2.1 Setting

Nepal exemplifies many low-resource settings that are struggling to expand mental health services due to policy, legislative and logistical barriers [10]. Nepal's current health system structure is reflected through the federal government systems, which consist of the federal, provincial and local levels of administration. At the federal level, the Ministry of Health and Population (MoHP) oversees overall health policy and planning and is responsible for the governance, planning and implementation of mental health services through various departments and units [11]. For a country of 31 million people, Nepal's only national mental hospital with 50-bed capacity is in Kathmandu, the capital city. Although healthcare systems have multiple tiers of structure (at federal, provincial, district and municipality levels), policy development, planning and resource allocations for mental health are carried out in top-down approaches.

2.2 Data Mapping Exercise

The first step in implementing FOCUS-MHS was to map relevant data readily available through online sources for each indicator under its five domains. Two researchers (D.G. and B.S.) charted information from reviews of mental health stigma literature [12] and mental health-related health and social policies [11] into the indicators using an Excel sheet. Next, we searched Google Scholar for further published peer-reviewed or grey literature for missing information or information requiring updates. Relevant government websites and their reports were also scoped to identify information for each indicator. The data mapping exercise was conducted from March 2024 to April 2024.

2.3 Key Informant Interviews (KIIs)

2.4 Data Management and Analysis

The percent endorsement of Red/Amber/Green for each indicator was analysed descriptively using MS Excel. The qualitative data analysis used the framework approach [19], where a priori themes based on topics explored and domains/indicators within the FOCUS-MHS framework were generated in an Excel sheet. Researchers (D.G., B.S. and B.A.) fluent in both Nepali and English transcribed and translated the recorded interviews. Three researchers deductively coded, charted and indexed interview transcripts in the Excel sheet that included the a priori themes in the rows (e.g., their understanding/experiences of mental health-related structural stigma, data gaps in terms of recency/relevancy/accuracy, the feasibility of getting information for the indicator, etc. for each indicator), and respondent ID in the columns. The team logged memos and discussed any issues, possible solutions and critical reflections during the processes.

3.1 Domain 1: Discriminatory Legal Framework and Policy Environment

The indicator ‘exclusion of mental health from universal health coverage’ had the most consistent rating (100% rated as amber). All participants indicated that the inclusion of mental healthcare services within the basic healthcare package and the government's health insurance scheme clearly showed attempts made at inclusion. However, they mentioned substantial work was still necessary for the indicator to turn green due to the lack of scheme coverage, geographically limited implementation of mental health programmes within primary healthcare services and the lack of comprehensive mental health services (e.g., psychosocial counselling, evidence-based therapies) within basic health and insurance packages.

Indicators such as ‘unavailability of mental health act and policies’ and ‘discriminatory language and provisions in mental health policies’ also had a higher percentage of amber rating, indicating some levels of discrimination but also attempts made to address them. For example, almost all participants mentioned the drafting of national mental health strategies and action plans as a huge step forward. Still, without a mental health law that impacts PWLE's services and rights, the indicator has yet to turn green.

3.2 Domain 2: Stigmatising System Infrastructure and Resources

Most participants rated ‘unavailability of trained human resources for mental health’ (75%) and ‘mental health indicators not included in national health information systems’ (83%) as amber, as in these areas, the government had realised potential gaps and initiated mechanisms such as prioritising the mental health training for primary healthcare workers, initiating health administrators training for the implementation of mental health programmes and incorporating additional indicators for better understanding national-level mental health data. Nevertheless, they raised concerns that much was needed to turn the indicator green, such as increasing the number of trained health workers at health facilities and creating specialist positions at district hospitals, ensuring not just clinical team availability but also the therapists/counsellor and nurses' availability to provide holistic psychosocial support and implementing the newly adapted mental health indicators in all tiers of healthcare.

The ‘differential quality of space/infrastructure’ indicator had a higher rating of red (75%). In comparison, some participants (17%) rated it between red and amber, given the unavailability of specified beds or in-patient units within the larger provincial and district hospitals. Similarly, participants highlighted that although there were infrastructural limitations at the primary healthcare facilities, some municipalities and health facilities with motivated health professionals and administrations had created ad hoc confidential spaces within their limited infrastructure for the provision of mental health services (e.g., meeting rooms within ward administrative offices or medication dispensary room), highlighting the lack of space as a stigma issue.

Participants seemed ambivalent towards indicators such as ‘insufficient funding for mental health’ and ‘unavailability of medications’, with ratings split between red and amber. In terms of funding, all participants agreed that the budget was insufficient to implement mental health policies and programmes. Some participants emphasised that mental health program funding mostly came from development partners, leading to drastic changes (increases or decreases) in mental health expenses from 1 year to another, as such funding is short-term and unsustainable.

Ratings for ‘unavailability of medications at health facilities’ were distributed between red and amber, mainly due to limited clarity on whether it was a supply chain problem (if so, are there supply issues also for medications for other health conditions) or is restricted supply chain management caused by issues with mental health-specific policies. While most agreed that including psychotropic medications in the free drug list was a huge step forward, its implementation was lacking as the medicines could only be supplied to health facilities with trained health workers. Even then, the supply of medicines was ad hoc, without proper evaluation of prevalence, catchment population or need (due to problems in recording mental health indicators), resulting in either an over- or undersupply of medications. A highlighted structural stigma issue was a policy on authorising the purchase of psychotropic medications. While most drugs on the free drug list could be authorised for purchase by local government (municipalities), purchasing psychotropic drugs was only permitted by the provincial and federal government bodies, leading to lengthy procurement and supply time.

Participants also seemed unsure about ‘unaffordable services compared to other chronic conditions’, with ratings distributed between red (58%), amber (16%), green (8%) and undecided (18%). Participants found the indicator difficult to rate due to a lack of aggregated representative data on out-of-pocket (OOP) costs for mental health compared with other physical health conditions.

3.3 Domain 3: Stigmatising Knowledge, Attitude and Practices of Individuals Affiliated With the Health Systems (Health Workers and Administrators)

These indicators had the most variable ratings compared to other domains, although most ratings were between red and amber. The ‘culture of not involving PWLEs in decision-making’ had a higher red rating (57% vs. 43% amber). In comparison, ‘negative aggregate attitude/behaviour of health systems staff’ (57% amber, 29% red, 14% green), ‘culture of stigmatising mental health staff’ (58% amber, 14% red, green and undecided each) and ‘less competency of health workers’ (42% amber, 29% for both red and undecided) had a higher proportion of amber rating compared to red and green. As most indicators in this domain had no aggregate data or information that could be mapped, the ratings were based on personal experiences, which varied between participants.

The ‘health systems staff is not aware/knowledgeable about human rights of PWLEs’ indicator had comparable ratings for red and amber (42%), while 16% were undecided between red and amber. The varying opinions reflect the participants' experiences. For example, a health coordinator thought the indicator was irrelevant because health professionals were not lawyers and did not need to know the human and legal rights of PWLEs. Some participants, however, believed that the indicator was very relevant and necessary and rated it amber as some development partners had conducted WHO Quality Rights training [22]. Some participants rated it red, as the lack of data or information for the mental health indicator was considered to reflect structural stigma, and a lot was needed to turn the indicator green.

3.4 Domain 4: Inequitable and Poor Quality of Care

Like the indicators in Domain 3, these indicators had limited data from published literature, so the information was mainly derived from qualitative studies conducted in specific geographic or healthcare settings. As such, indicators like ‘involuntary/compulsory treatments’, ‘separation of mental health from primary healthcare’ and ‘lack of collaborations’ had higher non-response rates compared to other indicators. Nevertheless, all indicators within this domain except ‘separation of mental health from primary healthcare’ were rated mainly as red, implying higher manifestations of structural stigma in this domain.

For example, 50% of participants rated ‘involuntary/compulsory treatment of PWLEs’ as red. This was corroborated by a mental health specialist rating the indicator red due to unnecessary sedation and overprescription, even by specialist practitioners. In contrast, 29% rated it amber by those who perceived improvement in restraining or compulsory admittance practices over time, occurring only when justified by the nature of the condition. About 21% had no response due to lack of information

Most participants also rated the ‘unavailability of evidence-based mental health services’ red, given the government not prioritising evidence generation (evidenced by no mental health data for use in resource allocation), using a political agenda rather than evidence to inform policies, and the information lacking for most indicators within the framework. Some participants, however, highlighted the adaptation and implementation of WHO's mhGAP program [23] as a noteworthy government step and therefore rated it yellow.

The ‘lack of multi-sectoral and inter-disciplinary team collaboration for mental health services’ had more varied ratings, with 50% red, 36% amber, 7% green and 7% undecided/no response. Those with amber or green ratings highlighted government attempts to bring together different stakeholders for policy and planning. They also highlighted that such collaborations were more common in mental health than other health conditions, making it a comparatively lesser structural issue for mental health. Nevertheless, some participants considered collaboration and communication between different tiers of government—federal, provincial and local—challenging for mental health and therefore rated this red. Medication supply problems were given as an example of lacking collaboration between different tiers of government bodies and institutions, such as the health directorate office, health training centre and health logistics management centre.

The ‘lack of clear referral pathway’ indicator, too, had mostly red ratings (57%). Participants highlighted lacking referral mechanisms between primary healthcare facilities and general or specialist hospitals, compared to safe motherhood programmes.

3.5 Domain 5: Negative Experiences of PWLEs

Most information mapped to these indicators came from either FGDs, expert consultations with PWLEs during the framework's development or qualitative studies on barriers to care in Nepal. Most participants (55%) rated ‘low satisfaction of mental health care received’ as red, while the rest rated it as amber. PWLE participants especially resonated with the information mapped on the indicator and shared their own experiences of dissatisfaction with the care received. This ranged from the unavailability of medications resulting in having to purchase medicines from private pharmacies, health workers not listening and communicating like a ‘robot’, lack of freedom during hospitalisations and being treated as prisoners, and the troublesome/annoying process of using services like insurance schemes.

Most participants (55%) rated ‘negative interaction with health system staff’ as amber, indicating some progress in this area. However, several initiatives were needed to turn the indicator green, given issues like the use of stigmatising words towards patients, teasing patients, making snide comments about the patient while talking amongst themselves, commenting on them looking normal when trying to access disability services, being ignored or not taken seriously and being judged. The government initiatives to integrate mental health into essential healthcare services, with primary healthcare workers trained to provide free basic mental health treatment, was the reason behind most participants' amber ratings for ‘high OOP expenses for services’ (55%) and ‘lack of ease of access to mental health services’ (64%). However, some PWLEs rated these indicators ‘amber’ due to experiencing difficulties in accessing care, such as accessing services from multiple providers and spending lots of money before receiving a diagnosis or treatment.

The final indicator, ‘PWLEs insufficiently informed about their conditions and treatment’, was rated red by most (73%). PWLEs shared instances where they observed no two-way communication between service providers and patients in a government hospital's psychiatric outpatient ward, where service providers prescribed medications and asked patients to attend follow-up without considering the patient's or their caregiver's opinions.

4.1 Feasibility and Applicability of Using the Structural Stigma Measurement Framework

One way of determining feasibility and applicability is the availability of information or the ability to retrieve the data without undue burden [18]. While indicators in some domains (such as Domain 1 and Domain 2) had readily available information from government reports and publications, most indicators had minimal information that could be identified through the data mapping exercise. The biggest challenges were the unavailability of the information, as it was never recorded, studied or published and gaps in recency, as some included information was from older studies considering very different healthcare systems and did not reflect the current trends [8, 9]. While the KIIs filled the gaps to some extent and provided nuances to the issue of structural stigma, these are reflections of personal experiences, and triangulating the information was complex.

For instance, most amber ratings for indicators (such as excluding mental health from universal health coverage, unavailability of trained human resources and unaffordable services) were due to the government's inclusion of mental health in—BHS, which involves primary healthcare worker training. However, there was no published information on the number of trained health workers, and the key informants were unsure of possible numbers. Nevertheless, the participants mentioned the government's recording system of trained health workers available from health training centres, which, if published annually, could be sourced without undue burden and be easily accessible. Similarly, little to no information was available on funding allocations and mental health service costs, although these were highlighted by most as essential indicators of structural stigma. These information gaps are reflected as a structural barrier in mental healthcare, as they indicate a lack of proactivity in healthcare systems governance to track progress or recognise challenges.

Some indicators, such as ‘less competency of health workers’, ‘involuntary/compulsory treatment of PWLEs’ and ‘negative interactions with health system staff’, received some non-responses due to limited information from the evidence mapping and the participants' lack of information/experience. This was specifically true when the domains and types of informants did not match. For example, policymakers struggled to rate domains 4 and 5 on the quality of care and PWLE experiences. Meanwhile, PWLEs and health workers found it difficult to rate or provide additional information on domains 1 and 2 regarding the legal framework and resource allocations.

Participants were also asked to reflect on using the RAG rating system as a visual analogue scale to rate each indicator's information and level of structural stigma in the healthcare system. Most participants liked the ratings using colour codes of red, amber and green, as this was universally understood and provided a simple visual representation of performance. For example, policymakers thought this could give easy signs of which indicators were performing poorly, and which were doing better, thus helping them with decision-making. This was also useful with PWLEs with literacy barriers, as they could easily pick up the colour code-based ratings. Similar visual analogue scales have been used to measure mental health outcomes among the low-literacy population in Nepal [27]. Nevertheless, participants felt that ratings based on only three criteria were restrictive, reflected in many undecided ratings. Some participants suggested a 5-point RAG rating for added nuance while retaining the visual analogue scale for ease.

Essentially, most participants perceived the indicators and the FOCUS-MHS framework to be acceptable and applicable in Nepal's healthcare systems to measure mental health-related structural stigma. Participants underlined the framework's usefulness for development agencies and PWLE advocacy networks in identifying structural barriers and discrimination and as an advocacy tool for policymakers/programmers. However, some participants mentioned the need for concise indicators, as some were difficult to rate or collect information on. Indicators such as ‘unavailability of evidence-based services’, ‘lack of multi-sectoral collaboration’, and ‘PWLE lack of ease of access to mental health services’ were examples where participants struggled with ratings due to unclear understanding of what the indicators were trying to measure.

Another suggestion regarding the framework's applicability was the need for comparator conditions for some indicators. As the FOCUS-MHS framework measures mental health-related structural stigma and discrimination, participants struggled to rate indicators like ‘unaffordable services’ and ‘differential quality of space/infrastructure’ as information mapped onto the indicators exhibiting structural barriers for mental health services. Nepal's health system, in general, is under-resourced and under-functioning, with poor coordination between government tiers, delayed fund releases, staff shortages and issues with medicine procurement [28]. In such a low system-functioning context, participants argued that the poor functioning of mental healthcare services does not reflect structural discrimination. Therefore, they suggested comparing the indicators with other chronic conditions with similar prevalence rates in the community.

4.2 Implications and Recommendations for the Application of the Framework in Healthcare Settings

The FOCUS-MHS framework and its implementation may be relevant in understanding the structural stigma prevalent in the mental healthcare system in other low-and-middle-income countries (LMICs). However, the findings from this study suggest few recommendations and modifications for implementing the FOCUS-MHS framework step by step to determine the structural stigma against PWLEs in other healthcare systems. A guide and template for adapting and implementing the FOCUS-MHS can be accessed in File S1.

Determining relevant indicators for different health system settings: Although the indicators within the framework were developed through a comprehensive Delphi exercise with diverse stakeholders representing high, low and middle-income country settings, not all indicators may be relevant or applicable to all cultural/health system contexts. Nevertheless, the five domains within the framework are meant to provide a comprehensive and nuanced portrayal of mental health-related structural stigma within the healthcare system and areas that need targeted interventions. These domains also help compare nature and level of structural stigma in different health system settings. Hence, the first step in implementing the framework would be to select the most relevant indicators in each domain.

Data mapping exercise: The data mapping exercise allows researchers to identify information already available from various indicators with information gaps. Information can be collected through multiple sources and methods for better data triangulation.

In Nepal's case, we mapped information from government reports, websites, published directives and peer-reviewed papers. The mapped information was triangulated with the stakeholder interviews, including collecting any missing information. Other sites could also use existing information sources, such as the WHO Mental Health Atlas [29], WHO Quality Rights reports [22], quality of care, audit reports and patient satisfaction of care reports that already exist in some healthcare settings [30]. Measures such as the ‘stigma cultures in healthcare’, if applied, can provide rich information that can be mapped onto domain 5 (negative experiences of PWLEs) [31]. Identifying data sources and methods of data collection is vital for the quality, validity and trustworthiness of information mapped into the framework.

As recommended by the study participants, a comparison of the indicators with other healthcare conditions would help understand whether the structural issues are disproportionate to mental health conditions or whether they reflect a generally low-functioning health system. Hence, the data mapping exercise should also compare the state of the indicators for other health conditions. The sites could compare the indicators to other physical health in general or particular health conditions such as diabetes, cardiovascular diseases or other chronic conditions.

Selection of stakeholders: As experienced while applying the framework in Nepal, different stakeholders better understood different domains. For example, domains 1 and 2 were better understood by policymakers, advocates and development agencies than PWLEs and healthcare workers, while domains 3, 4 and 5 were better understood and comfortably ranked by PWLEs, health administrators and health workers. This could be because many indicators were missing information, so the participants were asked to provide it during their KIIs. Participants had no issues providing ratings where rich information was available for the indicators.

Rating exercise: Rating indicators can be done through online surveys or in-person interviews. The rating may be carried out as a stand-alone aspect or merged with qualitative interviews, as done in this case example, to understand nuances and complexities within the indicators. This is especially important if the information gap identified from the data mapping exercise is significant. As participants in Nepal suggested, visual analogue rating scales support participants, such as policymakers and PWLEs, in better understanding the criteria and grading how the health systems domains perform regarding structural stigma. One suggested modification is using a 5-point visual analogue rating scale, like battery level, instead of a 3-point RAG rating scale (see Figure 3 for the modified visual analogue scale). This gives participants more flexibility to rate the indicators. The criteria for rating based on this visual analogue scale is explained in the recommended guide for the implementation of FOCUS-MHS (see File S1).