Impact of a Culturally Focused Psychiatric Consultation on Depressive Symptoms Among Latinos in Primary Care

Depression is a mental health condition that is often recurring and when left untreated may lead to significant distress and impaired functioning; it is the leading cause of disability worldwide (1,2). Although evidence-based interventions for treating depression are available, the condition is often undertreated (3). Among racial-ethnic minority groups in particular, a number of disparities in the treatment process serve to hinder care (4). Clinical interventions are needed to address disparities in depression treatment that affect Latinos, with a strong need for interventions that can be applied in primary care settings (5). To this end, this study examined the impact of a two-session culturally focused psychiatric (CFP) consultation service on depressive symptoms among Latino primary care patients.

Latinos represent the largest ethnic minority group in the United States (16%) (6), which underscores the public health risk of untreated depression in this population (7). Although lifetime prevalence of major depressive disorder among Latinos is similar to that of non-Latino whites (8,9), Latino ethnicity has been associated with significant barriers to both assessment and treatment (10). Specifically, Latinos are less likely to receive any depression care, and when they do receive care, services are more likely to be delayed and less likely to meet quality standards (11). Latinos are also less likely to remain in depression care (12).

When Latinos do seek mental health care, they are much more likely to do so through primary care than through specialty mental health care (13,14). However, depression treatment disparities also extend to the primary care setting, with disparities noted at each point of care from diagnosis to referral and receipt of care—even after analyses controlled for factors such as age, education, employment status, and medical illness (5,15). These findings underscore the importance of targeting primary care clinics in efforts to enhance depression management for Latinos.

A number of complex, multilevel challenges contribute to mental health disparities among Latinos in primary care (13). Providers may face difficulties in diagnosing depression among patients whose ethnicity is different from their own, because of varying symptom presentations and cues or vocabulary used to describe depression (16,17). Latinos with depression, for example, may give greater emphasis to somatic complaints (18). Signs of depression may therefore differ from what providers are typically trained to expect, resulting in misdiagnosis and poorer symptom recognition (19,20).

Patient-level factors may also contribute to depression-related disparities. Among Latinos, treatment preferences may differ from those of non-Latino whites (21) or from standard treatment guidelines (22). Latinos may have had fewer helpful psychiatric treatment experiences in the past, lowering their willingness to seek future treatment (23,24). Factors such as stigma about depression may have a negative impact on treatment seeking and utilization (25). Although depression-related disparities affect both English- and Spanish-dominant Latinos, disparities are greater among Spanish-dominant Latinos (26).

Although depression-related disparities have been documented in primary care, there is limited guidance on how to address these disparities through clinical interventions with Latinos. As one solution to bridge the gap between provider knowledge and patient needs, Kirmayer and colleagues (27) developed the Cultural Consultation Service (CCS) model. Results indicated that cultural misunderstandings were associated with increased risk of incomplete assessments, incorrect diagnoses, and inadequate or inappropriate treatment. Through cultural consultations and formulations based on an expanded version of the DSM-IV cultural formulation model (1), the CCS effectively improved diagnostic assessment and treatment in diverse patient populations (27,28).

Van Voorhees and colleagues (4) conducted a systematic review of studies that attempted to address disparities in depression outcomes between non-Latino whites and ethnic minority groups in health care settings, with the goal of identifying modifiable mechanisms and effective interventions. They found that multicomponent approaches, rather than single-component approaches (for example, screening only), and culturally adapted interventions, rather than standard psychotherapy interventions, are successful in improving depressive symptoms. However, they noted that further research is needed to help determine key components of effective interventions for ethnic and racial minority groups.

To address this need, we developed a CFP consultation service for Latino primary care patients with depressive symptoms. The CFP service entailed use of multicomponent consultations over the course of two visits to improve the recognition and treatment of depression. Intervention goals were to facilitate patients’ knowledge of and resources for receiving treatment and primary care providers’ ability to provide appropriate, culturally informed care. Our primary aims were to assess feasibility, satisfaction, and cost of this multicomponent intervention (29). This report is based on a secondary data analysis to test whether the intervention reduced depressive symptoms.

Methods

Participants

Participants included Latino adults attending a primary care practice affiliated with Massachusetts General Hospital (MGH). Eligible patients included those who were age 18 or older, who self-identified as Latino/Hispanic and screened positive for likely depression (Patient Health Questionnaire-2 [PHQ-2] score ≥2 out of 6) (30), who were able to consent to study participation, and who were free from unstable psychiatric illness precluding participation (for example, actively suicidal).

Primary care practices

Four MGH adult primary care practices agreed to participate in the study—one located at the main hospital in Boston and three located in Boston-area community health care centers. All participating clinics within each practice had a large number of Latino patients. At each practice, the mental health service was located separately from the primary care clinic; providers seeking to have their patient seen by a mental health provider would provide their patient with a referral. On the basis of their insurance, patients could also seek a mental health provider in the community.

Study design

Details of the study protocol and intervention have been published elsewhere (29) and are summarized below. The MGH Institutional Review Board approved all study procedures (Clinicaltrials.gov identifier NCT01239407). Study materials were available in English and Spanish.

Clusters of participating primary care providers at each of the four clinics were randomly selected to administer either the CFP intervention or enhanced usual care. Randomization was based on clinic, rather than patient, as a measure to prevent providers from being exposed to the intervention and then unintentionally providing it to enhanced usual care participants.

Participants were recruited between December 2009 and November 2010, with similar recruitment strategies applied for both groups. Recruitment was conducted as part of the patient’s routine primary care visit; patients were given a screening form at check-in, before meeting with the medical provider, that contained a study description, the PHQ-2, demographic questions, and an option to consent to be contacted by study staff. Completed screening forms were then available to the provider at the time of the medical visit and were collected afterward by study staff. Patients who met initial eligibility criteria on the basis of the clinic screen and who consented to be contacted were then contacted to confirm eligibility.

Patients in both arms completed an initial baseline and six-month follow-up psychosocial assessment, receiving $8 gift cards for each. The analyses reported here included data for all enrolled Latinos who completed the initial assessment.

Intervention

CFP consultation.

Components of the CFP consultation intervention were previously described (29). The intervention’s primary aims were to directly affect patient’s depressive symptoms and influence the primary care provider’s provision of culturally informed care.

During the baseline visit, study clinicians conducted a psychiatric interview and reviewed psychoeducation materials, provided in both a print version and an audio CD, which included a depression tool kit, cognitive-behavioral tools, and community resources. The DSM-IV-TR cultural formulation model (1,28) and Engagement Interview Protocol (31) were applied to engage the patient, inform culturally appropriate diagnosis, and provide tailored treatment recommendations. Study clinicians also had access to participants’ electronic health records.

Participants were also offered a second visit within two weeks, during which the clinician addressed participant questions and adherence to recommendations. The initial visit typically lasted 90 minutes and the second visit lasted 45 minutes. A summary of diagnostic and treatment recommendations was communicated to the provider and, as appropriate, to other members of the existing health care team (for example, psychiatrist or neurologist).

In addition, triage was conducted for participants who endorsed critical psychiatric issues that necessitated contacting the primary care clinic or communicating with the mental health service in order to facilitate care. To ensure protocol fidelity, study clinicians also participated in weekly supervision meetings.

Enhanced usual care condition.

In enhanced usual care, primary care providers were notified via e-mail of positive depressive symptom screens, and participants received usual care for depression through their primary care practice.

Measures

Initial screening.

The initial screening form included demographic factors (that is, age, race-ethnicity, gender, and language) and the PHQ-2. The two-item PHQ-2, a brief tool to assess depression likelihood (30,32), measures mood and anhedonia over the past two weeks with a four-item Likert scale (“not at all” to “nearly every day”). A cutoff score of ≥2 (out of a possible total score of 6) has been validated for diagnosing any clinical depressive disorder with an 82% sensitivity, 80% specificity, and positive predictive value of 48% (30).

Baseline demographic factors and mental health care use.

At baseline, participants completed a demographic questionnaire that assessed education, marital, parental, employment, and housing status; birth country (Puerto Rico was considered born outside the United States); and, if applicable, age at immigration. Four questions were used to assess current mental health care utilization through a primary care provider, a psychiatrist or mental health provider, any form of counseling, or psychopharmacology. Responses to each were recorded as yes or no. For this study, one categorical variable was used to indicate whether participants were receiving mental health services at baseline.

Depressive symptoms.

At baseline and six-month follow-up, depressive symptoms were assessed with the 16-item Quick Inventory of Depressive Symptomatology–Self Rated (QIDS-SR) (33). Total scores range from 0 to 27, with severity categorized as follows: none, 0–5; mild, 6–10; moderate, 11–15; severe, 16–20; and very severe, 21–27. In similar samples, the QIDS-SR was found to be equivalent to the clinician-rated version (34,35), with equivalence also shown between English and Spanish versions (36).

Statistical analyses

Bivariate analyses were first conducted to compare baseline characteristics of participants in the CFP consultation intervention and enhanced usual care. We also tested associations between baseline characteristics and depressive symptom severity. Finally, we tested differences in baseline characteristics between study completers and noncompleters (that is, those who did or did not complete the follow-up assessment). Relevant characteristics were subsequently included in our primary analyses.

To evaluate whether CFP intervention participants showed greater improvements in depressive symptoms from baseline to six-month follow-up compared with those in enhanced usual care, we tested a multiple regression model in which group condition (CFP intervention versus enhanced usual care) was estimated to predict depressive symptoms at follow-up, controlling for baseline depressive symptoms, primary care site, and relevant covariates. Analyses were conducted by using an intent-to-treat model, applying multiple imputation (five imputed data sets) to account for missing data. Missing data at baseline were limited to 3% for one covariate (employment) and to 15% for the depressive symptom outcome at the six-month time point. In the case of an independent association between group condition and six-month depressive symptoms, we planned follow-up tests to evaluate change in depressive symptoms within each group. All analyses were tested by using a two-sided p<.05 and conducted by using SPSS software, version 21.

Results

Baseline characteristics

Of the 1,975 Latinos who completed screening forms, 552 were eligible and agreed to be contacted. Of these, 120 enrolled, and two were excluded at enrollment because participation was not appropriate (for example, active suicidality). Participation rates differed by site, with one site accounting for a majority of participants (N=99, 84%), compared with the other three (N=10, 9%; N=2, 2%; and N=7, 6%). [A flowchart illustrating sample recruitment is presented in an online data supplement to this article.]

Baseline sample characteristics of participants in the CFP intervention and enhanced usual care (total N=118) are reported in Table 1. Participants were primarily Spanish-monolingual speakers (64%) and had migrated at 12 years of age or older (74%). The mean depressive symptom score on the QIDS-SR (14.6±5.0) indicated moderate severity (33). Slightly more than half of participants (54%) reported currently receiving mental health care. Rates of mental health care from various providers did not significantly differ between groups—a primary care provider: enhanced usual care, 28%; CFP intervention, 15%; a psychiatrist or mental health provider: 30% and 25%, respectively; other forms of counseling: 24% and 23%, respectively; and psychopharmacology: 37% and 41%, respectively (detailed data not shown).

CFP intervention participants were older than those in enhanced usual care (p=.03) and more likely to be male (Fisher’s exact test, p<.001), with no other differences in other baseline characteristics (Table 1). Also, among baseline characteristics, employment status was the only variable associated with depressive symptom severity (F=8.23, df=2 and 111, p<.001); those who were employed reported lower severity (mean±SD QIDS-SR score=12.5±5.0) than those who reported not being employed due to a disability (16.9±4.6) or who were otherwise not employed (15.4±4.7). Evaluation of study attrition indicated that study completers (N=100) (those who completed the six-month follow-up visit) were more likely than noncompleters (N=18) to be currently employed (N=45, 45%, versus N=2, 12%; χ²=8.00, df=1, p=.02), with no other baseline differences between groups. Age, gender, and employment status were, therefore, included as control variables in the primary model. Analyses also controlled for clinic site.

Change in depressive symptoms

A multiple regression model was evaluated to determine whether the CFP intervention group showed improvements in depressive symptoms at the six-month follow-up compared with the group receiving enhanced usual care, with the analysis controlling for clinic site, baseline depressive symptoms, age, gender, and employment (Table 2). Results indicated that participation in the CFP intervention predicted lower depressive symptom severity at follow-up compared with receipt of enhanced usual care (B=−3.09). Reduction in depressive symptoms was greater among CFP participants (change in mean QIDS-SR score of 3.46±5.48) than among those receiving enhanced usual care (.09±4.43). Depressive symptoms at follow-up remained in the moderate range for both groups (CFP intervention, 11.5±7.0; enhanced usual care, 13.8±5.4), although the average score for CFP intervention participants approached the mild depression range. In a separate analysis, among CFP participants only, time from baseline to provider receipt of consultation results (less than six weeks versus six weeks or more) was not associated with change in depression at follow-up.

Discussion

Our findings provide preliminary evidence that a brief psychiatric consultation with a cultural focus improved depressive symptoms among Latinos in primary care. Latinos who received the CFP intervention reported significantly fewer depressive symptoms at the six-month follow-up compared with study participants who received enhanced usual care alone. This finding expands on our previous findings indicating strong satisfaction with the CFP intervention (37).

These results add to a growing literature on the benefits of integrated or collaborative care, in which mental health specialists work with primary care providers to improve depression management and outcomes for minority populations (38–40). Although our study utilized a psychiatric consultation model, the intervention was conducted in the primary care setting, and, when appropriate, study clinicians communicated not only with primary care providers but also with patients’ other providers. In severe cases, study clinicians worked with the medical center's mental health service to schedule an initial intake. The impact of integrated mental health care on depression outcomes among Latinos has been mixed (41–44). On the basis of the CFP intervention model, one or two sessions with a highly experienced clinician (that is, a psychologist or psychiatrist) trained in culturally competent techniques holds promise for improving depressive symptoms.

Results are also consistent with Van Voorhees and colleagues’ (4) systematic review, in which they concluded that the most successful interventions for reducing depression-related disparities are multicomponent approaches that affect one or more domains, including system-, community-, provider-, or patient-level factors. For example, interventions targeting depression management in primary care have not been successful when only screening is provided or when only diagnostic data are given to the provider and the patient (45). Our results supported this finding in that patients receiving enhanced usual care, which consisted of informing primary care providers of positive depression screens, did not exhibit significant change in depressive symptoms. As described above, the CFP intervention included multiple components and was also designed to place minimal additional burden on the existing primary care system. Using Van Voorhees and colleagues’ framework (4), the CFP intervention included components that simultaneously targeted system-, provider- and patient-level factors; as a psychotherapy intervention, it also provided culturally tailored care that may have served to bridge primary care and specialty mental health care.

Limitations of the study should be considered. The primary study aims were to test intervention feasibility, satisfaction, and cost. The findings reported here are preliminary and need to be replicated in a larger trial. Differential enrollment is also a potential source of bias that could not be examined. For example, only a quarter of eligible patients participated, and most came from one site. In addition, participants, providers, and study personnel were aware of randomization status. It is possible that providers treated participants in the intervention arm differently. Similarly, at follow-up, participants who received the intervention may have endorsed improved depressive symptoms to show appreciation for study efforts. Mental health care use was not assessed over the study course. Therefore, we cannot determine whether the potential benefits of the intervention were mediated by service utilization. The number of clinic clusters and participation rates did not allow for analyses that controlled for clustering. Finally, study results may not be generalizable to primary care clinics with different or fewer mental health care resources.

These findings suggest areas for future research. Participants who received the intervention reported greater symptom reduction on a validated measure of depression, although symptom severity remained in the moderate range. More work is needed to understand which of the components were key to improving depressive symptoms, the extent that psychiatric service utilization was influenced by the intervention, and how best to enhance the intervention’s efficacy. Participants self-selected to be in the study; future studies could examine factors that influence participation in depression care, such as why significantly more men enrolled in the CFP intervention. In addition, intervention cost-effectiveness and scalability may be limited because of the use of highly trained and experienced bilingual clinicians. Further study is needed to evaluate intervention delivery by clinicians who may be more accessible in community primary care practices.

Conclusions

The CFP consultation intervention complements the current movement in psychiatry to enhance the treatment of depression in primary care settings. These results add to the growing literature on the benefits of psychiatric specialists working in collaboration with primary care providers to decrease mental health disparities. Patients, however, may benefit from a more intensive CFP intervention, given that participants who received the intervention reported improved symptoms at follow-up but continued to endorse moderate depressive symptoms. The positive findings for this group of Latino primary care patients, most of whom spoke only Spanish and were foreign born, also suggest the benefits of providing patients with psychiatric services that are multicomponent and address patients’ cultural and linguistic needs.