Clinician Descriptions of Communication Strategies to Improve Treatment Engagement by Racial/Ethnic Minorities in Mental Health Services: A Systematic Review

Abstract

Objective

To describe studies on clinician communication and the engagement of racial/ethnic minority patients in mental health treatment.

Methods

Authors conducted electronic searches of published and grey literature databases from inception to November 2014, forward citation analyses, and backward bibliographic sampling of included articles. Included studies reported original data on clinician communication strategies to improve minority treatment engagement, defined as initiating, participating, and continuing services.

Results

Twenty-three studies met inclusion criteria. Low treatment initiation and high treatment discontinuation were related to patient views that the mental health system did not address their understandings of illness, care or stigma. Treatment participation was based more on clinician language use, communication style, and discussions of patient-clinician differences.

Conclusion

Clinicians may improve treatment initiation and continuation by incorporating patient views of illness into treatment and targeting stigma. Clinicians may improve treatment participation by using simple language, tailoring communication to patient preferences, discussing differences, and demonstrating positive affect.

Practice implications

Lack of knowledge about the mental health system and somatic symptoms may delay treatment initiation. Discussions of clinician backgrounds, power, and communication style may improve treatment participation. Treatment continuation may improve if clinicians tailor communication and treatment plans congruent with patient expectations.

1. Introduction

The low treatment engagement of racial and ethnic minorities with mental illness is a major public health problem in multicultural societies. Engagement has been defined as how patients initiate, maintain, and participate in treatments such as pharmacotherapy and psychotherapy [1]. Disparities persist along this entire pathway as racial/ethnic minorities are 20–50% less likely to initiate mental health care [2–4], 40–60% less likely to fill prescriptions [5, 6] and 40–80% more likely to end treatment prematurely for mood, anxiety, and psychotic disorders [7–9].

The problems with engaging minorities into mental health treatment are multifactorial and occur at individual, organizational, and systemic levels. They include a lack of health information to make treatment decisions [1] (individual), unequal access to evidence-based practices [10] or language-matched services [11, 12] (organizational), and communities without the resources to fund health services (systemic) [13]. However, poor engagement continues even in cases of free or subsidized language-matched treatment after patients have initiated care [14–16], emphasizing the role of the individual patient-clinician interaction. Studies of patient-clinician communication demonstrate that the interaction can be improved since clinicians often involve racial/ethnic minority patients less in treatment participation activities such as clinical decision-making [17], rapport building [18], and friendly conversation [19]. Clinicians are also frequently unaware of patient cultural interpretations around preferred illness labels [20], perceived illness causes [21], and treatment preferences [22, 23]. Minorities who perceive that clinicians do not understand their cultural views have sevenfold higher odds of ending treatment [24], but participate in treatment with greater retention (OR=2.78, 95%CI= 1.33–5.79) and follow-up (OR=1.29, 95%CI=1.16–1.43) with clinicians who understand their views [25].

Patient-clinician communication may therefore act as a key mechanism of action for treatment engagement. Medical encounters can be divided into those that are intra-cultural when clinicians and patients share the same culture and ethnicity and those that are intercultural when their cultures and ethnicities differ [26]. Although definitions of culture have evolved throughout the social and behavioral sciences based on intellectual trends [27], we understand culture according to the latest definition created by the Cultural Issues Subgroup of the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, based on systematic literature reviews:

Culture refers to systems of knowledge, concepts, rules, and practices that are learned and transmitted across generations. Culture includes language, religion and spirituality, family structures, life-cycle stages, ceremonial rituals, and customs, as well as moral and legal systems. Cultures are open, dynamic systems that undergo continuous change over time; in the contemporary world, most individuals and groups are exposed to multiple cultures, which they use to fashion their own identities and make sense of experience [28].

One model of medical communication posits two targets for intervention: (1) the clinician’s elicitation of patient cultural interpretations of illness and treatment preferences, and (2) improved clinician communication behaviors such as proactively building rapport, answering patient questions, showing positive affect, and using open-ended questions [26]. This model can be seen as differentiating between the content of communication through the exchange of ideas about illness and treatment and the context of interpersonal and situational influences that affect the exchange of ideas. Studies suggest that communication interventions for physicians in different medical specialties [29] can help close disparities [30]. This paper advances this scholarship by reporting the results of a systematic review on the types of communication content and context affecting engagement in mental health settings among minorities.

2. Materials and methods

2.1. Search Strategy

This systematic review of communication interventions was undertaken in accordance with a five-step review protocol.

Table 1 lists the resources utilized to initiate the search. As a first step, we searched PubMed, PsycInfo, Embase, CINAHL, and the Cochrane Library which index the most clinical studies among mental health databases [31, 32]. Table 2 lists the combination of terms entered into databases. Terms were modified from systematic reviews on communication and treatment adherence [33] and interventions for racial and ethnic minority groups [34] to make the review evidence-based from the outset. Terms consisted of four categories with synonyms and closely-related concepts for: clinician communication (group 1) with patients having a mental disorder (group 2) from minority populations (group 3) to improve treatment engagement (group 4). All databases were searched from inception to November 2014. As a second step, we conducted hand searches with journals in cultural mental health, the social sciences and mental health, and health communication for relevant articles. Third, we searched prominent grey literature databases indexing health publications.

Table 1.

Search Resources

Databases	Journal Hand Search	Grey Literature
PubMed	Cultural mental health journals	New York Academy of Medicine Grey Literature Report
PsycInfo	Transcultural psychiatry	System for Information on Grey Literature in Europe
Embase	Culture, Medicine, and Psychiatry	Clinical Medicine & Health Research
CINAHL	International Journal of Culture and Mental Health	University of York Centre for Reviews and Dissemination
Cochrane Library	Cultural Psychology	Proquest Digital Dissertations
	Culture & Psychology
	Journal of Cross-Cultural Psychology
	Cultural Diversity & Ethnic Minority Psychology
	Social sciences and mental health journals
	Medical Anthropology
	Medical Anthropology Quarterly
	Anthropology & Medicine
	Social Science & Medicine
	Health communication journals
	Patient Education and Counseling
	Journal of Communication in Healthcare
	Communication & Medicine
	Journal of Intercultural Communication Research
	Journal of Health Communication
	Western Journal of Communication
	Communication Studies
	Health Communication
	Language & Communication

Table 2.

Search terms

All groups are combined by “AND”

Terms within groups are combined by “OR”

Group 1	Group 2	Group 3	Group 4
communicat*	psychiatric disorder	cultur*	treatment initiation
medical communication	psychiatric illness	multicultural	treatment entry
interact*	mental disorder	multi-cultural	adher*
interview*	mental illness	crosscultural	complian*
medical interview	psychological disorder	cross-cultural	concordan*
conversat*	psychological illness	transcultural	nonadher*
discourse		trans-cultural	noncomplian*
dialogue		ethnic*	nonconcordan*
relationship		multiethnic	activat*
alliance		multi-ethnic	attendance
shared decision making		race	retention
		racial	engag*
		multiracial	participation
		multi-racial	continuation
		ethnocultural	discontinuation
		sociocultural	drop out
		divers*	drop-out
		minorit*

Truncation was used to capture all variants of a concept, i.e. psychiatr* for “psychiatry” and “psychiatric”

Articles of interest were screened by reading titles and abstracts according to predetermined inclusion criteria. The inclusion criteria were: all studies of communication with original clinician data reporting how eliciting patient cultural interpretations of illness or treatment or improving communication context with mentally ill racial/ethnic minorities could improve mental health service engagement. Studies were screened in English, French, and Spanish. Studies were only included if they assessed the relationship between clinician communication and treatment engagement. Studies had to also address clinician communication strategies with minorities to improve treatment engagement in mental health settings, not just report patient interpretations of illness or stigma, which have generated substantial scholarship in cultural psychiatry, medical anthropology, and medical sociology [35–36] and may not help clinicians improve patient interactions. The exclusion criteria were: studies that did not present original data with clinicians (e.g. theoretical models or literature reviews); did not enroll clinicians; did not occur in mental health settings; only explored statistical associations of race/ethnicity to outcomes without reporting clinician communication strategies; focused on interpreter use at the organizational level rather than modifying the communication of treating clinicians; or repeated findings from an included article with the same dataset. We excluded single case studies or clinician reflection pieces since these may reflect individual idiosyncratic perspectives. Articles assessed for eligibility were retrieved in full.

After producing an initial set of included studies, we conducted forward citation analysis and backward bibliographic snowball sampling as the fourth and fifth search method steps, respectively, for each round of additionally included papers. Forward citation analysis was conducted by examining citations in Google Search since it covers more conference proceedings and international, non-English language journals than Web of Science or Scopus [37]. Backward bibliographic snowball sampling was conducted until reference lists produced no additional papers meeting inclusion criteria [38].

2.2. Screening and Data Extraction

The first two authors independently assessed each retrieved paper following inclusion/exclusion criteria and compared their results, compiling a single list of records. Authors were contacted when full texts were not available online and two attempts were made to contact each author through email or telephone. Characteristics of each study were abstracted regarding research design, aims, populations, intervention, comparison group, and outcomes in accordance with PRISMA guidelines [39].

The first two authors coded all articles together to minimize bias by copying text verbatim into the spreadsheet with communication and engagement codes. Communication codes were divided into (1) the content of communication such as patient cultural interpretations of illness and treatment and (2) the context of communication such as clinician interpersonal behaviors and the circumstances of communication [26]. Treatment engagement codes were defined as (1) treatment initiation – patients starting treatment in mental health settings, (2) treatment participation – patients collaborating with clinicians in devising a treatment plan, and (3) treatment continuation – patients adhering to treatment plans through appointment retention and full treatment utilization [1]. Both researchers undertook peer-debriefing activities to detect coding differences and met with the senior author to resolve differences by consensus. All researchers are physicians trained in social science methodologies.

The criteria for conducting a meta-analysis were not fulfilled due to variations in study designs, differing measures for communication and engagement, and low use of statistical procedures. Findings are synthesized narratively based on reports of how clinician elicitation of patient cultural interpretations or clinician communication context impact patient treatment initiation, participation, and communication.

2.3. Quality Assessment and Analysis

No “gold standard” design exists for studies of patient-clinician communication and many studies do not have interventions, control groups, or blinding. Therefore, we adopted Thompson and McCabe’s framework [33] that assesses the quality of communication studies along four dimensions:

1. Reporting – Do studies clearly describe aims, outcomes, sample characteristics, findings, and probability values? (Total/5)

2. External validity – Is the study sample representative of the entire population recruited? The sample would be considered representative if it were the entire source population, an unselected sample of consecutive patients, or a random sample. (Total/1)

3. Internal validity – Are the statistical tests appropriate to measure outcomes? Are the study measures for cross-cultural communication and treatment engagement validated and reliable? Was there adequate adjustment for confounding in statistical analyses? (Total/3)

4. Study design – To what extent can the study identify causality? Was the study design cross-sectional, prospective, or experimental design? (Total/2)

3. Results

Twenty-three studies were included in the review: 12 from the United States (US), 6 from Australia, 3 from the United Kingdom (UK), 1 from Spain, and 1 from several European countries. All studies that met inclusion criteria were retrieved electronically or from the authors. Figure 1 presents the PRISMA flow chart on the sources of these studies.

Figure 1.

PRISMA Flow Diagram of Included and Excluded Studies

Table 3 suggests uneven evidence in this field. Of 23 included studies, only 3 had quality ratings greater than 50% [33]. All studies exhibited relatively complete reporting though common limitations can be identified. Most studies relied upon purposive rather than randomized sampling, limiting external validity. Most studies also used qualitative methods to elicit patient cultural interpretations, but did not use validated semi-structured interviews or published interview guides to assess generalizability. Many studies postulated mechanisms between intercultural medical communication and treatment engagement, but the use of cross-sectional interviews with individuals or focus groups limits claims of causality.

Table 3.

Quality assessment ratings for included studies

Paper	Reporting (5 items)	External validity (1 item)	Internal validity (3 items)	Study design (2 items)	Study quality score %
Alegría et al. (2009)	5	1	3	2	100%
Barn (2008)	4	0	0	1	45%
Blignault et al. (2008)	4	0	1	1	55%
Burman et al. (2002)	4	0	0	1	45%
Choi and Gonzalez (2005)	4	0	0	1	45%
Conrad and Pacquiao (2005)	4	0	0	1	45%
Cross and Bloomer (2010)	4	0	0	1	45%
Flaherty and Donato-Hunt (2012)	4	0	0	1	45%
Garcia Campayo et al. 2006	4	1	0	1	55%
Gibbs et al. (2004)	4	0	0	1	45%
Gonçalves et al. (2014)	4	0	0	1	45%
Iltis (2002)	4	0	0	1	45%
Ito and Maramba (2002)	4	0	0	1	45%
Keating and Robertson (2004)	4	0	0	1	45%
Knox et al. (2003)	4	0	0	1	45%
Rowe (2013)	4	0	0	1	45%
Sandhu et al. (2013)	4	0	0	1	45%
Saw et al. (2013)	4	0	0	1	45%
Shea and Long (2013)	4	0	0	1	45%
Smith et al. (2013)	4	0	0	1	45%
Sun et al. (2014)	4	0	0	1	45%
Williams (2010)	4	0	0	1	45%
Yoshida (2013)	4	0	0	1	45%

Table 4 lists study characteristics.

Table 4.

Characteristics of included studies

Author (date)	Research aim	Study design	Study population	Findings related to engagement
				Initiation	Participation	Continuation
Alegría et al. (2009)	To examine communication pattern variations across racial and ethnic patient-clinician dyads during mental health intakes and their effects on treatment continuance	Observational study	18 clinicians and 34 patients in White dyads, 10 clinicians and 24 patients in Latino dyads, 19 clinicians and 35 patients in mixed dyads in the United States; Clinicians were 31% psychiatrists, 16% psychologists, 47% social workers, and 6% nurses	n/a	Compared to other dyads, Latino concordant dyad patients were more verbally dominant, had more patient-centered communication, scored higher on the WAI-O (all p<.05); had more positive affect and made more positive statements (both p<.001)	n/a
Barn (2008)	To study social worker perspectives of how Bangladeshi women experience mental illness and care	Individual interviews	10 social workers in Bangladesh	Initiation was limited by a lack of understanding and awareness of services, concerns about family honor and stigma, and accessing care only through indirect referrals from child services	Cultural models of the clinician relationship led to decreased treatment participation	Treatment is frequently discontinued when medications were prescribed and seen as unnecessary
Blignault et al. (2008)	To examine how sociodemographic factors influence help-seeking and service use of China-born migrants in Australia	Individual interviews through the MDEMS	11 mental health providers,1 caregiver, 13 community membersand 9 patients in Australia; Clinician breakdown not reported	Initiation was limited by concerns about expenses and a general lack of knowledge by patients and community members about the health system, particularly within mental health services	Barriers to treatment participation included immigration and visa problems, as well as being treated rudely for limited English proficiency	n/a
Burman et al. (2002)	To make recommendations improving self-harm services for South Asian women	Focus groups and individual interviews	8 service providers, 4 community groups, 7 South Asian women survivors in England; Clinician breakdown not reported	Clinicians reported that cultural issues may mask structural issues such as problems with domestic violence, immigration, and poverty in all phases of treatment engagement; Cultural issues may be overemphasized and gender issues may be underemphasized
Choi and Gonzalez (2005)	To explore barriers and contributors of continued treatment among Hispanic and Black geriatric mental health patients	Focus groups and individual interviews	15 clinicians in a focus group and 6 clinicians interviewed individually in the United States; Clinician breakdown not reported	Awareness of mental health problems may be lower among Latinos; Suffering may be seen as a religious necessity rather than mental illness	Blacks and Latinos may feel discomfort with psychotherapy; White therapists may need to establish rapport and trust with Blacks and Latinos	Stigma and shame may prevent Blacks and Latinos from continuing treatment
Conrad and Pacquiao (2005)	To explore cultural influences on depression and care outcomes among Asian Indians with depression	Focus groups and chart reviews	23 clinicians in focus groups with experience caring for Asian Indians and chart review of 20 Asian Indians admitted for depression in the United States; Clinicians were physicians, clinicians, and nurses, but exact numbers were not reported	Suffering may be seen as a religious necessity (karma) rather than mental illness; Somatic manifestations of illness may delay treatment	Need for treatment is perceived as crisis-based and episodic, not continuous; Patients may be unaccustomed to discussing emotions	Stigma and shame may prevent treatment adherence
Cross and Bloomer (2010)	To explore how mental health clinicians in Australia modify communication to reconcile cultural differences	Focus groups	53 clinicians in 7 focus groups in Australia; 70% nurses, 9% physicians, 7.5% social workers, 7.5% occupational therapists, 6% psychologists	Treatment initiation may be hindered by many culturally and linguistically diverse patients presenting with somatic symptoms to emergency settings	n/a	n/a
Flaherty and Donato-Hunt (2012)	To investigate familial and cultural contexts of help seeking among Australian clients with mental health or substance issues	Individual interviews	22 clinicians across 22 services and 56 patients in Australia; Clinician breakdown not reported	For minorities, culture may be expressed through family relationships which may stigmatize treatment compared to Anglo patients who may have little family connection	Cultural background was not understood to be significant in individual service experiences	n/a
Garcia Campayo et al. (2006)	To assess the attitudes of Spanish doctors toward immigrant patients	Cross-sectional survey	191 physicians and residents of whom 119 responded (62.3% response) in Spain	Asked about difficulties with immigrant patients, 36.9% named language, 27.7% named cultural differences, 15.1% named follow up, 2.5% named tropical diseases; 10% did not answer and 7.5% offered various reasons; Asked why immigrant patients are more complex, 48.7% did not answer, 24.3% named cultural differences, 15.1% named high visit frequency, 8.4% named distrust, and 3.3% named different “ways of being”
Gibbs et al. (2004)	To consider the views of Maori patients and their extended family on mental health professionals	Individual interviews	39 interviews with Maori minority patients, their psychiatrists, their case workers, and 5 extended family members per patient in New Zealand	n/a	Maori patients viewed case workers as friendly and trustworthy, but psychiatrists negatively as authority figures focused on medication; Case workers viewed patient involvement in cultural activities and identity exploration as central to engaging mental health services	n/a
Gonçalvez et al. (2014)	To understand provider difficulties in treating Portuguese-speaking patients with mental illness	Individual interviews	3 clinicians in the United States; 1 social worker, 1 assistant, and 1 physician	n/a	Speaking simplified non-medical language improves treatment participation, Stigma of mental illness around people of the same language group may prevent participation in services	Cultural taboos against taking medications may prevent treatment continuation
Iltis (2002)	To explore psychotherapists’ perceptions of salient cultural and ethnic characteristics of adult Dominicans in psychotherapy	Individual interviews	20 psychotherapists with experience caring for Dominicans in the United States; Clinicians were 10% psychiatrists, 30% social workers, and 60% psychologists	Awareness of mental health problems may be lower among Dominicans; Dominicans may present with somatic complaints; Stigma and illegal status may prevent Dominicans from participating in treatment	Dominicans may feel discomfort with psychotherapy	Treatment discontinuation may result from frequent trips to the birth country and therapist distance
Ito and Maramba (2002)	To report Asian American therapist practices with their clients and clients’ perceptions of mental illness and treatment	Individual interviews	16 psychotherapists in the United States; 80% of clinicians were social workers and 20% of clinicians were doctorate level	Asian clients did not see negative symptoms such as social isolation, low energy or excessive sleeping as psychological problems, limiting treatment initiation	American-born Asians prefer exploratory communication and foreign-born Asians prefer directive communication	Treatment continuation was established by reframing psychiatric symptoms as personal problems; Family involvement improves treatment participation with medications
Keating and Robertson (2004)	To explore the content and consequences of fear among African and African Caribbean adults accessing mental health services	Focus groups	45 mental health professionals and 16 focus groups with 29 patients, 19 families and caregivers in England; Of the professionals, 36% were nurses, 29% were social workers, 2% were psychologists, and the rest were support staff	Mental health services may be seen as a last resort; Fears of hospital admission may limit treatment initiation	African and African Caribbean adults may see mental health services as racist and discriminatory, especially when treatment is mandated; Clinicians may be seen as too powerful to challenge	n/a
Knox et al. (2003)	To understand how addressing or not addressing race in psychotherapy (with Black and White American clinicians) affects the therapeutic relationship	Individual interviews	12 licensed psychologists in the United States; 6 men and 6 women, 7 White and 5 Black	n/a	All therapists addressed race if it emerged as a barrier to treatment participation, though Black therapists addressed race even in same-race dyads; all therapists believed that addressing race improved trust and empathy, though Black therapists also discussed factors such as timing, method, and patient defensiveness for consideration	n/a
Rowe (2013)	To explore how substance use counselors balance cultural relevance with fidelity to psychosocial interventions among diverse Australians	Individual interviews	Number of counselors not reported; 24 patients from diverse language groups in Australia	Concerns about confidentiality may limit treatment initiation	Unfamiliarity with counseling may limit treatment participation; Questions perceived to be too personal may limit treatment participation unless trust is established	n/a
Sandhu et al. (2013)	Tp explore mental health professionals’ experiences of delivering care to immigrants across Europe	Individual interviews	34 mental health professionals and 14 administrators in 16 countries, 50% were psychiatrists, 26% were nurses, 15% were psychologists, 9% other therapists	Immigrants may distrust authorities or be unfamiliar with the health care system, limiting treatment initiation and participation		n/a
Saw et al. (2013)	To characterize the cultural needs of Asian immigrants with serious mental illness in smoking cessation	Individual interviews	3 clinicians who provide smoking cessation counseling in the United States; Clinician breakdown not reported	Language access may inhibit treatment initiation	East and Southeast Asian patients found discussions of diet and exercise not culturally relevant; Meditation and prayer were important to treatment participation	Cultural factors around relationships influence participation and continuation— poor clinician rapport may increase discontinuation, but familial relationships incentivize treatment continuation
Smith et al. (2013)	To describe reasons for treatment disengagement and develop practical guidelines for diverse individuals with serious mental illness in America	Individual interviews	25 clinicians from 14 agencies and 56 patients (55% Black, 25% Hispanic, 7% Caucasian, 13% other) in the United States currently treatment disengaged with either a history of criminal justice involvement oir mandated outpatient mental health treatment	n/a	n/a	Clinicians commonly identified family and cultural barriers to treatment continuation, specifically other family members not thinking that the patient needs treatment when the patient is receptive
Sun et al. (2014)	To identify service barriers for Chinese-American families caring for a relative with dementia from caregiver and clinician perspectives	Focus groups	6 clinicians and 6 family caregivers in the United States; 2 social workers, 2 administrators, 1 physician, 1 nurse	Initiation was limited by stigma, language access, and limited understanding of the mental health system amongst family caregivers	Caregiver tendency to minimize problems may inhibit treatment participation	Culturally-specific communication is crucial for treatment continuation
Williams (2010)	To understand the treatment experiences of middle class African-American women	Individual interviews	8 clinicians all in private practice in the United States (25% psychologists, 25% social workers, 50% other psychotherapists)	Beliefs that depression is existential and must be treated through faith limits treatment initiation; Perception of racial similarity led to treatment initiation based on appearance	Clinician interest in race and gender improves patient treatment participation; Therapist personal disclosure and physical contact may lead to treatment participation	n/a
Yoshida (2013)	To explore how race and ethnicity affect Asian therapists working with various clients	Individual interviews	10 clinicians in the United States (40% social workers, 40% psychologists, and 20% other)	Perceptions of ethnic similarity led to patient treatment initiation based on appearance	Counselors adopted directive communication to improve treatment participation and continuation

n/a – No applicable finding from the study

MDEMS – Mental Distress Explanatory Model Schedule

RIAS – Roter Interaction Analysis System

WAI-O – Working Alliance Inventory, Observer form

Investigators used mostly qualitative methods such as individual interviews (n=21). Focus groups were also used, with (n=2) and without (n=3) individual interviews. Mixed methods were used less frequently in controlled, observational designs (n=1), cross-sectional surveys (n=1), and a combination of focus groups and chart reviews (n=1). The controlled, observational study was the sole prospective study in the sample with the rest employing a cross-sectional design (n=22). Only two studies used objective measures not relying on self-report, one for communication content and another for communication context. Sample sizes ranged from 3 to 191. Fifteen studies provided information on treatment initiation, 18 on treatment participation, and 13 on treatment continuation, with some studies contributing information about more than one aspect of treatment engagement.

3.1. Treatment Initiation

Of 15 studies on treatment initiation, 13 reported on communication content, 2 reported on context, and 1 reported on both. The most common reason stated by clinicians for patient delays in treatment initiation (n=5 studies) was the interpretation among patients that the mental health system and its services could not help them. This was reported for Bangladeshi women immigrants in the UK [41], Chinese immigrants in Australia [42] and in the US [59], Asian and African immigrants throughout Europe [56], and elderly blacks and Hispanics in the US [44]. Somatic interpretations of mental illness were the second commonest reason for delays in treatment initiation (n=4 studies). Clinicians stated that patients interpreted neuro-vegetative symptoms such as poor energy or poor appetite as better treated in primary care settings or not warranting treatment at all. Somatic interpretations and presentations in non-mental health settings were common among patients with limited English proficiency in Australia [46] and South Asian [45], Dominican [51], and East Asian [52] immigrants to the US.

The third reason reported for delays in treatment initiation was stigma, though no common definition of stigma was used across the studies. Interpretations of stigma encompassed concerns about family dishonor among Bangladeshi women in the UK [41] and cultural minorities in Australia [47], illegal immigration status among Dominicans in the US [51], fears of involuntary hospitalization among African and Caribbean immigrants in the UK [53], and a lack of medical confidentiality among cultural and linguistic minorities in Australia [55]. Three studies reported that depression was interpreted as an existential form of suffering among African Americans [44, 60] and immigrants from India in the US [45] that required personal fortitude and religious introspection, not a treatable mental condition requiring services.

Three studies reported context-based communication studies. Two studies reported that patient perceptions of clinician ethnic resemblance encouraged treatment initiation among African American women [60] and Asian Americans [61]. Both studies also noted that delays in treatment initiation occurred based on patient concerns that ethnically different clinicians would not understand their cultural needs. Only one study on East Asian immigrants in the US reported that a lack of fluency in the language of clinicians delayed treatment initiation [57].

3.2. Treatment Participation

Of 18 studies on treatment participation, 17 focused on communication context and 1 focused on content. The most common finding was patient discomfort in discussing emotions with clinicians who were viewed as strangers (n=6 studies). This tendency was reported among cultural and linguistic minorities in Australia [55], Asian and African immigrants in Europe [56], immigrants in Spain [48], and Indian immigrants [45], Dominican immigrants [51], and elderly blacks and Latinos in the United States [44]. One study reported that caregivers of Chinese American patients with dementia minimized problems to clinicians, limiting treatment participation [59]. This study did not frame minimization in terms of discomfort or distrust, but as a cultural expectation in interactions with strangers.

Two studies discussed patient preferences for clinician communication styles. Clinicians noted decreased treatment participation among female Bangladeshi immigrants to the UK due to a preference for directive communication in which the clinician instructs patients rather than an exploratory style in which clinicians encourage patient introjection to uncover their own desires [41]. However, a study of Asians in the US found that those native-born preferred an exploratory style compared to those foreign-born who preferred a directive style, suggesting an acculturation effect [52]. Two studies discussed perceptions of clinician power. A study among Maori patients in New Zealand noted that case workers were viewed as friendly and trustworthy, but psychiatrists were viewed as negative authority figures focused on medication management [49]. A study of Africans and Caribbean immigrants in the UK found that clinicians were seen as intimidating based on historical legacies of discrimination [53].

Six studies discussed the role of clinician communication behaviors in patient treatment participation. The sole study utilizing a prospective observational design and not based purely on clinician self-report compared patient-clinician dyads in three ethnic categories: white dyads, Latino dyads, and mixed dyads [40]. Compared to other dyads, Latino concordant dyadic patients were more verbally dominant, had more patient-centered communication, scored higher on the Working Alliance Inventory (WAI) (p<.05), and had more positive affect with more positive statements as measured through the Roter Interaction Analysis System (RIAS) (p<.001) [40]. The RIAS is a method to code medical interactions based on linguistic-based techniques of conversation analysis [62] and the WAI measures patient-clinician bond, tasks, and goals from patient, therapist, and observer perspectives [63].

Other studies noted clinician attempts to improve treatment participation by changing the context of communication. Black and non-black clinicians working with black patients found that discussing race in psychotherapy improved trust and empathy [54], a finding also among middle-class, black women [60]. This approach turns the context issue of potential interpersonal mistrust into a content issue for explicitly exploring patient interpretations of difference. Clinicians working with Asians in the US have successfully adopted directive rather than exploratory styles of communication based on patient preferences, though clinicians relied on past experiences with this population rather than discussing preferences explicitly [61]. Clinicians of Portuguese-speaking patients have used simplified non-medical language to involve patients collaboratively [50]. Only one study of Chinese immigrants in Australia noted that patient perceptions of clinicians treating them unprofessionally due to limited English proficiency were responsible for decreased treatment participation [42].

Only one study focused on patient cultural interpretations of care relevant to the clinical relationship. East and Southeast Asian immigrants in the US believed that clinician-initiated discussions of diet and exercise were unnecessary in mental health contexts, preferring instead to discuss treatment modalities such as meditation and prayer [57].

3.3. Treatment Continuation

Of 13 studies on treatment continuation, 7 focused on communication content, 5 focused on context, and 1 focused on both. The most common finding was that patient interpretations of therapies that were deemed appropriate influenced decisions to continue or discontinue treatment and, specifically, medications (n=5 studies). Clinicians reported that Bangladeshi women in the UK [41], and Indian immigrants [45], elderly blacks and Hispanics [44], and Portuguese-speaking immigrants [50] in the United States discontinued medications perceived as unnecessary. However, black, Hispanic, and Asian patients in the US continued medications when clinicians confirmed with family members that the use of medications is a culturally-appropriate expectation of care [58]. The second commonest finding was that clinicians improved treatment continuation after framing communication according to patient interpretations. This occurred with Asian immigrants in the US when clinicians framed psychiatric symptoms as existential rather than biomedical problems [52] – especially among Chinese American caregivers of dementia patients [59] – and patient engagement as clinicians saving “face” in front of other clinicians [57].

The most common finding (n=2 studies) among studies focusing on communication context concerned fears of a lack of medical confidentiality if clinicians were from the same cultural group, among South Asian women in the UK [43] and Portuguese-speaking immigrants in the US [50]. In these studies, patients feared that clinicians would not maintain patient confidentiality in shared community settings. One study with Dominican patients in the US noted that treatment discontinuation resulted from frequent trips to the birth country since regular (weekly or bi-weekly) communication with clinicians was not seen as necessary [51]. In contrast, a major barrier to trust among African and Caribbean immigrants in the UK was the need to report on the same symptoms with different psychiatrists each visit, limiting continuation due to frustration with not having a single clinician [53]. In a reverse trend, clinicians in Spain expressed frustrations that immigrant patients with limited language fluency may come to clinicians more than scheduled to receive other types of social services, worsening rapport [48].

Only one study reported an attempt to improve treatment participation by changing the context of communication. As with treatment participation, Asian American therapists found with Asian American patients that directive rather than exploratory communication improved treatment continuation [61]. As before, this study relied on clinicians’ past experiences with this population rather than explicitly eliciting patient preferences.

4. Discussion and Conclusion

4.1. Discussion

This systematic review synthesizes scholarship on clinician data relating cultural communication to treatment engagement among minorities with mental illnesses. The 23 studies reflect significant international interest in this topic, predominantly from English-speaking countries with multicultural populations. Only 3 studies had quality ratings [33] greater than 50% with most opting for purposive sampling without using validated instruments. All but one used cross-sectional study designs, limiting inferences of causation between patient-clinician communication and treatment engagement. Some studies reported on clinician experiences to improve patient-clinician communication, though none reported randomized controlled interventions to target specific mechanisms of action. These findings suggest that clinician problems in communication with racial/ethnic minorities have been amply documented to warrant the development of interventions.

Despite these limitations, broad patterns can be identified. Studies on treatment initiation and continuation mostly reported on communication content compared to treatment participation that largely focused on communication context. The most common reasons for lower treatment initiation were patient interpretations that the mental health system could not help them, somatic symptoms were not part of mental illnesses, and stigma would interfere with care. These results can inform ongoing engagement initiatives. Community-based psychoeducational interventions are typically designed to increase awareness about mental illness and improve treatment initiation [64]. However, psychoeducational approaches assume that patients have decreased health literacy and that improved literacy increases service engagement. Our review challenges this notion since patients disengage from services when treatments are not interpreted as relevant. Instead of emphasizing only health literacy, psychoeducation can also adopt content-based communication strategies, recognizing that patients possess their own interpretations of illness and care. A better-examined phenomenon has been the integration of mental health and primary care services through collaborative care to avoid interruptions in clinician referrals. Group-based therapies for major depression in primary care over 12 weeks have improved symptoms and treatment adherence among depressed patients [65]. Collaborative care implementing mental health treatment within primary care has been effective for anxiety and depressive disorders [66] in different countries [67] and immigrant groups [68], though controlled studies have not been conducted for schizophrenia and bipolar disorders [69]. Rather than view collaborative care solely as a service model to prevent interruptions in care, clinicians can use this service model to reiterate that mental illnesses include somatic symptoms and that clinicians will protect patient confidentiality to avoid stigma.

Similarly, clinicians and administrators may improve treatment participation by improving communication context. Discomfort in discussing emotions with strangers, preferences for a particular style of clinician communication, and concerns about clinician power were the most common reasons for decreased patient treatment participation among minorities. Cultural expectations of communication affect each of these reasons: from the circumstances of where, when, how, and to whom patients narrate illness [35] to preferences for communication styles [70] and patient concerns about clinicians’ potential misuse of their power and authority based on clinician responses to patient narratives [71]. This review shows that some clinicians converted patient concerns about race and interpersonal differences, a context issue, into a content issue as a communication strategy. These differences occurred in dyads where patients and clinicians did not share the same backgrounds.

Consequently, the context finding that ethnically-matched Latino dyads improved treatment participation suggests a complex relationship among culture, ethnicity, and communication. Studies have shown that ethnically-matched dyads improve treatment duration and outcomes among ethnic minorities [72–74]. Apart from the sole study in this review [40], we know of no other studies on clinician communication behaviors comparing ethnic groups in mental health settings. Two studies [43, 50] reported patient concerns about medical confidentiality when clinicians belonged to the same cultural groups, suggesting that ethnic matching may not be a solution for all minorities given the limited number of minority clinicians. Instead, patients and clinicians in ethnically-matched dyads may be responding to culturally-accepted expectations of communication in which shared understandings about communication content and context, rather than just ethnicity, improve treatment participation. Culture is inextricably linked to medical communication in determining the topics deemed suitable for discussion (content) as well as the use of clinical vocabularies and linguistic registers that convey authority (context) [75, 76]. These aspects may be taken for granted in patient-clinician ethnic matching and have been the focus of attention in culturally-based interviews.

Finally, clinicians described three main types of communication content related to patient treatment continuation: patient interpretations that treatments were appropriate, clinicians framing discussions based on patient models of illness, and discussions about medical confidentiality. These findings correspond to communication strategies recently developed by cultural psychiatrists. Semi-structured interviews such as the DSM-5 Cultural Formulation Interview (CFI) [77] have encouraged clinicians to use patient vocabularies, explore patient interpretations of illness and care, and inquire about concerns in the patient-clinician relationship, targeting communication content and context. Patient satisfaction with medical communication has risen immediately after CFI use [78] though studies are needed that examine the relationship between improved patient satisfaction and treatment continuation. Another cultural interview known as motivational pharmacotherapy [79] elicits patient preferences for medication and appointment frequency and can assist clinicians with tailoring treatment plans. These interviews also elicit patient concerns about confidentiality, a barrier to treatment continuation identified among immigrants [80]. Turning a communication context issue such as confidentiality into a content-based discussion is an example of addressing patient communication concerns in line with this review’s findings, though it may not be possible to convert all context issues into content discussions since cultural aspects of communication are often implicit understandings [75]. Nonetheless, the context-content conversion is a clinician communication strategy that offers a starting point to develop interventions.

This systematic review has several limitations. First, we may have missed articles that could have been included, though this is a limitation of all systematic reviews and we used a standard method to avoid this. Second, the heterogeneity of study methods prevented us from completing a meta-analysis or estimating effect sizes. However, this limitation characterizes the field of cultural communication in mental health and is not limited to this paper. Third, this study relied upon codes and definitions for cultural communication and treatment engagement to interpret and synthesize all records. Others may differ in their definitions and interpretations for these terms. Nonetheless, we have been transparent about our definitions and coding process for reproducibility and generalizability.

Our literature review points to future research directions. Interventions can be designed around common challenges experienced across multiple cultural groups for maximal impact such as clear steps for clinicians to elicit and incorporate patient treatment preferences into treatment planning. For example, interpretations that certain treatments recommended by clinicians were not relevant to the patient’s illness were common among several racial and ethnic groups. Mental health clinics may erroneously assume that patients will adhere to clinician recommendations without first asking patients about treatment preferences. Instead, communication interventions can examine the exchange of ideas about illness and treatment and interpersonal and situational influences affecting this exchange of ideas in controlled trials as they relate to treatment initiation, participation, and continuation.

4.2. Conclusion

Distinctions between communication content and context occurred by treatment engagement phase. Decreased patient treatment initiation was most commonly due to interpretations that the mental health system could not address patient illness experiences, somatic interpretations of mental illness leading to treatments in other settings, and interpretations of mental illnesses as existential rather than biomedical problems. Delays in treatment initiation were related to patient perceptions that the mental health system does not address their interpretations of the illness experience and expectations of care. This theme continues in treatment continuation with most studies noting that patients disengaged from services due to differing expectations of care with clinicians. In contrast, treatment participation focused on communication context rather than the content of patient interpretations. Decreases were attributed to patient discomfort in discussing emotions with strangers, unmet expectations of communication style, and concerns about clinician power whereas increases were attributed to patient-centered communication and positive affect, discussions of background differences with minorities, use of simplified language, and communication tailored to patient preferences.

4.3. Practice Implications

Our results have practical implications for extant service settings. To close disparities, clinicians can improve communication content by incorporating relevant patient interpretations of illness and care into treatment planning and context by improving interpersonal behaviors and situations that promote patient satisfaction. Clinicians can adopt content-based communication strategies to improve treatment initiation by discussing services offered through the mental health system, links between somatic symptoms and mental illnesses, and ways to target stigma. Clinicians can adopt context-based communication strategies to improve treatment participation by discussing any patient discomforts around sharing emotions with strangers, exploring preferences for directive or exploratory communication styles, openly addressing cultural differences such as race and ethnicity, and simplifying language. Finally, clinicians can adopt content-based communication strategies to improve treatment continuation such as discussing treatments acceptable to patients rather than making assumptions, framing the content of communication based on patient models of illness, and addressing concerns about confidentiality. We have focused on the perceptions of clinicians who provide direct care to patients, though the contributions of administrators and patients themselves must also be areas of ongoing research and practice.

Highlights.

• Communication affects mental health treatment retention for minority patients.

• Communication can be divided into cultural content and interpersonal context.

• Cultural content is important for minorities starting and maintaining treatment.

• Interpersonal context is important for minorities participating in treatment.

• Clinicians can improve content and context communication skills.