Strengthening decentralized primary healthcare planning in Nigeria using a quality improvement model: how contexts and actors affect implementation

Quality improvement models have been applied across various levels of health systems with varying success leading to scepticisms about effectiveness. Health systems are complex, influenced by contexts and characterized by numerous interests. Thus, a shift in focus from examining whether improvement models work, to understanding why, when and where they work most effectively is essential. Nigeria introduced DIVA (Diagnose-Intervene-Verify-Adjust) as a model to strengthen decentralized PHC planning. However, implementation has been poorly sustained. This article explores the role of actors and context in implementation and sustainability of DIVA in two local government areas (LGAs) in Nigeria. We employed an integrated mixed method approach in which qualitative data was used in conjunction with quantitative to understand effects of actors and contexts on implementation outcomes. We analysed policy documents and conducted interviews with PHC managers. Then using the Model for Understanding Success in Quality (MUSIQ), we measured contextual factors affecting implementation of DIVA in the selected LGAs. The LGAs scored 117.42 and 104.67 out of 168 points on the MUSIQ scale, respectively, indicating contextual barriers exist. Both have strong DIVA team attributes, but these could not independently ensure quality implementation. Although external support accounted for the greatest contextual disparities, the utmost implementation challenges relate to subnational government leadership, management, financial and technical support. Although higher levels of government may set visionary goals for PHC, interventions are potentially skewed towards donor interests at lower (implementation) levels. Thus, subnational political will is a key determinant of quality implementation. Consequently, advocacy for responsible and accountable political governance is essential in comparable decentralized contexts.