Dr. Vega, Dr. Escobar, and Dr. Marin are affiliated with the Department of Psychiatry, Robert Wood Johnson Medical School, 151 Centennial Ave., Piscataway, NJ 08854 (e-mail: vegawa@umdnj.edu). Dr. Karno is with the Neuropsychiatric Institute, Dr. Kopelowicz is with the Department of Psychiatry and Biobehavioral Sciences, and Dr. Lopez is with the Department of Psychology at the University of California, Los Angeles. Dr. Alegria is with the Department of Psychiatry, Harvard Medical School, Cambridge, Massachusetts. Dr. Alvidrez is with the Department of Psychiatry, University of California, San Francisco. Dr. Bernal is with the Department of Psychology, University of Puerto Rico, San Juan. Dr. Escamilla is with the Department of Psychiatry and Cellular and Structural Biology, University of Texas Health Science Center, San Antonio. Dr. Guarnaccia is with the Department of Human Ecology, Rutgers University, New Brunswick, New Jersey. Dr. Jenkins is with the Department of Anthropology, University of California, San Diego. Dr. Lagomasino is with the Department of Psychiatry and Behavioral Sciences, University of Southern California, Los Angeles. Dr. Lewis-Fernandez is with the Department of Psychiatry, Columbia University, New York City. Dr. Loue is with the Department of Epidemiology and Biostatistics, Case Western Reserve University, Cleveland, Ohio. The National Institute of Mental Health assembled an expert consensus group for a workshop held August 8—9, 2005, Bethesda, Maryland. This report summarizes the work of that group.

This report is the outcome of a workshop convened by the National Institute of Mental Health, which assembled 15 senior researchers with a background in treatment and services research with the Hispanic population. The purpose of the workshop was to identify research issues most pertinent to improving quality and effectiveness of treatment for Hispanics experiencing persistent mental disorders, defined as psychiatric syndromes that are of sufficient severity and duration to cause long-term impairment in social and occupational functioning and significantly diminish quality of life. These disorders are often stigmatizing and require caregiver support and long-range professional treatment for stabilization and improved personal functioning. Persistent mental disorders do not refer to a specific diagnostic group or groups. Disorders such as schizophrenia, bipolar disorders, anxiety, obsessive-compulsive disorder, and major depression are obvious candidates for persistent mental disorders, but these disorders vary widely in severity and course. There are additional disorders that can impose a severe burden of illness.

Several recent reports and an expanding research literature have identified problems in access to and quality of care affecting Hispanics and other minority groups. These disparities in quality of care are not fully explained by patient characteristics, sociodemographic factors, or insurance availability. The 2006 National Healthcare Disparities Report found that Hispanics scored lower than European Americans on 90% of access measures (1). Focused research on priority areas in Hispanic mental health is needed to improve the training of clinicians, to further intervention development, and to improve systems of care in order to deliver more equitable services (2). Who will provide services to this population, and what new strategies will be used? There are serious "pipeline problems" in the creation of a behavioral health care workforce that is prepared to provide adequate treatment to Hispanics. For example, only 1% of licensed psychologists and American Psychological Association members are Hispanic, and there are similarly slim proportions of Hispanic psychiatrists and researchers receiving National Institutes of Health (NIH) research awards (3). It is unlikely that a majority of treatment services for Hispanics can be provided by Hispanic clinicians in the foreseeable future. Therefore, high-quality research is needed that can provide a resource for general workforce improvements to increase culturally appropriate and effective treatment across the life course.

A 2004 U.S. census update estimated that Hispanics numbered 41.3 million in the United States, a population larger than that of any Spanish-speaking nation other than Mexico. Today more than 40% of this population is foreign born, and 75% of the total Hispanic population comprises immigrants or children of immigrants (4,5). Consequently, the population has a relatively low level of educational attainment and income and high rates of poverty and uninsured status (6). These population characteristics reduce access to mental health care, create cultural and linguistic incompatibilities with existing systems of care, and result in problems of inaccurate diagnosis, poor treatment adherence and retention, and dissatisfaction with care received. Although these problems are now well documented, high-quality research to address them is lacking in almost every area (7,8,9,10). Indicative of these deficits is the failure to include Hispanics in statistically meaningful numbers in NIH-sponsored intervention development and clinical trials. Several reports issued by various government and private agencies have reached a central conclusion: minorities receive worse medical and psychiatric treatment even after clinical environment and patients' personal characteristics were accounted for (9,11). Identifying the priority research topics affecting treatment of Hispanics with persistent mental disorders and related co-occurring disorders motivated this review.

A multidisciplinary group of 15 senior researchers was convened August 8—9, 2005, to represent a spectrum of experience and expertise in mental health and psychiatric research. The meeting commenced with a summary of high-priority research issues selected from a recent review of the research literature (7). Using an open group process, we identified a list of research issues that was intended to be comprehensive. The topics were later integrated into five key areas: diagnosis, quality of care and culturally appropriate services, psychosocial intervention development, psychopharmacologic interventions, and access to health care.

The issue of diagnosis encompasses the initial detection of disorders and the impact of culture and language use on diagnostic accuracy and patient disclosure. Detection and accurate diagnosis of persistent mental disorders are problematic because of a combination of personal and provider factors (12,13). The problem requires careful examination in both primary care and specialty care settings (14). Stigma and low health literacy among many Hispanics, and inadequate provider training, decrease the likelihood that potentially persistent mental disorders will be accurately diagnosed or appropriately treated or referred, especially when the entry point into the health care system is primary care (15). Improved cross-cultural diagnostic training is needed in primary and specialty mental health care. The complexity of accurately recognizing cultural expressions, medically unexplained symptoms, and idioms of distress and distinguishing them from significant clinical signs, especially when language barriers exist, can lead to misdiagnosis and high rates of changed diagnoses (16,17). The DSM-V work groups of the American Psychiatric Association are confronting the problem of incorporating cultural information into the structure of the evolving DSM-V without the benefit of an adequate research base to provide useful guidelines for clinicians regarding the phenomenology of persistent mental disorders among Hispanics. A cultural formulation was included as an appendix in the DSM-IV for the purpose of stimulating research on treatment improvement, but insufficient research has been completed to supply an evidence-based model for clinical practice with Latinos.

Hispanic patients' level of competence with the English language varies from none to completely fluent, and clinicians have equally variable competence with Spanish, which affects to an unknown degree the quality of evaluation that Hispanic patients receive. Existing research, which is scant, has reported inconsistent language effects on diagnosis, and these effects may be more pronounced for some diagnoses than for others. For example, one research report found that use of Spanish by Hispanic patients resulted in disclosure of psychotic symptoms not revealed in interviews conducted in English. This finding was attributed to the degree of cognitive complexity required to be evaluated in a second language (English) and possibly more inhibition about disclosing personal information (18). Another study, however, reported that use of English oral diagnostic assessments among Hispanic patients with schizophrenia resulted in disclosure of more total pathology (19,20). Empirically supported reasons for either outcome remain obscure, and significant design limitations affected both studies.

In a well-designed and controlled study of bilingual Hispanic patients with schizophrenia, it was reported that clinical raters who spoke Spanish rated symptoms as more severe than did raters who spoke English. Patients who spoke both English and Spanish during the interview received the most severe symptom ratings, followed by those who spoke only Spanish and only English during the interview (21). However, complex effects of rater and patient language were reported for different disorders, such as depression, anxiety, and schizophrenia. Respondents assessed bilingually were more likely to be rated in English as having cognitive symptoms and in Spanish as having emotional symptoms, thus resulting in a rating of higher severity overall. A central conclusion reached was that language-ethnic matching resulted in clinical judgments of more severe psychopathology. On the basis of this limited literature, Lopez (22) seems justified in asserting that it is premature to draw conclusions about causes or consequences of language or cultural biasing effects on assessment and treatment.

Research is needed about language use and how acculturation of patients and clinicians influence symptom presentation and level of patient disclosure. Problems in detection and diagnosis can result in less effective and more costly treatment, clinical error, frustration of patients, and greater burden on caregivers (23,24). These problems will not be solved in the near term by ethnic and language matching because Spanish-speaking clinicians are in critically short supply, and we have no empirical grounds to conclude that their accuracy of diagnoses is superior to that of clinicians who speak only English and use translators, even if we strongly suspect this to be true. There is a need for research to determine the benefits of detecting and diagnosing persistent mental disorders through the use of trained interpreters for clinical evaluation when patients either request them or show limited fluency with English.

To address these issues, the work group suggests attention to the following research areas: first, the role of patient and clinician language use in the assessment of persistent mental disorders among Spanish monolingual, bilingual, and English monolingual patients; second, the effectiveness of properly trained interpreters in the diagnosis of persistent mental disorders; third, the diagnostic process in different clinical settings, including primary care and specialty mental health care, to determine rates and determinants of accurate detection of persistent mental disorders; fourth, elements of the diagnostic process associated with systematic misdiagnosis in psychotic spectrum and mood disorders that are potentially confounded by substance abuse problems, language usage, and cultural idioms of patients who present with putative psychotic symptoms; and fifth, the adequacy of health literacy related to mental illness and treatment among Hispanic patients and their families.

The President's New Freedom Commission on Mental Health highlighted the fragmented state of mental health care and the need for an improved model of mental health care that would be seamless and eliminate the many truncations of current service arrangements (9). The core elements of the report included patient- and family-centered care, the need for high-quality research to overcome disparities in treatment, and the goal of infusing cultural sensitivity into the entire system of service delivery. Hispanics have a long record of receiving insufficient quantity and quality of care even when insurance availability is not a factor (25). Furthermore, Hispanics are much more likely to rely on primary care doctors as the sole or preferred provider. This tendency has important implications for research, especially for improving consistency and cultural appropriateness of primary care in treating persistent mental disorders, augmenting treatment capacity, and improving coordination among primary and specialty care providers for Hispanics who are reluctant to use mental health services (26,27).

In mental health specialty care, Hispanics have high rates of dropout and missed appointments and poor medication adherence. They generally report lower patient satisfaction than patients from racial or ethnic majority groups (28,29,30). A recent epidemiologic report indicated that Hispanics are more likely to have persistent recurring disorders, which suggests inadequate treatment (31). The roots of this issue are multiple, and some have already been mentioned, such as linguistic incompatibility between clinicians and patients. However, little systematic research has been completed that examines the interaction between Hispanics, clinicians, and the environment of care at each step in the treatment process that could influence patient commitment to enter care, ability to remain in care, and adherence to clinicians' treatment expectations (32).

The phases of care encompass effective communication between patient and clinician, including recognizing patients' expectations about treatment and resolving cultural-linguistic disconnects, developing treatment plans that are culturally appropriate, engaging families or social networks in support of treatment and successful patient management and relapse prevention, and developing culturally tailored models for case management of patients (33,34). Although these areas of patient care are a concern for the entire patient population, the adverse impact on Hispanic patients is disproportionate because of linguistic isolation, poverty, and minority status. Interwoven with these factors is the potential influence of stereotyping of Hispanic patients by health care providers and subsequent effects on the quantity, type, and quality of treatment received. There is virtually no research available on this point, yet it is vitally important and can affect the full spectrum of care.

To address these issues, the work group suggests attention to these research areas: first, the role of stigma as it affects the willingness to seek care, access to care, and perceived barriers; second, development of new models in the provision of mental health care to overcome linguistic barriers and improve understanding of patients' conceptions of mental illness and acceptance of treatment; third, development of therapies and decision-making models that improve access to and quality of mental health care for Hispanics; fourth, the special cultural issues affecting quality and continuity of care for special populations, including elderly, rural, undocumented and refugee populations, and sexual minorities, with a focus on discrimination, clinician readiness, and risk of HIV-AIDS; fifth, the course and outcome of treatment for persistent mental disorders in relation to cultural features, such as patients' conception of illness and family response to medications; and sixth, the tailoring of existing interventions to improve their cultural suitability for Hispanics with persistent mental disorders and increase their use in clinical care.

The centrality of the family in the care of the persistently mentally ill in Mediterranean culture, independent of specific national origin, has been recently documented in the five-nation initiative called European Psychiatric Services: Inputs Linked to Outcome Domains and Needs (EPSILON) study of schizophrenia, in which it was found that "the way patients live splits the sites into two distinct groups: the northern European sites (London, Amsterdam, and Copenhagen), where many patients live on their own and where they have less contact with their caregivers, and the Mediterranean sites (Verona and Santander), where most patients live with their relatives and have more contact with their caregivers" (35). It is not surprising that U.S. Hispanics, who share cultural antecedents in Mediterranean culture, have greater familial interdependence than the general population (36) and that a sample of family caregivers in the Eastern United States indicated that 75% of Hispanic (Caribbean-origin) patients with persistent mental disorders lived with the primary caregiver's family compared with 33% of the Euro-American patients (34). Research must carefully consider how families differ and transform across generations in their disposition and capacity to care for members with persistent mental disorders. Nearly half of the Hispanic population is foreign born, with disproportionate representation of families with low income and education. This demographic profile suggests interpersonal problems associated with reunification of immigrant families, intergenerational social and cultural adjustments between foreign-born parents and U.S.-born children, and family stressors emanating from economic marginality and employment instability. To avoid stereotyping it is important to produce accurate information about how cultural, economic, and social factors affect the viability of Latino families to support the treatment of family members with persistent mental disorders and to retain them in home settings or support them in independent living situations.

Consideration of the likely special importance of the Hispanic family of origin in the sustenance of adult disabled members, as well as the family-centered investigations concerning the course and outcome of schizophrenia (later extended in the United States to the affective disorders) pioneered by British investigators (37,38,39,40,41,42), led to the investigation of both risk and protective factors within Hispanic families that might influence the outcome of persistent mental disorders afflicting family members. Although levels of expressed emotion have been found to be substantially lower among Mexican-American households coping with a family member with schizophrenia compared with both British and U.S. non-Hispanic white households, high expressed emotion when present has been found to be predictive of relapse among Mexican-American family members with schizophrenia (43). This finding has been repeated in studies from diverse national and cultural settings (44,45,46,47,48). Although a recent Southern California study found that high expressed emotion within the household did not predict relapse among a Mexican-American sample of families coping with schizophrenia (49), this is likely due to methodology that differed from that of earlier studies. Concerning the critical question of whether the expressed emotion construct represents a significant risk or a protective factor, comparative cross-cultural studies provide the empirical basis for theorizing expressed emotion as cultural in nature. The constellation of emotions, attitudes, and behaviors that are indexed by expressed emotion represent cross-culturally variable features of family responses to an ill relative (50).

Lopez and colleagues (51) have turned attention to prosocial or potentially protective factors that may be operative within Mexican-American families living with a family member with schizophrenia. The team has discovered that family warmth, such as spontaneous expressions of concern, interest, understanding, and positive regard, as coded from recorded family interviews, appears to serve as a buffer against relapse. Also, absence of warmth may represent a significant stressor in addition to the loss of protection (51). It is noteworthy that the research team emphasized the need to design interventions based on family strengths. Weisman (52) reported on recent efforts to integrate spiritual and other cultural elements into family-centered interventions with Hispanic families coping with schizophrenia.

Single-family therapy in the treatment of schizophrenia, when provided in modes of problem solving, psychoeducation, skills acquisition, and the reduction of high expressed emotion, has been shown to be effective in ameliorating the course of illness in both Hispanic and non-Hispanic cultural settings (53,54,55,56), as well as in the strongly family-centered Chinese cultural context (57,58).

McFarlane and colleagues' (59) pioneering work with the use of multiple-family groups in psychoeducational interventions in schizophrenia, has been extended to low-income Mexican-American families in Southern California and appears to be well received and effective in a public mental health clinic. This intervention conforms to the recommendations arising from the Schizophrenia Patient Outcomes Research Team (PORT) that family interventions continue for at least nine months in order to provide adequate education, support in crises, and problem solving for best outcomes with schizophrenia in the general population (60,61). The multiple-family approach clearly provides a greater economy of scarce professional resources, even if ultimately found to be no more effective than single-family interventions in long-term outcomes for persistent mental disorders. These limited findings need to be tested more widely with diverse Hispanic groups in different treatment settings.

Major unanswered questions concern the exportability of such family-centered interventions to non-research-oriented, public mental health facilities as well as the applicability of family support, education, and skill building in the management of other persistent mental disorders, such as major depression, bipolar disorder, and typically comorbid axis I disorders in Hispanic communities. Also unknown are the long-term outcomes for family-oriented approaches versus standard clinical care, as well as additional potential protective factors (such as spirituality, religion, and folk beliefs and practices) that may be operative within the family setting in various Spanish-speaking subcultures in the United States.

Nonpharmacological treatment interventions have been compared between U.S. Hispanics with depressive illness and cohorts from different ethnic groups. Miranda and colleagues (62,63,64) have reported on the use of cognitive-behavioral therapy and case management among outpatients seeking primary care treatment for depression with very low income and who were Hispanic, from another minority group, or Caucasian. They also have compared psychotherapy with medication in the treatment of young minority women with depression (63), and they have conducted a randomized, controlled study of the use of quality improvements versus standard care in the treatment of minority and European-American outpatients with depression (64). Hispanic patients in these studies had outcomes similar to those of European Americans and other minority patients, but the supplementation of standard care with quality improvements produced greater benefit among the Hispanic samples. A similar differential benefit has been reported for older, depressed Hispanic patients treated with above-standard care in a primary care setting (65).

It remains to be determined if the translation and cultural adaptation of the full range of nonpharmacologic intervention strategies shown to be effective against other persistent mental disorders in the general population—including severe manifestations of such anxiety disorders as posttraumatic stress disorder, obsessive-compulsive disorder, panic disorder, and social phobia—will be found to be of at least equal efficacy in Hispanic populations.

Psychoeducational efforts to increase awareness and knowledge of the symptoms of severe and persistent mental disorders and to help combat stigma and reluctance to seek professional care in Hispanic communities have received recent attention in an innovative exploration of the use of items of popular culture among Spanish speakers from both sides of the U.S.-Mexican border. Film and video clips and well-known popular song lyrics are being used to arouse and maintain interest while illustrating symptoms of psychosis and recommending use of mental health services (66). The National Alliance on Mental Illness' Family-to-Family curriculum, which has been translated into Spanish, also needs wider dissemination and assessment to evaluate its effectiveness when used with Hispanic families of differing nationalities.

The benefit of such public educational interventions is as yet unknown, as are potential public acceptance and responsiveness to more traditional psychoeducational efforts in Hispanic communities, such as the "Real Men. Real Depression" ("Estos hombres son reales. La depresion tambien") campaign (67) sponsored by the National Institute of Mental Health (NIMH) and earlier Spanish-language versions of pamphlets from not-for-profit (68) and industry-sponsored groups and NIMH. These pamphlets are designed to explain a range of mental disorders in everyday language and include brief illustrative clinical vignettes.

To address these issues, the work group suggests attention to several research areas.

The first recommendation is to examine the acceptability, adaptation, utilization, long-term outcomes, and cost-effectiveness of single-family and multifamily group interventions for Hispanic families coping with schizophrenia in a multisite, collaborative effort, using public specialty mental health outpatient resources. Such efforts might profitably also be mounted in international collaborative undertakings with established investigators in Latin-American nations.

The second recommendation is to explore the feasibility and efficacy of single-family and multifamily group interventions for Hispanic families that are coping with severe and persistent affective, anxiety, and comorbid disorders, including substance abuse and dependence, with the goal of conducting later outcome studies if found feasible and effective.

Third, researchers are urged to investigate the comparative efficacy of culturally and linguistically adapted interpersonal cognitive-behavioral, supportive, group therapeutic, family therapeutic, and brief psychodynamic psychotherapy among Spanish-speaking Hispanic patients in the treatment of both affective and anxiety disorders. Such studies might also best be established through multisite collaborative efforts, with the goal of later long-term outcome and cost-effectiveness assessments.

Fourth, researchers are asked to examine the receptivity of and responses to diverse psychoeducational materials concerning persistent mental disorders delivered through various media. Postexposure measures should include changes from preexposure levels of stigma as well as utilization of services in low-income Hispanic communities. Creative use of the Spanish language media in both urban and rural communities, with attention to the most common and disabling of the anxiety, affective, and psychotic disorders and their symptoms is needed. Again, initial single-site investigations could lead to multisite collaboration and longer-term outcome assessments.

Fifth, researchers must continue to explore potential protective factors, including religion and spirituality, operative in Spanish-speaking families in diverse U.S. subcultural groups, including families with Caribbean and Central-American origins whose members are coping with persistent mental disorders. In addition to the positive components of family communication assessed in studies of expressed emotion, more open-ended and anthropologically informed analyses of the roles of religion, spirituality, and folk beliefs and practices relevant to living with persistent mental disorders are needed in collaboration with clinical interventions at the family level.

The research literature is sparse regarding pharmacologic treatment trials among U.S. Hispanics compared with other ethnocultural populations. We have found small open-label studies but only one randomized placebo-controlled study of the treatment of depression (69) and only one prospective study of neuroleptic efficacy among persons with schizophrenia (70). The former included only 17 Hispanics out of a sample of 118; the latter was an open-label pilot study that included only eight Hispanics. These numbers are too small for meaningful comparisons. Extremely small numbers (3% of total samples) of Hispanics have been recruited into a four-year recent period of U.S. investigations concerning new pharmacologic agents.

Hispanic Americans have significantly higher than average rates of diabetes mellitus, excessive weight, and obesity. These statistics are important because many psychotropic medications can exacerbate these metabolic and cardiovascular risk factors. One recent study indicated a significantly higher rate of the controversial metabolic syndrome, considered a serious risk factor for type II diabetes and cardiovascular morbidity, for Hispanic versus non-Hispanic patients with schizophrenia treated with neuroleptic medication (71% versus 41%) (71).

Regarding the metabolism of psychotropic medications, no significant difference has been found between Hispanic and Caucasian Americans, although the information is still incomplete. A series of recent investigations has demonstrated that individual responses to psychotropic agents are strongly influenced by both genetic and environmental factors differentially associated with culture and ethnicity (72,73,74,75,76,77). The drug-metabolizing cytochrome P450 enzymes (CYPs), essential to the fate of most psychotropic agents in the body, are subject to genetic polymorphisms (mutant alleles), which often are distributed to different degrees among different ethnic groups. However, the available evidence indicates that the inhibiting (slowly metabolizing) alleles of the critical CYP2D6 enzyme are found among Mexican Americans at approximately the same low rates as those found among European Americans (77,78). The very important CYP3A4 enzyme's active (AA) genotype has been found at comparable levels among European-American and Hispanic males (79), although this CYP is particularly vulnerable to both inhibition and induction by a wide range of foods and medications (80) and therefore is also subject to ethnocultural influences upon its activity.

A handful of other studies also show little or no difference in CYP or transferase activity between Hispanics and European Americans (72,81,82,83). Mischoulon (83) has drawn attention recently to staples of the traditional Mexican and Central American diet—specifically, corn, grilled meat, and cabbage, as well as regular use of caffeine and tobacco—which may significantly influence psychotropic metabolism and thereby require adjustments to medication dosage and closer monitoring for potential adverse effects.

Medication adherence has been reported in one study as significantly lower among Hispanic and African-American patients than among European-American patients with schizophrenia who were treated with neuroleptics (84). Similar rates of adherence for Hispanic and non-Hispanic outpatients with schizophrenia were reported in two other studies (85,86). In another study only monolingual Spanish-speaking patients (with diverse disorders) were less adherent than non-Hispanic patients, suggesting that linguistic, cultural, and socioeconomic factors rather than ethnicity per se are more likely to influence the differences that have been found in this small number of studies. Adherence studies need to carefully consider the importance of improving provider-patient communication about medications, educating families about pharmacologic treatments, and increasing families' ability to monitor and support medication use by family members.

From the foregoing brief overview, it is clear that the culturally and genetically highly diverse U.S. Hispanic population has received scant attention from investigators, and more research is needed to be informative about the use of medication and tolerance to dosage requirements, toxicity, and efficacy of the full range of psychotropic medications being prescribed in specialty mental health and primary care practice settings.

To address these issues, the work group suggests attention to several research areas.

First, researchers should conduct genetic linkage studies (or genomewide association studies) regarding persistent mental disorders among Hispanics to complement the growing databases being assembled for non-Hispanic patients.

Second, further pharmacogenetic studies are needed among Hispanics regarding drug-metabolizing enzymes as well as risks of adverse medication side effects, including metabolic syndrome and diabetes mellitus.

Third, medication adherence should be investigated in relation to family and cultural beliefs, perspectives, and meanings concerning the use of psychotropic agents.

Fourth, the cultural influence in dose requirements, toxic responses, and subjectively perceived benefits and adverse effects of psychotropic agents should be studied in a Hispanic sample within the context of family and cultural beliefs and practices.

Fifth, researchers should investigate cultural and subcultural aspects of diet and use of herbal preparations that may affect efficacy, dosage requirements, and potential toxicity for psychotropic agents among Hispanic patients.

Finally, investigators must insist that their samples include adequate numbers of Hispanics to ensure statistical validity and generalizability in both government- and industry-supported psychopharmacologic research.

A critical issue in Hispanic mental health is health insurance, because one-third of Hispanics lack it, which is double the national average. Lack of insurance is a general problem we do not analyze in detail here, but it cannot be omitted when discussing access to care. Another issue that significantly affects both access and quality is poor English proficiency; 30% of Latinos report difficulty in communicating with health care providers. Besides this fact, there are multiple issues that affect the coverage of the mental health delivery system. We lack population research about prevalence and treatment rates for persistent mental disorders and the functional status of individuals with these disorders in the community. We are aware of problems encountered by Hispanics with persistent mental disorders in treatment settings, especially with language comprehension, yet the consequences of these problems and other threats to quality of care have not been carefully examined (87,88,89,90). Previous epidemiologic studies have shown that only about one in 20 immigrants who experienced a disorder in the past 12 months received treatment from specialty mental health providers (15,91). We know little about the course of untreated persistent mental disorders, the effectiveness of current treatments for Hispanics with persistent mental disorders, or the ultimate burden of failure to treat on families, communities, the health care system, and agencies, such as criminal justice, corrections, and drug rehabilitation. It is logical to assume that untreated mental disorders are creating a cost burden in other human service sectors, but these cross-sector impacts have not been estimated.

The New Freedom Commission report recommended improved coordination across diverse institutions and agencies inside and outside of the mental health care system so that referral networks are efficient and individuals can receive continuous care (9). There are many examples, but effectiveness is needed to guide development of organizational change and policy (33). There is very low recognition and treatment of disorders occurring at different stages in the life course, such as among children, adolescents, and Hispanic elderly persons (92,93,94,95).

Improving systems of care for Hispanic youth with subsyndromal and syndromal clinical features of persistent mental disorders is critical for preventing onset, improving the long-term course of illness, and reducing family and treatment burden. Early intervention requires effective detection methods and coordination of care and referral networks among schools, youth and family protective services, juvenile justice agencies and corrections, health care providers, community-based social agencies, and mental health care providers. Another example is the rapidly expanding adolescent and adult Hispanic prison population with persistent mental disorders who are either receiving minimal or substandard treatment and require specialty care before and after release (96). Often these individuals require integrated addiction and mental health treatment as well as selected community residential placement, rehabilitation, and vocational programs (97). Transcending the problems of co-occurring disorders and incarceration is the additional problem of HIV infection and living with HIV and associated chronic medical conditions, such as hepatitis C, that can be comorbid with psychiatric disorders. Few delivery systems today would qualify as "seamless" in delivery of multimodality services, especially in the context of language barriers and poverty. Effectiveness research is needed to identify access problems for these special populations and determine what enhancements or program and fiscal realignments to usual care in medical, mental health, and other treatment settings can improve this situation.

Research is needed to examine how individuals with persistent mental disorders are coping, what social assets they have available, and what types of interventions are effective, feasible, and sustainable to implement within their social environment. As noted in the previous section on psychosocial intervention development, models are needed for testing improved processes of care, including key problems of access and retention, for patients whose primary language is Spanish. Research should take heed of apparent increases in discriminatory behavior attributable to the recent national immigration controversies, especially as these translate into perceptions about and fears of Hispanics that potentially influence their disposition to seek treatment, or as these perceptions color Hispanics' personal experiences in the health care system. These models should incorporate family support systems into the treatment of family members with persistent mental disorders in usual care (98,99). Additional special populations include Hispanics with persistent mental disorders and physical disabilities and Hispanic rural populations who face many barriers to accessing and receiving regular and appropriate care. The needs of these special subgroups should be carefully examined with mixed methods of data collection and analysis, including community-based participatory research, in order to realign the mental health delivery system toward improved ease of access and effectiveness of care and improved use of technology and community outreach models (100,101).

To address these issues, the work group has several suggestions for future research. First, investigate the role of persistent mental disorders in increasing risk of HIV infection, and develop models for integrating HIV-preventive interventions, coordinated medical services, and family care models. Second, conduct studies of onset and course of untreated persistent mental disorders among Hispanics and the social and economic consequences for caretakers and for the health care, human services, and criminal justice systems. Third, investigate the feasibility of early detection and treatment of persistent mental disorders in childhood and adolescence, including early identification of prodromal psychosis in high-risk children and adolescents in school, juvenile justice, and detention facilities.

Fourth, investigate the feasibility and effectiveness of using innovative models, like telepsychiatry services, telephone-based psychotherapy and medication follow-up, and expert consultation and supervision for clinicians with limited experience in the diagnosis and treatment of Hispanics in rural areas. This research should especially focus on the coordination of regular sources of care (such as primary care) with psychiatric services when the latter are situated in distant locations.

Fifth, conduct epidemiologic and services studies in adult correctional institutions to assess the prevalence of persistent mental disorders and addictive disorders, the adequacy of medical and psychiatric treatment provided in detention, and the coordination of care and community placement postdetention. Sixth, use community-based methods of participatory research to inform the development of new models of community outreach in Hispanic low-income communities for supporting families with members who have persistent mental disorders, especially special populations such as the elderly and disabled populations, and people living with HIV, to achieve appropriate access to and regularity of specialty treatment services. Seventh, examine primary care as a site of treatment, coordinated care, and referral for Hispanics with persistent mental disorders.

Improved research concerning the mental health status and treatment of the United States' largest ethnic minority group will contribute to the improved science and improved services for this fast growing population. Studies recently supported by NIMH, such as the six-site Ethnicity and Diagnosis of Affective Disorders Study, or those already completed, such as the National Latino and Asian American Study (102), will help to develop a stronger evidence base for reforming the system of care. In addition, given the rich sociocultural and biological heterogeneity of Latinos, the proposed research directions also have the potential to lead to new discoveries of basic processes underlying persistent mental disorders and to improving mental health services for all.

The authors especially thank Thomas R. Insel, M.D., director of the National Institute of Mental Health (NIMH), for supporting this initiative and writing a commentary. The authors are greatly appreciative of Ernie Marquez, Ph.D., and Emeline Otey, Ph.D., the Office for Special Populations, and the NIMH Mental Health Disparities Team for their interest in fostering the expert consensus meeting and assisting in the completion of this report.

The authors report no competing interests.