Testing Family Intervention to Address Mental Health and Domestic Violence Among Women

The study was conducted with an objective to test a family intervention that engages three key family members (Woman, mother-in-law (MIL) and husband) to improve mental health while simultaneously reducing DV. The study team assessed the acceptability, feasibility, and preliminary efficacy of the intervention to reduce domestic violence and depression among married women. The study was conducted in collaboration with Women's Rehabilitation Center (WOREC), a national non-profit, community-based organization working on violence against women and children for more than two decades.

This study, was implemented in two phases. The first phase was conducted among young married women, aged 15-24 years at Dhanusa district of Madhesh Province in Nepal. After incorporating the learnings from this phase, the study team then conducted the second phase of the study with married women of reproductive age group at Dhanusa (Madhesh Province) and Morang (Koshi Province) districts.

For both phases, a single-arm intervention study was conducted, where all participants were provided with family intervention. The multi-component (MIL-DIL and BCT) family intervention engaged women, their husbands and MILs. The intervention comprises nine sessions with 11 hours total over 9 weeks. The first four sessions included the MIL-DIL sessions focused on skills building, enhancing connections with important relationships, and promoting resilience and empowerment. After MIL-DIL, the participants were provided with BCT sessions which concentrated on improving spousal relationships, fostering increased trust and caring behaviors between woman and her husband. The BCT sessions lasted for four weeks (each session lasting an hour). Finally, the participants participated in one hour triad session. This session reviewed key lessons from all previous sessions and prepared participants to address potential challenges in sustaining improvements in IPV and mental health.

To assess the feasibility and acceptability of the proposed interventions, the study used mixed-method (Quantitative + Qualitative) research design. The sampling unit was individual participant and the primary outcome of interest was depression. The study team used convenience sampling for quantitative component of the survey in both phases. The study adopted referral-based recruitment where social mobilizers, psychosocial counselors and WOREC program team who came into contact with potential research participants referred the women to the study team. The women were then screened using a screening tool following the clearly outlined screening process. The total sample size for the overall study was 258 participants (86 families), including both phases.

The quantitative data collection included administering of 17 pre-validated tools during first phase and 12 pre-validated tools during second phase which were administered at baseline (prior to intervention session), one month, three month and six month from the baseline assessment. The tools included:

1. Demographics (for women, their husband and MIL);
2. Patient Health Questionnaire (PHQ-9) to measure depression; also validated in Nepal (for women only)
3. Indian Family Violence and Control Scale (IFVCS) to measure violence (women only)
4. Adapted 12-item Conflict Tactics Scale (CTS)-2 to measure domestic violence (for women only).
5. PCL-5 to measure PTSD; has been used and validated in Nepal (for women only)
6. Women's Autonomy Scale (WAS), 5 items to measure freedom of movement of women (women, husband and MIL)
7. Fear Survey Schedule (FSS) subscales for injury fears and sex aggression fears (for women only)
8. Intimate Partner Violence Stigma Scale (IPVSS), a 20 itemed to measure perceived stigma related to IPV (for women only)
9. Rosenberg Self Esteem Scale (RSES), a 10-item scale to measure positive and negative feelings about self; has been used and validated in Nepal (for women only)
10. Marital Satisfaction Scale (MSS), a 30-item scale on marital relationship functioning (for women and husband)
11. Coping checklist, a 68 item scale to measure individual coping strategies (for women, husband and MIL)
12. Communications Patterns Questionnaire (CPQ), a 35-item tool for couple's behavior patterns (women and husband)
13. Multidimensional Scale of Perceived Social Support; 10- item tool (for women)
14. Sexual Relationship Power Scale (SRPS), a 8-item scale for decision-making power in the relationship (for women, husband and MIL)
15. Gender-Equitable Men Scale (GEMS), 4-item, 3-point response subscale (women, husband and MIL)
16. Depression Scale-Short Form (BADS-SF) to measure behavioral activation (for women only)
17. Acceptability tool to measure the acceptability and likeability of the intervention (women, husband and MIL).

For qualitative data collection, the team conducted serial in-depth interviews (IDIs) with 6-8 families in each phases (women, husband and MIL that gave us a total of 36-48 IDIs) who received MIL-DIL and BCT interventions, including both phases. The study team used purposive sampling (based on age, experience of violence, ethnicity) to select IDI participants from a subset of participants who participated in the quantitative component

Data Analysis Plan : Data from questionnaires and intervention fidelity assessments were collected and entered by trained staff on tablets that are password-protected and encrypted and securely transferred into a secure, password-protected REDCap data collection system. The study coordinators conducted weekly quality control checks under the supervision of the senior statistician to minimize data collection errors. IDIs were audio-recorded, with "debriefing" reports completed immediately after each interview and rapidly disseminated to the study team to highlight preliminary findings and emergent themes. Key findings were also captured in a results matrix that were updated from the debriefing reports. Recorded interviews were translated and transcribed at Possible in Nepal. All data were backed up on a separate password protected server. Our study coordinator reviewed debriefing reports and transcripts for quality control, with queries resolved on a weekly basis. Preliminary efficacy of the intervention for the primary outcome of women's depression was assessed by looking at the PHQ-9 score at baseline and comparing that with a 6-month follow-up. The study hypothesized that DILs had a PHQ-9 score less than 9 (mild depression) at 6-month follow-up than at the baseline. For the secondary outcomes of DV and PTSD, the study team compared the mean IFVCS and adapted CTS-2 tool and PCL-5 scores, respectively, between the baseline and six month follow-up, and expect the women in at 6-month follow-up report lower mean scores on both. The data analyst primarily conducted descriptive analysis. Besides, the data analyst also ran unadjusted analyses via t-test, as well as logistic regression and multiple linear regression models to look at the preliminary efficacy. Due to the small sample size, significance testing was de-emphasized. Acceptability was assessed among all participants using 5-point Likert scale ratings on the overall experience, self-efficacy, usefulness, and difficulties in engaging with the intervention, based on a scale in our current study. The qualitative team obtained additional insights on intervention acceptability to provide complementarity (e.g., "are there additional ideas on acceptability of the intervention that were not covered by the surveys?") and expansion (e.g., "why do you think participants gave this average score to the question about overall usefulness of the intervention?"). Feasibility were determined by: a) tracking number screened, eligible, and consenting; b) retaining > 80% of participants; c) completing > 90% assessments; and d) participants attending > 85% of intervention components as scheduled. In addition, during the IDIs, participants were asked about feasibility, barriers, and ease in engaging with the intervention.

Analyses of the IDI transcripts- The qualitative team used a thematic approach to coding and summarization using Dedoose qualitative analysis software. A preliminary codebook was developed from interview guides and rapid analysis content, then refined after all analysts code a set of transcripts independently and resolve discrepancies through consensus. The remaining transcripts were single-coded, maintaining intercoder reliability of at least 80% throughout the coding process using coding test features of Dedoose and subsequent kappa calculations. Analysts also reviewed and analyzed coded serial IDI content within each participant's dataset to look specifically for patterns over time to further understand the change process.

Triangulation of quantitative and qualitative data: Findings from the qualitative and quantitative analyses were triangulated to build a more comprehensive understanding of feasibility, acceptability and preliminary efficacy their relative weight in creating changes from baseline to 6-month follow-up. Both differences and similarities in the data sets were examined and presented to understand the relationship between the constructs in the conceptual model and to highlight any factors that may be uniquely identified by qualitative or quantitative methods. These data informed interpretation of the effectiveness findings and next steps for any refinements that may be necessary as the intervention is tested in a well-powered study.