Exploring community savings group mechanisms and their health impacts in Sub Saharan Africa: a scoping review using artificial intelligence

Achieving Target 3.8 of the Sustainable Development Goals for Universal Health Coverage (UHC) requires a robust transition from inefficient out-of-pocket (OOP) payments toward equitable pre-payment mechanisms. In Sub-Saharan Africa, where financial hardship and healthcare inequity remain critical challenges, traditional Microfinance Institutions (MFIs) often fail the rural poor due to a "mission drift" toward profitability. Community Savings Groups (CSGs) have emerged as a vital, decentralized alternative. With over 14 million members active across the continent as of late 2025, these autonomous, savings-led groups demonstrate significant potential for increasing healthcare utilization and enhancing financial resilience. However, while the implementation of CSGs is widespread, the evidence linking these diverse models to specific health outcomes remains fragmented and poorly synthesized. A preliminary search of the JBI and Cochrane databases confirms that high-level evidence synthesis in this area is scarce. This highlights a critical need for a scoping review to systematically map the existing literature, clarify the complex typologies of these groups, and identify exactly how their mechanisms, particularly dedicated health funds, influence financial protection and health access. This study aims to explore regional evidence on the impacts of savings groups on access to healthcare, health outcomes and financial protection in Subsaharan Africa. The study has a particular interest in dedicated health funds in savings groups. This study will be conducted in accordance with the Joanna Briggs Institute (JBI) methodological guidance for scoping reviews and reported following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) guidelines. To enhance the efficiency and reproducibility of the evidence synthesis, the review process will be supported by Artificial Intelligence (AI) and Large Language Models (LLMs).