Study design and setting
This was a cross-sectional analytical-quantitative study involving married or cohabiting women in Isingiro Town council in Isingiro district in southwestern Uganda. Isingiro district is located 309.2 KM from the capital city Kampala, with a total estimated population of 18,045 women . Data were collected from Mabona parish which was purposively selected because it had a group of married or cohabiting women who were organized in money generating group (“Bataka kweyamba group”). In addition, this group had a list of all married or cohabiting women. We enrolled women aged 18-45 years, who were part of a money generating group and consented to participate in the study. We excluded women who had a physical illness that could have impaired their ability to stand the length of the interview and those whose spouses did not give permission to participate to avoid misunderstandings in the family. Pregnant women, postpartum women, and those aged 45 years and above were excluded, due to hormonal changes in these women associated with higher prevalence of depression . Women above 45 years may be experiencing premenopausal or menopausal symptoms that may present as symptoms of depression hence overestimating the prevalence .
Sample size determination
A pilot study involving 50 married women was done in a similar setting to determine the prevalence of depression among married women. The prevalence of depression among women in the pilot study was 23% and this was used to determine the sample size of the current study using the Kish-Leslie fomula . A sample of 153 participants was reached based on a 95% confidence interval, and margin of error 5%. Consecutive recruitment was done until the required sample size was reached.
Data were collected by women research assistants who were trained in data collection methods. They administered the questionnaire in the local language and each interview took about 30 minutes to complete. The questionnaire had four parts. 1) Participants sociodemographic characteristics such as age in years, level of education, monthly income, number of children, employment status, history of alcohol and substance use, and history of mental illness; 2) the spouse’s reported characteristics (by participating women), such as age, level of education, employment status, extramarital relationship, history of mental illness, alcohol and substance use history; 3) intimate partner violence assessed using the Composite Abuse Scale (Revised) – Short Form (CASR-SF); and 4) depression among participants assessed using the Patient Health Questionnaire 9 (PHQ-9)
The Composite Abuse Scale (Revised) – Short Form (CASR-SF)
CASR-SF is a 15 item self-administered scale that measures intensity of intimate partner violence in the past 12 months. It measures three types of violence; psychological, physical, and sexual violence by assessing their frequency on a 5-point Likert type scale (0 = not in the past 12 months, 1 = once, 2 = a few times, 3 = monthly, 4 = weekly). Total scores range from 0 to 75, and a score of 25 and above indicates severe domestic violence. The scale had an internal consistency of 0.94 in a Canadian study; whose reliability and validity estimates are comparable to those obtained for the original 30-item Composite Abuse Scale .
The Patient Health Questionnaire 9 (PHQ-9)
The PHQ-9 is a 9-item self-administered tool that may help identify depressive symptoms. A recent Cronbach’s alpha of 0.71, a sensitivity, and specificity of 88% for major depressive disorder were obtained in a study by Carrol et al. 2020  based on a score above 10. It uses a Likert type scale where for every answer, not at all = 0, several days = 1, more than half the days = 2, and nearly every day = 3. The PHQ-9 has become an international gold standard measurement tool for depressive symptoms. The PHQ has been used in Uganda with excellent psychometric properties [32-35].
The study was conducted according to the ethical guidelines of the Declarations of Helsinki. It was approved by Mulago Hospital Research and ethics committee (MHREC 2044). Permission to collect data was obtained from the district, local councils of the parish, and villages.
The partners (head of the family) of the participants gave permission to the researchers to interview their wives about mental wellness. The spouses were not given the details of the study and what information may be asked. All participants were interviewed in a private place either within the home environment or away from home or any other pace of their choice not accessed by another person to ensure privacy and confidentiality. Participants were informed about the study and they provided written informed consent before enrollment in the study. Counselling was provided to all the women who were found to have depression by the psychologist on the team (BN).
Data were entered into Excel and exported to STATA version 16.0 for data analysis. Descriptive statistics were summarized using mean and standard deviations for continuous data and percentages for categorical variables. Bivariable and multivariable logistic regression analyses were used to determine association between depression and independent variables including participant and spouse sociodemographic characteristics, substance use, mental illness history, IPV, spouse extramarital relationships status. A p-value of less than 0.05 for the level of significance was considered.