A Systematic Review of Interventions to Improve Initiation of Mental Health Care Among Racial-Ethnic Minority Groups

Psychiatric conditions cause significant disability and cost more than $300 billion in the United States annually (1); yet, even for the most prevalent conditions, the majority of people in need of care do not receive effective, high-quality care (2). Shortages of mental health professionals, culturally divergent beliefs about the causes of mental illness, mistrust in the formal mental health system, negative attitudes about psychiatric treatment, and limited English proficiency contribute to disparities in mental health care for members of racial and ethnic minority groups (3–5). Between 2004 and 2012, racial-ethnic disparities in accessing mental health care increased in the United States, most prominently because of increased access to psychotropic medicine among whites, high uninsured rates among members of racial-ethnic minority groups, and difficulty detecting psychiatric distress among some members of minority groups (6).

Access to care comprises a complex set of attitudinal, behavioral, and resource-related factors, including availability, accessibility, affordability, and acceptability of health services as well as accommodation of services to patients’ needs (7). In addition, these factors interact with each other, creating additional layers of complexity. Although racial-ethnic disparities exist among high and low users of mental health services, much of the disparity appears to stem from lack of initial access to mental health services (8,9).

In the literature, episodes of mental health care are described as having a beginning, known as initiation; a middle, corresponding to the adequacy of follow-up visits; and an end, when treatment is terminated because of recovery or dropout (8–11). Evidence suggests that among African Americans and Latinos, disparities with whites have been greatest for the initiation of care compared with later episodes of care (9). According to these data, when there are racial-ethnic disparities in both the initiation and the adequacy of care, disparities related to adequacy of care appear to be less meaningful than those related to initiation. This suggests that intervening to improve the initiation of mental health care among racial-ethnic minority groups may reduce disparities. Equally important is understanding which interventions widen or fail to reduce disparities and the factors involved in their inefficacy.

The purpose of this systematic review was to synthesize current evidence on interventions intended to reduce racial-ethnic disparities in initial access to mental health care and summarize the extent to which the interventions were successful.

Methods

The Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) standard was used to ensure clarity and quality of the review (12). The first author (SL) wrote the systematic review protocol, including all screening and selection methods, and other authors (DC and PYC) revised and approved the protocol.

This systematic review did not involve recruitment and participation of human participants. An institutional review board was not involved, given that we examined publicly available published studies.

Inclusion and Exclusion Criteria

Studies that met the following criteria were included: the research sample consisted of members of racial-ethnic minority groups residing in the United States; the study design used an intervention that sought to improve the initiation of mental health care, rather than foster subsequent treatment, such as follow-up visits; and the study examined outcomes related to initial access, including visits to or new appointments with a mental health care provider, or willingness to initiate and actual initiation of interventions involving psychiatric medication or psychological treatment as a result of the study intervention. We restricted our search criteria to peer-reviewed journal articles written in English. We describe study selection criteria by population, intervention, and outcomes of interest.

Population.

The population of interest were members of racial-ethnic minority groups of all ages in the United States. We included participants with or without mental disorders, as long as a portion of participants had risk factors associated with mental disorders or symptoms or diagnoses of mental disorders. We included studies that compared outcomes between members of racial-ethnic minority groups and whites or studies that focused entirely on racial-ethnic minority groups.

Interventions.

We included interventions that sought to improve initial access to specialty mental health care, such as psychological counseling and pharmacological treatment. We also included interventions that sought to improve attitudes toward seeking mental health services. We included studies that assessed participants’ voluntary choice to initiate an appointment with a mental health provider. We excluded interventions that solely assessed follow-up visits to mental health services by participants from racial-ethnic groups, including interventions designed to improve adherence or quality improvement initiatives for ongoing mental health interventions.

Outcomes.

We included studies that measured initial access to mental health services, including visits to, or new appointments made for, consultation with a mental health care provider. Additional outcomes included pharmacological treatment initiation as a result of the study intervention and willingness to seek psychological or pharmacological treatment. Outcomes of interest were initiation of psychotherapy or pharmacologic treatment among participants motivated by a particular intervention, such as screening and referral, psychoeducation, or care coordination from a primary health clinic, school, nursing home, or homeless shelter. We excluded the studies in which care was initiated solely because of assignment to a mental health provider as part of the study protocol. In those cases, participants did not initiate care as a result of an intervention; rather they initiated care because they were assigned to receive care, and, therefore, study outcomes reflected other aspects of care.

Search Strategy

An informationist at the National Institutes of Health Library collaborated with the first author to develop the search strategy and filters. We conducted searches in February 2016 of the Medline (1946–2016), Scopus (1960–2016), and PsycINFO (1880–2016) databases. [A figure illustrating the search strategy is available as an online supplement.] A manual search of the bibliographies of included studies was conducted to identify additional studies for possible review.

Data Extraction

One author (SL) read all abstracts. Concurrently, additional authors and reviewers (JE, DC, and two research assistants) independently screened titles and abstracts of identified studies. On the basis of the screening of abstracts, three authors (SL, JE, and DC) read full texts of titles and abstracts from included records to assess eligibility. These authors discussed discrepancies in inclusion or exclusion of the articles after screening titles and abstracts and full texts. The authors also discussed discrepancies in low, unclear, and high risk of bias until they were resolved. If needed to achieve concordance, a third author not involved in the initial screening of abstracts or full texts (JE or DC) reviewed discrepant articles.

The flow diagram shows the number of articles identified from the electronic search and the number that were included and excluded [see online supplement]. We included 29 studies. We counted the full texts of multiple articles as one study if the same interventions were administered to the same study population at a given time. We counted studies separately if two full texts presented the same intervention but one manuscript used only a subgroup of a larger study population. Two pairs of studies fell into this category. The study by Chinman et al. (13) used a subsample of the multisite study population described by Horvitz-Lennon et al. (14). Ayalon et al. (15) used data from a study by Areán et al. (16) and Bartels et al. (17) (Primary Care Research in Substance Abuse and Mental Health for the Elderly [PRISMe]) and used a subset of data from the full multisite study.

Authors categorized studies on the basis of type of intervention used to improve initiation of mental health service use among members of racial-ethnic minority groups. We further assessed the intervention characteristics by participant age, race-ethnicity, analyses of outcomes by racial-ethnic subgroup, categories of mental disorder, intervention settings, and disparities in access to care. We collected information on the significance of findings comparing outcomes before and after the intervention and information about comparison groups, where available, as evidence of intervention effectiveness. We defined statistical significance as a p value below .05.

Given the heterogeneity of interventions, outcomes, and participants’ demographic characteristics, as well as the low number of studies per intervention, we did not conduct a meta-analysis of results. We qualitatively synthesized the results based on groupings by intervention type.

Quality Appraisal

Authors used the Cochrane Collaboration’s tool for risk-of-bias assessment (18). We examined the domains of random-sequence generation, allocation concealment, participant blinding, blinding of outcome assessment, incomplete outcomes data, and other sources of bias. We categorized each risk-of-bias domain as low, high, or unclear on the basis of data reported in each study. For each article, we considered the overall risk of bias to be low if the risk of bias was low in all domains, high if the risk of bias was high in any domain, and unclear if information on risk of bias was unavailable.

Results

Study Characteristics

Table 1 shows the characteristics of the 29 included studies (33 full-text manuscripts). Three studies enrolled only women (19–21). Participants in eight studies were, on average, age 60 or older, and two studies focused on youths who were age 21 or younger. In the majority of studies, the study sample was made up primarily (over 70%) of one racial-ethnic group, including Hispanics or Latinos (N=11), African Americans (N=5), and Chinese Americans (N=3). Other studies included a mix of participants from various racial-ethnic minority groups. Studies consisting primarily or entirely of Latinos, African Americans, or Chinese Americans were likely to test psychoeducation interventions that improved mental health literacy, attitudes toward mental health services, and motivation to seek care (19,20,22–26).

Participants were recruited from health clinics (for example, a primary care, a geriatric health, and an oncology clinic) (N=18), non–health care or community-based settings (for example, a school, a church, and a domestic violence center) (N=11), multiple sources (N=2), and a nursing home (N=1). Seventeen studies focused on participants with mood disorders or symptoms (for example, depression, dysthymia, and bipolar disorder), six focused on various mental illnesses (a mix of mood disorders, anxiety disorders, and psychotic disorders), and six did not require participants to have a diagnosis. Ten studies investigated intervention outcomes for collaborative care, seven for psychoeducation, five for case management, four for colocation of mental health service in existing social or educational services, and two for screening and referral. One study was a natural experiment involving a policy change in Medicare Part D drug coverage.

Studies varied in the assessment of initiation of mental health care, our outcome of interest. Nineteen studies measured antidepressant use or psychotherapy with health professionals, 11 studies measured attitudes toward mental health care or knowledge of mental health, and two studies assessed referral to mental health services or making appointments for specialty care.

Racial-Ethnic Differences in the Initiation of Mental Health Care

Ten of the included studies examined racial-ethnic differences in initiation of mental health care in response to an intervention (14–17,27–34) (Table 2). Disparities decreased in response to the intervention in seven of the 10 studies, including four collaborative care interventions (29–32), two studies of colocation of psychiatric care in primary care settings (15–17), and one screening and referral intervention (27). The collaborative care intervention described by Joo and colleagues (29) led to a statistically significant decrease in disparities between African Americans and white older adults in use of psychotherapy. The remaining studies demonstrating reduction of disparities (N=6) did not test for statistically significant differences in outcomes between racial-ethnic groups (15–17,27,28,30–32). Those six studies demonstrated that members of racial-ethnic minority groups who participated in the intervention had improved access to care compared with those in the control group. Among white participants, there was little difference in outcomes between the intervention and control groups. All seven studies demonstrating disparities reduction showed that compared with whites, African Americans increased antidepressant use, visits to a mental health provider, psychotherapy use, or counseling (15–17,27–29,31,32). Four of the seven studies that demonstrated disparities reduction showed that compared with whites, Latinos increased mental health service use, antidepressant use, psychotherapy visits, or counseling visits after completing the integrated care or collaborative care intervention (16,17,28,31,32). Table 2 summarizes the intervention design and outcomes of all 29 studies, categorized by whether the study demonstrated disparities reduction.

Three of the ten studies that examined racial-ethnic differences in outcomes reported increased disparities after the intervention. Adams et al. (33) found that lifting drug caps in states that had restricted the number of reimbursable prescriptions created wider disparities between African Americans and whites (p<.05). Kataoka et al. (34) assessed rates of initiating community- and school-based mental health services after receipt of a school-based suicide gatekeeper program. Latino students initiated community services at lower rates compared with non-Latino students, but there was no racial-ethnic difference in accessing school-based mental health services. Areán et al. (16) and Bartels et al. (17) found that Asians assigned to an integrated care group used fewer mental health and substance use services compared with other racial-ethnic groups. Horvitz-Lennon and colleagues (14) showed that disparities in access to mental health services among clients who were homeless did not change after participation in the Access to Community Care and Effective Services and Support (ACCESS) intervention.

Three of the 19 studies that did not examine racial-ethnic differences in outcomes enrolled participants primarily from a mix of racial-ethnic minority groups (13,35–37) (Table 2). These studies showed improvement in attendance of mental health clinics and reduced psychiatric hospitalization rates. Sixteen of the 19 studies that did not measure racial-ethnic differences in outcomes were unable to do so because the participants were entirely or primarily members of the same racial-ethnic minority group.

Cultural Adaptation

Of the 29 included studies, 23 incorporated some level of social, cultural, or linguistic adaptation of the intervention (Table 2). These adaptations involved incorporating feedback from participants from racial-ethnic minority groups on intervention design, hiring culturally and linguistically compatible interventionists, using images and phrases that fit participants’ cultural context, training providers in cultural sensitivity, meeting practical needs (such as appointment reminders and transportation), or modifying the intervention to align with participants’ conceptualization of mental health (for example, the belief in a connection between mind and body). Three of the seven interventions that demonstrated reduction in disparities used cultural adaptation, such as referencing diverse ethnic backgrounds in educational materials, training staff in culturally sensitive issues in depression treatment, and providing the interventions bilingually. Twenty of the 22 studies that did not demonstrate reduction in disparities used varying degrees of cultural adaptation, and all of the psychoeducation studies (N=7) were adapted culturally or linguistically.

Mental Health Interventions

Policy change as a natural experiment.

After implementation of Medicare Part D, the federal program that subsidizes the costs of prescription drugs and the drug insurance premium for Medicare beneficiaries, restrictions on the number of reimbursable prescription medications that Medicare enrollees could receive per month were removed in some states (that is, drug caps were lifted). Adams et al. (33) analyzed health insurance claims and enrollment data for a nationally representative sample of Medicare and Medicaid dual enrollees with comorbid depression and diabetes across 35 states. Receipt of antidepressants increased for both whites and African Americans in states that lifted drug caps, compared with states that had no change in prescription reimbursement policy. However, disparities in the receipt of antidepressants between African Americans and whites increased in states that removed drug caps because of increased receipt of antidepressants among nonelderly white patients.

Screening and referral.

Social workers and primary care providers were trained to screen for depression in nursing homes and primary care clinics (27,38). Nursing home patients were more likely to receive antidepressant prescriptions in study sites that had implemented mandatory screening (27). Screening reduced disparities in the receipt of antidepressants between African-American and white older adults (p<.05). Willingness to be screened for depression increased in primary care practices, which treated mainly Latino adolescents (38). Screening was provided by medical assistants and was followed by symptom confirmation by primary care providers.

Psychoeducation.

The seven psychoeducation interventions aimed to educate patients about psychotherapy entry and attendance, improve depression or psychosis literacy, reduce stigma associated with antidepressants, and increase access to help seeking or available mental health resources (19,20,22–26). Participants were mainly recruited from primary care clinics or community organizations (such as community clinics, schools, or churches). All psychoeducation studies consisted entirely or primarily of racial-ethnic minority groups and did not have a racial-ethnic comparison group. Most of the interventions were delivered in one session. All psychoeducation interventions were culturally or linguistically modified by using varied approaches. Interventions used promotoras or Spanish-speaking individuals for education and data collection, developed the intervention based on the target population’s perception of depression (for example, the belief among Chinese Americans in a mind-body connection), and incorporated photographs, scripts, and cultural icons familiar to members of racial or ethnic minority groups.

Psychoeducation improved the frequency with which patients raised concerns about treatment and discussion of ethnic, cultural, and religious issues (22). It also improved knowledge about depression or psychoses (19,23–25), reduced stigma associated with antidepressants (19,23), improved attitudes about depression treatment (20,26), and increased participants’ willingness to recommend seeking professional mental health services (24).

Colocation of mental health services.

Mental health services were provided in primary care clinics as part of the PRISMe project (15–17). This multisite, randomized controlled trial compared an integrated care intervention with an enhanced-referral model in which psychiatric services were provided in a separate location. The services in the primary care setting included assessment, care planning, counseling, case management, psychotherapy, and pharmacological treatment. The PRISMe studies showed that Latinos and African Americans in the integrated care group were more likely to seek mental health providers compared with Latino and African Americans in the enhanced-referral group. Among whites, there was no change in seeking mental health care by treatment condition. Asians who received integrated care were significantly less likely to initiate mental health care compared with Asians in the control group. However, there were no statistically significant differences in care initiation outcomes by treatment condition when race-ethnicity was used as an interaction term.

Chong and Moreno (39) found that providing telepsychiatry at a community health center increased mental health appointments, antidepressant use, and willingness to pay for telepsychiatry among Latinos. Kataoka et al. (34) reported that a school-based youth suicide prevention program that trains school staff as crisis team members improved help seeking from community mental health services; however, Latino students were less likely than non-Latino students to seek care from community mental health services. There was no racial-ethnic difference in the use of school-based mental health services.

Care or case management.

Of five studies involving care or case management, two reported findings from the ACCESS program, a federally funded demonstration of the effectiveness of a systems-integrating strategy using assertive community treatment (ACT) for homeless adults with mental illness (13,14). ACT provides intensive case management and includes a multidisciplinary team approach, frequent and intensive contacts, small caseloads, and assertive outreach. In both studies, the investigators found significant improvement in the receipt of psychiatric care at 12 months and over time, but initiation of psychiatric care did not differ by race-ethnicity.

The Sandy Cove Bridge Project provided cultural sensitivity training to nurses and primary care practitioners treating Chinese-American patients (40). Elements of the training included common illness beliefs and attitudes toward mental disorders and mental health service use. The providers rehearsed explaining mental disorders to patients without eliciting resentment. The intervention led to increased referrals for psychiatric evaluation and greater acceptance of referrals by the patients.

Uebelacker et al.’s (41) telephone-based care management intervention for Latinos with depressive symptoms resulted in no difference in outpatient visits or receipt of antidepressants between the intervention and the control groups. Boyd and colleagues (35) showed that a behavioral health referral intervention based on identification of barriers, effective solutions, systematic follow-up, and psychoeducation with a case manager led to improvement in scheduling appointments for mental health services and in appointment attendance.

Collaborative care.

The Prevention of Suicide in Primary Care Elderly: Collaborative Trial was the only study that showed a statistically significant reduction of racial disparities in psychotherapy use between African Americans and whites with minor depression after the participants completed the primary-care based intervention (29), which was intended to improve depression outcomes. No reduction was found in racial disparities in access to care among participants with major depression.

Three collaborative care interventions, all administered mainly in primary care settings, reduced racial-ethnic disparities in the receipt of antidepressants or psychotherapy, although the results were not statistically significant. The Improving Mood–Promoting Access to Collaborative Treatment (IMPACT) study, a multisite, randomized trial of a collaborative, stepped-care treatment for depression, found reduced disparities in receipt of antidepressants or psychotherapy visits among participants from racial-ethnic minority groups, especially Latinos (22). Among African Americans, participants in the IMPACT group improved in initiating psychotherapy but not in receipt of antidepressants compared with the control group.

Wells and colleagues (30) conducted a five-year follow up of Partners in Care (PIC), a quality improvement intervention in which case managers support primary care providers with patient evaluation, education, medication, psychosocial treatment, and linkage to mental health services. PIC marginally improved receipt of psychotropic medications among African Americans and Hispanics. Youth PIC, for participants between the ages of 13 and 21, reduced disparities in use of psychotherapy or counseling among African-American and Latino children and young adults compared with usual care (31,32).

Community Partners in Care (CPIC) was a group-level, randomized, comparative-effectiveness trial of Community Engagement and Planning, a community-engaged depression quality improvement program. The study took place in community-based settings with vulnerable populations with depression, including mental health and substance abuse services, faith-based programs, hair salons, and exercise clubs. Participants were primarily African Americans and Latinos. The intervention reduced hospitalizations for behavioral disorders among African Americans and Latinos at six-month follow-up, whereas the initiation of outpatient and specialty care visits did not change (36,37). However, the reduction in hospitalizations for behavioral disorders for African Americans and Latinos was less evident at 12-month follow-up.

Alleviating Depression Among Patients with Cancer (ADAPt-C), adapted from the collaborative, stepped-care IMPACT model, took place in community public safety-net and oncology clinics (42–45). In ADAPt-C, a cancer-depression clinical specialist provides psychotherapy in addition to care management under psychiatrist supervision. Compared with patients who received enhanced usual care, the intervention group received more depression treatment and were more likely to receive treatment at one-year follow-up if major depression recurred. At two-year follow-up, however, there was no difference between the groups in the receipt of antidepressants or counseling. Among Latino patients with diabetes and depression at one-year follow up, those in the intervention group was more likely than the control group to receive antidepressants, although no difference was seen in the receipt of psychotherapy.

Two collaborative care interventions, both with sample sizes under 150, were administered in primary care settings. The Blacks Receiving Interventions for Depression and Gaining Empowerment (BRIDGE) study compared a standard model of collaborative care for depression with a patient-centered and culturally tailored collaborative care intervention (46). All participants were African American. The BRIDGE intervention marginally improved patient attitudes towards readiness for depression treatment; however, it did not change receipt of antidepressants. In another study, Chinese-American adults received Culturally Sensitive Collaborative Treatment, which uses culturally sensitive psychiatric assessment to explore patients’ illness beliefs. The intervention improved uptake of psychiatric assessment but did not reduce stigma (47).

Risk of Bias Assessment

Fifteen studies had a low risk of bias (Table 3). Most (N=12) studies with a low risk of bias were randomized controlled trials of collaborative care or colocation of mental health services. One natural experiment arising from a change in Medicare medication reimbursement policy and two studies with a historical control group (a study of psychoeducation and a study of case management) also had a low risk of bias.

The most common reason for a high risk of bias was the absence of a control group. Ten studies lacked a control group. In addition, two randomized controlled trials of psychoeducation had a high risk of bias because of other methodological concerns. Two randomized controlled trials had an unclear risk of bias because of incomplete data about allocation concealment or blinding.

Discussion

This systematic review synthesized current evidence about interventions intended to improve initiation of mental health services among racial-ethnic minority groups and assessed the extent to which these interventions were effective. Of the 29 studies that met inclusion criteria, seven studies provided evidence that screening and referral, colocation of primary care and mental health services, and collaborative care interventions not only improved mental health outcomes but also contributed to disparities reduction in initiation of care. Notably, six of the seven interventions that reduced racial-ethnic disparities in initiation of mental health services primarily used an integrated care model. Each of the six interventions resulted in increased uptake of psychotherapy or antidepressant use among members of racial-ethnic minority groups compared with white participants.

The case for promoting integrated mental health care in primary care settings is well established (48–50). This model may have particular benefits for disparities reduction, also. The practical convenience and privacy in seeking care for mental illness in these settings may enhance help seeking by members of racial-ethnic minority groups (51). A study of African-American primary care users suggested that primary care settings could provide exposure to mental health services that familiarize people with mental health care, allowing African-American patients to “try before they buy” (52). Integrated care settings may also provide a greater variety of providers who deliver mental health services, thus extending services in racial-ethnic minority communities. These providers played an integral role in many of the interventions we reviewed. Depression screening by medical assistants, social workers, and primary care doctors improved receipt of antidepressants, provider attitudes toward depression identification, and patient attitudes toward screening (27,38). In communities with large populations from racial-ethnic minority groups and shortages of mental health providers, telepsychiatry also may extend access to mental health care in community and primary care settings. However, implementation of telepsychiatry faces numerous legal, ethical, and logistical challenges, despite evidence of its effectiveness and potential for reducing disparities (53).

More than one-third of the included studies involved collaborative care interventions, which have been shown to improve access to mental health care and clinical outcomes among racial-ethnic minority groups (54). In fact, the body of evidence supporting the effectiveness of collaborative care for treatment of minority groups is larger than for any other intervention. Recent policy changes make nationwide dissemination and implementation of collaborative care increasingly likely. The Centers for Medicare and Medicaid Services modified the physician fee schedule to reimburse physician and nonphysician providers of collaborative care model in clinical practice (55). In our included articles, several components of collaborative care appeared to be effective in improving initiation of treatment among racial-ethnic minority groups, including management of mental illness in primary care settings, use of culturally and linguistically sensitive psychoeducational materials for patients, and encouraging the use of psychotherapy to patients in need.

The challenge of reducing disparities in mental health care initiation will necessarily encompass more than implementation of effective models of care. Access to care involves understanding the interplay of context, illness characteristics and course, social networks, and systems of care, and the factors that predict disparities in initiation of care operate at multiple levels (for example, individual, neighborhood, and policy levels) (56,57). Being educated, living in an urban area, being female, and having mental and general medical complaints all predict greater initiation of care, whereas being black or Latino, being publicly insured, not being a citizen, and being healthier predict less initiation (57). Indices of disadvantage, such as living in a neighborhood with fewer college-educated residents, are associated with less initiation of mental health care. Even in settings where the density of mental health care providers may be greater, provider preferences, insurance coverage, and population perceptions of the services may explain low initiation of care among racial and ethnic minority groups (57). These findings suggest that efforts to influence health policies (for example, policies affecting insurance coverage and reimbursement) and social policies that reduce disadvantage could contribute to disparities reduction and that these efforts should be pursued simultaneously with efforts to expand the delivery of clinical interventions.

Among the contextual factors that influence initiation of mental health care, reducing cultural and social barriers to mental health care could also facilitate initiation of care and thereby reduce disparities, although evidence of the effectiveness of this approach is limited (58). Twenty-three of 29 studies in our sample mentioned using some form of cultural adaptation as part of their intervention. Of the seven interventions associated with reduced disparities in access to care, three studies involved interventions that had been culturally and linguistically adapted and four studies did not mention cultural adaptation. Many interventions, especially psychoeducation interventions, used scripted stories and visual representations to depict mental illness among members of participants’ own racial-ethnic groups. Other studies in our sample incorporated clients’ illness beliefs in culturally sensitive training of providers. However, these studies lacked comparison groups, and the added benefits of cultural adaptation for disparities reduction were not assessed. There is little question that providing psychoeducation or counseling in a client’s preferred language is effective for those with limited English proficiency. It is likely that using a more comprehensive, personalized approach to understanding clients’ values, preferences for care, and perceptions of need as well as facilitating communication with providers could more effectively reduce barriers and disparities associated with care initiation (59).

Although this systematic review fills an important gap, it had several limitations. First, the included studies were heterogeneous, and we were unable to conduct a meta-analysis of results and effect sizes. The interventions were diverse in terms of type, setting, sample size, and severity of symptoms or diagnoses. Second, our target outcome, the initiation of mental health services, was sometimes difficult to ascertain. Many studies with multiple outcomes did not make a clear distinction between initial access to mental health care and subsequent follow-up visits (quality improvement or engagement). Third, few studies (only 10) recruited both members of racial-ethnic minority groups and whites or another comparison group, thus allowing for measurement of disparities within the sample. Fourth, the populations represented in this review included African Americans, Latinos, whites, and Chinese Americans. There was no explicit inclusion of other underrepresented minority groups, such as American Indians or Native Hawaiian or Pacific Islanders. These limitations have been echoed in other reviews of populations at risk of mental health care disparities (60).

Conclusions

To our knowledge, this is the first systematic review that applies rigorous methodology to report on interventions that can potentially reduce racial-ethnic disparities in the initiation of mental health care. We found evidence that integrated care models hold promise in reducing these disparities. Importantly, effective interventions to reduce disparities in care must be widely disseminated and implemented to be of value.

This knowledge could be enhanced by studies that examine the mechanisms of disparities reduction as well as alternative models of integration, such as the use of eHealth or mHealth technologies. At the same time, we note that the current disparities literature could be strengthened by improving meaningful inclusion of participants from racial-ethnic minority groups and by conducting subgroup analyses when assessing intervention outcomes. An examination of how to study and implement interventions that target multiple levels of predictors of poor access would enable a more holistic approach to disparities reduction. The lack of clear policies and guidelines for reduction of mental health care disparities is partially due to the dearth of published studies on effective interventions targeting the complexity of factors affecting access to care among racial-ethnic minority groups.