Mental health, social functioning and social capital in Rwanda.

Primary outcome measure is mental health, as measured by use of the Self Reporting Questionnaire 20 items (SRQ-20). This is a structured interview developed by the World Health Organization (WHO) as a screening tool for common mental disorders in primary health care settings, especially in developing countries. When patients are literate it can be self-administered, but in developing countries it is usually administered by lay interviewers. The instrument consists of 20 yes/no questions about mood, thinking capacity, feelings of anxiety and physical well-being. The SRQ-20 has been used in numerous settings, also as a screening instrument in community samples. Cut-off points vary considerably depending on setting (community, primary care, hospital) and culture. A cut-off point of 8 is widely used.
During our pilot study we validated the instrument locally. We interviewed 99 respondents, who were also clinically assessed by clinicians who were blind for the SRQ-scores. We established a local cut-off score for caseness/non-caseness. The capacity of the SRQ-20 to identify probable
psychopathology in this setting proved to be sufficient. The AUC was 0.76. When analysed separately for men and women the SRQ-20 showed to perform equally well in men (AUC=0.74) and women (AUC=0.76). In evaluating the instrument as a potential screener for psychiatric disorder the most appropriate cut-off point is a trade-off between a high sensitivity and an acceptable specificity. From our analyses the optimal cut-off point for the SRQ-20 appeared to be 10/11 (sensitivity 0.69; specificity 0.79). However, when men and women were analysed separately the cut-off points differed. The optimal cut-off point for men is 8/9 (sensitivity 0.69; specificity 0.65), while the optimal cut-off point for women is 10/11 (sensitivity 0.81; specificity 0.80). Reliability is considered to be good: alpha=0.83 (men: alpha=0.81; women: alpha=0.85).

1. Social functioning, as measured by the Byumba Social Functioning Questionnaire (BSFQ). This instrument contains questions which have been derived from a list that resulted from previous research in Rwanda. The actual questions have been compiled and listed during focus group interviews with people living in Byumba. Questions concern common daily activities for women and men, respectively, such as dressing oneself or taking care of the children. Data from our pilot study revealed that the instrument’s reliability is sufficient to good: alpha=0.77 (men: alpha=0.77; women: alpha=0.73).

2. Social functioning will also be measured with the Medical Outcomes Study (MOS) Social Functioning Scale 36 items (SF-36). This internationally widely used instrument is based on the WHO definition of health. Within this definition, health is divided into three domains: physical, mental, and social health. The SF-36 contains questions about functional impairment related to physical or psychological problems. In this study the instrument will be used to further found the validity of the BSFQ. For use in Rwanda, items of the SF-36 will be adapted if containing contextually inadequate phrases ('pushing a vacuum cleaner', 'walking blocks', 'playing golf'). Alternatives will be determined during discussions with local community members.

3. Social capital, as measured by use of the Social Capital Assessment Tool, short adapted version (SA-SCAT). This instrument contains questions used to get an impression of the respondents’ connectedness to their social environment. They have been derived from a list of questions that resulted from previous research in various countries. The actual questions were compiled and listed during focus group discussions with people living in Byumba. This has resulted in a locally adapted version. Questions concern the way people feel connected to others in their living area, feel supported by others, get along and undertake common actions with others. Data from our pilot study revealed that the locally adapted version consists of five factors: Active Membership, Cognitive Social Capital, Individual Support, Group Support, and Participation. These factors are congruent with previous research and theory. Internal consistency of the factors ranges from 0.59 (Cognitive Social Capital) to 0.95 (Active Membership). In an effort to raise the instrument’s psychometric qualities, a version will be used during this study with items added to its pilot study version, and yes/no answering options changed to Likert scales. Items to add will be determined during another round of discussions with local community members.

4. Alcohol use, as measured by use of the Alcohol Use Disorders Identification Test Consumption (AUDIT-C). This instrument contains 3 questions assessing the number and frequency of having ‘drinks containing alcohol’, a formulation which is suitable for use in a country where alcohol may often not be consumed in identifyable units like commercial bottles or standard glasses. The AUDIT-C has shown to be an effective screening instrument for alcohol misuse on primary care level. Subsequent validation in American clinical outpatients and in European samples, respectively, yielded different optimal screening thresholds. The instrument has not been validated locally during the pilot phase of this study. Therefore, after translation and back-translation, the instrument will only be used to explore change in alcohol use over time.

5. Partner violence, as measured by using elements of the Revised Conflict Tactics Scales, short form (CTS2S): the
‘negotiation’, ‘psychological aggression’, ‘physical assault’ and ‘injury’ scales. (The remaining scale concerns sexual coercion, a subject inacceptable to address in a short, one-time interview in Rwanda.) The CTS2 is the most widely used and men, respectively, such as dressing oneself or taking care of the children. Data from our pilot study revealed that the instrument’s reliability is sufficient to good: alpha=0.77 (men: alpha=0.77; women: alpha=0.73). The CTS2 is the most widely used instrument for measuring family violence. The short form enables the CTS2 to be used when testing time is limited. There is only a clear indication so far that the validity of the CTS2S is comparable to the CTS2, and the CTS2S has not been validated locally during the pilot phase of this study. Therefore, the instrument will be considered to only provide an indication of partner violence. The instrument classifies respondents into three severity levels: ‘none’, ‘minor only’ and ‘severe’, and creates mutuality types: ‘male partner only’, ‘female partner only’, and ‘both aggressive’. The latter is relevant as in this study only one of the partners of possible couples will be interviewed. As the 7-options-per-item answering system of the CTS2S is considered too complicated for the study population to provide valid answers, an adaptation to this will be designed and piloted after translation and back-translation.