Given this, the objective of this review is to synthesise qualitative evidence on the experiences of receiving and providing care and treatment for mental health conditions in non-specialist settings in LMICs, and the factors that influence the provision and uptake of such services.

Study selection and analysis

This review is a systematic review of reviews. The review protocol was registered on PROSPERO (CRD42022315291), and the review adheres to Preferred Reporting Items for Systematic Reviews and Meta-Analyses reporting guidelines.

Search strategy and eligibility criteria

This review included reviews that synthesised qualitative evidence on the experiences of receiving or providing care and treatment for mental health conditions in non-specialist health settings in LMICs, and factors influencing the uptake/provision of such services. The search strategy was structured around terms reflecting: population (service users/caregivers/providers) AND intervention (care for mental health conditions in non-specialist settings) AND context (LMICs) AND outcome (service uptake, experience/provision of care; barriers/facilitators to care uptake/provision) AND qualitative research AND systematic reviews. There were no restrictions on publication language or date. Table 1 outlines the eligibility criteria for these terms, see online supplemental appendix 1 for the full search strategy.

Table 1

Review eligibility criteria

The searches were conducted in March 2022 in the following databases: PubMed/MEDLINE, EMBASE, PsycINFO, CINAHL, Scopus, African Index Medicus, Global Index Medicus (Index Medicus for the Eastern Mediterranean Region, Index Medicus for the South-East Asian Region, Latin American and Caribbean Health Sciences Literature and Western Pacific Region Index Medicus). These database searches were supplemented by sourcing potentially relevant reviews for inclusion through screening repositories of systematic reviews protocols (PROSPERO, Open Science Framework and Cochrane), and contacting authors of potentially relevant protocols to hear if their work had been published and could be screened for relevance for inclusion this review of reviews.

Study selection

Search results were initially screened based on their title and abstracts (English abstracts were available for screening all articles), followed by full text screening of the articles deemed potentially relevant for inclusion. For articles published in languages in other than English, Google Translate (with added input by a native language speaker) was used to translate the full texts for screening. Two authors (PCG and AM) independently screened the first 20% of the records at both stages. Discrepancies were resolved through discussion, with a third author (NC) serving as arbitrator if needed. Once consistency in screening was achieved, the remaining records were shared for screening between the two authors.

Data extraction and synthesis

Data from eligible articles were extracted into a predefined template capturing review design, methods, setting, population, intervention details and relevant outcomes (service uptake, experience/provision of care, barriers/facilitators to care uptake/provision). For outcomes, relevant data for extraction and subsequent analysis were review findings labelled as ‘results’ or ‘findings’, as reported in the reviews’ results, discussion and abstract sections.

To minimise errors in data extraction, the data extraction template was piloted by two authors (PCG and AM) using an example review prior to the main data extraction, after which the extraction template headings were refined for clarity. The data extraction process was shared between two authors (PCG and AM). Each author extracted details from half of the included studies, after which the other author reviewed the extracted data, followed by a discussion of the process (clarification and consolidation of potential discrepant views) to reach final agreement on the extracted data.

The evidence synthesis drew on both deductive and inductive approaches. A framework analysis approach11 was used as an initial guide for the coding structure, based on domains in the Grading of Recommendations Assessment, Development and Evaluation Evidence to Decision framework12 (eg, considerations of health equity, equality and non-discrimination related to service receipt or provision, feasibility of service uptake/delivery, sociocultural acceptability of interventions). This initial deductively driven grouping was refined further through an inductive approach drawing on thematic synthesis principles.13 This involved building on the initial coding structure through examining the data groupings, and developing these further through pruning and grouping in an axial coding process to develop analytical themes reflecting inductive, thematically coherent concepts in the data.

Quality appraisal

We assessed the methodological quality of the included reviews using the Checklist for Systematic Reviews by the Joanna Briggs Institute (JBI).14 This critical appraisal checklist is specifically developed for systematic reviews of reviews and is suitable for assessing qualitative systematic reviews. The critical appraisal of the included reviews was conducted independently by two reviewers (PCG and AM); discrepancies were resolved through the same approach as was applied during the study selection process to establish final agreed ratings.

Findings

The search produced 3264 records after removing duplicates, of which 42 were considered for full-text screening. Nine studies met review inclusion criteria; figure 1 illustrates this study selection process.

Figure 1

PRISMA flow diagram of study selection process.

Table 2 details the design and methods of the included reviews, and table 3 provides an overview of their focus in terms of key domains of interest (population (service users and healthcare providers), healthcare intervention, mental health condition). In brief, most were mixed-methods reviews (n=7)15–21 whereas one considered qualitative, quantitative and mixed-methods evidence separately22 and one was focused on qualitative evidence only.23 In terms of setting in focus, four reviews focused on LMICs generally.17 19–21 Two articles had a country-specific focus,15 16 and the remaining publications focused on a specified range of countries (eg, sub-Saharan Africa).18 In terms of mental health condition, most reviews focused on mental health conditions generally (n=7)15 16 18 19 21–23 and two reviews focused specifically on schizophrenia17 and substance use.20

Table 2

Included reviews—design and methods

Table 3

Included reviews—key domains of interest

In terms of methodological quality, the included reviews were scoring moderately to highly on the JBI quality appraisal tool; see table 4 for overview of scores. The quality domains in which most reviews lacked was whether the likelihood of publication bias was assessed, whether the critical appraisal of articles included in the review was conducted by two or more reviewers independently, and whether there were methods used to minimise errors in data extraction. All included reviews met the criteria of a clear and explicitly stated review question, appropriate inclusion criteria, appropriate quality appraisal criteria and appropriate directives for new research.

Table 4

Scores on the quality appraisal checklist for systematic reviews by the Joanna Briggs Institute

No articles were excluded based on their quality rating, but the quality assessment was considered in the results synthesis. When synthesising the findings, insights from the articles with the weakest methodological quality ratings15 20 were not highlighted as singular source for any conclusions, but rather were included as examples alongside evidence sourced from other articles of more robust methodological quality.

The included reviews reported on a range of experiences and influencing factors related to the provision and uptake of mental healthcare by non-specialist health workers in LMICs. These findings are synthesised in the following six themes: (1) health worker competency, (2) availability of resources, (3) recipient-related and provider-related characteristics, (4) service accessibility, (5) sociocultural acceptability and (6) vulnerable groups for whom barrier to care were potentially exacerbated.

Health worker competency

The competency and work experience of healthcare workers was highlighted as one factor influencing mental health intervention delivery. Self-perceived competency (eg, ability to deliver intervention, communicate core concepts, counselling competency, remembering guidelines, developing treatment plans) is one part of the healthcare worker competency that comes into play.19 Competency from an external perspective was also a factor, including scepticism or negativity towards task-sharing workforce by other healthcare workers, non-specialist mental health personnel not being regarded as an important part of the healthcare system, and health system managers unwilling to prioritise counselling due to lack of understanding.19 Challenges related to competency were suggested as mitigated through training, refreshers, supervision and networking with others in same role.19

Availability of resources

The availability of resources was a further key influencing factor for the provision and uptake of mental healthcare. The impact was present on several levels. On a structural level, the lack of investment for mental health was discussed as a challenge to planning and delivering mental health services at primary care level.16 On a health-systems level, a lack of funding also impacted infrastructure, such as the availability of transport for home visits, or suitable locations for service delivery.19 Additionally, a lack of funding could result in barriers to providing training and supervision, which is particularly important for task-sharing interventions involving non-specialist staff, and the sustainability of these efforts.22 23 Limited resources also in terms of specialist staff can be a challenge, if interventions with non-specialist health workers also require specialist staff to prescribe medications and supervise treatment.22 It was suggested health worker resource availability could be improved by minimising the burden of their involvement with further interventions and programmes, for example, through integrating contributions to non-specialist interventions into existing commitments (eg, ongoing training), or assigning different staff different duties.19 Limited availability of the non-specialist healthcare work force was also noted, with possible barriers to intervention feasibility including shortages of counsellors, scarcity of people for task-sharing roles, availability of suitable people (eg, literacy requirements), high cadre turnover and competing responsibilities (eg, family).19 At the caregiver level, intervention feasibility and sustainability were reported to be influenced by caregivers’ ability to join due to employment, family members’ inability to join home-based programmes or family workshops, inconvenience for family members to attend group sessions due to lack of interest or other responsibilities (eg, childcare), and unsuitable session timing.17 Availability of personal financial resources was another participant-level factor, with costs (affordability of medication and treatment) a potential barrier to access/service engagement.20

In terms of resources, there were also some indications of task-sharing interventions reducing costs: two examples using trained and supervised lay health workers (non-medical personnel instead of health workers) reduced the implementation cost of psychoeducation programmes, which made their implementation more feasible.17 21

Recipient-related and provider-related characteristics

Recipient-related characteristics could also affect the engagement with treatment for mental health conditions. For example, participants’ education and literacy levels can affect how much the participants are able to engage with the intervention if it involves writing-based tasks, or information and instructions inappropriate for the literacy levels of the care receivers or their caregivers.17 Other factors related to the care providers, and included the perceived safety risks for service providers doing home-visits (especially regarding the schizophrenia intervention),17 lack of private office/space to carry out intervention, and restrictions on abilities to prescribe medication.19 Additionally, reaching the eligible participants or getting the buy-in of carers to sustain the participation in treatment23 could be another barrier for the service providers.

Service accessibility

Practical considerations of service accessibility were another key factor in intervention uptake and delivery. Travel related barriers (eg, time and cost),19 affordability of travel to receive care, reaching/retaining participants who are very mobile (eg, moving out of district/family home),17 challenges to make time due to demands of occupation (eg, causal work, daily pay),20 and inconvenient hours of operation15 could all act as barriers both for lay healthcare providers and service recipients. Informational accessibility was identified as another important factor. Limited awareness about existing service facilities could contribute to the lack of accessibility.16 Additionally low ability to impact potential recipients’ interest,20 lack of motivation21 and perceived levels of usefulness of the interventions could result in additional challenges.17 Similarly, negative perceptions of help could be a barrier.18 It was mentioned that these barriers could be overcome through providing mental health services and screening within primary care settings also to reduce stigma and raise awareness.18

Sociocultural acceptability

When considering factors influencing service provision and uptake, the importance of sociocultural acceptability of mental health programmes was clearly expressed. For example, inadequate considerations of the cross-cultural applicability of how services were provided could lead to lack of satisfaction in the treatments implemented.19 22 Also, when interventions had been sociocultural adapted, and when interventions were perceived as appropriate for the culture and target group, participants had a more favourable response to them and their content and medium of delivery received more positive feedback from service users and caregivers.17 19 21 23 Breaches to sociocultural acceptability were evident in reflections on stigma-related concerns and experiences: poor treatment and stigmatisation by health providers and confidentiality issues were a concern for access to the services.17 18 20 21 Beyond cultural sensitivity, considerations of appropriateness regarding age, sex and language were highlighted as important to increased intervention acceptability and accessibility.18 19 21 22

Vulnerable groups

It was reported that the factors influencing service uptake and specifically barriers to care were exacerbated among some groups of potential recipients. The main groups affected included women, people with low levels of education and literacy, or people from low resource settings. Women were reported to face particular access challenges17 20 through their inability to travel away from their own locality, and stronger concerns regarding stigma and shame (noted for substance use particularly). This disparity might also be particularly pronounced for caregiver interventions, where women are less often able to participate due to other caring responsibilities.17 Another vulnerable group was people with low levels of education or literacy17 who might struggle with writing-based tasks and taking instructions compromising the utility of psychoeducation materials. Lastly, people from low-resource settings were considered to be particularly impacted by affordability and cost challenges.20 22 23