Summary

Five good quality systematic reviews of interventions for diabetes mellitus identified studies that were mostly conducted in single-race populations. The outcomes studied included glycemic control measures, patient knowledge and satisfaction, dietary habits, physical activity, self-management activities, emergency room visits and hospital admissions. Process measures included the use of eye exams, microalbumin testing, HbA1c monitoring, foot care, and exercise counseling. We identified a primary intervention study tested on a multiethnic population of Veterans that examined the effects of care coordination and telemedicine intervention on glycemic control. We also reviewed a randomized controlled trial of the effects of cultural competency training on both clinician awareness and glycemic control among black and white outpatients. There was some evidence of benefit for interventions focused on community health workers, care managers, and culturally tailored health education for patients. Provider-focused interventions reported improvements in process measures, although computerized reminders for physicians resulted in negligible or negative results. Studies on the long-term effects of diabetes mellitus interventions on process and outcome measures are lacking. Heterogeneity between studies in subjects, settings, study design, and multiple aspects of the interventions limit the comparisons that can be made across studies. One small single center VA study suggests a telemedicine/care coordination intervention may reduce disparities in black Veterans with diabetes; this finding warrants further research. A study of cultural competency training significantly increased clinician awareness of racial differences in diabetes care, but no effect was observed on reducing disparities between white and black patients on glycemic control targets, LDL cholesterol, or blood pressure 12 months after the intervention.

Details

We identified five good-quality systematic reviews of interventions for diabetes mellitus,^6-11 including a Cochrane review⁷ that was subsequently published as a journal article.⁶ We found one primary intervention study conducted in VA settings that examined race/ethnic disparities for diabetic Veterans.

The interventions identified by existing reviews included health education interventions;^{6, 7} the use of community health workers;¹⁰ primary care interventions (including case management, patient counseling, and the use of reminder cards for providers);⁹ and self-care interventions aimed to change behavior using culturally tailored techniques.¹¹ One review included a variety of interventions including patient-targeted interventions, physician provider-targeted interventions, health care organization interventions (including case management, community health workers, and pharmacist-led medication management), and multi-target interventions.⁸ Most of the studies identified by the systematic reviews were conducted in single-race populations.

A review of health education interventions determined that culturally tailored health education was more effective than usual care in improving HbA1c and knowledge for up to one year of follow-up, although clinically important long-term outcomes were not examined.^{6, 7} Another review of health education interventions that sought to improve dietary habits, physical activity, or self-management activities reported that using interpersonal interventions such as peer support and nurses/nutritionists/ health educators more often had positive findings than computer-based patient education.⁸

One review examined self-care interventions that aimed to change the behavior of patients, rather than simply educating them. Four of the 12 included studies were designed using cultural tailoring techniques such as the use of focus groups and specific recipes for the ethnic group being studied. Improved glycemic control tended to occur in studies in which baseline glycemic control was markedly poor (A1C>10%). The author of the review concluded that culturally tailored interventions appeared to be successful among African Americans and Latinos, and that self-care interventions may be effective in more difficult-to-treat patients.¹¹

One review examined five provider-focused interventions (reminder systems, practice guidelines, continuing medical education, in-person feedback, and problem-based learning) and reported improvements in process measures such as the use of eye exams, microalbumin testing, HbA1c monitoring, foot care, and exercise counseling. By contrast, computerized reminders for physicians resulted in negligible or negative results. None of the physician-targeted studies included provider communication, cultural competence, or shared decision making.⁸

Interventions that used community health workers, care managers, and other non-physician providers showed evidence of benefit, although comparisons across studies were limited by the heterogeneity in the interventions, settings, and types of providers. A review of health care organization interventions found that nurse care managers were effective in improving quality of care as well as patient outcomes, including diabetes control and onset of retinopathy. Telemedicine case management had more modest results than on-site nursing staff. Positive findings were found for other non-physician interventions, including the use of community health workers and medical assistance programs that provided free medications.⁸ A review of case management delivered by specialist nurses found improvement in glycemic control and cardiovascular disease risk factors including blood pressure and total cholesterol.⁹ A review of community health workers serving in a variety of roles showed some improvements in patient knowledge, behavior and satisfaction, and decreases in emergency room visits and hospital admissions; increases in retinopathy screening and glycemic control monitoring by providers were also noted in a few studies.¹⁰

A primary study conducted at an urban VA Medical Center (VAMC) examined the effects of a care coordination and telemedicine intervention on glycemic control among older Veteran diabetics.⁴ Patients aged 60 and older (mean age 72 years SD=6) were enrolled for at least nine months. Although the study population was small (n=41), it did recruit black (n=14; 34%), white (n=21; 51%) and Hispanic (n=6; 15%) Veterans. The study was designed as a pre-post intervention evaluation. Patients used a telephone-based, in-home messaging device to transmit blood sugar data and answers to clinical questions to the care coordination team, comprised of a nurse practitioner, social worker, administrative assistant, and geriatrician. The data were used to both risk stratify patients and guide ongoing clinical advice from the care coordination team. A comparison of pre- and post-intervention data found that glycemic control improved significantly in black, but not in white or Hispanic Veterans. However, several methodological issues limit the validity of study findings including the single-site observational design and high attrition rate (28 originally eligible Veterans did not complete the study).

Summary

One fair quality systematic review of behavioral interventions for arthritis in racial and ethnic minority populations found limited evidence from a single randomized controlled trial that exercise interventions may be effective in improving pain and disability. Compared with a health education control program, exercise programs were comparably effective between whites and African Americans in improving pain and disability. In addition, one primary VA study investigating an educational intervention provides evidence of improving knowledge and expectations related to total knee replacement; however, it does not improve willingness to consider total knee replacement surgery.

Details

A fair-quality systematic review¹² that included 25 randomized controlled trials of psychosocial interventions in patients with osteoarthritis or rheumatoid arthritis identified only randomized controlled trial (represented in two publications^{13, 14}) that directly compared the effectiveness of an intervention between racial groups. The trial compared an aerobic exercise program and a resistance exercise program with a health education control program; patients were not blinded to the treatment assignment. The study reported that the exercise programs were comparably effective in improving pain and disability between whites and African Americans.

One primary study involved Veterans with moderate to severe knee osteoarthritis. The study assessed whether an educational video and total knee replacement decision aid affected expectations about postoperative pain and physical function following total knee replacement, as well as their willingness to undergo total knee replacement surgery. The study included 31 white and 33 African American Veterans. Although the video and decision aid improved expectations significantly among African American Veterans (bringing them in line with white Veteran expectations), expectations for both groups were still lower than post-total knee replacement outcomes reported by individuals who had undergone total knee replacement surgery. In addition, despite improving knowledge about total knee replacement pain and functional outcomes, the study found little change in participant willingness to consider total knee replacement surgery after the intervention. Although reported baseline and post-intervention willingness runs counter to what is found in previous studies (willingness was higher for African Americans than for white Americans in this study), the educational intervention did not alter subsequent willingness for either group. Despite improving disparities in knowledge and expectations about total knee replacement surgery, this intervention failed to demonstrate a similar improvement in African American Veterans choosing total knee replacement surgery.⁵

Summary

We identified the greatest number of reviews in preventive and ambulatory care interventions. 14 good quality reviews of single-race and comparative studies encompassed several preventive health areas, including cancer screening, smoking cessation, physical activity and diet. Little research explicitly focuses on reducing gaps in screening, treatment and outcomes for minority compared to white adults. Several reviews note the lack of sufficient number of studies to compare similarly-configured interventions or specific components of multifaceted interventions. There is some evidence that community health workers may improve rates of preventive health service utilization. Overall, improvements in preventive and ambulatory care for minorities are inconsistent. The overwhelming majority of reviews focused on improving screening and process of care measures for race/ethnic minorities; however, there are fewer studies evaluating the effects of interventions on health outcomes.

Details

Eight reviews examined interventions to improve cancer screening rates among various combinations of white, Hispanic, Asian/Pacific Islander, Native American and African American study participants. Two reviews focused on smoking cessation interventions for African Americans and Mexican Americans. Four reviews examined physical activity and diet interventions for African Americans, Hispanics and Japanese Americans. The use of community health workers (also referred to as lay health workers, peer navigators, and promotoras in studies among Hispanics), and various examples of culturally tailored health education and counseling were the most commonly evaluated interventions. We found no primary intervention studies conducted in VA settings that examined race/ethnic disparities in preventive and ambulatory care.

Cancer Screening

Eight reviews evaluated interventions designed to improve breast, cervical and colorectal cancer screening and treatment in the United States.^15-22 The best quality breast cancer screening review included 43 studies of primarily African American, Hispanic and white providers and patients.¹⁸ Included studies examined patient-focused (i.e., reminder letters, telephone calls, patient education and counseling), or provider-focused interventions (i.e., clinical reminders, provider education) to improve mammography screening rates. Only two studies examined interventions specifically designed to reduce disparities. These studies found that comprehensive case management interventions that combined health education, counseling, and assistance with health system navigation demonstrated promise in reducing time to initiation of cancer treatment. However, the review also found that interventions aimed at improving screening were more effective for white, educated populations, suggesting that the interventions may exacerbate disparities. Another systematic review also identified health education dissemination for Hispanic women by community health workers to have promising results in improving rates of breast cancer screening.¹⁵

Three reviews assessed interventions aimed at improving cervical cancer screening.^{16, 19, 20} The best and most recent of these reviews examined 18 studies including access enhancement, community education, individual counseling, mass media and community health worker interventions.²⁰ A meta-analysis of these studies found an overall improvement in cancer screening rates in the intervention groups among African American (d=0.146 [95% CI=0.028, 0.265]) and Asian (d=0.177 [95% CI=0.098, 0.256]) women, but not among Hispanic (d=0.116 [95% CI=-0.008, 0.240]) women. To attempt to explain the neutral findings for Hispanic women, the authors cite potential contamination of treatment and control groups in addition to low education and consequently, low health literacy for this specific ethnic group. Additional analyses found that all intervention types were associated with improvements in Pap screening, with access-enhancing interventions associated with the biggest improvements, while community health workers rendered the smallest effect. The investigators also note that the use of culturally matched materials and culturally matched intervention delivery was associated with improved Pap screening for minority women.

Three reviews examined interventions to increase colorectal cancer screening rates.^{19, 21, 22} One review included 15 randomized controlled trials of interventions in multiethnic populations.²² All interventions were associated with increased screening rates whether they were low intensity outreach programs (i.e., by telephone or mail) or more comprehensive community education programs (i.e., system navigation, risk counseling, cultural self-empowerment). The review found no differences between race/ethnic groups in screening uptake, though five studies did not report outcomes by race/ethnic group in some cases because these data were not collected. Another review offered insights into improving screening among African American patients, such as using personalized materials to educate and remind patients to improve screening adherence.²¹ Although this review suggested that the most successful interventions were tailored to address important barriers to screening (i.e., lack of knowledge, perception of risk), the review authors note that only a few studies directly addressed these barriers in the intervention design and evaluated the subsequent effect of interventions on these barriers.

Smoking Cessation

Two good quality reviews offer insights into smoking cessation interventions for minority adults in the US.^{23, 24} Each review focused on cessation programs aimed at African American and Hispanic populations, respectively. For African American adults, one meta-analysis of 20 quasi-randomized or randomized controlled trials indicated the odds of quitting were 40 percent higher for intervention programs of stand-alone or combinations of pharmacological, individual/ group/telephone counseling, targeted print materials, community outreach and media campaigns, compared to usual care or placebo. Although treatment setting moderated intervention effectiveness (i.e., church and community over clinical settings), treatment intensity did not. Interestingly, culturally specific interventions were only effective insofar as smokers indicated readiness to quit.²³

One good quality review included in the “Cardiovascular Disease Interventions” section below also examined 13 tobacco cessation intervention trials in minority populations.²⁵ The findings for that systematic review are largely in agreement with the systematic reviews presented here. In particular, patient-level interventions found that pharmacologic therapy – especially when combined with counseling – was effective in increasing quit rates. Culturally tailored, patient-level interventions produced mixed results. Of note, one study tested physician training and patient enrollment in a culturally tailored cessation program and found a 21 percent quit rate at seven months. Another study found culturally tailored health education was more effective than motivational interviewing.

A systematic review and meta-analysis of 17 studies involving Mexican Americans also indicated good evidence for pharmacological, and moderate evidence for community health workers (promotoras) and counseling (both group and telephone) interventions for smoking cessation. However, concerns with methodological design, length of follow-up and generalizability to other Hispanic subgroups limit conclusions based on these findings.

Physical Activity and Diet

Four reviews involved interventions to improve physical activity and nutritional education among minorities.^26-29 We identified one systematic review as the most comprehensive and best quality.²⁶ This review included 29 studies in African American populations, 15 of which were randomized controlled trials. Intervention modalities included telephone counseling, community health worker counseling, structured exercise programs, group exercise sessions and unstructured exercise programs. Most of the trials found a neutral effect of interventions. Those interventions that were associated with benefit demonstrated only short-term improvements in physical activity behavior change.

A second review of 19 culturally tailored weight loss intervention trials in African American, Hispanic, and Japanese American adults found that though most found a benefit from the intervention, the benefits were short-lived.²⁷ The authors also indicate that even the short-term weight loss benefits did not extend to African American women.

In a third review, tailored nutrition programs showed marginal benefit, although there were few studies that examined race/ethnic group differences explicitly and little exposition of tailored education program details.²⁸ This meta-analysis included multiple ethnic groups, including white, African American and Hispanic adults in 16 quasi- or randomized controlled trials. Tailored nutrition education interventions were implemented in the form of face-to-face interactions, email or print materials and appear effective in improving dietary intake over the long-term (six months or more) for priority minority groups. However, review authors do not comment on minority versus white intervention effectiveness.

One review focused on Hispanic Americans and found the use of community health workers (promotoras) for peer nutrition education improved diabetes disease management, though the authors emphasize the need for longer-term trials to further evaluate the effectiveness of such interventions.²⁹

Summary

We identified three systematic reviews that examined cardiovascular health care interventions. Most studies were conducted in single-race populations and could not test the ability of interventions to reduce disparities. Those comparative studies with mixed populations did not test for differential intervention effects based on race/ethnicity. The largest body of literature focused in the areas of hypertension and smoking cessation. On the whole, nurse-based interventions were associated with improvements in proximal health outcomes (e.g., blood pressure, lipid level, body mass index) for minority populations, but the addition of community health workers provided limited gains. Culturally tailored education approaches to lifestyle change interventions appear promising. Several small trials suggest intensive nurse led multicomponent care management interventions may reduce hospitalization in minority patients with heart failure.

Details

We found three systematic reviews examining interventions related to cardiovascular health. ^{25, 30, 31} The Davis et al. review was methodologically the most rigorous of these, and included studies from 1995 to 2006 of a broad variety of health care delivery interventions focusing on cardiovascular risk factor management (hypertension, hyperlipidemia, smoking, obesity), and management of cardiovascular conditions including myocardial infarction and heart failure. Studies were included if they included >50 percent minority populations, or were subgroups of larger trials for which race/ethnicity subgroup data were reported. Studies were excluded if the intervention had no connection to a health care setting. Most studies included minority-only populations and therefore could not test the impact of the interventions on disparities. The authors did find, however, a number of interventions tested in minority populations summarized below:

• Twenty-seven studies of interventions focused on hypertension, most of which were intended to change the structure of care delivery. Only nine studies evaluated patient-level interventions. Nursing interventions – either using home nurses alone or in combination with community health workers – were assessed in eight studies, most of which found these interventions to be successful in lowering blood pressure. Some of these also successfully lowered lipid levels. Pharmacist and community health worker interventions were not well studied. One clinic reorganization intervention was effective for both African Americans and whites, while two others were either ineffective or demonstrated only short-term gains. Of note, one of these was a VA study which found that chart-based reminders failed to improve physician adherence to hypertension guidelines. Patient-level interventions such as salt restriction were effective in some studies, and the Dietary Approaches to Stop Hypertension diet was significantly more effective in African Americans than in other racial/ethnic groups.
• Four patient-level hyperlipidemia intervention studies produced largely negative results, though studies using culturally tailored recipes in African Americans did find very modest improvements in lipid levels.
• The review included 13 trials examining tobacco cessation interventions in minority populations, and these results are described under the “prevention” section above.
• Only three trials assessed interventions promoting physical activity; high drop-out rates limited the conclusions that could be drawn from these studies.
• Four trials found nurse-led care management interventions featuring patient education and close follow-up appeared to reduce heart failure hospitalizations in minority subpopulations.

One poor quality review, which covers an identical search period as Davis et al., did not present study results, but included a qualitative critique of the literature.³⁰ The authors' interpretation of the literature indicated that the location of health care delivery matters, with community based approaches being particularly promising. They also noted that studies did not clearly show intensive interventions to be more effective than less intensive ones, and the intensity of intervention may have contributed to the high attrition rates seen in some studies. They noted that group based interventions were associated with high rates of recruitment and retention.

Finally, a third review focused on Native Hawaiian and other Pacific Islander populations and found only three intervention studies which were all limited by significant methodological weaknesses.³¹

Summary

No intervention studies were specifically designed to reduce disparities. However, evidence suggests that behavioral interventions can be effective in improving HIV/AIDS service utilization and health care outcomes for African American and Hispanic populations. A number of studies consistently found that behavioral interventions can reduce risky sex behavior and sexually transmitted infection rates. In particular, gender and culture specific interventions focused on empowerment were effective in at-risk African American female populations. The reviewed studies did not address organizational barriers and only targeted behavioral intervention efficacy. Based on this scant evidence, there is insufficient data to suggest that these interventions would be effective in reducing disparities in HIV/AIDS. Moreover, none of the reviews focus on reducing disparities among Veterans.

Details

We identified four good quality meta-analytic reviews that examined behavioral interventions to reduce HIV risk behaviors and incidence of sexually transmitted infections among African Americans and Hispanic Americans. None evaluated the effectiveness of behavioral HIV/ sexually transmitted infection risk reduction interventions in reducing racial and ethnic disparities in HIV service utilization or health outcomes between racial/ethnic groups.

One systematic review evaluated the efficacy of HIV behavioral interventions for African American women.³² The majority of participants were low income women who were unemployed or received public assistance. Most studies contained multiple intervention components aimed at reducing the risk of heterosexual transmission of HIV. All interventions provided information to increase HIV/sexually transmitted infection knowledge. Skills training components usually took specific forms, including correct use of male condoms, or negotiating safer sex practices through demonstration or role-playing. Several common constructs of behavioral change theories were addressed, including motivation; positive attitude toward condoms; normative influence; self-efficacy for protective behavior; personal responsibility to protect oneself, family, significant others, or community; and personal risk or vulnerability. Most interventions were delivered in small groups, had more than one session, and lasted longer than 240 minutes.

In 33 studies including 11,239 patients, interventions overall reduced unprotected sex rates by 37 percent (OR=0.63; 95% CI=0.54 - 0.75), and in 17 studies interventions reduced sexually transmitted infection diagnosis rates by 19 percent (OR=0.81; 95% CI=0.67, 0.98; n=8760). Efficacious interventions were those delivered by women, and focused on self-efficacy, assertiveness, and negotiation skills intended to empower women to seek equality in their relationships. Additionally, the success of HIV behavioral risk interventions may be more dependent on the quality than number of intervention sessions. Culturally tailored interventions with fewer sessions and skills training were as efficacious as multiple session interventions in reducing HIV risk behaviors.

Another systematic review by the same author evaluated the efficacy of behavioral interventions in reducing unprotected sex and sexually transmitted infection incidence among African American and Hispanic American sexually transmitted infection clinic patients, and found similar positive results.³³ The number of intervention sessions ranged from one to eight, were commonly delivered in small groups, took 10 minutes to 16 hours to deliver, and spanned from less than one day to six months. Beneficial intervention effects were seen in trials regardless of participants' characteristics (i.e., sexually transmitted infection date at baseline, specifically targeting African Americans or Hispanic Americans, 90 percent of participants being African American or Hispanic American), methodological quality of trials (i.e., participation rate, retention rate, reporting generation of randomization sequence or allocation concealment), or intervention features (e.g., intervention contents or setting, unit of delivery, total time to deliver intervention). In addition, the review found that interventions using facilitators that were ethnically matched to patients were more efficacious.

Despite being of lower quality, we included the third review to provide limited discussion of the effectiveness of behavioral interventions in reducing HIV transmission among African Americans.³⁴ The analyses indicated that, overall, sexual risk reduction intervention participants improved condom use but neither increased nor decreased the number of sexual partners compared with controls. Interventions less than 13 weeks long achieved greatest results when the intervention content included intrapersonal skills training (i.e., self-management). Interventions lasting from 13 to 43 weeks were more effective when the interventions included: (a) more HIV+ participants or men who have sex with men and fewer intravenous drug users, (b) higher retention rates, (c) tailored content to participants, (d) more sessions of longer duration, (e) interpersonal skills training (i.e., partner negotiation), and (f) no counseling and testing. Interventions of longer duration (i.e., 43 to 152 weeks) were more effective when they: (a) included a sampling of more HIV+ participants, younger people, and females; (b) had higher retention rates; (c) tailored content to participants; (d) offered more sessions; (e) included interpersonal skills training; and (f) did not include counseling and testing.

The fourth review focused on HIV/AIDS or sexually transmitted infection prevention interventions seeking to reduce the HIV risk behaviors of Hispanic Americans residing in the US or Puerto Rico.³⁵ In summary, participants in the intervention groups experienced a 25 percent reduced odds of engaging in sex risk behaviors (OR = .75, 95% CI = .66–.85); a 56 percent increased odds of condom use (Reverse =1.56; OR = .64, 95% CI = .54–.75); a 25 percent reduced odds of unprotected sexual intercourse (OR = .75, 95% CI = .63–.89); a 25 percent reduced odds of number of sex partners (OR = .75, 95% CI = .66–.86); a 31 percent reduced odds of new sexually transmitted infections (OR = .69, 95% CI = .54–.88); a 17 percent reduced odds of engaging in injection drug use; and a 27 percent reduced odds of sharing injection paraphernalia (OR = .73, 95% CI = .63–.85). Efficacious interventions: (a) did not use peer outreach, p < .01; (b) were delivered by non-peers such as health care providers, counselors or other professional facilitators, p < .05; (c) comprised of four or more sessions, p < .05; (d) included problem solving skills, p < .01; (e) discussed barriers to condom use, p < .01 and sexual abstinence, p < .05; (f) used peer norms to encourage behavior change, p < .05.; and (g) targeted either females only or males only and were successful in reducing sex risk behavior, p < .01. Interventions that utilized ethnographic formative interviews, p < .05, or addressed the Hispanic traditional gender norm of machismo, p < .05, were more efficacious than those that did not.

Summary

There is good evidence suggesting that multicomponent chronic disease management interventions including case management strategies and care coordination are helpful in reducing health disparities related to depression. There is insufficient research investigating the effectiveness of culturally tailored psychotherapeutic and preventive interventions in reducing depression health disparities; however, the preliminary evidence in this area indicates that these types of culturally tailored interventions hold promise. No good quality primary studies designed to reduce health disparities in Veteran populations were identified; however, two primary studies provide initial support for the feasibility of using technology-based interventions with ethnic minority Veteran populations. There were no good quality reviews examining disparities reduction interventions for mental health conditions other than depression. Though there is insufficient evidence for psychopharmacological, psychotherapeutic, and preventive interventions in ethnic minority populations, preliminary research on a variety of interventions suggests that such interventions can be effective for this population, particularly when they are culturally tailored and include a care coordination or case management component.

Details

We identified two good quality systematic reviews examining interventions aimed at reducing health disparities in mental health care in settings not limited to the VA. Additionally, we found two primary studies conducted within VA settings that addressed mental health care disparities interventions.

One systematic review examined 20 interventions related to depressive disorders.³⁶ Of the 20 reviewed interventions, 14 were randomized controlled trials and six were observational studies. This group of studies was comprised of 12 studies which the authors classified as “chronic disease management” research including case management and collaborative care approaches; the remaining eight were classified as “culturally tailored interventions” and included treatment programs, preventive interventions, and psychoeducation.

Multicomponent chronic disease management interventions were successful in reducing or eliminating disparities; in most cases, ethnic minorities obtained a greater benefit than non-Hispanic white participants, though outcomes for minority populations often remained below those of majority group. The authors identified the following components of successful chronic disease management interventions: practice redesign, patient education, expert consultation or decision support, feedback information, active case management by a trained provider or layperson, and adequate tailoring to patient and provider unique factors. Systems-level interventions included enhanced access to care (including patient cost reduction as well as integrated mental/physical health care), screening, and process improvements (including meetings, patient reminders, progress reviews, and expert team leaders). It was generally impossible to identify the individual parts of these multicomponent interventions that were more or less efficacious. However, physician reminders and screening tools alone were not effective in reducing disparities. Among patient-level intervention studies, case management was the most commonly used strategy. Case management was provided by a range of providers and laypersons, and focused on improved access and adherence to care, mental health care stigma, and psychotherapy focused on management of challenges; this was often accompanied by reading, electronic, and culturally tailored educational materials.

Culturally tailored psychotherapy and preventive interventions were described as showing promise, though few randomized controlled trials were identified. Promising components of cultural tailoring included culturally specific explanatory models of illness (e.g., family structure, autonomy, and time), educational and intervention materials, problem-solving approaches, recruitment of participants, and participant or provider ethnicity (e.g., ethnicity specific groups or providers).

A second systematic review examined 10 studies (7 of which were randomized controlled trials) investigating the effectiveness of psychopharmacological management, psychotherapy, and combination psychotherapy and religion or psychotherapy and case management interventions.³⁷ There was inconsistent evidence for relative effectiveness of one type of intervention versus another within or across ethnic groups. The authors highlight specific examples of differential response to treatment across ethnic groups, and though group differences were inconsistent across studies, there was consistent evidence supporting the effectiveness of psychotherapeutic and psychopharmacological interventions in ethnic minority populations, particularly when included as part of a case management or care coordination intervention.

In spite of evidence supporting the effectiveness of these interventions in ethnic minority groups, there was inconsistent and insufficient evidence documenting reductions in health disparities related to depression. Many interventions were conducted with only ethnic minority populations, making comparisons to majority group populations impossible. Though some interventions examined multiple ethnic groups, the findings across studies were inconsistent in terms of relative effectiveness of particular intervention types for specific ethnic groups.

We identified two primary intervention studies focused on mental health disparities reduction in Veteran populations. One study investigated a screen-phone assisted care coordination intervention for caregivers of Veterans with dementia.² Caregivers were provided with a screen-phone that included visual resources related to caregiving for individuals with dementia, as well as at least monthly phone calls from a nurse care coordinator. Outcomes including burden, depression, coping, quality of life, knowledge, and satisfaction with the intervention were assessed pre- and post-enrollment. There was no control group. The only significant difference on pre-test measures for African American versus white participants was related to burden, with African American participants endorsing significantly more burden than white participants. There were no significant changes on any outcomes comparing pre- and post-test scores. Results related to intervention satisfaction presented in aggregate format across racial/ethnic groups indicated patient satisfaction with the intervention, with 92 percent of participants recommending the intervention. Finally, cost analyses indicated a significant cost savings to the VA of approximately 50 percent compared to pre-intervention service utilization costs.

The other primary study investigated the acceptability of videoconferencing versus in person administration of a psychiatric assessment with an American Indian Veteran population.³ Though this study did not investigate an intervention, we chose to include it because it investigated a novel method of service delivery designed to reduce mental health disparities in an ethnic minority, Veteran population. This study used a no-control, test-retest design. American Indian participants were administered a culturally tailored, structured clinical interview over videoconferencing and in person. There were no statistically significant differences between the two methods of administration on process and satisfaction measures, though 45 percent of participants indicated a preference for the in-person interview, while only 20 percent indicated a preference for the videoconferencing.

Summary

There is good evidence that cultural competence interventions can improve provider knowledge, attitude, and skills, but there are few good quality studies of effects on patient outcomes. Overall, interventions designed to improve the standardized delivery of care for all patients are effective; however, most interventions to reduce disparities between minority and white patients are characterized by poor quality. One small single-site VA study provided very limited initial evidence that care coordination and multiprofessional home-based primary care programs can improve process of care measures for an African American cohort.

Details

We found five good quality reviews conducted in settings not limited to the VA,^38-42 as well as one primary VA study¹ concerning interventions that cut across clinical categories. Of the five reviews, four focused on cultural competency interventions and one focused on interventions to improve quality of care delivered in primary care settings. One VA study examined the effects of home-based primary care on improving outcomes for Veterans with multiple chronic conditions.

Cultural Competence Interventions

We identified four good quality systematic reviews investigating the effectiveness of cultural competence interventions. One review examined whether cultural competence interventions improved provider knowledge, attitudes, and skills; as well as patient adherence, satisfaction, and health status outcomes.⁴¹ This review identified 34 studies: 2 randomized controlled trials; 12 non-randomized controlled trials; and 20 non-randomized, non-controlled, pre-post studies. All but three of the included studies focused on provider outcomes and found good evidence that both general and culture-specific cultural competence training improved provider knowledge. There was consistent, fair-quality evidence that cultural competence training improved provider attitudes including cultural self-efficacy, attitudes toward community health issues, and interest in learning about patient and family backgrounds. There was consistent, fair-quality evidence that cultural competence training improved provider skills, including outcomes such as communication skills, community involvement, social interactions, and facility of treatment implementation.

A more recent review focused specifically on patient outcomes and identified seven studies with patient outcomes, including satisfaction, self-efficacy, health status, and patient assessment of provider competence.³⁸ The studies ranged from poor to fair quality and included two quasi-randomized, two cluster randomized, and three pre-post study designs; four of the studies included in this review were not included in the Beach, 2005 review⁴¹ and investigated outcomes including patient satisfaction, resourcefulness, service access, health status, trust, ratings of physician cultural competence, and treatment adherence. The authors indicate that in spite of low quality and inconsistent results among existing studies, the general trend among studies suggests the potential for cultural competence training to have a positive impact on patient outcomes.

Another review examined comparative interventions designed to facilitate culturally competent health care including as outcomes patient satisfaction, service utilization, and health status.⁴² No studies on recruitment and retention of ethnically diverse providers were identified. Two studies provided insufficient evidence for the effectiveness of interpreter/bilingual services in improving patient treatment receipt and adherence. One study provided insufficient evidence for the effectiveness of cultural competence training in improving patient treatment adherence. Four studies provided insufficient evidence for the effectiveness of culturally tailored educational materials in improving treatment receipt and adherence, as well as in terms of patient satisfaction. No studies on culturally specific health care settings were identified.

One review investigated cultural competence interventions in mental health care with a specific focus on evaluation of cultural competence models, and found only nine poor quality observational studies.³⁹ Overall, evidence for the effectiveness of cultural competence models in mental health care was insufficient and low quality.

Two well-designed randomized controlled trials identified by our technical expert panel may provide additional insights about culturally sensitive approaches to interventions. One study found that culturally tailored peer mentoring on advance directives elicited a significant impact on advance directives completion among black end-stage renal dialysis patients. By contrast, peer mentoring had no effect on advance directives completion among white end-stage renal dialysis patients in the trial.⁴³ Another randomized controlled trial found that providing clinicians with data reports on disparities along with cultural competency training significantly increased their awareness of racial differences in diabetes care, but there was no effect on reducing disparities between white and black patients on glycemic control targets, LDL cholesterol, or blood pressure 12 months after the intervention.⁴⁴ The findings of this study suggest that cultural competence training for clinicians, while effective in raising clinician awareness, may not be sufficient to have a measurable impact on disparities in health outcomes.

Quality Improvement

One good quality systematic review examined health system organization elements, as well as provider education efforts to improve rates of preventive services and quality of care for racial/ ethnic minorities.⁴⁰ Consisting of 27 studies (20 were randomized controlled trials and seven were case-control studies) predominantly in primary care settings, the authors found mixed evidence regarding various intervention components. Tracking and reminder systems were very effective in increasing rates of preventive service use, particularly for cancer screening and advance directives. However, these effects weren't demonstrated for all clinical areas, limiting their potential for addressing overall disparities.

The review authors identify several additional promising interventions. These include interventions that bypass the physician to offer standardized services directly to patients, use of remote simultaneous translation for patients with limited English proficiency, and using structured questionnaires of patients to assess health behavior risk. Fair evidence was found for bypassing physicians in order to provide preventive services to patients. However, there are no indications that language proficiency may be a relevant issue for Veteran populations; therefore, remote simultaneous translation services may not be applicable to VA settings. The authors also note the dearth of studies to explicitly evaluate the ability of interventions to reduce disparities between minority and white patients.

Home-Based Primary Care

In one primary study, 130 African American (71%) and 53 white (29%) Veterans from an urban VA setting were enrolled in home-based primary care for at least six months.¹ The home-based primary care program involved a multiprofessional team including a medical director, nurse practitioners, registered nurses, social workers, pharmacists, a registered dietician, a dental hygienist and a program director. The single-site study was designed as a retrospective chart review of enrollees, with pre-post intervention evaluation of a number of physical and mental health measures.

Patients were older (mean age 73.6), mostly male (95.6%) and had an average of six comorbidities per patient. The intervention was associated with a significant decrease in hospital admissions (43.7% reduction, p=0.001) and total number of days hospitalized (49.9% reduction, p=0.001), but not in emergency room use. Unfortunately, the study did not analyze outcomes according to race/ethnic group and though the results for this majority African American patient population were encouraging, the intervention's impact on health disparities reduction is uncertain.¹