کارآیی و کارآمدی ترکیبی کارآزمایی تصادفی شده فعالسازی رفتاری برای لاتینی ها با افسردگی

Depression presents a significant public health burden for Latinos, the largest and fastest-growing minority group in the United States. The current study performed a randomized controlled trial of Behavioral Activation (BA) for Latinos (BAL, n = 21), with relatively minor modifications, compared to treatment as usual (TAU, n = 22) in a community mental health clinic setting with a sample of depressed, Spanish-speaking Latinos. TAU was a strong comparison condition, taking place at the same clinic, under the same guidelines and clinic protocols, with similar levels of ongoing consultation, and using the same pool of therapists as BAL. Results indicated that BAL performed well with respect to treatment engagement and retention. Regarding acute treatment outcomes, an interaction emerged between number of sessions attended and condition. Specifically, only BAL clients who were engaged in treatment and attended more sessions demonstrated significant reductions in depression and improvements in quality of life and mental health functioning. Results are discussed in terms of the balance of efficacy and effectiveness issues addressed in this trial.

Latinos are the largest minority group in the United States, comprising over 50 million individuals in 2010 (Passel, Cohn, & Lopez, 2011), with projections that the population will double in size by 2050 (Passel & Cohn, 2008). Depression, identified by the World Health Organization (2008) as one of the most burdensome diseases in the world, presents a significant public health problem for Latinos in the United States (U.S.), with U.S. Latinos reporting comparable rates of major depression diagnoses and possibly higher levels of depressive symptoms compared to non-Latino Whites (Mendelson, Rehkopf, & Kubzansky, 2008). Many contextual factors are important to the etiology of depression among U.S. Latinos (Cabassa et al., 2007 and Martinez-Pincay and Guarnaccia, 2007). Research has related the onset of depression among U.S. Latinos, for example, to stressful immigration experiences (Grzywacz et al., 2010), the process of acculturation and adapting to a new environment (Organista, Organista, & Kurasaki, 2003), separation from children and family (Miranda, Siddique, Der-Martirosian, & Belin, 2005), and overrepresentation in low socio-economic status brackets (Bruce et al., 1991 and Vega et al., 1998). Other contextual factors, including experiences of racism and discrimination, stressful interactions with agencies, and language barriers, also are seen as important (Santiago-Rivera, Arredondo, & Gallardo-Cooper, 2002). In line with these factors, U.S. Latinos tend to conceptualize depression as having contextual origins, rather than in other terms (e.g., biological or cognitive; Martinez-Pincay & Guarnaccia, 2007). A primary obstacle to the psychotherapeutic treatment of depression in U.S. Latinos has been engaging and retaining Latinos in treatment (Fortuna, Alegria, & Gao, 2010). Variants of cognitive-behavior therapy (CBT) have been developed (Muñoz and Mendelson, 2005 and Muñoz and Miranda, 1986) and evaluated (Voss Horrell, 2008) to address this issue. In these trials, improvements in retention rates have been observed, often after the introduction of supplemental, resource-intensive interventions such as additional case management, psychoeducation for those unfamiliar with depression, provision of child care services and transportation, and cultural competency training for therapists (Miranda, Azocar, Organista, Dwyer and Areane, 2003, Miranda, Cheng, et al., 2003 and Miranda, Duan, et al., 2003). Depression treatment outcomes have also improved in these studies, which typically involve complex, multifaceted CBT interventions sometimes as part of larger, collaborative care models (Wells et al., 2004). More research is needed to produce culturally sensitive, resource-efficient interventions that are disseminable to community settings in order to improve treatment engagement, retention, and outcomes for U.S. Latinos with depression. Kanter, Santiago-Rivera, and colleagues (Kanter et al., 2008, Kanter, Santiago-Rivera, Rusch, Busch and West, 2010 and Santiago-Rivera et al., 2008) proposed that Behavioral Activation (BA; Kanter et al., 2009 and Martell et al., 2001) may represent a promising treatment option in this regard. BA, identified as an empirically supported treatment for depression (Mazzucchelli, Kane, & Rees, 2009), primarily involves identifying and scheduling personally meaningful activities to reduce depression for clients by addressing avoidance and other obstacles to activation (Kanter, Manos, et al., 2010). BA may be a good fit for Latinos for several reasons. First, the rationale for BA treatment focuses directly on contextual factors such as those described above and thus may be seen as relevant and acceptable to Latinos, who tend to prefer present-focused, active treatment strategies (Santiago-Rivera et al., 2008). Second, because BA treatment targets are collaboratively determined based on the client’s identified values and life goals, BA may be easily adapted to be consistent with Latino cultural values without compromising its mechanism of action (Kanter et al., 2009 and Santiago-Rivera et al., 2008). Third, because of BA’s parsimonious and relatively straightforward set of treatment techniques, many authors have suggested that BA holds promise as easy to train and disseminate, especially to clinicians working with populations that have barriers to access to quality mental health care, such as recent immigrants who are not familiar with Eurocentric models of treatment (Kanter, Puspitasari, Santos, & Nagy, 2012). Thus, it was hypothesized that BA, with relatively minor adaptations, would be feasible, appropriate, and effective when administered in Latino community mental health clinic settings. Results of an open trial of BA for Latinos (BAL) at such a clinic provided preliminary support for the feasibility and effectiveness of BAL in terms of treatment adherence, retention, and outcomes (Kanter, Santiago-Rivera, et al., 2010). The current study extended the open trial of BAL (Kanter, Santiago-Rivera et al., 2010) to a randomized controlled trial of BAL compared to treatment as usual (TAU) at a bilingual (Spanish-English) community medical/mental health clinic. Several features of this trial, which combined elements of efficacy and effectiveness research consistent with the treatment development model of Weisz (2004), are noteworthy. First, the study took place at a comprehensive community medical/mental health center targeting the medically underserved located centrally in a Latino enclave of a large midwestern urban center. Many possible obstacles to treatment utilization and engagement by Latinos (Alegría et al., 2002) are successfully addressed by this center as part of their TAU, including access to behavioral health specialists (all therapists were master's or Ph.D. level), the presence of multilingual therapists, easy access (located on a major arterial in the center of the Latino community), financial affordability (acceptance of Medicaid and uninsured/undocumented), and cultural competence (the center has earned a strong reputation in the community as culturally competent). Thus, the community clinic setting for this trial was a model for adherence to many of the recommendations for addressing Latino health disparities as the U.S. adapts to the increasing Latino population (McGuire & Miranda, 2008). A demonstrated improvement in treatment retention or outcomes over TAU at such a clinic would be meaningful. A related noteworthy feature of this study is the nature of the TAU comparison condition. In the past, concerns about the use of TAU comparison conditions in randomized trials have been raised, because in some trials TAU has consisted of nothing more than referral to unspecified and untracked outside providers, selected in advance to function as TAU therapists, while the therapists in the experimental treatment arm were hand-selected, were required to pass significant training criteria, and received significant amounts of supervision and monitoring over the course of the study, among other differences (Wampold et al., 2011). Such research design decisions confound the treatment-TAU comparison and bias results against TAU. The current study, in contrast, employed a strong, ecologically valid TAU condition, with procedures in place to ensure equivalence between BAL and TAU on variables such as treatment setting, treatment affordability, treatment availability, availability of consultation, and institutional reputation. This was done to ensure that the research results would be informative with respect to improving real-world community clinic functioning. Likewise, the training of BAL therapists was kept to levels that would be potentially implementable in community clinic settings, and a 12-session protocol was chosen in line with feasibility of implementation in this setting. The two primary hypotheses were that BAL would improve retention rates and treatment outcomes compared to TAU at the end of acute treatment. Secondary outcome measures included health functioning and quality of life. An additional aim of this study was to explore outcomes over a 9-month follow-up period. However, an unexpected outcome of this study was a large amount and differential pattern of attrition to follow-up that co-varied with depression change over the course of therapy, rendering the follow-up outcomes unreliable and possibly misleading. These observed patterns of attrition are presented and discussed in lieu of follow-up outcomes.

Table 1 presents demographic and clinical characteristics of the sample at baseline. Significant differences between the treatment groups on randomization or other demographic and clinical characteristics were not observed. Table 1. Baseline Characteristics of Participants BAL (n = 21) TAU (n = 22) Full Sample (N = 43) Female: n (%) 16 (76.2) 18 (81.8) 34 (79.1) Age: M (SD) 38.7 (11.7) 37.5 (10.1) 38.1 (10.8) Unemployed: n (%) 11 (52.4) 12 (54.5) 23 (53.5) Income*: n (%) ≤ $10,000 9 (42.9) 11 (50.0) 20 (46.5) $10,000-20,000 6 (28.6) 5 (22.7) 11 (25.6) > $20,001-$30,000 3 (14.3) 5 (22.7) 8 (18.6) Married or common law: n (%) 11 (52.4) 12 (54.5) 23 (53.5) Not born or raised in the U.S.: n (%) 18 (85.7) 16 (72.7) 34 (79.1) Country of Origin**: n (%) Mexico 14 (66.7) 15 (68.2) 29 (67.4) Puerto Rico 6 (28.6) 3 (13.6) 9 (20.9) Other 1 (4.8) 3 (13.6) 4 (9.3) Pre-treatment BDI-II: M (SD) 34.4 (9.2) 29.4 (10.1) 31.8 (9.9) Low severity (BDI-II 14–28): n (%) 6 (28.6) 11 (50) 17 (39.5) High severity (BDI-II 29–63): n (%) 15 (71.4) 11 (50) 26 (60.5) Post-traumatic Stress Disorder: n (%) 7 (33.3) 7 (31.8) 14 (32.6) Any other comorbidity: n (%) 7 (33.3) 8 (36.4) 15 (34.9) Note. Results of independent samples t-tests for continuous variables and chi-square tests of independence for categorical variables demonstrated no difference between the treatment groups. *Missing for four people (3 from BAL). **Missing for 1 person (TAU). BDI = Beck Depression Inventory; HRSD = Hamilton Rating Scale for Depression. Table options Significant attrition to research assessment at posttreatment was observed. Of the 43 clients in the original sample, 15 (34.9%) clients did not provide posttreatment data, including 5 (23.8%) BAL clients and 10 (45.5%) TAU clients. This led to concerns about outcome analyses at posttreatment possibly being biased by differential attrition. To explore this, post-hoc analyses examined several key variables, specifically pretreatment BDI-II and HRSD scores, number of sessions attended, and final session BDI-II before dropout, for possible differences by condition between those who provided data for the posttreatment assessment and those who did not. To be conservative, these analyses were not corrected for multiple comparisons. Findings suggested that there were no interactions between condition and postintervention assessment attrition on pretreatment HRSD scores, but trends were observed for both BAL and TAU clients with respect to pretreatment BDI-II scores. Specifically, BAL clients who did not provide posttreatment data tended to have higher pretreatment BDI-II scores (M = 40.80, SD = 9.94) compared to BAL clients who did provide posttreatment data (M = 32.38, SD = 8.26), F(1, 19) = 3.62, p = .07, while TAU clients who did not provide posttreatment data tended to have lower pretreatment BDI-II scores (M = 25.30, SD = 7.01) compared to TAU clients who did provide posttreatment data (M = 32.83, SD = 11.16), F(1, 20) = 3.42, p = .08. However, this possible difference disappeared over the course of therapy, before the clients were lost, as per last session BDI-II scores, which reversed the pattern observed at pretreatment and were not significantly different. Therefore, evidence for a biased impact of differential attrition on primary outcomes at posttreatment was not found. At follow-up, significant attrition to research assessment also was observed, with 18 (41.9%) clients who did not provide follow-up data, including 10 (47.6%) BAL clients and 8 (36.4%) TAU clients. Post-hoc analyses suggested that pretreatment BDI-II scores and BDI-II scores over the course of therapy predicted a differential pattern of attrition to the follow-up assessment between BAL and TAU clients. Regarding pretreatment BDI-II scores, TAU clients lost to follow-up had significantly higher pretreatment BDI-II scores (M = 35.63, SD = 12.15) compared to TAU clients who provided data at follow-up (M = 25.86, SD =6.80), F(1, 20) = 5.94, p = .03. Furthermore, the BAL clients who were lost at follow-up had significantly lower last available BDI-II scores before they were lost (M = 6.00, SD = 6.63) compared to BAL clients who provided data at follow-up (M = 20.56, SD = 17.19), F(1, 17) = 6.18, p = .02, while the TAU clients who were lost at follow-up had significantly higher last available BDI-II scores (M = 38.43, SD = 20.09) compared to TAU clients who provided data at follow-up (M = 17.85, SD = 9.62), F(1, 18) = 9.81, p = .006. These findings are depicted in Figure 2. In other words, it appeared to be the case that most of the clients who were doing relatively well in BAL (according to their last available BDI-II scores) were lost to the follow-up assessment, while clients who were doing relatively poorly in TAU were lost to follow-up, and this pattern represented a reversal from other findings at posttreatment. Thus, because of significant concerns about the biasing effects of these patterns of differential attrition on follow-up data, follow-up results were not presented. Full-size image (39 K) Figure 2. BDI-II Scores for BAL and TAU Participants Who Provided Data at Follow-up and Who Were Lost to Follow-up.