Sehore sub-district is a predominantly rural area in Madhya Pradesh, with a population of 427,432 (14). 31.7% live below the poverty line and agriculture is the mainstay of the local economy (15). General health indicators are below the national average (16), literacy rates are 81% for males and 58% for females (14), and 88% of residents have completed only primary education or less (17).
The PRIME programme (Programme for Improving Mental Health Care) aimed to implement and evaluate district-level Mental Health Care Plans (MHCP) (18). The MHCP for Sehore focussed on depression, psychosis and alcohol use disorders and was implemented through community health centres between August 2014 and October 2016 (19).
Prior to implementation of the MHCP, outpatient and inpatient services were provided through Sehore District Hospital, by one psychiatrist and one clinical psychologist who are employed under the District Mental Health Programme and provide their services on alternate days, with periodic “outreach camps” (20). No psychotropic medication or psychosocial interventions were available in primary care facilities, there were no psychiatric social workers or psychiatric nurses, and primary care workers were largely untrained in identifying and treating mental disorders. After the plan was implemented, depression treatment was available at three Community Health Centres with psychological interventions delivered by case managers and pharmacological treatments prescribed for severe cases by medical officers. Community awareness activities were conducted to encourage service uptake, such as community meetings and proactive case finding in the community by the case managers. They also screened patients in Community Health Centres. The study area (20), Mental Health Care Plan (21), and PRIME evaluation plan (22) have been described in more detail elsewhere. The term “implementation area” will be used to refer to those villages where MHCP activities were fully implemented.
This report is a secondary analysis of data from a population-based, cross-sectional community survey carried out with the primary aim of estimating the change in treatment-seeking among adults with probable depression, before and after implementation of the MHCP. This secondary analysis focuses on characterising treatment-seeking patterns for adults with probable depression in both rounds.
The study design, sampling plan, and data collection have been described in detail elsewhere (21, 22). Briefly, data collection for the first round took place prior to Mental Health Care Plan implementation, in two waves (May-June 2013 and January-March 2014), and the second round after implementation of the plan (October-December 2016). The target population was adults (aged 18 and above) residing within the implementation area, with participants selected from voter lists through systematic random sampling. Inclusion criteria were fluency in spoken Hindi, residency in the selected household, willingness to provide informed consent, and absence of cognitive impairments that would preclude informed consent or ability to participate.
Across both rounds, 6,203 adults were recruited, 6,134 (98.9%) consented to participate, and 4,297 resided within catchment areas of the de facto implementation area, where treatment was made available in the Community Health Centres. Of the 4,297, 568 adults (289 in round 1, 279 in round 2) screened positive for depression and comprise the primary sub-sample for this secondary analysis. No difference in the probability of treatment-seeking for depression was observed between rounds, so for the purposes of the current analyses, data from both rounds were pooled to increase statistical power (23). In order to compare use of health services by adults with and without depression, for this analysis we also included the 3,531 community survey participants who resided within the implementation, did not screen positive for depression, and who did not report equivalent symptoms within the past 12 months. The sample size was calculated for the parent study based on the numbers required to detect a difference in contact coverage between rounds (the proportion of people with depression and alcohol use disorders who sought treatment for their condition), as described elsewhere (21).
Interviews were administered orally, in Hindi, by trained local fieldworkers who recorded participant responses using a questionnaire application programmed on Android tablets. The structured questionnaire included sections on socio-demographic details, health care use, barriers to using health services, depression symptoms, treatment-seeking for depression, alcohol use and related treatment, disability, internalised stigma related to depression and alcohol use, suicidal ideation and behaviours, and mental health knowledge and attitudes.
The Patient Health Questionnaire (PHQ-9) consists of 9 items on depression symptoms which are summed to generate a symptom score (24). We used a cut-off point of ≥ 10 to indicate probable depression (25, 26) which has previously been validated in India (27, 28). Participants were also asked if they had experienced equivalent symptoms for any 2 week period in the past 12 months.
Barriers to the use of health services were based on the Study on Global Ageing and Adult Health (SAGE) (29). We added one question in round 2 on distance to health services. These barriers were not specific to depression.
We chose factors to investigate based the Andersen socio-behavioural model (30, 31), which groups factors associated with health service utilisation into; (a) need factors, which include both objective and subjective assessments of health status, (b) predisposing factors, covering both demographic characteristics and attitudinal factors such as health beliefs, and (c) enabling factors, which refers to structural determinants such as financial situation, transport and social support.
Predisposing factors included gender, religion, education, age, caste, marital status and internalised stigma (measured using questions from the Internalized Stigma of Mental Illness (ISMI) scale (32)). Enabling factors included land ownership, housing type, employment status, discussing depression symptoms with someone, and reporting cost and travel barriers to health care. Need factors included symptom severity, disability (measured using the 12-item World Health Organization Disability Assessment Schedule (WHODAS 2.0)(33)), perceived need for health care, probable alcohol use disorder (measured using the Alcohol Use Disorders Identification Test (AUDIT) with a cut-off of ≥ 8 (34–37)), suicidal thoughts (measured using the Composite International Diagnostic Interview (CIDI) suicidality module (38)), and PHQ-9 item-specific symptoms of depression.
Treatment-seeking was measured after completing the PHQ-9 questionnaire by asking “Did you seek any treatment for these problems at any time in the past 12 months?”. Thus, in this report, “treatment-seeking for depression” refers to seeking treatment for the symptoms listed in the PHQ-9. Participants who answered affirmatively were asked to specify the type of provider consulted. In the section on health care utilisation, participants were asked “In the last three months, have you visited any health facility or provider for any health problem?”, and in which sector. Details of all measures used, and how these were treated in the analysis, are presented in the supplementary material.
First, we describe the sociodemographic and clinical characteristics of the sub-sample of adults with probable depression, using unweighted counts and weighted percentages to account for the sampling design.
To estimate proportion of adults with probable depression who consult different types of treatment providers for depression symptoms and for general healthcare, we present the frequency of self-reported treatment-seeking for depression symptoms and general health care use, using weighted percentages and unweighted counts. We also present the frequency of general health care use by adults without depression (excluding those who reported depression symptoms over the past 12 months) and compare these proportions using Chi squared tests.
We next measure the prevalence of self-reported barriers to health service use by adults with probable depression, by presenting percentages on the frequency with which each barrier was reported, again using weighted percentages and unweighted counts.
To assess the association between perceived need, predisposing and enabling factors and treatment-seeking for depression, we present the proportion of adults with probable depression who sought treatment for depression by each characteristic, along with prevalence ratios and 95% confidence intervals, and tested the association between each variable with the outcome of treatment-seeking for depression using univariable log-linear regression analyses. For brevity, we present only the results for factors where this association reached a significance level of p < 0.05, but a full table is included in the supplementary material. Since these analyses were intended to be descriptive and hypothesis-generating, rather than causal and hypothesis-testing, we did not conduct multivariable analyses to control for potential confounders. In order to interpret the findings on the effect of discussing depression symptoms (presumed to be a proxy measure for social support), we also examined participants’ self-reports on who they discussed symptoms with, but the numbers in each group were too small to treat as separate variables.
All analyses were conducted using Stata/IC 15.1 (39). Frequencies are reported as observed, while percentages, regression coefficients, 95% confidence intervals, and P-values are design adjusted.
Researchers explained the purpose of the survey to potential participants, read out the contents of study information sheets, and answered potential participants’ questions. Informed consent was indicated with either a signature or a thumbprint. All screen-positive participants who were not receiving treatment were referred to the nearest public health facility where depression treatment was available.
Ethical approval was provided by the World Health Organization Research Ethics Review Committee (Geneva, Switzerland), the Sangath Institutional Review Board (Goa, India), and the London School of Hygiene & Tropical Medicine Observational Ethics Committee (London, United Kingdom) (ref: 10439).