Understanding the effect of a participatory intervention with women's groups to improve maternal and neonatal health in rural Nepal

In: Doctoral thesis, UCL (University College London).

Abstract

Background: Rigorous evaluation of community based interventions is necessary to build the evidence base for maternal and child survival in developing countries. Randomised controlled trials may be the most reliable method of determining effectiveness of interventions, but they are unable to evaluate the impact of context and implementation, and explain how an outcome occurred. This thesis explores the outcome of a cluster randomised controlled trial using women's groups in Nepal. Intervention areas experienced a 30% reduction in neonatal mortality, and increases in good care behaviours. I describe how communities experienced the intervention and explore how the implementation process and community context affected the community response. Methods: I used qualitative research methods, purposefully sampling stakeholders from two intervention areas, and one control area. Photoelicitation was used, and research assistants conducted semi-structured interviews, group interviews, focus group discussions, and observations. Data were tape recorded, transcribed and translated into English. Data were analysed using framework and NVIVO qualitative analysis software. Data were also fed back to respondents to increase validity. Results: Themes emerging from the data revealed a lack of trust in local health facilities, plurality in care seeking during illness, and a positive attitude towards working with non governmental organisations. Women's groups created a learning forum for group members, and this knowledge was disseminated in communities. Strategies to address problems were of particular significance in enabling dissemination of information and enabling community participation. The intervention increased social networks and built community capacity. Discussion: The effect of context on the intervention is considered in order to assess the generalisability of the intervention. Potential mechanisms of effect are also discussed which help to explain the trial outcome. This thesis supports calls for the integration of concurrent process evaluations within randomised controlled trials, and contributes to the evidence for community-based interventions for maternal and newborn survival.