Interventions to improve cultural competency in healthcare: a systematic review of reviews

In November 2011 the following databases and electronic journal collections were searched from 2000 to 2011: Medline, Cinahl, Eric, PsycINFO, Proquest (Dissertation/Theses), Scopus and the Cochrane Systematic Review Database. Reference lists were hand-searched for other reviews. Key websites (i.e. http://​www.​diversityrx.​org, http://​www.​nccc.​georgetown.​edu, http://​www.​minorityhealth.​hhs.​gov, http://​www.​ceh.​org.​au, http://​www.​hrsa.​gov, http://​www.​nice.​org.​uk) were also searched. In July 2012 the search was updated to include recent studies published up to June 2012. See Additional file 1 for search terms used.

As cultural competency did not achieve popularity until the late 1990s and government policies mandating cultural competence did not occur until the early 2000s [11], a search timeframe of 2000–2012 was chosen.

A review was considered eligible for inclusion if it met the following criteria: i) included quantitative, qualitative or mixed methods studies, ii) was written in English, iii) included studies involving: health care/service providers/practitioners/clinicians, health administrators, support staff and patients/clients/health service users, iv) included studies utilizing any strategies or interventions to improve cultural competency (e.g. training programs or workshops or educational courses), v) included studies involving intervention settings or services related to the health sector (e.g. hospitals, community health services, educational institutions teaching health related courses), vi) included studies that used one or more outcome measures at an individual level (e.g. survey), organizational level (e.g. programs) or system level (e.g. policies).

Reviews were excluded if they described cultural competency in other non-health settings (e.g. education system), were conducted prior to the year 2000, or did not contain a methods section that included information on: search strategy, number of included studies, and details of studies.

Each review was critically appraised independently by two authors using the health-evidence.org tool for reviews [12]. This tool consists of ten questions to assess the quality of the review using commonly accepted evidence-informed principles. Reviews were given a quality assessment rating of weak, moderate or strong.

The total search identified 6,830 results. Based on the inclusion and exclusion criteria, titles and abstracts were screened for eligibility by the first author. Full texts were retrieved for all reviews where inclusion was in doubt. To reduce the potential for bias in the screening process, the second author independently screened 10% of the total identified titles and also extracted data for 10% of the reviews that met the inclusion criteria. There was no difference in agreement between reviewers. See Figure 1: PRISMA flow diagram for a flow chart summary of the search and inclusion/exclusion process.

Data extracted from the included reviews was entered into an Excel 2011 spreadsheet under the following headings: author(s), year of publication, health care setting/context, definition/concept/framework of cultural competence, method of review (e.g. database(s) searched), inclusion/exclusion criteria, number of included papers, types of papers included, study quality assessment, major findings, recommendations and quality/critical appraisal of review.

Data extracted from the reviews was descriptively analyzed using Excel 2011. Meta-analysis was not conducted due to the heterogeneity of the reviews and their included studies. Analysis focused on: types of interventions and study outcomes.

Searching yielded a total of 6,830 titles, of which 19 met the inclusion criteria and were extracted for analysis (Table 1) [3‐8, 13‐25]. The main reasons for exclusion were: articles were commentary or opinion pieces, articles were of primary studies, review articles examined cultural competency assessment tools and review articles but did not include any studies with interventions. Six review articles were excluded for not providing information on search strategy and details of included studies [26‐31].

The majority of reviews (n = 15) were published between 2007 and 2012. Reviews focused on a range of health care settings/contexts, including: health professionals, community rehabilitation, nursing and health systems. A range of study designs were included in the reviews, including randomized controlled trials, pre and post designs as well as qualitative studies. Most reviews provided a definition of cultural competency or related concept. The number of studies included in each review varied between 5 and 38. Smith et al.’s [19] review consisted of two meta-analyses, of which only the second meta-analysis (n = 37) met the inclusion criteria. (The first meta-analysis consisted of retrospective survey studies that did not report outcome measures.) Thirteen reviews assessed the quality of studies using critical appraisal tools such as the Oxford Centre for Evidence Based Medicine [13].

Types of interventions to improve cultural competency included in the reviews were: training/workshops/programs for health practitioners (e.g. doctors, nurses and community health workers), culturally specific/tailored education or programs for patient/clients, interpreter services, peer education, patient navigators and exchange programs.

Seven of the 19 reviews focused solely on healthcare provider cultural competency interventions [6, 7, 13, 14, 17, 19, 23] whilst six reviews examined only culturally competency interventions aimed at patients/clients [8, 18, 20‐22, 24]. Three reviews included studies that examined organizational level interventions such as culturally adapting health programs for patients and employment of bilingual community health workers [5, 15, 16]. One review focused primarily on interventions directed at health care personnel and/or organizations, although interventions targeting both health personnel and patients were also included [25]. Another review looked at the structures and processes that support the development of culturally competent practices [3]. Evaluated models of cultural competence in mental health were reviewed by Bhui et al. [4].

There were three main categories of study outcomes amongst the reviews: provider-related outcomes, patient/client-related outcomes and outcomes related to health service access and utilization. Evaluations of implemented models of cultural competency [4] and cost-effectiveness [6] were also examined.

Measured provider outcomes focused on knowledge, attitudes and skills related to cultural competency. In Beach et al.’s [6] review, knowledge refers to information about general cultural concepts such as the impact of culture on the patient-provider encounter or culture-specific knowledge such as traditional cultural practices. Attitude outcomes measured by studies included cultural self-efficacy (assessing learner confidence of knowledge and skills in relation to ethnic minority patients), attitudes towards community health issues, and interest in learning about patient and family backgrounds [6]. Skills included communication skills or use of treatment plan. In contrast, Smith et al.’s [19] review used multicultural counseling competence as their main outcome measure while Kokko [17] included studies examined nursing students’ cultural knowledge, personal growth and nursing practice.

There were a variety of patient/client outcomes reported, including physiological outcomes such as blood glucose, weight and blood pressure [8] as well as outcomes such as patient satisfaction and trust [7], knowledge of cancer screening and knowledge of health conditions [15]. Behavioral outcomes such as dietary and exercise behaviors were also examined in three reviews [15, 18, 21]. Other reviews looked at primarily patient-focused interventions to improve breast and cervical cancer screening among women [24] and to improve participation in cancer treatment processes [22].

Outcomes related to health service access and utilization included use of bilingual community health workers, interpreters, and patient navigators. These interventions were designed to influence individuals’ ability to access the resources of health care organizations by bridging the cultures of the organizations and those of the target communities [15].

Cost-effectiveness of interventions was considered in three reviews [6, 8, 24]. Beach et al. [6] noted that only 4 of 34 studies included in their review addressed the costs of cultural competence training. No studies in Hawthorne et al.’s [16] review measured the cost effectiveness of their interventions, although some included a rough estimate of costs. Two studies in Lu et al.’s [24] review reported cost information. These reviews noted this as an important limitation of studies they examined.

The majority of reviews noted methodological limitations of studies. This limited conclusive statements about the effectiveness of interventions to increase cultural competency. The main methodological criticisms of the studies by the reviews were: small samples [13], poor methodological rigor [7, 13, 15], no or few long-term studies [8, 18], no economic analysis of interventions [6, 8], reliance on self-report measures [19], lack of detail about interventions [7, 19], lack of patient outcome measures [4‐6, 15] and lack of objective provider measures related to change in practice [14, 17].

Some reviews reported the quality/strength of evidence supporting the outcomes measured [5‐8, 15, 25]. For example, Beach et al. [6] graded the strength of evidence for each outcome type based on its quality, quantity, and consistency (grades A – D). In their review, evidence of impacts on provider knowledge were graded A compared with provider attitudes which were graded B.

Twelve of the nineteen reviews concluded that further research (e.g. more rigorous trials and evaluations) was required to determine the effectiveness of interventions to improve cultural competency for providers and patients/clients. The reviews found that many of the studies were difficult to compare as different frameworks of cultural competency were used and studies often lacked a standardized and validated instrument to measure cultural competence [6]. Most reviews concluded that training had positive impacts on provider outcomes. However, it was difficult to determine exactly what types of training interventions were most effective in relation to particular outcomes [6, 13, 19]. A need for research into long-term outcomes [8, 18] was identified along with the need to consider other factors that facilitate cultural competency, such as links with community organizations [3, 15]. It was also recommended that cost-effectiveness be assessed [8, 24].

Some of the reviews focused on one type of intervention such as diabetes education for patient outcomes [8] or health provider cultural competency training [14]; one type of study outcome such as patient outcomes [7]; one type of study design such as randomized controlled trials [25]; or a particular study population such as Hispanics [20], Asian women [24] or nurses [17]. Although it may be more feasible for a review to focus on a particular group of health providers or type of health care setting, it limits generalizability and applicability of the findings. Many studies were heterogeneous in outcome and interventions, making statistical synthesis and analysis difficult [13]. According to Smith et al. [19], their meta-analysis was limited by: studies with single-group pre-to post-test assessments (e.g. [32, 33]), studies rarely reporting disaggregated data, and predominantly self-report measures (e.g. [34, 35]).

It is also difficult to determine the extent to which knowledge and skills learnt from training/programs are translated into practice and how they impact on patient/client outcomes [5, 17]. Provider outcomes determined by self-report are subject to multiple threats to internal validity [36] and hence limit the conclusions made regarding impact on provider practice [19] and ultimately on patient outcomes.

The literature includes a diverse range of populations (e.g. African American, Hispanic/Latino, and Asian), health care settings (e.g. community centers, hospitals and academic medical centers) and interventions (e.g. culturally tailored programs for particular racial/ethnic groups and provider training). However, the majority of studies were based in the United States. Two reviews limited their included studies to only those conducted in the United States [15, 21].

Meta-analysis was not conducted in this review of reviews due to the heterogeneity of the reviews and their included studies. Intervention effects were also difficult to determine as only some reviews described outcomes in terms of statistical significance and effect sizes [5, 6, 8, 13, 19]. Some reviews noted that studies rarely provided sufficient information on the curriculum or format, or details of the providers involved (e.g. age, race, gender, prior training) making it difficult to conclude from studies which types of training interventions were most effective for which groups and in producing which particular outcomes [6, 7].

All reviews were critically appraised by two authors using the health-evidence.org tool for reviews [12] and companion tool dictionary [37]. This process was reliant on the author’s prior knowledge and experience of the topic, research principles and study design methods. There were minor disagreements between authors and consensus was reached through discussion. Reviews were predominantly of moderate-strong quality (overall assessment of review quality is included in Table 1).

All reviews had a clearly focused question in relation to the population, intervention and outcomes. Appropriate inclusion criteria to select primary studies were used by the majority of reviews. The majority of reviews described comprehensive search strategies, although some were slightly limited in scope [4, 14, 15, 20]. For example, Kehoe et al. [18]’s search strategy consisted of only two electronic databases. The number of years covered by the search strategies was 20 years or more by the majority of reviews. Two reviews [16, 17] covered 10–11 years, one review searched between 2005–2011 [23], and one review [14] did not provide this information.

The methodological rigor of studies was identified and described in thirteen reviews [3, 5, 7, 8, 13, 15, 16, 18, 19, 22‐25]. The methodological rigor of primary studies using an assessment tool/scale was conducted by all these reviews except for two [18, 19]. Nine reviews reported the use of two or more reviewers to assess each study for methodological quality [6‐8, 13, 15‐17, 22, 24]. Most reviews used appropriate methods for combining and comparing results across studies. However, Pearson et al.’s [3] results were not well presented.

Cross-cultural interactions are likely structured and shaped by the worldviews and past experiences of not only the staff and clients but also the culture of the organization, which is embedded in and produced by policy frameworks, organizational arrangements and physical settings of the organization [38]. Interventions to improve cultural competency need to consider the individual and organizational contexts and the interplay between them. Training programs may need to be tailored to particular groups, for example physicians would need particular knowledge and skills specific to their clinical tasks that would be inapplicable for reception staff.

It is likely that cultural competency training as a stand-alone strategy is insufficient to improve patient outcomes without concurrent systemic and organizational changes [7, 9, 39]. Embedding cultural competency in organizational policy documents such as position statements and strategic plans are more likely to result in sustained change within organizations. There should be a commitment among the leadership of the organization and embedded key performance indicators supported by allocated resources.

There is some evidence of a relationship between the cultural competence of health practitioners and the cultural competence of organizations [39, 40]. Providers may be influenced by their organization’s commitment and actions in relation to cultural diversity and vice versa. A study found that providers with attitudes reflecting greater cultural motivation to learn were more likely to work in clinics with more culturally diverse staff and those offering cultural training and culturally adapted patient education materials [40].

In recognition that different components of the health system influence health outcomes, some models of cultural competency advocate a multi-level approach [2, 41]. Although some studies have shown that culturally competent practices among organizations are adopted to varying degrees [42, 43], more research is needed in this area. Grol et al. [44] found that empirical evidence of the effectiveness and feasibility of most theoretical approaches to produce change in health care was limited. Dreachslin et al. [45] found a paucity of research on organizational behaviour in the healthcare and general management literature. A more recent review by Parmelli et al. [46] showed limited available evidence regarding effective strategies to change organizational culture in health care. Barriers and incentives to organizational change should be considered when designing and implementing an intervention to increase the likelihood of success and sustained change [47, 48]. Issues related to organizational readiness for change and innovation also require consideration before implementing organizational cultural competency interventions [49]. Understanding the reasons for adoption and spread of innovation can assist with addressing the difficulties of organizational change [50]. Planning and implementation of cultural competency interventions should acknowledge the interaction between an intervention and the setting. Organizational cultural competence involves an understanding of the strengths and weaknesses of the health care organization and the unique needs of the people it serves.

Self-assessment was the most common approach to assessing cultural competency, which is a subjective measure subject to a range of biases [36]. The assessment tools were mostly process and survey tools, including patient satisfaction and provider self-assessment questionnaires as well as self-administered organizational checklists. Many of these tools have not been validated [51]. Self-rating at the individual level may be affected by the respondents’ level of cultural awareness and is subject to biases such as social desirability [52]. Broader organizational and systemic approaches to cultural competency should consider assessments of cultural competency that include objective measures such as document review [52]. Moving beyond self-assessment is a necessary step towards developing a stronger evidence base for the use of cultural competency related interventions to improve patient/client health outcomes. In addition, more research is needed to determine how well both individual-level and organizational-level guidelines for cultural competency are followed by those directly involved in service delivery [53].

Academics have asked whether cultural competency can be achieved without focusing on issues related to racism and white privilege [54, 55]. Concepts related to racism, bias and discrimination were noted in some reviews [3, 5, 6, 14, 15, 23], although none were measured as outcomes in studies. Two of the 34 studies in Beach et al.’s (2005) review included mention of these concepts in their education content. Factors such as structural inequalities and racism may have a greater impact on health disparities between particular groups than cultural differences [56].

Self-reflection and awareness of one’s professional and personal culture is an important component of cultural competency [57]. Of the seven reviews that focused on provider outcomes, four discussed these concepts [13, 14, 17, 23]. This is despite these self-reflexive elements being critical to cultural competency. Cultural awareness alone is inadequate for addressing the effects of structural and interpersonal racism on health disparities. Cultural awareness training has been criticized for increasing stereotyping and reinforcing essentialist racial identities [58]. Reflexive antiracism training is a promising alternative to cultural awareness training that reflects upon the sources and impacts of racism on society whilst avoiding essentialism and negative emotional reactions associated with White guilt [59].

There is a tendency within healthcare to equate culture with essentialized notions of race and ethnicity, which can lead to practices that separate culture from its social, economic and political context [10]. Narrow conceptualizations of culture and identity may limit the effectiveness of particular approaches, and a focus on specific cultural information may inadvertently promote stereotyping. Care must also be taken to avoid over-focusing on ‘culture’. Although cultural differences may worsen the problem of differential access and discrimination, broader factors such as poor education and poverty may play a greater role in the poorer health outcomes of some individuals and groups in the community [60]. However income and race/ethnicity as risk factors for health disparities can overlap and discrimination is often a driver of socio-economic disparities [61].

A limitation of this review is that there may be primary studies in the field that are not included by existing reviews in the literature as this was a review of reviews rather than a review of primary studies. There were only 22 primary studies cited by more than one review. This is likely due to the relatively narrow focus of some reviews e.g., by health condition, minority groups or type of outcome (i.e. patient or practitioner). Another limitation is that only reviews published in English were included. Given that the timeframe for this review was 2000–2012 it is possible that reviews published prior to 2000 were not included. However, cultural competency did not achieve popularity until the late 1990s and government policies mandating cultural competence did not occur until the early 2000s [11].

The search strategy could have been improved by adding more patient-related terms (e.g. migrant and refugee) and applying a more sensitive search filter for systematic reviews. It is possible that some reviews were excluded as a result, however it is unlikely that the overall findings would be significantly different as this paper includes reviews from various health care contexts and health care provider and patient populations as well as different types of studies and cultural competency interventions.

It is difficult to make a general statement about the strength of the effect of interventions as the reviews assessed their studies differently. Different critical appraisal tools were used by authors and a majority of reviews noted methodological limitations of studies in their reviews, which limited their ability to make conclusive statements about the effectiveness of interventions. In order to determine the strength of the effect, an assessment of all primary studies of the included reviews using a single critical appraisal tool to determine the effectiveness of their interventions and an assessment of quality and bias of each individual study is required. However this is beyond the scope of this paper.