Depression Stigma and the Marriage Market in India

From 2017-2019, we implemented a randomized trial to evaluate the impact of community-based pharmacotherapy for people with depression. Angelucci and Bennett (2024A) provide the details of this study. We conducted this study in 506 localities in the Doddaballapur, Korategere, and Gauribidanur taluks near Bangalore.

The psychiatric care (PC) intervention provided eight months of free psychiatric care through the Shridevi Institute of Medical Sciences and Research Hospital. Shridevi is an accredited private hospital in Tumkur, Karnataka, near the study area. The facility has 750 beds, 80 percent of which are allocated for pro bono care of disadvantaged patients. The hospital sometimes receives patients from GASS. The initial visit included a diagnosis, an explanation of the significance of mental illness, and an individualized course of medical treatment. Patients returned for monthly follow-up visits. The most commonly prescribed anti-depressants were Selective Serotonin Reuptake Inhibitors (SSRIs). These drugs are generally not under patent and are available inexpensively in India. They are widely used and have relatively few well-tolerated side effects (Ferguson 2001, Cascade, Kalali and Kennedy 2009).

The livelihoods assistance (LA) intervention provided two group meetings and personalized livelihoods assistance during the first two months of the study.

2016-11-01

2018-11-01

Marriage market preferences about depression within the community: tradeoffs between marrying a depressed spouse (or a spouse with a depressed mother) versus dowry size and level of spousal education.

Marriage market beliefs about depression within the community: first and second-order beliefs about depression.

Marital status: realized marriages and marriage timing over the previous seven years among household members aged 18-35 who were unmarried during our baseline survey.

Marriage characteristics: dowry, fertility, marital satisfaction, partner age, education, caste, and income. The gap in education, age, and caste with the partner. Distance from the partner's locality of origin. Physical and mental health. Violence.

Marriage market preferences about depression for original study participants: tradeoffs between marrying a depressed spouse (or a spouse with a depressed mother) versus dowry size and level of spousal education.

Marriage market beliefs about depression for original study participants: first and second-order beliefs about depression.

We used a cluster-randomized design to cross-randomize PC and LA by locality. Before starting recruitment, we stratified the randomization by district and terciles of a locality socioeconomic index based on the 2011 Census of India. Both the modal and median number of participants per locality is 2. This design minimized spillovers and cross-arm contamination. Treating few people per locality limited information leakages, protecting patient confidentiality. We randomized localities into four intervention arms using a factorial design. The arms were PC/LA, PC, LA, and Control.

We sampled participants through a door-skip pattern in which the skips were proportional to locality size. Once at the household, surveyors randomly chose an available adult to screen for eligibility. We screened people for depression symptoms with the PHQ-9 depression severity scale. To obtain a sample of mildly or moderately depressed people, we recruited subjects with PHQ-9 scores of 9-20.

Computer.

Randomization was conducted by locality.

Yes

506 clusters based on the localities of the original study participants. However, our primary analysis is likely restricted to the 470 localities with men and women aged 18 to 35, extrapolating from our baseline data.

We estimate that there are 569 people who were unmarried at baseline and who are now aged 18 to 35. We will conduct the discrete choice experiment among 500 people from the original study localities. We will also follow up with the 1000 original study participants.

For the pooled data (18-35 year-old men and women), our baseline indicates that we have 569 participants (from 192 localities) in the control arm, 325 participants (from 95 localities) in the PC arm, 280 participants (from 92 localities) in the LA arm, and 310 participants (from 91 localities) in the PC/LA arm.

For females who are 18-35 years old, our baseline data indicate that we have 293 participants (from 156 localities) in the control arm, 162 participants (from 80 localities) in the PC arm, 145 participants (from 76 localities) in the LA arm, and 153 participants (from 78 localities) in the PC/LA arm.

For males who are 18-35 years old, our baseline data indicate that we have 276 participants (from 155 localities) in the control arm, 163 participants (from 78 localities) in the PC arm, 135 participants (from 73 localities) in the LA arm, and 157 participants (from 69 localities) in the PC/LA arm.

The power calculations for AIT estimates of impacts on marriage market outcomes are as follows. Pooled data: Assuming a 10% attrition rate for each arms in the follow-up survey, the minimum detectable effect (MDE) for the comparison of any of the intervention arms with the control arm (e.g. PC/LA vs. control) is 0.18 SD. This calculation is based on the assumptions of 80 percent power and 95 percent confidence. For a comparison of pooled PC arms with the control arm (e.g. PC/LA and PC vs. control), the MDE is 0.14 SD. Assuming a 20% attrition rate for each arms in the follow-up survey, the minimum detectable effect (MDE) for the comparison of any of the intervention arms with the control arm (e.g. PC/LA vs. control) is 0.19 SD. This calculation is based on the assumptions of 80 percent power and 95 percent confidence. For a comparison of pooled PC arms with the control arm (e.g. PC/LA and PC vs. control), the MDE is 0.15 SD. Separately by gender: (men and women have similar sample size, thus the MDEs are also similar): Assuming a 10% attrition rate for each arms in the follow-up survey, for the comparison of any of the intervention arms with the control arm (e.g. PC/LA vs. control) is 0.23 SD. This calculation is based on the assumptions of 80 percent power and 95 percent confidence. For a comparison of pooled PC arms with the control arm (e.g. PC/LA and PC vs. control), the MDE is 0.19 SD. Assuming a 20% attrition rate for each arms in the follow-up survey, the minimum detectable effect (MDE) for the comparison of any of the intervention arms with the control arm (e.g. PC/LA vs. control) is 0.25 SD. This calculation is based on the assumptions of 80 percent power and 95 percent confidence. For a comparison of pooled PC arms with the control arm (e.g. PC/LA and PC vs. control), the MDE is 0.20 SD.