Randomized Control Trial of the Bandebereho intervention

Challenge

The Bandebereho intervention was first implemented by the Rwanda Men’s Resource Center (RWAMREC) and  Promundo-US, in collaboration with the Rwanda Ministry of Health and local authorities between 2013 and 2015. The intervention invited men and their partners to participate in small group sessions set up to stimulate a reflection on gender norms and how these norms shape their lives. The participants were also invited to rehearse equitable and non-violent behaviors in a comfortable space with supportive peers. The goal was to promote men’s engagement in maternal and newborn health, equitable caregiving, healthier couple relations, and the prevention of violence against women and children.

RWAMREC and Promundo-US have partnered with Laterite since 2015 for the impact evaluation of the intervention. The primary goal was to assess the impacts on men’s and women’s behaviors and experiences, including parenting, caregiving and family dynamics. A second goal was to see if the intervention had a long-term impact on the couples’ children, specifically on their cognitive and social-emotional development.

Laterite conducted a two-arm Randomized Controlled Trial (RCT) in 2015-2016 with 1,199 couples, with about half as control. The couples were recruited to participate in the Bandebereho intervention via the male partner in early 2015 and were later randomized to either the intervention or control arm. The control arm received no group intervention, though it did have access to community activities and campaigns related to the broader MenCare+ project.

Laterite collected data pre-intervention in February-March 2015 and followed up with a second round in November 2015 and a third round at 21 months in November 2016. The findings from this last data collection were published in PLoS ONE and demonstrated significant impacts of the intervention on multiple gender and health-related outcomes.

In 2021, Laterite went to the field again for a six-year follow-up, 71 months after the intervention ended. The study included surveys with men enrolled in the original RCT and their female partners, and child assessments with a sub-sample of 800 children aged 4 to 7 years to assess child development outcomes.

Innovation

The first challenge was to reach out to the original 1,200 participants couples of the RCT. Building on RWAMREC’s initial work, Laterite identified more than 1000 couples eligible for follow up, thanks to a combination of screening phone calls and visits.

Field work for this study was conducted in two rounds over the summer of 2021, to comply with the movement restrictions enforced by the Government of Rwanda to curb the spread of COVID-19. A team of 60 enumerators and supervisors, divided in sub-teams, was deployed in Karongi, Musanze, Nyaruguru and Rwamagana districts. Deploying such a large team to reach out to vulnerable populations, including children, called for the enforcement of strict precautions.

We put a lot of care into verifying respondents’ identities before collecting data. It is not unusual for other individuals to come to the study site to participate in the study in order to receive a transportation stipend, or for respondents to send relatives to replace them when they are unavailable at data collection time. In addition to requiring participants to bring their IDs to study sites, enumerators would ask them for basic information about their household (family members names, number of children, location, etc.) that we could compare to data collected at earlier stages. Enumerators investigated further suspicious cases by contacting local leaders or community health workers. These identity checks also confirmed that overall, information about children’s dates of birth gathered during field preparation was more accurate when collected from female respondents.

Another factor adding to the complexity of this assignment was the diverse nature of the survey instruments. Surveys directed to adult respondents covered gender-sensitive topics, which called for adequate precautions. Enumerators were careful to set up interviews with respondents in places where they could not be overheard to ensure their privacy. Our teams also followed a risk of harm protocol to report any adverse events (such as intimate partner violence) where they suspected the respondent could be at risk of harm. We reached out to male respondents first to both recruit them for the interview and let them know we would also be reaching out to their partner.

Children were assessed using the International Development and Early Learning Assessment (IDELA). To do this, Laterite deployed a team of 8 female enumerators experienced in conducting early childhood assessments and trained by Save the Children experts. The IDELA assessment involved inviting children to participate in short activities and answer questions on stories, shapes, letters, numbers. We put a lot of care into ensuring that children were comfortable throughout the assessment and felt free to stop at any time, even after providing assent to participate. In addition, to ensure the tool was implemented consistently, more than 10% of study children were assessed simultaneously by two enumerators to check that the enumerators shared a common understanding of the assessment.

Outcomes

The follow-up study of the Bandebereho intervention was a complex data collection project, involving a large team, multiple survey instruments, and an extra level of precautions to comply with our COVID-19 protocol and the Government of Rwanda guidelines. Here are the lessons we will take for future projects:

• Planning ahead. We found that all the time invested in planning, researching and discussing logistics with the field teams paid off. In fact, any setbacks we did face during fieldwork happened when planning was rushed.
• Focusing the teams. We minimized the COVID-19 disruption on movement by assigning one field team per district. We also found that during a pandemic it pays off to have additional enumerators on stand-by to cover for colleagues in contact with suspected cases.
• Cross-referencing instruments. We found that there is a lot to be gained in terms of quality assurance by comparing metadata information from the different instruments administered to the same household, such as their location or their children’s identity and age. For future projects, we will dedicate more time to testing and tracking complex survey changes and data monitoring tools.
• Inter-rater reliability is your friend! IRR is a measure of the agreement between two independent observers and is especially useful to assess data quality when the instruments rely on subjective observations. We made great use of IRR to improve the training of the field team and to provide continuous feedback to IDELA assessors. Read more about the uses of IRR as a tool to reduce bias in surveys.