Proactive community case-finding to facilitate treatment seeking for mental disorders, Nepal

Abstract Problem Underutilization of mental health services is a major barrier to reducing the burden of disease attributable to mental, neurological and substance-use disorders. Primary care-based screening to detect people with mental disorders misses people not frequently visiting health-care facilities or who lack access to services. Approach In two districts in Nepal, we trained lay community informants to use a tool to detect people with mental, neurological and substance-use disorders during routine community service. The community informant detection tool consists of vignettes, which are sensitive to the context, and pictures that are easy to understand for low literacy populations. Informants referred people they identified using the tool to health-care facilities. Three weeks after detection, people were interviewed by trained research assistants to assess their help-seeking behaviour and whether they received any treatment. Local setting Decentralized mental health services are scarce in Nepal and few people with mental disorders are seeking care. Relevant changes Out of the 509 people identified through the community informant detection tool, two-thirds (67%; 341) accessed health services and 77% (264) of those individuals initiated mental health treatment. People in the rural Pyuthan district (208 out of 268) were more likely to access health care than those living in Chitwan district (133 out of 241). Lessons learnt The introduction of the tool increased the utilization of mental health services in a low-income country with few health resources. The tool seems beneficial in rural settings, where communities are close-knit and community informants are familiar with those in need of mental health services.


Introduction
Globally, underutilization of mental health services is a major barrier to reducing the burden of disease attributable to mental, neurological and substance-use disorders. 1 Service underutilization has been attributable to lack of awareness of service availability; lack of recognition of mental, neurological and substance-use disorders in oneself or one's family; stigma against seeking mental health care; and perceived ineffectiveness of treatments. 2 Routine or indicated primary health-care screening has been proposed to tackle this challenge, but this approach misses people who rarely use primary health-care services. In areas with high poverty levels and/or long travel times to health facilities, large portions of the population access primary care infrequently. Moreover, many low-and middle-income countries lack resources for widespread screening, especially in populations with high illiteracy that require health staff to administer screening tools.
An alternative approach to increase utilization is community case detection, which employs a gate-keeper model where people with regular community engagement are taught to identify and refer people for assessment and treatment in primary health care. However, community case detection has received limited attention for mental health.
To address this challenge, we developed a community informant detection tool, which we piloted in Nepal. 3 The tool facilitates detection of people with depression, alcohol-use disorder, epilepsy and psychosis and helps identified people to seek care. The disorders were selected based on prevalence, burden of disease and responsiveness to evidence-based treatments, and have been confirmed for Nepal through an expert priority-setting study. 4 The tool is developed on the premise that people who are intimately connected within the community, such as community health workers (CHWs), are in a position to identify those in need of care, if they are provided with a tool for identification. The structured tool contains vignettes, which are sensitive to the context, rather than symptom checklists and uses pictures that are easy to understand for low literacy populations. Trained lay community informants (e.g. CHWs or civil society women's groups), use the tool during daily routine activities, where they check the extent to which people match paragraph-long vignettes using a four-point scale. The pictorial vignettes are designed to initiate help-seeking for mental health treatment in primary care settings. The community informants do the vignette matching based on their observation of people as part of their interactions during their regular responsibilities. If the person fits well with the description, they will ask additional questions on need for support or impairment in functioning. The questions are an integral part of the tool with yes/no responses, functioning as a decision flowchart. In the case of a positive reply to either of the two questions, the informant encourages the person (possibly through their family) to seek help in health-care facilities where mental health services are being offered and the person can be evaluated by trained health professionals. No stigmatizing psychiatric labels are used Problem Underutilization of mental health services is a major barrier to reducing the burden of disease attributable to mental, neurological and substance-use disorders. Primary care-based screening to detect people with mental disorders misses people not frequently visiting health-care facilities or who lack access to services. Approach In two districts in Nepal, we trained lay community informants to use a tool to detect people with mental, neurological and substance-use disorders during routine community service. The community informant detection tool consists of vignettes, which are sensitive to the context, and pictures that are easy to understand for low literacy populations. Informants referred people they identified using the tool to health-care facilities. Three weeks after detection, people were interviewed by trained research assistants to assess their help-seeking behaviour and whether they received any treatment. Local setting Decentralized mental health services are scarce in Nepal and few people with mental disorders are seeking care. Relevant changes Out of the 509 people identified through the community informant detection tool, two-thirds (67%; 341) accessed health services and 77% (264) of those individuals initiated mental health treatment. People in the rural Pyuthan district (208 out of 268) were more likely to access health care than those living in Chitwan district (133 out of 241). Lessons learnt The introduction of the tool increased the utilization of mental health services in a low-income country with few health resources. The tool seems beneficial in rural settings, where communities are close-knit and community informants are familiar with those in need of mental health services. and encouragement for help-seeking is targeted to observable behaviours and signs of distress.
Previous studies demonstrated that the tool has an accuracy comparable to primary health-care screening in high-income countries and better than standard screening tools in Nepal (positive predictive value of 0.64 and negative predictive value of 0.93). 3,5 Here we determine whether application of the tool increases help-seeking behaviour among people who would otherwise be unlikely to seek care. We also assessed how many of the referred people pursued primary health-care services and started on treatment.

Local setting
Decentralized mental health services are scarce in Nepal 4 and less than 5% of people with alcohol-use disorder and less than 10% with depression seek treatment (Luitel et al., Transcultural Psychosocial Organization Nepal, unpublished data, 15 March 2017).
The study took place in two Nepalese districts. Chitwan district in southern Nepal has been the implementing site for the Programme for Improving Mental Healthcare (PRIME) since 2011. 6,7 The district is densely populated, is relatively well resourced and at the time of the study had 12 health-care facilities with mental health services. Pyuthan is a more remote and poorer hill district, and was the site for the Mental Health Beyond Facilities (mhBeF) initiative from 2013 to 2015. When the study was conducted, the district had six facilities providing mental health services.
Both mental health programmes were implemented by the nongovernmental Transcultural Psychosocial Organization (TPO) Nepal.

Approach
In 2014, community informants residing in the study areas were selected based on their interest in participating. In addition, the district public health office recommended CHWs. The informants received two-days of training in TPO's offices in both districts. The training consisted of how to use the tool and ethical issues associated with case-finding, confidentiality and how to encourage, but never impose help-seeking. In their routine work, the informants then used the tool to identify people with mental disorders proactively. For the purpose of the study, people identified were given a referral slip to visit a health facility with staff trained in mental health services (following the mhGAP intervention guide). 8 The informants filled in the contact information for the identified person and themselves on referral slips, while information on the location of the appropriate health facility was provided verbally. The visits were free of cost.
All 674 people, identified with the tool between April and May 2014, were scheduled to be visited by a research assistant three weeks after the date of detection. The people, who the informants were able to reach after three weeks and who provided consent to participate in the study, were asked whether they had visited a health-care facility in the past three weeks. The participants who answered "yes" were also asked the following questions: Who or what determined whether you sought help (including referral through the tool as one of the options)? What problem did you seek help for? Was treatment initiated, if so what treatment? Additionally, we asked participants for sociodemographic characteristics. For the participants that accessed health care, we cross-checked their answers with their clinical diagnosis and treatment records and we checked the clinical records for completeness.
We obtained ethical approval for this study from the Nepal Health Research Council.

Relevant changes
Out of the 509 participants, 67% (341) accessed a health-care facility after being referred as a result of the proactive detection approach. We excluded 17 participants that accessed health care, but who did not explicitly mention this was because of the tool (Fig. 1). Among the 341 participants accessing care, 264 (77.0%) received diagnoses and started treatment for mental illness: 34.8% (92) received diagnoses for epilepsy; 31.1% (82) for psychoses; 15.9% (42) for depression; 14.0% (37) for alcohol-use disorder; 2.3% (6) for anxiety; and 1.9% (5) for being bipolar (total exceeds 100% because of comorbidity). In Chitwan, 55.2% of participants (133 out of 241) accessed care, while the percentage in Pyuthan was 77.6 (208 out 268). The mental health services offered by trained primary health-care workers included pharmacological and psychosocial interventions. 7 Participants were mainly referred by female community health volunteers (84.9%) and civil society women's groups (14.5%).
Those who accessed health care versus those who did not had similar age, gender, education and marital status ( Lessons from the field Community case detection for mental disorders in Nepal Mark JD Jordans et al. nearest health-care facility than those not accessing care (P < 0.001; t-test).

Discussion
Our results combined with data from our previous study 3 demonstrate a two-thirds by two-thirds effect of identifying and treating people with mental disorders, when using the tool. First, community informants accurately detected in two-thirds of the cases. 3 Second, two-thirds of those detected initiated help-seeking and went on to access health care. These results indicate that the access gap for mental health care can be reduced by intervening on the demand-side, suggesting that this tool could be useful in other low-and middle-income countries experiencing low-treatment coverage for mental illness (Box 1). 9 Participants from Pyuthan were more likely to access health care than those living in Chitwan, even though they had longer distances to travel to health facilities. This result suggests that the tool is especially beneficial in more rural settings, where communities are close-knit and CHWs and other liaisons are familiar with those in need of mental health services. In rural communities, residents may be more likely to trust and follow the recommendations of CHWs. In addition, Pyuthan had no local access to treatment for mental disorders before the implementation of the mhBeF programme, so the tool may perform better when health services are newly initiated.
This study has limitations, of the people initially identified as potentially having a mental disorder according to the tool, 24% were excluded, mostly because they could not be reached after three attempts by research staff. Also, this study depended on participants' recall of what triggered them to seek care.
This study demonstrates that the structured and context-sensitive detection tool supports community informants in proactive case-finding. The informants' tacit knowledge and awareness of who is suffering in the community helps them to identify people in need of mental health care. This approach is useful in places where mental health care is newly established and heavily stigmatized due to lack of awareness; it also helps to support disadvantaged groups who face more barriers due to social and economic vulnerabilities. 10,11 The Nepalese government has included the tool in national health care packages 12 and the approach has been scaled-up to other districts during the emergency response following the 2015 earthquakes.
To increase coverage of mental health care in low-and middle-income countries, efforts to overcome supply-side and demand-side barriers should occur simultaneously; this includes shifting tasks from mental health professionals to CHWs. 13 Whereas the supply-side requires increased service delivery models, the demand-side requires stimulatingdemand, for example through proactive community case-finding. Inclusion of the community informant detection tool or similar case detection interventions in service delivery models, could address the treatment and

Box 1. Summary of main lessons learnt
• Using a community informant detection tool decreases the access gap to treatment for mental disorders. • The tool seems beneficial in rural settings, where communities are close-knit and community informants are familiar with those in need of mental health services. • The task of case identification can then be shifted to community members outside the health-care system as a way to broaden access to mental health care.