Mental Health Care for Latinos: A Research Agenda to Improve the Accessibility and Quality of Mental Health Care for Latinos

Even though there have been important methodological developments in current psychiatric epidemiological research on Latinos, current studies lack a prospective design to follow persons' help-seeking behavior in local communities. Longitudinal studies can contribute in two important ways: by identifying with increased precision the factors that both enable and restrict utilization, and by increasing the reliability of the assessment of service use. With a cross-sectional approach it is difficult to disentangle the processes that underlie help seeking. For example, on the basis of the Los Angeles Epidemiologic Catchment Area (LA-ECA) study, Wells and colleagues (9) found that Mexican Americans who were less acculturated relied more on general medical providers for mental health care than on mental health specialists. However, it is unclear whether this finding reflects a tendency for Mexican Americans to express mental health problems in somatic terms (10), the greater availability of medical care providers than mental health care providers, stigma, or other possible reasons. A longitudinal design would provide the best method for discerning the key factors associated with help-seeking. Better identification of enabling or restricting factors will result in interventions that have a greater chance of succeeding in improving the accessibility of services.

Current cross-sectional studies are also limited by problems of recall that can affect both the assessment of service use and predictors of service use. The extent to which a person has used various types of services for a specific emotional problem during the previous six or 12 months can be difficult to determine from a single interview. Inconsistencies between the service use data of the National Comorbidity Study (NCS) (11) and the Mexican American Prevalence and Services Study (MAPSS) (12), for example, raise questions about the reliability of service use data across studies. In the MAPSS, proportionately fewer Mexican Americans (8.8 percent) sought services from mental health specialists than was reflected in the NCS data (12.4 percent), yet the MAPSS suggested that more Mexican Americans sought services from their general medical provider (18.4 percent) and any provider (28.1 percent) than was suggested by the NCS data (7.9 percent and 24.7 percent, respectively).

It is curious that Mexican Americans use relatively fewer mental health services but use more services from other service providers. It may be that the pattern of service use suggested by both the NCS and the MAPSS is indeed valid. An alternative interpretation is that methodologic differences between studies, as well as the limitation of retrospective accounts, contributed to the inconsistencies in the results. A longitudinal approach—one that prospectively assesses service use over smaller time intervals, such as one to three months—would likely enhance the reliability of this important measure. Given that the assessment of service use is the foundation for studying service accessibility, it is important that this measure include as little error variance as possible.

To advance our understanding of service use and the provision of high-quality care it is critical that we give at least equal attention to broadly defined social factors as we do to clinical factors, such as level of impairment and treatment adherence. There are many ways to include an assessment of social factors in studies of service accessibility and the quality of care. One such way is to consider social factors as predictors of service use, such as economic factors, cultural models, and organizational factors. Pescosolido and Boyer (13) have also provided a framework for the social context.

A frequently studied barrier to care is the economic status of individuals and families. For example, Flisher and colleagues (14) found that children and adolescents from families with fewer financial resources were more likely to have unmet needs: youths identified as being in need of services were less likely to obtain services if their families received public assistance or if the youths were not covered by health insurance. Similarly, the MAPSS (15) found that among adults of Mexican descent who had a mental disorder, those with private rather than public insurance were more likely to visit a mental health provider. The fact that Latinos as a group are poor—23 percent live below the poverty line (7)—and many (35 percent) do not have health insurance (16) suggests that financial barriers to obtaining mental health care are considerable.

Another set of social factors that can serve as predictors of help seeking pertains to cultural models of disorders and conditions. Cultural models of mental health-related problems are particularly salient in the study of nervios, a prominent idiom of distress among Latinos. Jenkins (17) found that many Mexican-American families construe ill relatives' schizophrenia as nervios, which she argues helps reduce the stigma associated with the illness and promotes family support and cohesiveness.

Guarnaccia and associates (18) found evidence that among Puerto Ricans, primarily women, ataques de nervios often reflect a lack of power and disrupted social relations, such as divorce and death. Salgado de Snyder and associates (19) discuss how conceptualizations of distress, including nervios, and the social context—for example, the availability of social and professional support—are related to help seeking among rural Mexicans. Both groups of researchers point out the importance of nervios and related cultural models for advancing our understanding of how individuals and families cope with the identified problem, including seeking help.

Organizational factors of mental health services are other social predictors of care. To date the best test of the importance of practitioners' ethnic and linguistic background—key organizational characteristics—was carried out by Sue and associates (20). Among their many findings was that Mexican-American patients with practitioners who were ethnically similar to themselves were more likely to continue in treatment and to improve their level of functioning than Mexican Americans who had ethnically dissimilar providers. Thus ethnic match is one of many organizational factors to be considered. Ortega and Rosenheck (21) provide an alternative perspective. Financing of the mental health system is another aspect of the delivery system that needs significant attention.

The social world and its elements are not only relevant as predictors of service use and quality of care but also can be conceptualized as consequences or outcomes of mental health care or a lack thereof. In particular, it is important to examine the economic and social costs of mental illness to society, systems of care, families, and individuals. Considerable attention has been given to assessing the costs of various treatments (22). Much less attention has been given to assessing the costs associated with persons who need services but are not receiving them. Among the economic costs are lost wages of individuals and families as well as the costs of lost productivity to society. There are also social costs, such as the burden to the family and the fact that the individual is not fulfilling various social roles and obligations.

In addition to advancing a more complete understanding of service use in Latino populations, studying Latinos' social world has important practical implications. Such implications include the potential to garner support from policy makers, because social and economic outcomes are tangible and easily understood by the general public. Also, data on the social context of mental disorders serves to remind providers to address social factors—for example, employment and attitudes toward mental illness and its treatment—in addition to the clinical aspects of illness. Such data could inform the development of interventions to improve the accessibility and quality of care and to address social as well as clinical outcomes.

Treatment studies constitute a third approach that has considerable potential to improve access to care and the quality of the services provided to Latino persons. The value of treatment studies is that their findings can be readily exported to clinical settings so that practitioners can be trained to provide services according to specified guidelines.

The most ambitious intervention study carried out thus far with Latinos involved the treatment of primary care patients with depression in 46 clinics in six managed systems of care, including one in San Luis, Colorado, and one in San Antonio, Texas (23). Latinos accounted for nearly a third (30 percent) of the enrolled sample of 1,356 patients. The purpose of the study was to assess the effects of programs to improve the quality of care for depression. One program involved the administration of medications, and the other involved cognitive-behavioral treatment provided by psychotherapists. Although analyses related to ethnicity have yet to be published, initial findings indicate that compared with usual care, the quality-improvement programs had significant effects on treatment process, clinical outcome, and even some social outcomes, such as employment.

The implications of this research for Latino populations are great. First, the data build on previous findings (24) that mental health treatment within a primary care context improves the accessibility of mental health services for a broad spectrum of people, in particular Latinos, who make greater use of medical services than mental health services (3). Second, this research team demonstrated that a cadre of health and mental health care professionals can be trained at several sites across the nation to deliver effective mental health care. Finally, future analyses of these data will be able to identify what constitutes usual mental health care for Latinos with depression in a primary care setting and whether efforts to improve the quality of care will have a positive impact on functioning among Latino patients. Assessing whether interventions within existing services improves care for Latino patients will help identify specific ways in which services can be modified.

The study by Wells and associates (23) applied existing treatment models—medication management and cognitive-behavioral therapy—to the treatment of Latino patients. Other studies of specific interventions for specific disorders have targeted disorders such as schizophrenia (25,26) and depression (27). Some of these investigators pointed out the importance of the cultural adaptations they made in carrying out the specific treatment (28). For example, Kopelowicz (25) included families in skills-training sessions that were originally developed for patients only. As we build on this general line of treatment research in which cultural themes are implicated, it will be important to test directly whether specific cultural adaptations or specific treatment components of existing models are related to positive or negative outcomes. Such research will help us determine whether evidence-based treatments require cultural adaptation for Latino populations and, if so, what specifically should be altered.

To complement research on models of treatment for specific disorders, it will be important to test generic cultural competency models or guidelines (29,30,31,32,33,34,35), which have been developed to apply across treatment modalities and across disorders and presenting symptoms. The advantage of generic cultural competence models is that they can be incorporated across a wide range of practitioners, regardless of their chosen treatment approaches or of the disorders of their patients. Although a number of models or guidelines have been developed, little attention has been given to evaluating them. Testing both generic cultural competence models as well as specific treatments for specific disorders are two lines of research that have considerable potential to improve the quality of treatment for Latino persons (36).