Reducing relapse in children after recovery from severe acute malnutrition in Mali: participatory development of a theory of change for post-treatment monitoring including SQ-LNS supplementation

Introduction Up to 76% of children treated for acute malnutrition relapse within 6 months from discharge. Despite its known negative impact on children's health and the drain on health system's scarce resources, few studies have explored ways to reduce relapse. In order to design an intervention that can work, there is a need to thoroughly understand the drivers of relapse in a context. Objective We aimed to develop a robust theory of change (ToC) for how relapse can be reduced in Mali and propose an intervention package that is feasible for the health system to adopt with minimal external support. Methods We applied a four-step approach to develop the ToC including: (1) An evidence review of potential drivers of malnutrition and relapse locally, based on surveys, gray literature and scientific articles, (2) Schematization of the most important immediate and underlying causes of relapse based on step 1 in a problem tree style and development of a ToC for a program that aims to address the causes, (3) Validation of the ToC with local stakeholders, and (4) Identification of facilitators, barriers and assumptions. Most work was done through on-line workshops including participants from local Ministry of Health (MoH), a local university (USTTB) and both global and local IRC nutrition staff. One in-person workshop was organized with local stakeholders from technical partners and different levels of local health authorities to finalize the ToC. Results Three direct causes of relapse were identified: (1) Inadequate initial treatment, (2) Inadequate dietary intake post-discharge, and (3) Frequent illness episodes combined with inadequate treatment seeking. To reduce relapse post-treatment, three aims were identified: (1) Improving initial acute malnutrition treatment, (2) Improving nutritional intake after discharge, (3) Decreasing incidence of illness and increasing use of health services. The proposed intervention was built around a program for the post-treatment monitoring of children at the treatment sites including health checks and the provision of small-quantity lipid-based nutrient supplements for children. Inputs included initial and on-going training, supplies and equipment, and supervision and coordination. Conclusion The development of a comprehensive ToC to prevent relapse allowed us to design an intervention that tackles the key perceived and evidenced drivers of relapse in the context. This ToC will underpin a process evaluation that will need to accompany an impact evaluation to determine if and how the intervention reduces relapse.