Program overview

This study is a longitudinal study of a mental health intervention. This project was carried out over a 3-year period. It started with a 6-month preparation phase (April to September 2017 – Figure 1) to develop the digital tools, purchase the hardware, design the behavior-change communication (BCC) materials, recruit healthcare workers, train them, as well as to conduct some of the baseline surveys (door-to-door survey to evaluate prevalence [A], and barriers to treatment [B]; quasi-experimental design “before/after” to evaluate KAP among CHWs [C], GPs [D] pre- and post-training and among the general population prior to any awareness activities [E1]). CHWs and GPs were trained by psychiatrists from the Myanmar Mental Health Society, based on the WHO mhGAP intervention guide. The initial training on psychotic disorders, depression and epilepsy was a one-day session for CHWs, and five half-day sessions for GPs. Additional half-day sessions were conducted for both groups on the use of the digital devices, and for CHWs on how to conduct awareness and psychoeducational activities. Digital devices (Android based) provided to health workers were equipped with applications to assist with the identification, diagnosis, management and follow-up of people with mental disorders and epilepsy. This preparation phase was followed by an implementation phase in the field which lasted for 2 years (October 2017 to September 2019 – Figure 1) and included activities implemented by CHWs and GPs to raise awareness, identify, diagnose and manage patients. The application on the CHWs smartphones included interactive screening questionnaires to help them decide which people to refer to GPs for diagnosis (Patient Health questionnaire (PHQ9) for depression, which has not been validated in the context of Myanmar, Psychosis Screening Questionnaire (PSQ) for psychotic disorders, and questionnaire from the Manual on Epilepsy for Medical Officers). GPs, when seeing one of the patients referred to them, could access, via their tablet, the information collected by CHWs; they could also refer to the electronic version of the WHO mhGAP intervention guide for the diagnosis and management of patients and use the e-medical record function to better monitor and follow-up patients. The last phase was an evaluation phase which lasted for 6 months (October 2019 to March 2020 – Figure 1) with an observational design including analysis of the data collected regarding the patient flow between CHWs and GPs [F], treatment prescribed by GPs [G], and treatment adequacy assessed retrospectively by a psychiatrist [H], in addition to various end-line surveys (KAP of CHWs, GPs, and general population [E2]).

This project was conducted in Hlaing Thar Yar, a township located in the Western part of the Yangon region, in Myanmar (Figure 2). This area has a majority of inhabitants in the low socioeconomic stratum, as it has attracted waves of migrants from various rural parts of Myanmar because of its perceived abundance of jobs. There are many illegal slums and shacks built on public land. According to official census data, there were at the time 515,570 inhabitants living in this township, which is made of 9 village tracts and 20 wards.

GPs did not receive any remuneration to participate in the study. CHWs received a token amount approximately equivalent to USD 25 per month.

The project protocol in Appendix 1 provides additional information regarding the various interventions and surveys.

Study population

Participants aged 18 years and over, who had lived in Hlaing Thar Yar for a minimum period of 6 months at the time of the survey, were included in the general population surveys [A] and [E1], while for [B], people had to be identified with a psychotic disorder, depression or epilepsy. People who refused to participate, who were seriously ill or unable to speak were not included. A total of 3,100 people was the minimum sample needed for [A], [B] and [E1] surveys.

In the patient flow analysis [F], suspected cases who were identified by participating CHWs as potentially having a mental disorder or epilepsy and were referred to a GP, and patients who were diagnosed by GPs between October 2017 and September 2019 were included. In the analyses about treatments prescribed [G], and treatment adequacy [H], patients diagnosed between October 2017 and September 2019 with a psychotic disorder, depression or epilepsy by participating GPs and who had visits recorded in the database were included, except people who refused to participate.

In KAP surveys for CHWs [C] and GPs [D], all CHWs and GPs who took part in the Myanmar Medical Association Hlaing Thar Yar mental health project and who completed the mental health and epilepsy training in September 2017 were included, and those who refused to participate or left the project were not included. To detect the effect of awareness sessions on KAP in the [E2] survey, 205 subjects were required.

The sample size calculation and sampling procedures are detailed in Appendix 1.

Data collection

The screening questionnaires used in survey [A] were based on the PHQ9 for depression, on the PSQ for psychotic disorders, and on the Manual on Epilepsy for Medical Officers by Epilepsy Initiative, Myanmar for epilepsy, and modified according to local context. Socio-demographic characteristics of participants and data regarding their KAP on mental disorders and epilepsy [E1] and [E2], as well as their perceptions regarding barriers to treatments [B] were collected through face-to-face interviews.

The results from the screening by CHWs were captured through the questionnaires available on their smartphone; data regarding visits, diagnosis made by GPs and treatment prescribed at each visit, were captured by GPs on their electronic tablet. Treatment adequacy was retrospectively assessed by a senior psychiatrist. The psychiatrist determined, based on the information available in the database (clinical history, symptoms, diagnosis made, time of diagnosis, treatment prescribed and dosage, severity of the disease and any other detail recorded by the GP at each visit) whether the treatment prescribed was: adequate / not adequate / not possible to evaluate.

In survey [C], the KAP self-administered questionnaire used with CHWs was the same questionnaire as the one used in the general population [E1] & [E2].

The self-administered questionnaire filled by GPs for survey [D] included 106 questions based on the content of the training course and regarding their knowledge and attitudes about epilepsy, psychosis and depression.

Several endpoints were evaluated: number of patients identified by CHWs, diagnosed and managed by GPs, percentage of concordance between CHWs’ screening and GPs’ diagnosis, percentage of adequate treatments prescribed, and the evolution in KAP after training for CHWs, and GPs, and after the 2-year intervention for health workers and the general population, for which an increase in score was a sign of improved KAP.

More information on the details of the questionnaires and the calculated scores can be found in Appendix 1.

Statistical analysis

A statistical analysis was conducted using the Statistical Package for Social Sciences (SPSS 20). For descriptive analysis, numbers and percentages were used for qualitative variables, and means and standard deviation for quantitative variables.

Chi-square and Fisher exact tests were used to compare qualitative variables. Mann-Whitney test was used to compare means between two independent groups. Wilcoxon test was used to compare means between two dependent groups. Pearson correlation was used to evaluate association between two quantitative variables.

A coefficient kappa was used to measure the agreement between CHWs’ screening and GPs’ diagnosis. A comparison between scores and sub-scores before and after training was done using Wilcoxon signed-rank test. For the comparison of scores post-training and post-intervention for CHWs and GPs, the Mann-Whitney test was used.

Variables with a p-value <0·2 in the univariate analysis were included in the multivariate analysis and a backward linear regression model was performed to identify factors associated with KAP score in the general population.

A reliability of instruments used in this study was evaluated by Cronbach alpha. Shapiro-Wilk test was used to test a normality of quantitative variables.

A p-value ≤ 0·05 was considered significant throughout the study. Detailed statistical analysis can be found in Appendix 1.

Ethical considerations

Ethical approvals were granted by the Research Ethics Committee from the Department of Medical Research (2017/090) and by the Institutional Review Board from the Myanmar Medical Research Department (2020/025). All participants were informed about the research and signed a consent form.

Role of the funding source

Sanofi Global Health was involved in the design of the study, in the interpretation of data and the writing of the manuscript. However academic authors retained full control of data analysis, interpretation and decision to submit the paper for publication.

Implementation phase: patients’ identification, diagnosis and management

Screening and diagnosis

During the implementation phase, 1,378 people were referred by CHWs to GPs, including 154 for whom there was no specific mental disorder identified. Among them, 1,186 (86%) saw a GP, and 1,088 patients (92%) were diagnosed: 528 with psychosis, 305 with depression and 255 with epilepsy. Of the total, 938 had subsequent visits with a GP, and the adequacy of their treatment was evaluated by a senior psychiatrist at the end of the implementation phase (Figure 3). When comparing CHWs screenings with GPs diagnoses, overall, 75·6% of the findings were concordant, with 85·5% concordance for epilepsy, 80% for depression and 75·8% for psychosis. The Kappa coefficient (0·49) showed a moderate agreement between CHWs screenings and GPs diagnoses (Appendix 5, table S.4).

Evolution of KAP