Aim

The aim of this review is to synthesise the evidence for complex interventions to support access and/or engagement of ethnically underrepresented communities in UK health services. The synthesis will align to Levesque et al.’s patient-centred access framework which encapsulates five service domains: approachability, acceptability, availability, affordability, and appropriateness [14]. Data extraction and analysis will also include equity factors guided by the PROGRESS-PLUS framework [15]. The findings will enable 1) the identification of gaps for future research and 2) preliminary guidance for policymakers, clinicians and public health bodies on how best to support healthcare service access and engagement of ethnically underrepresented communities.

Eligibility criteria

Studies of interventions to improve ethnically underrepresented communities access to and/or engagement with adult UK healthcare services were included. These included interventions targeted at both patients and healthcare professionals (e.g., cultural competency training). The UK Medical Research Council (MRC) define complex interventions as having several interacting components and/or the properties itself being complex [17]. The MRC also describe how complex interventions may require flexibility or permit tailoring, be dependent on the behaviours of those delivering and receiving the intervention and involving multiple systems (e.g., individual, organisational, societal) [17]. This definition has been applied to the studies within this review. Ethnically underrepresented communities were defined as all ethnic groups except White British [18]. Studies focused on testing existing healthcare interventions on ethnically underrepresented communities but not aiming to improve access and/or engagement in health services were excluded. Studies exploring underrepresented groups where ethnically underrepresented communities were not > 50% of the participants were also excluded.

Searches

CINAHL Ultimate, Medline, Psychology and Behavioral Sciences Collection, APA Psych Info, Cochrane Library and Pedro were searched on 8^th April 2024. The search strategy and keywords were developed using the Population (ethnically underrepresented communities), Concept (access/engagement), Context (National Health Service [NHS]/UK health services) framework (S1 Appendix) [19]. Searches were limited to English language and by publication date (01/01/2014–08/04/2024). The search was limited to studies published from 2014 onwards to ensure relevance to the current NHS context, as earlier studies may not reflect current policy, service delivery changes or population needs; this timeframe also limited the number of results which aligned with the available resources for this project.

The initial searches were completed on CINAHL and Medline exploring combinations of search terms to identify the most effective terminology before being completed on the remaining databases. Reference and citation lists of included articles were also searched.

Selection

All search results were exported to the Rayyan database for de-duplication, screening, and selection against eligibility criteria by two independent researchers (JRP, SC, SN, CP, TK) [20]. Conflicts were resolved by consensus.

Data charting items and process

Data from the included articles was extracted to a standardised Excel data chart by an individual researcher (JRP) and 10% of the data was independently charted by another researcher for comparison (SN) [21]. Extracted data included study title, author, year of publication, study design, location, population, healthcare setting and intervention (S2 Appendix). The intervention components were mapped to the Patient Centred Access Framework [14].Eligibility criteria, baseline characteristics and subgroup analyses were mapped to the Cochrane and Campbell Collaboration Equity methods group’s PROGRESS-Plus framework [15,22]. PROGRESS is an acronym for Place of residence, Race/ethnicity/language/culture, Occupation, Gender, Religion, Education, Socioeconomic status and Social Capital [15,22]. PLUS enables the framework to capture other factors which impact equity, such as 1) age 2) disability 3) features of relationships and 4) time-dependent relationships [15,22]. Using the PROGRESS-Plus framework in the inclusion/exclusion analysis provided a structured and transparent way to map equity characteristics of each study, allowing a better understanding of the diversity of the populations represented under the broad term “ethnically underrepresented communities”. This approach helped identify which specific dimensions of disadvantage (e.g., language, socioeconomic position, language) were considered across the studies, highlighting variation and gaps in how equity has been defined and operationalised.

Synthesis of results

Numerical content analysis was used to summarise study designs, locations, populations and healthcare settings. The remaining data analysis was guided by Braun and Clarke’s [23] Thematic Content Analysis and reflexive thematic analysis [24]. Analysis involved deductively categorising components of interventions to Levesque et al’s patient-centred access framework domains [14]. The software Trello was used to extract and code the themes of each domain into categories (combining similar themes or those with overlaps). Trello was utilised to support organisation, visualisation and coding of themes, whilst it is typically an organisational tool, it has also been used in other qualitative studies to support reflexive thematic analysis [25]. The included papers were re-read to inductively assess for any additional themes not encompassed in the patient-centred access framework. After the information on Trello was finalised, a Word document of themes and categories was then developed and reviewed (by JRP and GW) (S4 Appendix).

Participants

A total number of 2,761 participants plus 102 stakeholders/service providers were included in the 35 studies. Sample size ranged from three [55] to 800 participants [28]. Four studies did not report the number of participants, two due to the use of aggregate health data [46,48] and two reported on programmes that did not collect numbers of participants they engaged with [29,32]. Studies included different populations including those identifying as Gypsy, Roma and Traveller groups, South Asians, Pakistanis, Chinese migrants, Jewish communities, migrant mothers broadly, self-identifying Muslims, Bangladeshis, Black British people, Black Caribbean or African communities, and Black, Asian and ethnically underrepresented communities more broadly.

Health services and location

The studies covered various health service areas, with 22 (63%) primary care, 11 (31%) mental health, 1 (3%) dental health and 1 (3%) maternity care.

Geographically, 40% of the studies were in London, 16% in Manchester, 13% were UK wide (or multiple sites across the UK), and the rest distributed between Bradford, Bristol, Coventry, Northern Ireland, Leeds, Middlesborough, North of England, Salford, Scotland and South East England.

Effectiveness and acceptability of interventions

Effectiveness of interventions was explored by 12 studies [31,36–38,43–45,48,56–58,60]. The studies that assessed effectiveness completed interventions using telephone reminders, GP incentives, culturally tailored programmes for depression, schizophrenia, diabetes, asthma and dental health, educational videos about heart disease and viral hepatitis screening, and a cultural consultation service. Positive results were reported by all studies, with only two studies reporting a neutral effect in some of their measures [44,45]. Table 2 provides further details of interventions and their effectiveness.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Effectiveness of included interventions.

https://doi.org/10.1371/journal.pone.0340079.t002

Acceptability of interventions was explored by five studies [33,38,40,49,58]. The HEAL-D intervention in person [33] and the adapted online version [57] were both well attended and acceptable to Black Caribbean, African and Black British participants. Moore and colleagues [58] found that participants attending their cooking intervention also had high levels of satisfaction. A written and posted information leaflet for dementia awareness in Black African and Caribbean populations was found to have high levels of acceptability (93.6% n = 47) [49]. A culturally adapted behavioural activation programme for depression was found to be acceptable to Muslim participants [40]. Finally, the “Positive Health Programme” was also reported as acceptable to participants in a feasibility study [38].

PROGRESS-Plus framework

Analysis using the PROGRESS-Plus framework focused on study eligibility criteria (inclusion and exclusion), baseline characteristics and subgroup analyses (S3 Appendix). For eligibility criteria, under Place of residence one study excluded care home residents [49], another excluded those living in temporary accommodation [47]. In relation to Race/ethnicity/culture two studies excluded people that were not 1^st generation South Asian immigrants [36,44], another excluded people that were not 1^st generation immigrants from the Democratic Republic of Congo [47] and two studies excluded those not self-identifying as African, Caribbean or Black British [31,49]. Age exclusions were utilised by several studies, with seven studies excluding people below 18 years [31,33,36,38,45,51,61], and two studies excluding people below 16 years [42,62], one study excluded those below 25 years [35] and another study those below 50 years old [49]. Many Disability exclusion factors were identified, two studies excluded those with severe physical or learning disabilities [38,43], three studies excluded those with severe mental health problems (including depression and psychosis) [33,38,53]. Four studies in community mental health excluded people that were actively suicidal or deemed a high risk to themselves or others [31,38,43,45]. One study excluded those with mobility problems preventing attendance of the programme and “complex clinical needs” [33]. Another excluded those with chronic conditions that would limit engagement in the diet and activity behaviour modification intervention [58]. Finally, two studies excluded people with cognitive impairments, dementia diagnosis or those lacking capacity to consent [31,49]. The Plus aspect of the framework identified one study that excluded people that were pregnant and those participating in another research study [33] and four studies excluded patients due to language proficiency [31,44,58,61].

In relation to baseline characteristics, Place of residence was reported as “rented, owned or shared” by one study [45]. Race/ethnicity/culture was well reported in baseline characteristics with groups as follows, Roma and Traveller communities [34], Black African, Caribbean or Black British [28,31,33,49,54,57,58,60], Asian Caribbean [49], “Asian” [28], South Asian [44,50,51,60], Eastern European [60], White British [51], “Other” [58,60], Arab, Mixed Arab and Turkish [51], Vietnamese, Chinese, Albanian, Nigerian, Afghan, Yemeni [39]. Generation in the UK (1st, 2nd, 3rd, 4th) was reported by two studies [38,45]. Followed by Occupation, where four studies reported employment status [31,33,38,49], three studies reported occupations from a list or classification [34,38,50], one study reported if participants were sick, disabled or in receipt of welfare benefits [58].

Three studies included subgroup analyses of their participants [28,45,58]. One study using peer outreach workers to increase organ donation in ethnically underrepresented communities had a higher sign-up rate of Asian people (11%) compared to Black people (2%) [28]. A study delivering an asthma intervention to South Asians [45] found that ethnicity moderated the intervention effect (P = 0.046), “Other South Asians” (Indian, Pakistani or Sri Lankan) had the best outcome (hazard ratio of 1.55), followed by the Bangladeshi group (hazard ratio of 1.11) and the poorest outcomes were seen in the British Bangladeshi group (hazard ratio of 0.72). They found that the role of immigrant generation in the UK also moderated the intervention effect, with first generation immigrants in the intervention group having the best outcomes (hazard ratio of 5.55), followed by first generation immigrants in the control group (hazard ratio of 4.56) and then higher generation immigrants in the intervention group (hazard ratio of 3.12). Finally, they explored the role of English fluency and found weak evidence that this moderated the intervention effect (P = 0.083). One other study looked at a healthy eating and cooking intervention for African Caribbean groups [58] and found those belonging to African Caribbean ethnic groups had a better response to the intervention than those from other groups. They did not find a significant difference in intervention response when comparing educational levels across the whole group and within ethnic groups [58].

Thematic analysis

This review analysed intervention studies using the five service domains in the patient-centered access framework [14], (S4 Appendix). There was a thorough spread across the five domains of the patient-centred access framework in the complex interventions that were discussed in the articles. Approachability and acceptability had the most supporting sources (30 and 27 respectively). With affordability having the least (16 supporting sources), followed by appropriateness (20 supporting sources) and availability and accommodation (22 supporting sources). The first step of accessing and engaging with a health service is knowing that a particular service exists (approachability) [14]. Eight themes aiming to increase awareness of health services were identified were supported by 30 sources (outreach, education about health services, sign posting to other services, ambassadors/community champions, collaboration with community links, advertisements, screening through health services, communication via phone). The most common activity to increase approachability was outreach to community spaces/organisations (n = 12) [28,32,33,41,42,44,47,51,53–55,58], places of worship (n = 4) [32,33,44,54], an Asian bridal show [28], dementia roadshows [50], community activities such as yoga [28] and support groups [53]. Advertising in Asian supermarkets [32], community centres [33], on the radio [29] and on social media was also used (n = 5) [30–32,51,63]. Community ambassadors/champions were developed by nine studies to form stronger community links and raise awareness [32,33,42,47,52–55,61]. Two studies employed telephone outreach workers for NHS health checks [27,48], while another study utilised outreach workers with Gypsy, Roma, and Traveller communities to improve vaccination uptake [34]. Other studies (n = 9) utilised existing health service structures to identify and contact patients that through General Practitioner/primary care contacts, or those who presented to urgent care or Accident & Emergency departments [27,33,44–46,48–50,60].

The second domain is acceptability [14], which refers to cultural and social factors needed for people to accept services. Thirteen themes were identified supported by 27 sources (cultural adaptations, building trust and relationships, gender split sessions, culturally matched staff, using religious spaces, familiar staff, incorporating social support, activities with minimal social barriers, engagement/information film, content verified by native speakers, consideration of literacy levels, adaptation of language, incorporating cultural and religious values). Religious spaces were utilised to increase cultural and social acceptability of interventions [32,38,41]. Cultural adaptations of programmes were used in nine studies [30,31,38,40,41,45,49,57,58], for example culturally adapted dietary advice to support diabetes, incorporating Caribbean examples instead of Western ones, to enhance relevance and participant adherence [33,41,57]. Another eight studies discussed building trust, therapeutic alliance and relationships [30–32,41,47,50,54,55]. Gender split interventions was considered by three studies [32,54,55], however seven studies targeted women specifically [35–39,42,48]. Matching staff and service users by ethnicity, cultural backgrounds or religious views was extensively discussed [28,29,35,37,40,48,50,51,56]. Cultural and ethnic matching was investigated by two studies, revealing differing views among service users, leading to the provision of a choice in the studies [40,55]. One study with Jewish communities [55] considered language adopting terms such as “well-being” and “self-improvement” rather than anxiety or depression which are less accepted by these communities. Another study [32] employed positive language, referring to participants as “learners” to encourage engagement, rather than emphasising low literacy levels. Studies also discussed aligning interventions with existing community groups, considering religious and faith barriers such as dates of religious events [31,32,55].

Availability and accommodation, the third domain in the framework [14] was identified with six themes supported by 22 sources (flexible appointments, local spaces utilised, childcare facilities provided, transport provided, intervention delivered online and home visits – including delivery at an immigration centre). This domain pertains to the physical accessibility of health services, including geographic location, timing, appointment duration, and flexibility concerning work or childcare. Local community spaces such as places of worship, community centres and local parks were used in multiple studies [32,33,35,36,41,44,47,51,54]. Three studies discussed having flexible appointments to meet patient preferences such as work and family commitments [32–34]. Home visits were completed by three studies: one looked at dental care of Chinese undocumented migrant mothers [37], another exploring postnatal depression [38] and the final one provided a specialist maternity service to migrant mothers in an immigration accommodation centre [39]. Three studies delivered interventions in a community centre with childcare services [38,42,43]. A diabetes intervention was delivered online to African Caribbean populations [57,59] and another study designed a text messaging intervention of diet and activity advice [61].

The fourth domain of the framework is affordability [14], which relates to people’s ability to spend money, resources and time accessing health services. This domain identified 10 themes supported by 16 sources (flexible appointments, local community spaces, childcare facilities, local community spaces used, intervention in initial accommodation centre, intervention delivered online, home visits). Direct overhead costs such as transportation was discussed in four studies [32,38,39,43]. Travel costs were eliminated by utilising home visits in two studies [37,38]. Location of interventions was rarely discussed with regards to affordability for participants however, many studies utilised community spaces which may have indirectly supported affordability. There was minimal discussion of indirect participant costs. Three studies discussed how childcare was provided in the same community centre as the intervention [38,42,43]. Two studies reported the interventions were free [32,35] or free for the first six months [32]. One intervention was delivered online which was categorised in affordability [57,59].

Appropriateness, the fifth domain [14], is the fit between services and client needs, such as timeliness and care given to assess health problems and determine correct treatment. This domain identified 10 themes supported by 20 sources (cultural training of staff, multilingual facilitators, consideration of religious beliefs, ethnically matched staff, personalised care, self-referral, joining up of services, intervention provided after health event and use of interpreters). Staff cultural training was discussed by five studies [30–32,34,56] which included for example appropriate body language and culturally appropriate dress [32]. Two studies directly discussed the use of interpreters [39,51]. Six studies discussed the use of multilingual facilitators [27,32,36,38,42,50]. Four studies addressed low literacy levels by delivering interventions through pictures and/or verbal means [32,36,49,58]. Two studies delivered tailored or personalised care [37,40].

The thematic analysis identified three themes that did not fit into the patient-centred access framework, supported by four sources. These included public engagement, longevity of programmes and community involvement in planning healthcare interventions [31–34,47]. Dyson et al. [34] discussed an intervention with multiple components which supported each other and therefore success relied on all components being incorporated. Another study discussed the longevity of programmes, since short term interventions do not support ethnically underrepresented communities long term [32]. This is especially relevant considering many interventions targeted long-term conditions such as diabetes. Public engagement and community involvement in planning and delivering the intervention was discussed by several studies due to its importance in developing relationships and ensuring relevance to the target population [30–32,34,41,47].

Main findings

Healthcare access and engagement for ethnically underrepresented communities is a complex and multifactorial issue requiring complex interventions. This scoping review synthesised the published literature and mapped intervention components to the five domains of the patient-centred access framework. Overall, 35 studies were identified which were predominately based in primary care/mental health services with interventions more frequently addressing approachability and acceptability from the patient-centred access framework with affordability and appropriateness less frequently addressed. The evidence was based on inclusive samples with equity-based factors limited in eligibility criteria.

Comparison to other research

Interestingly, a scoping review of international studies exploring access and engagement [12] found that the majority of the included interventions targeted approachability, availability and affordability, in contrast to this review, which found that approachability and acceptability were the domains supported by most sources. This perhaps may relate to the fact that USA healthcare is not free at point of access in comparison to the NHS, altering the needs of non-UK populations. International studies discuss designated staff for outreach-based roles, including care co-ordinators and integrating teams, such as linking primary and secondary care for patients receiving care from multiple services [12]. Another scoping review [64] explored the role of community care navigators in Canada and whilst it found supporting evidence, only USA literature was identified. The role of care coordinators has not featured in UK literature in this review and may be a worthwhile focus of future research. This could be especially pertinent as early onset multi-morbidity has been found to be more prevalent in ethnically underrepresented communities in the UK compared to White British groups [65]. Many of the international studies [12] had larger sample sizes than UK studies included in this review. For example, one study explored creating a managed care plan for USA uninsured patients and had a sample size of 23,143 people [66]. This perhaps identifies that UK literature is in its infancy in comparison to USA literature of complex interventions to support health equity.

Systematic literature reviews in this area are dominated by studies completed in USA, despite being completed by authors in other high-income countries with publicly funded healthcare. A review of interventions for care-seeking behaviour in migrants by Swedish authors identified 16 studies – one Canadian and the rest USA [67]. They identified promise with community-based interventions but concluded that further research using robust study designs is required. Similarly, a review of Indigenous communities in USA and Australia identified lack of effectiveness assessment and called for more rigorous study designs [68]. However, a Canadian review of barriers and strategies for Indigenous communities identified the following strategies: financial support, education and employment, culturally sensitive information and systems and involvement of Indigenous communities and elders in policy processes [69]. A review of cultural adaptations for behaviour change (smoking, diet, physical activity) from the Netherlands identified 17 studies, all USA [70]. Another review of interventions for culturally and linguistically diverse populations included 24 studies (19 USA and one each in Canada, UK, Australia, Switzerland, Mexico) concluded that culturally trained bi-lingual health workers should be considered as part of potential solutions [71]. A systematic review from Finland identified Nordic studies but the review was focused specifically on cultural competence and outcomes were aimed at staff confidence/competence rather than that patient experience or outcomes [72].

In recognition of the lack of high-quality research in this area a Delphi study was completed in Denmark recommending more co-designed interventions, culturally and linguistically appropriate services, and policies targeted towards health literacy, linguistic, cultural and social differences [73]. Other literature (not reviews) includes a qualitative project looking at dementia in Norway which concludes that inclusive policies to promote patient centred approaches in ethnically underrepresented communities are required for appropriate dementia care [74]. A culturally tailored dementia education programme in Denmark also demonstrated success, however they did not assess long term knowledge post intervention [75].

Interpretation

This review has demonstrated that complex interventions have been applied through multiple healthcare services, targeting ethnically underrepresented communities and have a key role to play in health inequalities. However, it is challenging to draw conclusions about the interventions included in this review due to effectiveness only being assessed in 12/35 studies (34%). Whilst the 34% that describe effectiveness mostly describe positive results (with few describing neutral results and none reporting negative results), most studies in this review do not examine how effective their interventions are. This is especially challenging when it is understood that whilst interventions aim to reduce inequalities, they also have the potential to generate them [76]. In a review of intervention generated inequalities in primary care, Gkiouleka and colleagues found a lack of robust evidence to investigate the outcomes of complex interventions [76]. They concluded that there are five factors that interventions in primary care should consider: being connected, intersectional, flexible, inclusive and community centred [76]. Many of these factors are echoed in the components of interventions included in this review supporting the role of complex interventions in helping to improve health equity.

Health inequalities arise from multiple intersecting factors beyond socioeconomic status (e.g., gender and ethnicity) and are shaped by structural elements such as education systems, political frameworks, housing policies and power imbalances [77]. Therefore, interventions aiming to reduce these inequalities require multifaceted components that are designed to target these intricate elements in particular groups, considering the complex interplay of factors (intersectionality). This is reflected in cultural adaptation of programmes, featured in several studies in this review. For example, interventions demonstrated positive effectiveness and included cultural adaptations for 1^st generation Bangladeshi women [36], South Asians with postnatal depression [38], South Asians with asthma [45], Black African, Caribbean or Black British people with diabetes [33,57,58] and ethnic minority women broadly [48]. Acceptability of cultural adaptations is also supported by studies in this review, further supporting the role of adapting services or programmes [40,57]. One significant challenge of “cultural adaptations” of interventions is the often limited detail in reporting of how these adaptations are completed, lacking information about how the adaptations make the intervention more “culturally acceptable”.

This limits potential learning from studies, reproducibility and application into future intervention development.

It also is important to consider the wider political and power structures that contribute or drive health inequalities and the impact that this has on communities. These wider power structures give a background to the importance of using community centred approaches to build trust and relationships. Trust, therapeutic alliance and relationship building were directly discussed by five articles in this review [30–32,47,55]. However, utilising community links, community ambassadors or outreach was used by 14 studies [28,32,33,41,42,44,47,51–55,58,61] and demonstrates the importance that many interventions place on working with target communities. Edge and colleagues were the only study describing relationship building that evaluated their intervention effectiveness [31]. They demonstrated a positive effect of their culturally adapted family intervention for Black African Caribbean people with schizophrenia [31]. Participatory methods were also utilised by several studies in this review, perhaps a methodology chosen to help to build trust and devolve power through the study process by involving target populations [33,34,47]. Whilst the importance of trust and relationship building was stressed [32], the matter of how to build trust was not reported on in detail. This leaves the question: how do we build trust and therapeutic alliance with ethnically underrepresented communities? And does this need to be approached differently for different ethnically underrepresented communities? Trust building with communities may also be challenging to research as it may involve activities that are difficult to document and evaluate. In their review, Cu et al. [78] also found that numerous studies lacked details on trust-building and engaging service users. Building trust is directly linked to the longevity of interventions. This is especially challenging with long term conditions such as diabetes which will involve lifelong behaviour and lifestyle changes. However, few projects evaluated outcomes beyond the short-term. It is therefore uncertain whether the interventions effectively ensure the trust built is preserved in the longer term. This challenge was highlighted by Such and colleagues who identified several short-term physical activity interventions but also highlighted how the three interventions they used as case studies also all aimed to build capacity to support long term delivery of interventions in the community [32].

Since only 12/35 studies assessed effectiveness, the potential for evidence-based adaptation is limited. Therefore, participation of target populations in intervention designs or adaptations becomes even more important to ensure that interventions will meet the needs of the communities they aim to serve. This highlights the further importance of participatory methodologies such as co-design to involve service users within intervention design and adaptation. Three studies included in this review utilised participatory methodologies [32,48,55], however 14 studies utilised community links, ambassadors or outreach [28,32,33,41,42,44,47,51–55,58,61]; highlighting the importance of participation within intervention (and research study) design and development.

It is well documented that structural systems of power contribute towards health inequalities [79,80]. Racism is pervasive and embedded within structures of society [81]. Structural racism is normalised and is embedded within societal norms, laws, cultures and institutions [80,82]. The Institute of Health Equity [79] highlights racism as having three key contributions towards poorer health over the life course. Firstly, racism directly damages physical and mental health. Secondly, racism may be linked to socioeconomic disadvantage and negative effects of social determinants of health. The third way, most relevant to this review, racism damages health through the operation of systems including health care system. Given this damaging effect that healthcare systems themselves can have on health inequalities it is a significant gap that none of the literature identified in this review of ethnically underrepresented communities mentions or discusses racism. The absence of discussion about racism across the included studies is especially striking given that many were authored by individuals working within institutions that both shape and sustain health inequalities [82]. Power often exists invisibly to those who hold it but is felt acutely by those who lack it; future research must make power visible and critically examine how it influences both research and service design itself.

The patient-centred access framework [14] provided an excellent framework for this scoping review. Selected for its comprehensive analysis of access and engagement in health services, this framework is commonly employed in studies focusing on underrepresented groups’ access [12,83,84]. However, some factors are not included in this framework and were not identified in the inductive thematic analysis that was independent of the framework. This includes the issue of power dynamics, an important factor when considering research and engagement with underrepresented communities and those with issues trusting research and/or healthcare. Interventions may more effectively address power imbalances by collaborating with communities through engagement activities and participatory methods, which may require effort to support engagement [2]. An Australian group discuss in detail the steps that they took to optimise engagement of ethnically underrepresented communities in the co-design process of two of their cancer trials which included inviting seldom heard voices and providing adequate resources [85]. Considering power dynamics is especially pertinent when considering the structural and systemic causes of health inequalities. Whilst the studies utilising participatory methodology will have considered the role of power within their research design, other studies did not discuss this. For example, Flanagan and colleagues provided an incentive to doctors in primary care to increase recruitment of “migrant” patients to viral hepatitis screening [60]. Whilst this study did see an increased uptake in participating GP practices this intervention fails to address the complex factors surrounding health inequalities such as barriers to accessing healthcare, discriminatory practices, or policies that marginalise specific groups [77]. More holistic approaches should be considered such as the complex intervention (HEAL-D) developed through extensive community engagement and participatory methods for Black African, Caribbean or Black British people with Diabetes. This intervention included components across all five of the patient centred access framework domains [33]; was shown to be highly acceptable [33], and to significantly improve participant confidence in controlling their diabetes and managing their diet in the adapted online version.

Understanding which ethnically underrepresented communities were targeted by these studies helps to appreciate how generalisable the results may be to other groups and communities. Overall, the studies employed wide eligibility criteria for participants. Of note was the specific nature of eligibility criteria which is seen less frequently in the wider evidence. For example, ethnicity determined by place of birth (first generation Bangladeshi) or heritage (South Asian). This is worth considering alongside the differences in intervention effectiveness between ethnic groups or generation of immigration in two studies in this review [45,58]. Moore and colleagues reported a better response in African Caribbean participants, which is logical considering that the cooking intervention utilised specific African Caribbean recipes so may have been less relevant to those from other cultural backgrounds [58]. However, it is curious that Griffiths and colleagues’ report the best results in 1^st generation South Asian immigrants in the intervention group, followed by the same population in the control group and a poorer response in higher generation immigrants in the intervention group [45]. Unfortunately, the study does not explore why the authors think this response was seen and therefore future studies may benefit from further exploring intervention responses between generations of migrants.

Culturally tailoring or adapting services has shown promise when applied to specific ethnically underrepresented communities and could be applied for example in service evaluation projects.. Whilst this review has highlighted key strategies for supporting ethnically underrepresented communities, only one third of studies have assessed the effectiveness of their interventions, limiting the impact of any conclusions drawn from these studies.

Policy recommendations

This review has highlighted several areas for policymakers to consider in improving access to and engagement with health services for ethnically underrepresented communities. Central to this is the need to recognise and address power imbalances, as policies that overlook structural inequities risk reinforcing the very systems of exclusion they aim to dismantle. Making power visible – by engaging underrepresented communities in decision making processes, supporting the devolution of power, and applying structural analysis – can help transform the conditions that perpetuate unequal access. Aligned with this, trust emerged as a key theme across several studies in this review [33,34,47], underscoring the importance of working collaboratively with underrepresented communities through establishing community links. Building trust and tailoring policies to the cultural context of underrepresented communities is essential as culturally adapted approaches have been highlighted in this review as a promising way of supporting ethnically underrepresented communities. As this review focused on literature and interventions specifically within the UK, policy recommendations are also focused on the UK setting. However, there may be elements that may be relevant to other high-income countries with public health systems. For example, these recommendations align with existing work within Canada and Australia, partnering with Indigenous communities to share power in developing research and policy priorities [86,87].

Future research

Future complex interventions in ethnically underrepresented communities should endeavour to assess the effectiveness of their work to provide conclusive results of strategies with ethnically underrepresented communities. Most of the studies in this scoping review were in mental health or primary care and therefore studies in other health services are needed. This is especially important for health services that may have specific cultural barriers, such as physiotherapy which requires enhanced engagement for success. Addressing the intricate challenges in healthcare access may necessitate multifaceted interventions covering all five domains of the patient-centred access framework. Particularly, future efforts should focus on enhancing affordability and appropriateness, identified as the least addressed components in this scoping review. Further, identified studies explored service-based interventions; it may be important for future studies to consider patients’ abilities to: perceive, seek, reach, pay and engage [14].

UK based interventions may also benefit from methods used in non-UK studies, such as care-coordinators and streamlining services. Future clinical research may support this approach due to the higher prevalence of early multimorbidity in ethnically underrepresented communities, and therefore the likelihood of those communities’ increased needs for multiple services.

Future research could make more efforts to increase visibility of power. This could be achieved by using positionality/reflexivity statements in outputs, adopting anti-racism approaches/interventions [88] and incorporating frameworks for structural analysis such as intersectionality [89] or critical race theory. Centring underrepresented voices through engaging with and compensating underrepresented communities for their contributions are other possible strategies. Other strategies include using participatory methodologies and developing opportunities for underrepresented communities to take positions of power within projects, for example as co-researchers [90]. Another approach is providing contextualised data by reporting on language, ethnicity and other intersecting factors. Ensuring that studies explore experiences and not just lack of engagement will provide further context to help understand the complexity and nuance of the challenges faced by underserved groups. Providing fair compensation for community contribution and redistributing power through providing opportunities for communities and underrepresented groups to take positions of power within projects.

Limitations

Whilst this scoping review attempted to mitigate limitations by providing a clear and transparent audit process of the search strategy and analysis, some limitations remain. Whilst six electronic databases, reference lists and citations were searched, grey literature sources were not searched due to resources. The six databases chosen for this search were selected due to their strong coverage of interdisciplinary areas and to be inclusive of psychology literature which is a known area with relevant research to this topic. Following searches, reference lists were reviewed alongside citation tracking for further identification of relevant literature. However, it is acknowledged as a limitation that other relevant evidence indexed in other non-overlapping databases such as PubMed may not have been included. Searches were limited to the years 2014–2024. This is in line with the timings of the previous NHS long-term plan, “NHS Five year Forward View” [91], and with UK worsening inequalities in the last 10 years [92]. The time limit also aligns with UK population distribution changes and an increase in ethnically underrepresented communities between the 2011 and 2021 UK census [93]. However, it may have led to an underestimation of the relevant evidence.

This scoping review focused on ethnically underrepresented communities. However, it is recognised that factors such as socioeconomic status or religious beliefs, can also impact access and engagement directly or in intersection with ethnicity. Ethnic origin was selected for its common usage in healthcare access literature [8,10] and enhanced accessibility in UK electronic health records compared to other factors [93]. This review used the term “ethnically underrepresented communities” to acknowledge the structural inequalities in representation and access, with the aim to be a more inclusive and context-sensitive term than “ethnic minority”, which may imply inferiority and overlook systematic exclusion. While the growing use of the term “global majority” as a powerful reframing is recognised, it is not yet widely adopted in academic literature and “ethnically underrepresented communities” better reflects both the focus of this study and the realities of marginalisation within specific institutional and geographical contexts.

Only 10% of the themes were independently charted and cross checked by two researchers in the team, which can lead to subjectivity [78]. This is particularly relevant for components, such as the use of multilingual facilitators which were classified as appropriateness but arguably could be classified as availability or acceptability. The PROGRESS-Plus framework, endorsed by Cochrane whilst providing a lens to analyse the inclusivity of research studies, has also received some criticism for disregarding complexity and failing to incorporate intersectionality [94].