Using Realist Methods to Unpack Communication in Community Engagement: Learning from Maternal and Newborn Health Programming

Introduction: Community engagement has been a crucial component of programming that aims to improve maternal and newborn health (MNH) in low- and middle-income countries. In Zambia, less than half of pregnant women complete the MNH continuum of care. To address this, community engagement initiatives such as the Safe Motherhood Action Groups (SMAGs) are used to mobilise communities and support MNH care-seeking. However, there is a lack of documented evidence on how community engagement is implemented across diverse settings and specifically how communication contributes to this. This thesis used two sequential realist studies to unpack how, why, to what extent, and for whom communication in community engagement works for MNH programming, using the SMAGs in Zambia’s Eastern Province as a case study. Methods: First, a realist review of academic and grey literature was conducted to develop and refine initial programme theories (IPTs). The first phase of this process involved developing five IPTs to inform the realist review by conducting an initial scoping of the literature, considering the researcher’s own experiences, and receiving feedback from an expert advisory committee. In the second phase, a search strategy was executed across seven academic databases, two grey literature databases, Google Scholar, and general web searches. Documents were appraised for their relevance, richness, and rigour. Data from included documents were extracted in the form of context-mechanism-outcome configurations (CMOCs), which were analysed and synthesised for demi-regularities to inform theory refinement. Second, to further refine and test the programme theories (PTs), a realist evaluation used two iterative, multi-methods case studies of the SMAG programmes at Chibolya Urban Health Clinic and Jerusalem Rural Health Clinic in Zambia. Data were collected from observations, facility metrics, interviews with SMAGs (n=10), local leaders (n=2), healthcare providers (n=4), and Ministry of Health stakeholders (n=4), photovoice interviews with pregnant women (n=10), and four focus group discussions with community members (n=20). Participants were purposively sampled. Using NVivo, CMOCs were identified in the data and linked to relevant/emergent theory areas. These CMOCs were analysed and synthesised to refine PTs using abductive and retroductive reasoning. Results: In the realist review, 45 CMOCs were extracted, analysed, and synthesised from 11 documents to inform five refined PTs, which were further tested and refined in the Zambian case studies. In the realist evaluation, 50 CMOCs informed the refinement of PTs. The resulting seven refined PTs explain that the SMAG programme in Zambia works when: communities are actively involved; messaging is relevant and relatable; SMAGs are trusted; SMAGs are familiar; the programme and facility are well-coordinated; communities perceive benefits from messaging; and SMAGs are motivated. For example, the first PT emphasises that involving communities in co-creating community engagement and its communication (e.g. identifying SMAGs or prioritising issues they address) makes the programme more relevant to their needs and more accepted because they have a sense of ownership over it. These seven PTs were further interrogated with existing evidence and substantive theory to inform four middle-range theories (MRTs) aligning with four elements of a communication model: sender, receiver, message, and channel/feedback. Conclusion: Applying realist methods enabled a deeper understanding and exploration of generative causation of ‘how’ and ‘why’ communication in community engagement works. Through the iterative testing and refinement of theories, this thesis makes a novel contribution to policy, programming, and theory by suggesting practical and transferable evidence-based guidelines in the form of refined PTs and MRTs which can enhance the design and implementation of communication in community engagement.