. health care system seems to be moving increasingly toward a model in which health organizations and, to a certain extent, individual providers will be judged based on their ability to provide care that is effective, safe, timely, patient-centered, equitable, and efficient.¹⁴  “Equitable” refers to the non-differential delivery of care regardless of patient characteristics. Addressing all six of these factors may seem like an impractically tall order for busy practitioners who have existed to date in an environment where revenue is primarily based on volume and fee-for-service as opposed to being based on performance. Nonetheless, these new potential standards reflect a growing recognition of the need for change in the way health care delivery is provided and evaluated.

What are some of the potential causes of disparities in diabetes care, and what has been done to address them? To answer these questions, this discussion focuses on both components of a patient-provider dyad, highlighting characteristics that may partially explain the observed disparate trends in diabetes care. Both members of this dyad have been shown conceptually to have an important and interrelated impact on disparities within the larger context in which health disparities are believed to exist.¹⁵  These two entities are also most relevant to those who regularly attend to patients with diabetes, whether in primary care or specialty care settings.

Patients with diabetes are known to play a vital role in their individual care in that much of their success hinges on the daily self-management of their chronic illness. Although it is true that differential access to care and patient-specific demographic factors such as education and income level partially account for disparate trends in care, there are still residual disparities in quality even after controlling for these variables.¹  From a practical standpoint, when individual patients are nonadherent to treatment regimens, scheduled appointments, and instructions regarding diet and exercise, they are obviously more likely to have poorer diabetes-related outcomes.

However, what providers often attribute to someone being a “difficult patient” very well could represent a manifestation of other intrinsic patient factors such as self-efficacy, disease knowledge, or health literacy or quantitative skills (i.e., numeracy). In 2004, the Institute of Medicine defined health literacy as “the degree to which individuals have the capacity to obtain, process, and understand basic information and services needed to make appropriate decisions regarding their health.”¹⁶  Numeracy is a component of health literacy and has simply been defined as “the ability to understand and use numbers in daily life.”¹⁷  There are more than 90 million Americans with basic or below-basic literacy skills and more than 110 million with basic or poor quantitative skills.¹⁸  Both health literacy and numeracy are important skills needed for successful diabetes self-management because patients are often required to interpret and apply dietary instructions, measure and dispense insulin, and quantify carbohydrate intake, among other tasks.¹⁹ 

Several studies have linked low health literacy and numeracy to poorer diabetes knowledge and symptom recognition, poorer glycemic control, and lower self-efficacy or confidence in one's ability to self-manage.^20-22  Also, low health literacy has been shown to be an independent predictor of how much patients benefit from a comprehensive diabetes management program.²³  These “other” patient factors are not intuitively obvious, particularly during routine clinic visits, but they can be easily measured even in the time-pressured clinical setting.^24-26  Additionally, patients have been shown to be amenable to such assessments without subsequent decreases in satisfaction.²⁷ 

Peek et al.²⁸  recently published a comprehensive review of the literature summarizing interventions targeted at each of the two aforementioned members of our diabetes care dyad with the overall goal of reducing disparities in diabetes care. Each of the patient-targeted interventions in their analysis was an educational program centered on improving self-management behaviors, physical activity, or dietary practices.

For example, Piette et al.²⁹  enrolled English- and Spanish-speaking adult patients with diabetes into a randomized controlled trial in a primary care setting. They compared to usual care the added benefit of an automated telephone assessment system with nurse-guided patient follow-up to address identified barriers to self-care. Patients in their intervention arm demonstrated greater self-management activities, less frequent problems with medication adherence, and better glycemic control compared to the control group. In a brief subanalysis, the Spanish-speaking intervention patients in this study demonstrated even greater benefit (i.e., improvement in A1C, success in reaching goal A1C, and fewer diabetes-related symptoms) compared to the Spanish-speaking controls, suggesting ethnicity as a potential effect modifier.²⁸ 

Anderson-Loftin et al.³⁰  evaluated the effect of an educational intervention among a small cohort of rural African Americans with type 2 diabetes. Their dietary intervention incorporated culturally tailored educational sessions and follow-up by a nurse case manager. Recipients of their intervention demonstrated significant improvement in BMI, with a net weight difference of 3.7 kg at 6 months compared to the control group. There were also significant reductions in dietary fat intake, with nonsignificant trends towards improvement in A1C and lipids.²⁸ 

Overall, Peek et al. found that patient-centered educational interventions that used peer support or one-on-one interactions were more likely to yield positive results, and those interventions that were culturally tailored resulted in greater reductions in A1C compared to interventions geared for the general patient population.

Primary care physicians are known to care for the majority of U.S. patients with diabetes.^31,32  They are, therefore, logical targets for advancing efforts to eliminate disparities in diabetes care. But could they also be a source of the problem? A national survey of physicians conducted by the Kaiser Family Foundation in 2001 found that physicians were less likely than the general public to believe that disparities in health care occur “somewhat often” in all areas except insurance status.^33,34  Even when physicians agree that disparities in care are ubiquitous, they are more likely to attribute the causes to patient factors and less likely to believe these problems exist among their own patients or are related to provider factors such as trust and communication. This has been dubbed the “not me” phenomenon.^35-37 

Still, there is growing evidence that even well-meaning providers can be subject to unintentional and unconscious biases that manifest in differential care of patients. Sequist et al.³⁸  examined a cohort of primary care physicians representing 14 ambulatory care centers and caring for nearly 300,000 adult patients with diabetes in eastern Massachusetts. Embedded within their electronic medical record system were components such as physician-directed support tools, capacity for team management, and the ability to generate patient mailings regarding needed health services, such as cholesterol screenings. They sought to determine whether variations in diabetes outcomes by race occurred at the level of individual providers.

The researchers observed that white patients were significantly more likely to achieve A1C, LDL cholesterol, and blood pressure control than African-American patients. Interestingly, a greater proportion of these differences were explained by within-physician effects (i.e., differences within the same physician's panel) rather than socio-demographic factors, and these differences were not linked to overall provider performance or the number of African-American patients in a given panel. Although not explicitly clear, it is plausible that these observed differences were the result of unconscious racial bias or perhaps differential levels of communication between patients and providers.

Unconscious biases and stereotyping have been postulated to manifest in the clinical setting as “cognitive shortcuts” that occur during the decision-making process inherent within the patient-provider encounter. These heuristic tools are thought to be more likely to occur in situations in which there is time pressure, fatigue, stress, or the need for multitasking—all hallmarks of the day-to-day environment in busy medical practices.³⁹  Although the majority of physicians are aware of their egalitarian responsibilities as providers of health care, many do in fact unconsciously categorize patients based on stereotypes, and there is evidence to suggest that this behavior can potentially influence both the quality of the encounter and physician behavior (i.e., recommendations, counseling, ordering of tests, and prescribing patterns).^40-43 

Several conceptual frameworks have been put forth in the literature as explanations of how physicians may contribute to disparities in health care.^40,42  One important element that is potentially remediable is that of communication. Ashton et al.⁴⁰  have proposed that differential communication between patients and providers is often the result of differing levels of congruence in each member's “explanatory model” of illness. For example, if a patient believes that his diabetes was caused by stress or a traumatic event, he may be less likely to adhere solely to the dietary or medication recommendations of his provider.^40,44-47 

Racial and ethnic concordance between patients and providers has been shown to be associated with improved communication as well as increased satisfaction, trust, and perception of the quality of care.^48-51  Although diversification of our provider workforce is important, it is impractical to believe that complete racial/ethnic concordance between patients and providers in our health care system is necessary or that it would completely alleviate patient-provider barriers in communication. Interestingly, Street et al.⁵²  reported that physicians who employed greater patient-centered communication skills were able to overcome the barriers of racial/ethnic discordance in terms of patient satisfaction, trust, and intent to adhere. Furthermore, physicians' knowledge of their patients' limited health literacy can aid providers in tailoring their delivery of health information and potentially affect both diabetes management and outcomes.⁵³  Therefore, providers of diabetes care would be prudent to consider and evaluate the quality of the communication that occurs during their clinical interactions, particularly with minority patients and regardless of racial/ethnic concordance.

All of the interventions geared toward providers of diabetes care in Peek's analysis used reminder systems and provider education, such as practice guidelines, personal feedback, and continuing medical education programs. These studies demonstrated significant improvement in several process measures, such as measurement of microalbumin, foot care, and exercise counseling. Only one study met inclusion criteria for the meta-analysis on reduction in A1C, and this study showed a nonsignificant reduction in A1C of 0.47% using problem-based learning as an intervention. Surprisingly, none of these intervention studies used cultural competency or Spanish-language training.²⁸ 

What can providers do for their patients and themselves to aid in the reduction and eventual elimination of disparities in diabetes care? We have seen that patients' underlying health literacy and numeracy levels can be associated with worse diabetes outcomes. Providers should not hesitate to assess these skills in their patients,^25,27,54  which can provide useful information and serve as a guide for how to best disseminate information to patients with diabetes. In addition to what is already available through organizations such as the American Diabetes Association, many researchers have developed and validated patient educational materials specifically designed for patients with diabetes and sensitive to the issues of literacy and numeracy.^26,54-56  Furthermore, many of these materials are both well-suited to and practical for use in the clinical setting. There are also resources available for providers to learn skills to improve their communication style, particularly with low-literacy patients.^23,57-61 

It has been shown that providers often experience uncertainty and apprehension when attempting to care for and respond to the needs of patients with racial/ethnic backgrounds that differ from their own.⁶²  This can potentiate clinical inertia, which has been shown to be more prevalent regarding minority patients.^63-66  Providers should be aware of the possibility of clinical inertia, especially with regard to their racial/ethnic minority patients and should, for example, consider adjustments to therapy, such as earlier addition of insulin, when clinically appropriate.^65,66 

Improved cultural competency can also serve as a means of addressing the uncertainty often experienced during clinical encounters. Beach et al.⁶⁷  showed that cross-cultural training can improve provider attitudes, knowledge, and skills, although there has been less evidence to support the positive impact that this has on patient outcomes.^67,68  Nonetheless, with time and more rigorous research in this area, we are confident that improvements in the delivery of culturally competent care will prove to be a useful tool in addressing disparities in diabetes care.

Providers are encouraged to incorporate these recommendations into their quality improvement projects, which are now requirements of accreditation entities such as the American Board of Family Physicians and the American Board of Internal Medicine. A summary of provider approaches to addressing disparities in diabetes care is provided in Table 2. Finally, providers must continue to lobby collectively for necessary changes in the larger health systems in which they work and which often hinder them in providing and hinder patients in receiving the highest possible quality of care.

In conclusion, disparities in diabetes care continue despite significant clinical advances in understanding the management of this chronic illness. Providers of diabetes care can play a key role in diminishing these disparities through understanding and addressing patient factors such as health literacy and focusing on improved patient communication and cultural competence. If this is done, the U.S. health care system very well may begin to turn the tide and make important strides toward equitable diabetes care and improved outcomes for all.

Dr. White is supported by Pfizer's Fellowship in Health Literacy & Clear Health Communication and Meharry Medical College's Clinical Research Education and Career Development Program (NIH NCRR CRECD 1R25RR17577). Dr. Beech is supported by Vanderbilt University's Diabetes Research and Training Center (P60 DK020593). Dr. Miller is supported by a Building Interdisciplinary Research Careers in Women's Health (BIRCWH) award (5 K12 HD043483-05).