Despite the overarching goal of Healthy People 2010—to reduce or eliminate disparities by 2010 (U.S. Department of Health and Human Services 2000a)—most disparity studies do not explicitly address the reasons for disparities in health care. There are suggestions that the provision of lower quality of care could be an important influence in disparities, but literature that explores quality improvement interventions targeting reductions in racial/ethnic or other types of health care disparities is still nascent (Chin et al 2007; Peek et al 2007; Saha et al 2008). Most current studies are "first generation"; (i.e., descriptive studies examining the extent of disparities) or ";second generation"; (i.e., studies examining only the association between race and potential factors that cause disparities without determining whether those factors truly mediate or cause disparities by race/ethnicity) (Saha et al 2008). Without detailed examination, it is challenging to ascertain whether potential mediating factors that influence disparities, identified and targeted through disparity-reducing interventions, truly mediate actual
A recent review of health care interventions to reduce racial/ethnic disparities focused on interventions for diabetes care at the patient, provider and health care organization or multitarget level (Chin et al 2007). The review is particularly relevant in the context of the conceptual model for this study because each intervention category reflects various points highlighted in the model, whether the intervention targeted individual (e.g., patient level) or ecological (e.g., provider, health care organization/system) levels. For each intervention level, the review highlighted areas that seemed most promising in terms of improving the quality of diabetes care for racial/ethnic minorities (e.g., through improved processes of care, such as regular physical activity) and in terms of improving diabetes-related health outcomes overall (e.g., reduced diabetes complications and intermediate outcomes such as mean glucose levels), although whether the interventions led to long-term reductions in diabetes care racial/ethnic disparities remains an open question (Peek et al 2007).
At the patient level, effective interventions that targeted racial/ethnic populations focused mainly on improving patients' diet, physical activity and self-management. Interventions that involved one-on-one interactions or peer support were more effective at improving health among racial/ethnic minorities than those using computer-based patient education. In meta-analyses, culturally-tailored interventions to improve general knowledge and health behaviors also had a more positive effect than general quality improvement efforts (Chin et al 2007; Peek et al 2007). At the provider level, "problem-based education"; targeted at physician providers, such as reminder systems and practice guidelines, continuing medical education, computerized decision-support reminders and in-person feedback, were most effective at generally improving processes of care and outcomes among patients with diabetes (e.g., improved rates of eye examinations) (Peek et al 2007), but these interventions did not necessarily provide information on whether
At the health care organization and multitarget level, there was strong evidence that interventions incorporating both the organization and the community reported large magnitudes of process measure improvements in general (e.g., improvements in HbA1c testing) (Peek et al 2007). Many organization-level interventions used a registered nurse as a case manager or clinical manager and incorporated a community health worker for peer support or community outreach and treatment algorithms targeting glucose, blood pressure and lipid control. Clinically significant patient outcomes included control of diabetes, hypertension and dyslipidemia (Peek et al 2007). The combination of nurse and community health worker was more effective than either used alone. Interventions targeting a combination of patients, providers, multiple heath care organizations and health care systems were also effective in improving process of care and outcomes among diabetics (Chin et al 2007). Multi-target interventions often mobilized multidisciplinary teams and patient registries and included many types of interventions: patient education, treatment algorithms, community outreach with community health workers, continuous quality improvement and nurse case management.
One multitargeted intervention formally measured and demonstrated a reduction in racial/ethnic disparities. The REACH 2010 project (Chin et al 2007; Peek et al 2007) consisted of a broad coalition of health care and academic institutions, community-based and faith-based organizations, civic groups, libraries, professional associations, government, businesses and media. It targeted all levels of intervention possible: patients (e.g., education strategies), providers (e.g., audits/feedback) and health systems (e.g., diabetes registries, community-based case management and continuous quality improvement teams) (Peek et al 2007). The study evaluating REACH 2010 found that previous racial disparities in process measures (e.g., HbA1c testing, eye examinations, lipid profiles, microalbumin testing) were eliminated after two years of implementation (Jenkins et al 2004; Peek et al 2007).
The limited state of interventional studies to reduce disparities in diabetes care also applies to cardiovascular and depression care. Only recently have two large-scale reviews of interventional studies in cardiovascular disease and depression been published in the literature, and both focus on interventions to reduce racial/ethnic disparities. In cardiovascular care, Davis et al (2007) comprehensively reviewed interventions aimed at reducing disparities in cardiovascular risk factor management, and found that hypertension and tobacco use received the most attention. Hypertension interventions targeted patients with sodium restriction promotion and have been somewhat successful, although other interventions targeting exercise or weight loss were not as effective. At the provider or community level, nurse-led interventions were commonly found to be effective in controlling blood pressure (Chin et al 2007; Davis et al 2007). Tobacco cessation interventions were the next most common cardiovascular interventions, with pharmacologic interventions (e.g., bupropion) shown to be effective in Blacks. Culturally tailoring education programs on smoking cessation to Black and Hispanic populations, whether targeted directly at the patient or through provider-targeted education programs, also met with success, although results were mixed. These interventional studies were limited because they did not formally assess whether disparities were reduced (Chin et al 2007; Davis et al 2007).
In depression care, Van Voorhees et al (2007) found that, like diabetes intervention studies, those that targeted multiple patient-level, provider or health care organization factors were most effective in improving depression care among racial minorities, such as improved receipt of cognitive behavioral psychotherapy and antidepressant medication treatment. Examples of successful interventions include those that culturally tailored programs to include bilingual providers, language-appropriate educational materials and case management for low-income persons.
While previous authors have acknowledged the extent of first- or second-generation disparity studies, the literature is still limited (Saha et al 2008). However, recent reviews of interventions that appear to hold promise in reducing racial disparities offer a hint about where targeted interventions seem to have the greatest effect within our conceptual model. Reviews of these interventions, whether for diabetes, cardiovascular care or depression care, point to multitarget interventions as perhaps the most effective, suggesting that causes of disparities are complex and wide-ranging for older adults with priority conditions.