An official website of the United States government
NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
Kathleen Ell, DSW, María P. Aranda, PhD, Shinyi Wu, PhD, Hyunsung Oh, PhD, Pey-Jiuan Lee, MS, and Jeffrey Guterman, MD, MS.
Author Information and AffiliationsDepression has negative effects on patient self-care and social stress management. The negative effects of depression disproportionately affect low-income Latino patients with chronic medical illness.
To evaluate the effectiveness of A Helping Hand (AHH; Programa Mano Amiga in Primary Care) for patients with depressive symptoms and comorbid medical illness.
Patients with significant depressive symptoms (9-item Patient Health Questionnaire score ≥10) and coexisting diabetes or heart disease were randomized to AHH or usual care (UC) in 3 Los Angeles County Department of Health Services (LAC-DHS) safety-net clinics that were implementing patient-centered medical home (PCMH) models. The AHH intervention supported patients, families, and care providers by facilitating self-care management skills and activating patient communication with clinic medical providers. Community based, bilingual promotoras delivered the intervention in 6 weekly in-person or telephone sessions, followed by 3 monthly booster sessions. From April 2014 to May 2015, we screened 1957 and enrolled 348 depressed patients, of whom 296 (85%) had diabetes, 14 (4%) had heart disease, and 38 (11%) had both diseases. All participants received care management materials and community resource information. An interviewer blind to intervention assignment assessed outcomes at 6 and 12 months. Baseline and outcome data include depression, mental health assessments, treatment receipt, comorbid illness self-care, social relationships, and environmental stressor assessments.
Study participants were predominantly female (85%), Latino (99%), and born outside of the United States (91%). Overall study retention rate was 70% (121 AHH and 121 UC). Baseline characteristics did not vary significantly between retained and attrition groups. Half of AHH patients received 4 or more promotora sessions. Promotoras made 12 referrals to LAC-DHS providers and 154 referrals to community resources (most frequently requested community services: community/senior/wellness center, 88 occurrences; transportation, 33; food bank, 25). During the trial period, LAC-DHS activated health care improvements, including adding community health workers into UC clinics. Depression outcomes did not vary significantly between intervention and usual care groups (UC as the reference group; at 6 months: mean difference = 0.01; 95% CI, −1.3 to 1.3; at 12 months: mean difference = −1.1; 95% CI, −2.5 to 0.2); however, we found significant improvements in most assessed physical and mental health outcomes for each study group.
No significant differences existed in primary depression outcomes between the AHH intervention and the PCMH usual care study groups.
The challenges were to maximize intervention attendance and minimize study attrition given the high representation of immigrant, Spanish-speaking, safety-net population in the sample. The effects of the intervention were confounded by major quality improvement initiatives in the participating clinics. Future work is needed to provide a more definitive test of the AHH promotora model, while addressing these potential confounders.
Activated in 2014, the Affordable Care Act (ACA) health insurance mandate included an aim to significantly reduce racial and ethnic disparities.1 Unfortunately, there is a concern that Latinos and African Americans will continue to have problems accessing and using high-quality health care, especially in states that are not expanding Medicaid eligibility as provided by the ACA.2-4 In the United States, public health safety-net organizations are facilitating integrated behavioral health services through biopsychosocial team preferences, language proficiency, and self-care skills. It is noteworthy that California is expanding community health centers and promoting patient-centered medical home (PCMH) models.
Depression is a common mental health disorder among low-income patients; depression with concurrent comorbid chronic illnesses can increase patient depression relapse and recurrence, morbidity, and mortality. At the same time, depression can negatively impact patient self-care. Difficulty in managing concurrent comorbid illness can also trigger depression.5,6 For example, low-income patients with diabetes are at high risk of clinically significant depression over time, and depression can negatively affect both depression and diabetes self-care management.5,7-9 A recent study among heart failure patients found that depression independently predicted increased use of health care resources and mortality, and suggested improved management of depression may improve outcomes.10-14 However, among minority heart failure patients, depression is often persistent and severe, whereas perceived emotional and informational support is associated with better self-care maintenance and self-care management.15 Depression and anxiety have been related to an increased risk of mortality in coronary heart disease patients.16
Integrated biopsychosocial team care is increasingly recognized as essential for patients with depression and other chronic illnesses.17,18 Unfortunately, depression care that is tailored to low-income, ethnically diverse patient populations is not readily available.19 Safety-net patients with major depression plus a concurrent chronic illness face significant barriers to patient depression care receipt, motivation, skills, and confidence that equip patients to become actively engaged in their health care.20 Patients with poorer health literacy, lower levels of social support, and more severe depression are also more likely to have poorer medical illness outcomes. Concurrent self-care challenges include (1) managing overlapping concurrent symptoms (eg, depressed mood, pain, anxiety, fatigue, adverse reactions to medications, overall medication management); (2) making daily decisions that affect overall health, including potential negative effects on concurrent illness treatment management and adherence; (3) addressing sociocultural and economic influences, such as day-to-day coping demands,21 ongoing or intermittent social relationship distress or abuse,22-24 and economic stress; and (4) managing family and caregiver communication regarding depression and chronic disease symptoms management, depression treatment preferences and stigma concerns, depression potential negative effects on comorbid concurrent chronic illness treatment uptake, and dietary and exercise management. Disparities in patient-centered care management are troubling given evidence that depression care for low-income, minority patients is effective in reducing depression symptoms and increasing self-care adherence and depression treatment retention, including among those with concurrent physical illness.5,6,25-30
Safety-net populations are also likely to encounter significant self-care management stressors triggered by patient difficulties in communicating with primary care providers,27 navigating multiple specialist care providers, managing uncoordinated treatment plans, and navigating existing community organization resources.28 At the same time, time-pressured safety-net medical providers are responsible for synthesizing multiple and complex health-related information, managing prescriptions, and communicating with multiple providers, including specialists and hospital/ER providers. Primary care providers often find that engaging patients with major depression is a significant challenge and, not surprisingly, is even more difficult when the depression is accompanied by a concurrent chronic illness, because engagement requires conducting initial and follow-up routine depression assessment and management over time.29,30
Biopsychosocial care models include medical, psychological, and social emotional stress, such as depression, whereas the social component investigates how different social factors such as socioeconomic status, culture, poverty, technology, and religion can influence health and depression.31 In a philosophical sense, the biopsychosocial model states that the workings of the body can affect the mind, and the workings of the mind can affect the body.4 This means both a direct interaction between mind and body as well as indirect effects through intermediate factors. Health is best understood in terms of a combination of factors rather than solely in biological terms. This is in contrast to the biomedical model of medicine that suggests every disease process can be explained in terms of an underlying deviation from normal function, such as a virus, a gene or developmental abnormality, or an injury.
Low-income patient self-care is also affected by culture, literacy, and language and financial barriers that often exceed provider skills in patient communication.32 Among Latinos (a significant majority in Los Angeles County Department of Health Services [LAC-DHS] clinics), group orientation/harmony in families can encourage or deter depression or physical illness care.33,34 In Latino culture, depression treatment is more effective if it aligns with contextual stress-related needs and cultural values and if it considers preferences for counseling over depression medication.17 The project A Helping Hand (AHH) assessed the comparative effectiveness of 2 diverse multiple-team providers: AHH integrated thoroughly trained English–Spanish bilingual promotoras into the standard Department of Health Services (DHS)-PCMH team of physician, nurse, and medical assistant. We evaluated the effects of promotora-mediated patient assistance in reducing patient depression symptoms, activating overall health management over time, reducing depression and comorbid illness treatment barriers, and optimizing patient care uptake and satisfaction in low-income Latinos. This study is the first to incorporate promotoras in a public health safety-net care system to reduce racial and ethnic disparities in depression and self-care management among Spanish-speaking patients with diabetes and/or heart disease.
Promotora is a commonly used Spanish term for what is referred to in English as a community health worker (CHW), navigator, outreach worker, lay health advisor, patient advocate, etc. His or her expertise is commonly used in health care, although not exclusively, as a promotora can be deployed to other workforce areas—eg, housing, education, social services, environmental advocacy. Like other CHWs, promotoras are commonly deployed in areas with high representation of limited-English speakers, in low-income and racially and ethnically diverse communities, in rural areas, in communities with shortages of health providers; in places with migrant groups, and in communities regarded as “hard-to-reach” due to public distrust of formal health care providers.35,36 Thus, the integration of promotoras has strong appeal across several types of target areas and communities. Implementation of promotoras in health care presents the same funding issues that any new service or program encounters. Funding for promotoras beyond grant-funded research is an ongoing challenge and the most frequently cited barrier to sustaining promotora-led programs.37 Promotoras and CHWs are prepared to engage patients and enhance their health literacy, address a wide range of community health issues, and provide advocacy and leadership development; however, significant DHS safety-net system challenges exist, including costs related to community health planning, management, and policies that include developing sustainable strategies (eg, they have recently activated a focus on preventive care). DHS PCMH is a path toward health equity, as it provides care that does not vary in quality due to patient characteristics—ie, gender, ethnicity, socioeconomic status, and geographic location—and that focuses on community engagement. The Los Angeles County DHS PCMH model (LAC-DHS-PCMH) continues to activate and encourage multiple practice and research opportunities. These opportunities are aimed at reducing fragmented care via biopsychosocial health care providers, patients, and family member stakeholders, and are subsequently aiming to implement and disseminate PCMH effective care management throughout LAC-DHS care network. In the United States, patient-centered care is emerging as a key element for improving the quality of health care, as it strengthens the patient–clinician relationship via coordinated care.
The University of Southern California and the Los Angeles County Department of Health Services (USC-DHS) research teams have previously conducted comparative effectiveness randomized and quasi-experimental clinical trials as well as qualitative studies on major depression. We conducted A Helping Hand (AHH) in collaboration with the DHS Ambulatory Care Network, the second largest safety-net care system in the United States. AHH is based on previous USC-DHS depression studies among DHS patients with diabetes, heart failure, and coronary heart disease.5,6,25,27-29
Engaged community stakeholders included nonstudy patients from previous DHS trials who were similar to the current study patients; California community organization Visión y Compromiso (VyC) promotoras and leadership; and DHS Ambulatory Clinics providers and administrative staff, including physicians, nurses, and clinic directors. Initially, the research investigators met with nonstudy patients (who were recruited from the same subject pool via telephone), provided promotora training, and held meetings with the clinic directors and their staff.
As a result of patient and stakeholder participation, we developed a community resources list of available services, programs, and products. All participants were socioeconomically disadvantaged, and many were facing significant financial strain, immigration issues, and daily life pressures. Other study design changes included augmenting the availability of completing intervention and interview sessions in the home rather than on site in the clinics. In addition, this participation informed the qualitative interview guide questions regarding quality of interactions with clinic staff and providers and degree to which matched-provider ethnicity and language availability was important to the patients.
The study population consisted of low-income, predominantly Latino patients (n = 348) with depression and concurrent diabetes and/or heart disease from 3 DHS safety-net community clinics. After completion of the baseline assessment, study participants were randomly assigned into AHH enhanced innovative promotora patient depression and care self-management support (AHH) or standard clinic depression treatment provided by the DHS PCMH multiple provider primary care model clinic team usual care (UC). Thus, the AHH trial compared 2 diverse multiple-team providers: AHH, which added a thoroughly trained promotora provider team member, to the standard DHS clinic usual care team of providers with its PCMH physician, nurse, and medical assistant.
Be based this randomized controlled trial (A Helping Hand/Programa Mano Amiga in Primary Care) on the assumption that personal, socioeconomic, cultural, and health systems processes are key elements in predisposing, reinforcing, and influencing patient-centered depression and chronic illness self-care management. Patient self-care management and treatment adherence are influenced by individual choice, cultural beliefs, stigmas, practical barriers, and physical health in combination with contextual supports/barriers (eg, missed work time, transportation to clinic costs, difficulty in communicating with providers who do not speak Spanish, low health literacy). Generally, low-income patients have fewer resources to support treatment participation.38 Culturally normative patterns of physical and psychological symptom definition, response, and help-seeking are critical antecedents of self-management behavior and care utilization, but they are not independent of environmental and care provider/system factors. Evidence exists that patient self-efficacy and provider–patient communication influences self-care management and treatment adherence. Culturally competent/tailored self-management education may enhance treatment adherence and care management.39 Thus, the study provided to both AHH and UC participants and their families culturally adapted depression educational booklets, including a photo storybook fotonovela40 and brochures on specific chronic illness treatments and self-care. All educational materials were available in English and in Spanish.
For the primary outcomes, we hypothesized reduced depression symptoms at 6 and 12 months among AHH vs UC patients, as well as improved concurrent illness self-care management. The secondary research question addressed medical care utilization: whether AHH reduced hospitalizations and ER visits and improved patient care satisfaction and quality of life. We also conducted qualitative assessments of patients, promotoras, DHS medical providers, and clinic organizational leaders.
The study was approved by the University of Southern California-Health Sciences Institutional Review Board and conducted in collaboration with LAC-DHS at 3 community clinics within similar geographic neighborhoods and with similar demographics (ie, race/ethnicity predominantly Latino with more than 50% monolingual Spanish speakers, 90% younger than age 65, low household income with a significant number living below the poverty rate, and 30% to 35% receiving Medicaid as health insurance). Inclusion criteria were aged 18 years and older, English or Spanish speaking, can communicate by phone, screened with clinically significant depression (ie, having at least 1 cardinal symptom of depressed mood or loss of interest more than half the days to nearly every day plus the 9-item Patient Health Questionnaire (PHQ-9) score of 10 or more), and concurrent diabetes and/or heart disease (coronary heart disease or heart failure). Exclusion criteria were current suicidal ideation; a score of 2 or greater on the CAGE alcohol assessment,41 recent use of lithium or antipsychotic medication, and cognitive impairment precluding informed consent.
Bilingual study recruiters trained in cultural competence described the study to eligible patients and obtained written informed consent at designated private areas in the recruitment clinics near the waiting room. From April 2014 to May 2015, 1957 patients were screened, 421 met study inclusion criteria, and 348 completed enrollment with study baseline assessment (see CONSORT Figure 1). The study recruiters recorded depression, chronic illness status, study enrollment, and baseline interview data via an iPad linked to the study's secure online data center. The study group assignment was randomized following completion of the baseline interview. We used the common computer-assisted randomization to generate equal probabilities of the study groups. For patients to feel included in the process of study group assignment, we asked patients to pick a number from 1 to 4 that were internally coded as 2 UC and 2 AHH in a random order. The coded group information was concealed from the interviewer and participant. A new randomization of the groups was generated at each study group inquiry. After the patient chose a number, the corresponding group information was revealed and displayed on the iPad screen, and the patient was informed of his or her group assignment.
Study CONSORT.
In view of known barriers to participation in clinical trials among low-income minority populations, we consistently made the following efforts to facilitate recruitment and acceptance of the intervention and to minimize study attrition: (1) Spanish-speaking promotora staff and study and intervention materials in Spanish were adapted to varied literacy levels and language idioms; (2) interventions were offered through in-person or telephone visits and allowed for evening or weekend appointments; (3) outcome data were collected by telephone; and (4) a $10 gift card incentive was given to each participant for each outcome interview completion.
Although recent studies have assessed and proposed diverse forms of self-management in patients with chronic illness, few studies have facilitated depression care management among patients with concurrent chronic illnesses. Based on the biopsychosocial care model, multiple chronic illnesses management requires multiple and complex components, such as patient–provider and provider–provider interactive management to facilitate patient self-management. To improve patient-centered care, safety-net medical and behavioral health providers must operate as multidisciplinary provider teams. Previous studies found that racial/ethnic minorities with lower socioeconomic status were less likely to provide essential information about progress of diseases and unusual symptom change during clinics visits.5,6,25 Failure to inform physicians might contribute to poorer patient health outcomes such as diabetes complications and increased mortality among heart patients. Therefore, we developed a culturally competent practice model utilizing bilingual promotoras in safety-net clinics.
AHH intervention was contextually guided by the Chronic Care Model,42 in which patient-centered needs, health care–provider delivery, and patient outcomes are seen as a product of the interaction between the patient, provider, and 7 health care process components: (1) delivery system design, (2) patient self-management support, (3) patient care management preferences, (4) community navigation resource linkages (5) provider team and decision support, (6) shared clinical information system, and (7) health system organization. DHS patients have relatively high rates of depression (about 20%), concurrent chronic illnesses (up to 80%), limited health literacy, and high rates of contextual sources of depression and overall medical stressors. However, few clinical trials among diverse populations have simultaneously examined racial/ethnic patient self-management care,30,32 health literacy limitations, patient preferences and uptake plus the patient-provider care management relationship and communication.43-46 In the study we followed a problem-solving framework using self-care management of chronic care conditions as a specific skill-building focus. Thus, the framework closely followed a key skill-building strategy involved in problem-solving treatment for primary care, which was the rational problem-solving component.47-49 The AHH intervention provided problem-solving modeling and opportunities for patients to use to activate self-care and communication with their medical provider about treatment access, health care concerns, and self-care management. AHH aimed to assess the effects of promotora-mediated patient assistance in reducing patient depression, activating overall health management over time, reducing depression and comorbid illness treatment barriers, and optimizing patient care uptake and satisfaction.42
The intervention protocol called for 6 sessions (on average about 45 minutes each) to be provided during the active intervention phase, with 3 brief booster sessions (on average about 15 minutes via telephone; see Figure 2). The first session included a rapport-building introduction followed by psychoeducational elements. Psychoeducation included the provision and review of written materials on depression, diabetes, and/or cardiovascular disease tailored to the participant's individual presenting symptoms and provider recommendations. The first session also included an introduction of problem solving, the rationale for the intervention, expectations, and an interactive session to develop an action plan. All subsequent sessions included evaluation of the action plan and development of a new plan for the subsequent week. All sessions included a participant-identified pleasant activity to activate along with their action plans. Community resource referrals were also provided when needed.
AHH—Participant Flow.
We recruited a total of 3 female bilingual AHH promotoras in close partnership with the host promotora organization, Visión y Compromiso (VyC), a leading nonprofit organization dedicated to promotora-assisted quality health care services.37, 50-58 The study promotoras received an initial orientation from VyC. The VyC-sponsored training was meant to introduce the role of the promotora from an empowerment perspective in order to value this role and its contributions to the health of the community. In the initial training, promotoras engaged in knowledge building, skill development, and consciousness raising based on popular education for marginalized groups.59 Hiring eligibility criteria included the following: (1) at least a high school education; (2) 1 year of prior promotora work experience in a health promotion program; and (3) fluency in verbal and written English and Spanish.
Each promotora received group-administered (1) orientation and intensive training, (2) booster training sessions, and (3) supervision throughout the active intervention study phase from the second author (MPA), a licensed clinical social worker and national trainer in problem-solving treatment for the bilingual/bicultural behavioral health workforce sector. Promotoras were trained to engage patients at home or in the clinics and to enhance their health literacy, address a wide range of community health issues, and provide advocacy and leadership development.37, 50-55 Held over several days, the promotora training included intensive training on patient engagement, the rationale for problem-solving therapy (PST) and its components, study protocol and procedures, and interventionist self-care. Learning formats encompassed the following: didactic lectures; role playing; videos; reading material; self-study; practice with patients; and guest lectures. The same trainer provided booster trainings on an ongoing basis to support any content or skill development needs.
Promotora supervision sessions were provided weekly or biweekly to offer booster training, sessions on promotora self-care, and a review of participant caseloads and progress. Ongoing in-person supervision sessions were dedicated to case presentations whereby the promotoras presented information about each of their assigned patients, the process and content of their interactions, integration of sociocultural strategies, and recommendations for each patient's upcoming session. Supervision sessions also involved electronic research records which the promotoras documented verbatim for each participant's problem list, action plan and pleasant activity. Any corrections during the supervision sessions were reviewed in the following session to evaluate changes in process or approach and need for additional training or enhancements.
Guided by our earlier studies, we based promotora training on (1) introductory modules that can be used across all patients (“getting to know one another”; introduction to depression and chronic disease); (2) wellness modules based on overlapping symptom management (that cut across diagnoses) such as dysphoric mood, pain, fatigue; (3) problem-solving and action-planning modules that serve as core activation strategies for depression and multiple chronic illness management and medical and community provider resource navigation issues; (4) a personalized menu of modules based on concurrent chronic illness-specific symptoms or consequences; and (5) training on research study protocols, human subjects protections, and documentation. Written health education materials were distributed based on patient health conditions and guideline-concordant national standards for illness self-management.
Promotoras scheduled to meet with patients over 6 visits to engage in 6 tasks: engagement (ie, initial rapport building); problem formulation (ie, targeted problem list); education (ie, self-care management strategies and health information); action planning (ie, developing action steps and implementation, community resource navigation, referrals to providers); and evaluation (ie, feedback). Sessions were administered primarily face-to-face in the patient's home or other preferred setting, which was aimed at fostering an engaged and supportive relationship with patients, followed by 3 monthly booster sessions. At each intervention session, promotoras documented patient progress in building problem-solving skills, and, where needed, any medical care concerns and referrals to care provider and community resources. Either during these sessions or by follow-up via telephone or email, promotoras could inform the AHH promotora supervisor or the PCMH team about a patient's worsening symptoms, medication side effects, and provider access concerns.
All participants continued to have access to depression care services from their health care providers. Because this was a real-world trial, the AHH intervention was not meant to modify patients' usual and typical receipt of care at the county clinics, which might include psycho-pharmacology and specialized mental health services.
Participants in the usual care (UC) arm received DHS PCMH clinic team usual care from their respective county health clinic providers. The DHS PCMH usual care multiple-providers care team included its PCMH physician, nurse, and medical assistant. The PCMH model had available DHS medical providers and social workers for depression care, who referred patients, when indicated, to community mental health clinics. Problem-solving therapy was available in some of the participating clinics. During the trial period, LAC-DHS activated health care improvements, including adding community health workers (CHWs) into UC clinics. The CHW had a similar role as that of the AHH promotora.
Study recruiters collected study participants' demographic data (ie, age, ethnicity/race, nativity, years in United States, marital status) at baseline. In addition, we collected patient mental health and chronic illness history, socioeconomic stress, and health literacy assessments at baseline before study group randomization. An independent study interviewer who was blind to patient study groups conducted the 6- and 12-month follow-up interviews. Promotoras logged their contacts with patients on service delivery and intervention session adherence, as well as patient self-care concerns and referrals to community resources, if any.
We assessed the primary depression outcomes using the 9-item Patient Health Questionnaire (PHQ-9) and the 20-item Symptom Checklist (SCL-20) depression scores. The PHQ-9, which establishes provisional depressive disorder diagnosis as well as grades depressive symptom severity, scores each of the 9 DSM-IV criteria as 0 (not at all) to 3 (nearly every day), with possible scores ranging from 0 to 27. PHQ-9 scores of 5, 10, 15, and 20 represent the thresholds for mild, moderate, moderately severe, and severe depression. We used a validated Spanish version of the PHQ-9.60-62 We used the SCL-20 depression scale because it has been shown to be sensitive to change over time in primary care trials of depression (Cronbach α = .91).63,64 The Self-efficacy for Managing Chronic Disease Scale is a 6-item self-reported measure that assesses symptom control, role function, emotional functioning, and communicating with physicians.65 The short form Patient Activation Measure is a 13-item instrument that assesses patient self-reported knowledge, skills, and confidence for self-management of one's health or chronic condition.66,67
Secondary outcomes included the following: To capture the extent of health service utilization, patients were asked to respond to questions about their outpatient (medical and mental health), inpatient, and ER service utilization, including use outside of DHS as well as depression care and treatment receipt. In addition to self-reported health care utilization, we determined service utilization using data from DHS web registry outpatient, pharmacy, ER, and hospitalization visits. We assessed patient satisfaction with care using a single-item ordinal scale of 1 to 5 (indicating poor to excellent). We assessed patient use, acceptance, and satisfaction with promotora assistance in the AHH group. To measure overall functioning and quality of life, we calculated the Medical Outcomes Support Short-Form Health Survey (SF-12)68 Physical and Mental Component Summary norm-based scores standardized to the general US population with a mean of 50 and an SD of 10. The Sheehan Disability Scale assessed illness interference that was derived as an average of 10-point Likert scales (10 indicating inability to carry out any activity) in 3 domains: work, family life and home responsibilities, and social life.69 We also assessed anxiety using Generalized Anxiety Disorder 7 (GAD-7)70; somatic symptoms using PHQ-1571; depression remission using the Remission Evaluation and Mood Inventory Tool72; social and economic stresses using selected items from the Hispanic Stress Inventory73modules (eg, financial, employment, family, community violence worry); stigma using the Latino Scale for Antidepressant Stigma74,75; chronic pain defined as having pain most of the time for 6 months or longer in the past year; and pain impact items from the Brief Pain Inventory.76
Power calculation in the planning of the study to estimate sample sizes needed to adequately evaluate the trial effects of AHH vs UC as a comparison condition was based on our previous studies; in which small to medium effect sizes were detected for depression outcomes in SCL-20 and PHQ-9 scores. Because the AHH system was a pioneering work and its main effect would be investigated in this project, we assumed a small to moderate size of the treatment effects in calculating statistical power. Given the health conditions of the study population, we estimated an attrition rate of 10% at each follow-up phase or 20% over the 12-month study duration, based on our previous studies on depression and diabetes or cancer DHS populations. Power to evaluate program effects was determined with G*Power (V3.1) software.77 We conducted all power calculations at 2-tailed alpha = .05. We aimed for power = .80 or higher, and were guided by Cohen's conventional standards for effect size (f index measuring the ratio of population SDs in analysis of variance) with f = 0.1 for small effect size and f = 0.25 for medium effect size. With the proposed sample size of 350 at baseline and 20% overall attrition rate, the final analytical sample of 280 would have sufficient statistical power to detect a small effect size around f = 0.14 of the AHH intervention.
We used simple descriptive statistics to summarize participants' demographics and baseline characteristics, with count and percentage presented for categorical variables and mean and SD for continuous variables. To examine comparability of study groups and investigate potential bias to group randomization, we conducted bivariate analyses to compare demographics and clinical characteristics between intervention and usual care patients at baseline, using a chi-square test for categorical data and t test for continuous data. To investigate risk factors to study attrition, we compared baseline values between patients who completed the 12-month end-of-trial assessment vs those who did not complete it.
To assess the intervention effect over the 12-month study duration, we conducted an intention-to-treat analysis of repeated measures (baseline, 6-month and 12-month follow-ups) for each dependent variable. We fitted mixed-effects regression models implemented in SAS Proc Mixed procedures for continuous variables and mixed-effects logistic regression models implemented in SAS Proc GLIMMIX procedures for dichotomous variables. We converted the logit model-predicted outcomes through an inverse link into predicted probabilities for ease of interpretation. The mixed-effects model allows time-dependent, within-cell covariates and operates on incomplete (ie, missing data) matrices. We accounted for missing values in the mixed model, which uses a likelihood-based estimation procedure. This resulted in in nonbiased estimates by imputation of missing responses based on the surrounding responses and the modeled covariance structure. We specified the unstructured covariance in the mixed-effects models to account for within-patient correlations of repeated observations over time, and we examined fixed effects of time, study group, and their interactions. In addition, because of the high attrition in the study population, we conducted a sensitivity analysis on outcomes using the last-observation-carried-forward method. Because LAC-DHS activated health care improvements, including adding community health workers (similar role as that of AHH promotoras) into UC clinics during the trial period, we also examined outcomes over time for each study group individually. We conducted all statistical analyses at .05 significance level (2-tailed) using SAS software version 9.3 (Cary, NC: SAS Institute).
Promotoras provided timely documentation of each encounter with study participants for research purposes and for participant caseload follow-up. The promotoras electronically entered intervention problem-solving session notes. We coded and analyzed each data entry using the summated content analysis method,78 which allows for identifying, quantifying, and summating words or context from text to understand individuals' responses to particular questions or prompts based on a priori areas in the existing literature. A trained bilingual qualitative interviewer reviewed all action plan entries and manually coded the responses, which were then coded into larger categories or classifications. Based on content analysis, we summarized the data extracted from the open-ended fields that reflected verbatim responses of the following fields for each participant: (1) the problem (or situation) selected by the participant to be the focus of the session and, as a result, the focus of the practice homework; (2) the action plan to practice for the upcoming week; and (3) the pleasant activity to activate for the upcoming week.
Of 1973 target patients identified from April 2014 to May 2015, 1957 (99%) were screened (96% diabetes, 3% heart disease, and 1% both illnesses) and 22% met criteria for clinically significant depression (female 25% vs male 15%; chi-square = 22.96; P < .0001). Excluding 10 women and 8 men who met exclusion criteria, a total of 354 patients agreed to participate and provided study consent (enrollment rate = 84%; female 85% vs male 78%; P = .11). Of them, 6 patients did not complete the baseline assessment. The recruitment yielded an analytical study sample of N = 348, with 178 (51%) randomized to the AHH intervention group and 170 (49%) to the PCMH usual care group (Figure 1).
Tables 1 to 3 present sample demographic and baseline characteristics in each study group. At baseline, demographics and baseline values did not vary significantly between study groups. Study participants were predominantly female (296; 85%), Latino (344; 99%), and born outside of the United States (318; 91%), and were an average age of 56.7 years old (SD = 8.3). Most of them were from Mexico (238; 68%), El Salvador (32; 9%), and Guatemala (25; 7%); had been living in the United States for 10 years or more (93% of immigrants); preferred to speak Spanish (315; 91%); and had not completed high school (261; 75%). More than 60% of patients (225; 65%) were overweight with body mass index (BMI) 30 or greater, and more than 40% had poor blood sugar levels (A1C more than 9%; see Table 1).
Sample Demographic and Baseline Physical Characteristics.
Regarding medical conditions, 297 (85%) had diabetes, 13 (4%) had heart disease, and 38 (11%) reported both chronic illnesses. More than two-thirds of patients had other comorbid medical conditions including arthritis, retinopathy, gastrointestinal problems, kidney disease, lung disease, stroke, cancer, and urinary tract or prostate problems. Nearly half (162; 47%) of the study participants had chronic pain, and 87% patients rated their health as fair (181; 52%) or poor (121; 35%). The average Sheehan Disability Scale was 5.4 (SD = 3.2), and the physical summary score was 38.3 (SD = 10.8) as assessed by the SF-12 health survey with score profiles standardized to the general US population with a mean of 50 (SD = 10).
At baseline, depression severity assessed by the PHQ-9 was moderate (score 10–14) for 38% of the sample; 147 patients (42%) scored moderately severe (score 15–19), and 68 patients (20%) scored severe depression (score 20 and higher). Self-report depression history indicated that 147 patients (42%) had a history of major depressive disorder, and only 26% had been prescribed antidepressant medication. More than half of the study population had a cut-point of 10 or higher (moderate to severe) on GAD-7 anxiety (185; 53%) and PHQ-15 somatic symptoms (216; 62%). This population's average mental component summary of SF-12 was 30.9 (SD = 9.2) vs mean value 50 for the US general population. Participants' psychometric measures did not vary significantly between study groups at baseline (Table 2).
Psychological Health at Baseline.
Table 3 presents patients' self-care management and social relationships at baseline assessment. Again, we found no significant group difference. For the Self-Efficacy for Managing Chronic Disease scale, on a scale of 1 to 10, group average was 6.2 in AHH vs 5.8 in PCMH UC (P = .17). The group mean score was 3.3 in AHH vs 2.9 in PCMH UC (P = .23) for the Latino Scale for Antidepressant Stigma, with a possible scale range of 0 to 14. The other baseline measures were similar between the AHH and PCMH UC study groups. We transformed the social support measure scored by the 8-item modified Medical Outcomes Study Social Support Survey to a 0 to 100 scale, with higher scores indicating more support. The study group had a mean score of 49.9 (SD = 33.7) in instrumental support, 54.7 (SD = 31.5) in emotional support, and 52.3 (SD = 28.9) in overall social support. This study population had an average of 3.6 (SD = 2.0) stressors assessed from the 10-item Hispanic Stress Inventory. Most patients (322; 93%) self-reported having financial problems, having difficulty in paying bills, or having no money left over at the end of the month. Other summary stress domains were work/employment (69%), marital/family conflicts (64%), and cultural conflicts and immigration issues (26%). In summary, we found no significant group difference with respect to any assessed variables at baseline.
Care Management, Social Support, and Stress Assessment at Baseline.
Over the course of intervention—that is, 6 weekly sessions followed by 3 boosters—113 (63%) patients started an initial session with a study promotora and 98 (55%) continued participating in the AHH intervention; 89 (50%) received 4 or more sessions (Table 4). Of the 65 patients who did not begin the intervention, 32 declined (9 refused promotora sessions, 8 were out of the state or country, 7 had no time, 2 had work conflicts, 2 were busy as a caregiver to a family member, 2 were seeing another therapist, 1 preferred follow-up with the doctor, 1 claimed not depressed); 21 could not be contacted by phone (12 not locatable from no answers and no return calls, 4 with wrong number, 3 phone out of service, 2 hung up the phone); and another 12 passively declined (6 with broken session appointments; 5 failed to set up appointment; 1 refused by daughter).
Receipt of AHH Intervention (N = 178).
A total of 755 problem-solving sessions were provided throughout the active intervention phase of the project. Our participants were actively engaged in their problem-solving sessions, as evidenced by the extensive documentation of their problem-solving components: problem formulation, action plans, and pleasant activities. Overall, we analyzed 1348 entries for problem formulation, 949 entries for action plans, and 753 entries for pleasant activities using content analysis. The sessions were person centered, wherein each session was tailored to the participant's unique needs, preferences, and learning style. Thus, the presenting problems covered a vast array of situations and problems that affected the participants' psychological and physical health. These centered on medical and psychiatric functioning; interpersonal conflicts, primarily with family; day-to-day problems such as financial woes; and the navigation required to apply for—and receive—benefits and community resources, to name the most frequently endorsed categories. Turning to the action plans, participants focused on applying for benefits and services, followed by communicating their needs to their provider or their health care team. In order to address their medical illnesses, they turned to following medical instructions or advice, taking care of themselves through changing their nutrition or diet, and increasing exercise or physical activity. They also addressed their low economic status by seeking employment or making extra money on the side. Communicating their needs to family members was also a priority. Participants understood the difference between action plans and pleasant activities; thus they were readily able to provide a behavioral action activity that fit their individual preferences—ie, engaging in an activity that used to bring them joy and contentment. They typically engaged in activities that did not entail a cost (or had a minimal cost), and preferred such activities as going to church, engaging in exercise/physical activity, socializing with family, gardening, dancing and listening to music, arts and crafts, and going out. In sum, the intervention was feasible, acceptable, and easily understood by participants and promotoras. Engagement with the providers and health care team continues to be a challenge for this population; many reported getting appointments and information about their medical care from health care providers as chronic challenges.
During the intervention period, the most-requested assistances and referrals were community resources (frequently inquired: community, senior, or wellness center, 88 occurrences; transportation, 33; food bank, 25; see Table 5). The project protocol included a component for promotora-clinic provider communication, but due to the fact that the clinic providers' workloads doubled with their transition to the patient-centered medical home model and new electronic health record implementation, minimal linkage was established between the providers and the project's promotoras. Promotoras made 12 referrals to LAC-DHS providers for patients' unmet needs, such as no longer receiving prescription for antidepressants, needing to see a psychiatrist, and a long wait for a surgery or an exam.
Referrals Made by Promotoras.
The study sample consisted of 348 patients with depression and concurrent diabetes and/or heart disease from 3 DHS safety-net community clinics. Consistent efforts were made to minimize attrition given the high representation of immigrant, Spanish-speaking, safety-net population in the sample. The proportions lost to follow-up were not significantly different between study groups (Figure 1). Study attrition of those who did not complete the follow-up interview at 12 months was 30% (AHH 31% vs UC 28%; P = .51). A total of 104 patients were lost to follow-up, including 80 (23%) whom we were unable to locate due to phone issues (phone not in service, number disconnected, wrong number); 21 (6%) patients declining further study participation; 1 in a nursing home and unable to take calls; 1 due to a medical reason (in and out of hospital, now on dialysis); and 1 death. We compared demographic and baseline clinical characteristics, including depression and quality-of-life measures, between patients who completed the 12-month interview and those who did not complete in the AHH and UC combined sample. We did not find any baseline variable associated with study attrition (Table 6).
Patient Baseline Characteristics and Study Attrition.
We evaluated the over-time group differences in depression and disease self-care management scores in linear mixed-effects regression models (Table 7). There was no significant over-time group difference in these primary outcomes as reflected by the nonsignificant P value for time by group interaction, as well as no significant cross-sectional group difference at each follow-up wave. However, AHH patients had a slightly better group mean in lower depression at 12 months and better self-care management at each follow-up wave compared with UC patients. These improvements were not statistically or clinically significant.
Primary Outcomes on Depression Score and Disease Self-care Management Analyzed in Linear Mixed-Effects Models.
We obtained medical records from LAC-DHS electronic medical services records for 168 UC and 174 AHH patients who authorized permission to release their health information. Health care facility utilization in clinic visits, ER, or hospital admission was similar between study groups (Table 8).
Secondary Outcomes on Health Care Utilization Analyzed in Mixed-Effects Models.
Although the LAC-DHS electronic health record (EHR) Orchid system had been implemented for more than 1 year, medical records had not been completely integrated into the new system. Some data were unavailable when we made requests, including outpatient visit clinic name and activity, mental health referral, and class with social worker. Thus, we reported mental health care receipt based on patients' self-reports. Trial data fitted in the mixed-effects logistic model showed that patient depression treatment acceptance increased over the 12-month trial (38%, 63%, and 68% in AHH and 46%, 65%, and 70% in UC at baseline, 6 months, and 12 months, respectively), and that more patients had been prescribed antidepressants since baseline (21%, 36%, and 33% in AHH and 26%, 33%, and 31% in UC at baseline, 6 months, and 12 months, respectively). The leading barriers to depression care reported in this population were difficulty in finding a depression care provider who spoke their language (23%), difficulty in finding a facility in the community for depression care (18%), and difficulty with clinic hours (7%). Further, fear of addiction to antidepressant medicine was a barrier reported among UC patients who did not seek help for depression care; that concern was not a barrier to the AHH patients.
More than half of the study patients had PHQ-9 depression scores of less than 10 (a cutoff score for minor depression) at follow-up assessment (AHH 60% at 6 months, 65% at 12 months; UC 56% at both 6 and 12 months), and about half of participants in both study arms reached clinical significant improvement of depressive symptoms defined as a 50% or more score reduction since baseline (AHH 53% at 6 months, 55% at 12 months; UC 51% at 6 months, 49% at 12 months).
Table 9 presents quality-of-life outcomes on physical well-being, stress, and social support. We did not find statistically significant group differences at any time point nor in time by group interactions.
Secondary Outcomes on Physical Well-Being, Stress, and Social Support Analyzed in Linear Mixed-Effects Models.
Because LAC-DHS activated health care improvements, including adding community health workers (similar role as that of AHH promotoras) into UC clinics during the trial period, we also examined outcomes over time for each study group individually. We found significant improvements at 6-month and 12-month follow-ups in almost all assessed outcomes (ie, depression, physical health, self-care management, psychometric measures, socioeconomic stress, and social support) in each individual group (Table 10 for UC and Table 11 for AHH), and both study groups performed evenly well. Because of the high attrition in the study population, we also conducted a sensitive analysis on outcomes using the last-observation-carried-forward (LOCF) method. We found similar results with LOCF.
Over-Time Changes Analyzed in Linear Mixed-Effects Models for Patients in the PCMH Usual Care Group (N = 170).
Over-Time Changes Analyzed in Linear Mixed-Effects Models for Patients in the AHH Intervention Group (N = 178).
Comparing lab test results before and after trial, both study groups had a statistically significant improvement in blood sugar level A1C (AHH reduced 0.51% in average absolute change, P = .0001; UC reduced 0.37%, P = .01). In both AHH and UC, systolic and diastolic blood pressures and all lipid profiles except HDL cholesterol level declined (Table 12).
Before and After Trial Lab Test Results in Each Individual Study Group.
Two patients were hospitalized (1 UC, 1 AHH), 1 AHH patient was admitted into a nursing home, and 1 UC patient death was reported by a family member when the outcome interviewer called for outcome interviews; otherwise, no adverse events existed.
We conducted a total of 25 in-depth interviews by telephone from a random sample of intervention and usual care patients. A trained bilingual/bicultural (English–Spanish) interviewer followed a prepared interview guide that included questions across the following domains: general knowledge of depression and depression treatment; stigma and disclosure related to depression; overall impressions and experience with the promotora-mediated counseling sessions; relational issues with health system providers; and language and cultural factors related to satisfaction of care. A review of these qualitative interviews indicates that our patient participants experienced changes in their health care due to changes in the system (eg, pharmacy/medication pick-up procedures, delays in receiving clinical behavioral health services, changes in designated primary care providers, unsatisfactory treatment experiences). The review of these qualitative interviews did not indicate impact on study operations and study quality (eg, recruitment and retention, study outcomes and findings).
Promotoras engaged very well with patients, but few clinic physicians had time to communicate with them in light of time pressures in the clinics; also, the physicians had begun taking on responsibility in adopting the PCMH model with other stakeholders (eg, nurses, CHWs).
We conducted postintervention stakeholder interviews with the study clinic providers—including approximately 25 physicians, 20 nurses and care managers, and a clinic medical director—to obtain feedback on our study. Specifically, we presented our study results for comments and discussed what changes were implemented in the clinics during the study period (eg, competing interventions) for chronic disease self-management or services for depression. We also discussed challenges the providers foresee in addressing patients' health care needs (eg, multiple chronic conditions, income and other resource needs, population health management, EHR, DHS mandates, ACA mandates, integrated care mandates, lack of community or family support for patients).
When presented with our finding that both the intervention group and the control group improved significantly over time, the clinic providers responded that they had implemented “a lot” of quality improvement initiatives during the study period (ie, 2014-2016). Based on the qualitative data, many competing interventions to the AHH promotoras intervention may have led to the positive outcomes for both groups. First and foremost, the study clinics have been implementing the patient-centered medical homes (PCMHs) using team-based care for about 2 to 3 years, during roughly the same time of the AHH study. Their Cerner (called Orchid in LAC-DHS) electronic health record system was implemented about 1.5 years ago. These clinics also have linked each patient is to a primary care provider (PCP), although the actual implementation found somewhat inaccurate or missing information, such as the name of patient's PCP not displayed at the Cerner system.
The clinic has a large number of diabetes patients (with comorbidities such as coronary heart disease and hypertension), and an estimated 70% of visits are for patients with diabetes. Similarly, the vast majority of AHH study participants have diabetes. In the past few years, the clinics have been highly focused on diabetes care and supporting patient needs. The information system will allow the clinics to patient health and identify high-risk patients. The clinics send their patients to the diabetes clinic for intensive treatment, and are conducting many other initiatives. For example, they send high-risk diabetes patients to a clinical pharmacist to review medications and educate them about compliance. They made 3 referrals for the high-risk patients—to (1) a nutritionist, (2) a health educator, and (3) a family resource center. The providers follow up with patients to determine if they made it to the referrals and encourage them to do so. They also screen patients for depression and actively prescribe antidepressants. The clinics have been testing a community health worker program since 2014-2015 that is very similar to the AHH promotoras' roles. Patients are randomly assigned to the program to evaluate its effectiveness.
Additionally, there are clinic-wide care management activities for diabetes patients. For patients with an A1C value of 10 or higher, care managers (with RN qualification) manage these patients and follow up with them in 2-week intervals through a face-to-face nurse visit or via telephone calls. If the patient's A1C is greater than 8 but less than 10, caregivers in the clinic follow up with patients so the workload is more manageable for the care managers. Care managers are the leads for these caregivers. In the PCMH model, there is also SMART goal setting for patients to set up a plan for exercise, weight loss, etc. Patient workshops in the clinic teach them how to cook diabetic meals and how to perform certain physical activities, using relatable approaches so they can learn to make these behavioral changes. In these ways, the care management team is helpful with patient diabetes self-management, and has facilitated better communication with PCPs.
Because of the implementation of the Cerner electronic medical records system, the study clinic's providers can report a patient's A1C level to care managers in near real time. The care managers follow up with patients to ensure they are educated about diabetes, receive their medication, and follow recommended diet and exercise programs. The care managers are “very efficient,” according to one physician.
In summary, for diabetes patients (which make up the majority of our sample), the providers believe that they are on a journey of continuously improving health outcomes because the clinic has been highly active in improving patient care.
Since the implementation of PCMH and ACA, which requires providing preventive services such as depression screening, cancer screening, etc. (National Committee for Quality Assurance [NCQA] requirement), in the past 2 to 3 years, the clinics have also had a great focus on depression care. Providers are required to screen patients for depression. They first use the PHQ-2, which has been implemented for 3 to 4 years. As of 2016, they are also required to use PHQ-9 if a patient's PHQ-2 score is high. Once a patient is screened and identified as having depression, providers engage in conversations with patients about their condition and prescribe medication and/or refer patients to mental health services.
The providers prefer the same kind of support for depression as that currently available for diabetes. The difficulty is that the providers screened patients using both PHQ-2 and later PHQ-9 but couldn't provide adequate care for because they lacked the supporting workforce or referral channels. Their training and time can afford them to only screen patients and prescribe medications, not to complete intakes, counsel patients, for mental health, or link patients with social or community resources. Providers need help to manage patients with depression and other mental health problems. Many LAC-DHS patients with depression also have comorbid conditions such as bipolar disorder, anxiety, other severe mental illnesses; drug addiction; incarceration experiences; poor social support; and poor family dynamics. Because the providers do not have the time to “open the can of worms,” they tend to send the patients to public mental health clinics to treat their nonmedical issues. Providers are concerned that if they begin to deal with depression-related problems, each medical encounter could easily take them 2 hours, rather than the expected 15 minutes (with half of that time dedicated to charting in the Orchid EMR system—an increase from about 3 minutes in the old electronic medical record system).
The providers do recall the AHH team coming to the clinic to introduce the study, and they recall the study recruiters and had good experiences working with them. However, providers were unaware of the study promotoras; they did not have contact with them or hear about them from their patients. The providers strongly felt that the promotoras might be the helping hand that they need to take care of patients with depression. They view the promotoras as the personnel they need to engage patients, formulate problems, educate patients how to rely on self-help strategies using down-to-earth approaches, hold educational workshops on depression and diabetes in the clinics for easy access, help patients determine plan actions and pleasant activities, help patients deal with their mental distress, and follow up with patients regularly.
Unfortunately, because the providers' workloads doubled with the PCMH and Orchid EHR implementation, no linkage was established between the providers and our AHH promotoras. The providers also do not think consumer health technology, such as mobile appss, would be helpful to only a small number of their patients because of the population's poor access to technology and low health literacy.
The providers identified the following strategies to facilitate higher-quality depression care:
The trial outcomes did not vary significantly between the promotora-assisted and LAC-DHS usual care groups, and both groups fared better over the active study period on most measures. The null result was likely a function of high refusal and drop-out rates in the AHH intervention arm as well as the enhanced LAC-DHS PCMH model, which was enacted during the study period as a result of the ACA and Medicaid expansion. These clinical improvements included (1) depression screening by the primary care provider; (2) referral to clinic staff including community health workers; (3) behavioral health specialty care referrals, which included receipt of PST in some clinics by social workers; and (4) activated patient care management. In essence, components of the experimental design may have been offered to the usual care group as part of regular practice, thus decreasing the potential significance of the experimental, promotora-assisted group (which trended better on many outcome measures). Furthermore, the high subject drop-out rate may have prompted a differential response to the intervention, depending on who remained in the center and who dropped out.
As a result of these 2 factors and perhaps others, the study was not a strong test of the promotoras model. The decisional context for informing patients, clinicians, and other key stakeholders about the model is not certain at this time. Additional research is required to arrive at a definitive judgment about the value of the AHH promotora model in primary care.
This study is the first to incorporate promotoras in a public health safety-net care system in order to reduce racial and ethnic disparities in depression and self-care management among Spanish-speaking patients with diabetes and/or heart disease. It evaluated promotora-assisted depression and self-care management among patients with diabetes and/or heart disease based on a 12-month randomized clinical trial within LAC-DHS community clinics. Promotoras are increasingly at the front line of health care organizations.
It is important to note that the group comparisons involved depression management in the context of a patient-centered medical home with or without the addition of promotoras who were trained to provide counseling. The PCMH model with or without promotoras has the potential to promote better health, functioning, and quality of life among Latino safety-net health care consumers.
This study took place while the LAC-DHS safety-net care system was experiencing unprecedented changes brought on by the ACA and by LAC-DHS uptake of the PCMH model, including increasing disease management within its primary care clinics and, recently, enhanced health information technology system to better track how patients are doing and to identify high-risk patients. LAC-DHS also began to include depression screening, based on the ACA and implementation of the 2008 Mental Health Parity Act, which requires insurers to cover treatment for depression just as they would cover treatment for a physical illness.
Given that our study did not yield a strong test of the model at this time, we advise additional research to determine the comparative effectiveness of the promotora model in public-sector health care before our study results are implemented.
Nevertheless, there remains a significant need to engage Latinos in public safety-net health care that is consistent with patient preferences and personalized care planning.79 The challenges to implementation are significantly and uniquely relevant among diverse low-income populations as well as among safety-net providers, community organizations, and stakeholders. Patient perceptions of care coordination problems are associated with both poorer self-care activation and health outcomes. This is particularly relevant to the complexity of patient self-care that is inherent to depression and concurrent chronic illness management.80 Safety-net patients with major depression plus a concurrent chronic illness face significant barriers to patient activation, motivation, skills, and confidence—the very attributes that equip patients to become actively engaged in their health and health care.5,6,26,81-84 Aside from health system financing barriers, the study encountered barriers to locating a study population that was financially impoverished, highly mobile, highly stressed due to unemployment, and experiencing poor access to public benefits and community resources.
The sample itself is reflective of highly impoverished patients with historically low rates of health care access. Whether the results are generalizable to other non-Latino populations—or even to the retained study population—is unknown.
The study results cannot be used at this point to differentiate among specific subgroups or subpopulations in the sample.
Study limitations include the following:
High (30%) attrition rates in both arms of the study led to uncertainty about how to interpret the results, since those participants who remained in the study may have had different study outcomes than those who dropped out of the study. Future studies should include strategies to address patient attrition at all observation points.
Future research should consider a more robust test of the promotora model, taking into account dynamic changes in the health care system of interest and addressing potential drop-out or attrition rates of the sample with innovative retention strategies. The drop-out rate in particular could be a signal for a misleading study result. To avoid these problems, future research should employ study procedures such as more frequent communication with the PCMH team through the EMR portal and more intensive education at the informed-consent phase about study participation expectations.
Promotora-assisted depression and chronic disease self-management care was not superior to usual care for any of the primary or secondary outcomes. We found that both the AHH promotora intervention and the LAC-DHS patient-centered medical home model were associated with significant improvements over time. The result of nonsignificant differences between the 2 intervention groups is likely to be a function of the LAC-DHS PCMH model, which emerged with several patient-centered care improvements, as well as the high dropout rate. Because our study is not a definitive test of the promotora model, our results should not be invoked as a reason to abandon the model altogether. Our study calls for additional studies of the promotoras model in depression care; these studies should integrate innovative strategies for subject engagement into the study design in order to avoid these potential confounders.
This study is supported by the Patient-Centered Outcomes Research Institute (PCORI) and activated by Principal Investigator, K. Ell. Trial Registration: NCT02147522, clinicaltrials.gov/ct. The corresponding author at the Suzanne Dworak-Peck School of Social Work, University of Southern California is Kathleen Ell (ude.csu@lle). Coinvestigators, Drs María Aranda and Shinyi Wu, had significant study roles. Dr Hyunsung Oh assisted while working on his doctoral dissertation, and Pey-Jiuan Lee provided data management and analysis. Dr Jeffrey Guterman, the chief of research and innovation officer in the Ambulatory Care Network of the LAC-DHS, provided consultation.
Research reported in this report was [partially] funded through a Patient-Centered Outcomes Research Institute® (PCORI®) Award (AD-1304-7364). Further information available at: https://www.pcori.org/research-results/2013/working-bilingual-community-health-worker-promotoras-improve-depression-and-self-care-among-latino-patients-long-term-health-problems
Disclosure: No conflict of interest, financial or otherwise, exists.
Ell K, Aranda MP, Wu S, et al. (2018). Working with Bilingual Community Health Worker Promotoras to Improve Depression and Self-Care among Latino Patients with Long-Term Health Problems. Patient-Centered Outcomes Research Institute (PCORI). https://doi.org/10.25302/3.2018.AD.13047364
The [views, statements, opinions] presented in this report are solely the responsibility of the author(s) and do not necessarily represent the views of the Patient-Centered Outcomes Research Institute® (PCORI®), its Board of Governors or Methodology Committee.
This book is distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License which permits noncommercial use and distribution provided the original author(s) and source are credited. (See https://creativecommons.org/licenses/by-nc-nd/4.0/
Your browsing activity is empty.
Activity recording is turned off.
See more...
