How did the context of COVID-19 affect the implementation and mechanisms of participatory learning and action to address type 2 diabetes? Mixed-methods research in rural Bangladesh.

Abstract

Objectives: Research indicates the effectiveness of participatory interventions to address rapid rises in type 2 diabetes in low-income countries. Understanding their transferability to different contexts is a priority. We aimed to analyse how the COVID-19 post-lockdown context and adjustments to a participatory learning and action intervention affected theorised mechanisms of effect in rural Bangladesh and to examine the broader implications of this context and intervention adjustments for developing optimal contexts for participatory interventions.

Design: Mixed methods using longitudinal qualitative and quantitative observation data, focus group discussions and interviews with group and community members and project personnel. We used descriptive content analysis, guided by realist evaluation research questions about context, implementation and mechanisms. We used team reflection to enhance the rigour of our analysis.

Setting: Cluster-randomised trial in Alfadanga upazila, Faridpur district, in the central region of Bangladesh. The intervention was implemented between January 2020 and December 2022, during the COVID-19 pandemic.

Participants: Participatory group members, community members and project personnel (n=32). Structured observations of participatory groups (n=1820) and unstructured observations of groups and their environments (n=15).

Interventions: Participatory learning and action community groups of men and women implemented by community-based facilitators.

Results: Due to COVID-19, the participatory learning and action (PLA) intervention was not implemented as planned, which had major effects on the time available to develop the intervention with communities. Communities learnt about diabetes and were motivated to address its causes at an individual level, but community action was a more challenging mechanism to trigger. The post-pandemic context made it difficult to build community rapport, and strategies to engage communities through home visits were challenging. Communities' prior negative experience in working together and in working with non-governmental organisations led to low community cohesion and low motivation to address diabetes collectively. This also resulted in expectations that the implementing organisation would implement community actions and incentivise attendance at meetings. This misalignment of expectations further disabled relationship building, and community strategies addressing the social causes of diabetes were largely not enacted.

Conclusion: PLA has optimal effects when time is available to build trust and social cohesion. These are contextual elements and mechanisms that need to be activated to enable critical reflection and community action to develop an enabling environment to address type 2 diabetes.

Trial registration number: ISRCTN42219712.