Abstract
BACKGROUND: The Pradhan Mantri Jan Arogya Yojana (PM-JAY), a publicly funded health insurance scheme, was launched in India in September 2018 to provide financial access to health services for poor Indians. PM-JAY design enables state-level program adaptations to facilitate implementation in a decentralized health implementation space. This study examines the competency, organizational, and leadership approaches affecting PM-JAY implementation in three contextually different Indian states. METHODS: We used a framework on implementation drivers (competency, organizational, and leadership) to understand factors facilitating or hampering implementation experiences in three PM-JAY models: third-party administrator in Uttar Pradesh, insurance in Chhattisgarh, and hybrid in Tamil Nadu. We adopted a qualitative exploratory approach and conducted 92 interviews with national, state, district, and hospital stakeholders involved in program design and implementation in Delhi, three state capitals, and two anonymized districts in each state, between February and April 2019. We used a deductive approach to content analysis and interpreted coded material to identify linkages between organizational features, drivers, and contextual elements affecting implementation. RESULTS AND CONCLUSION: PM-JAY guideline flexibilities enabled implementation in very different states through state-adapted implementation models. These models utilized contextually relevant adaptations for staff and facility competencies and organizational and facilitative administration, which had considerable scope for improvement in terms of recruitment, competency development, programmatic implementation support, and rationalizing the joint needs of the program and implementers. Adaptations also created structural barriers in staff interactions and challenged implicit power asymmetries and organizational culture, indicating a need for aligning staff hierarchies and incentive structures. At the same time, specific adaptations such as decentralizing staff selection and task shifting (all models); sharing of claims processing between the insurer and state agency (insurance and hybrid model); and using stringent empanelment, accreditation, monitoring, and benchmarking criteria for performance assessment, and reserving secondary care benefit packages for public hospitals (both in the hybrid model) contributed to successful implementation. Contextual elements such as institutional memory of previous schemes and underlying state capacities influenced all aspects of implementation, including leadership styles and autonomy. These variations make comparisons across models difficult, yet highlight constraints and opportunities for cross-learning and optimizing implementation to achieve universal health coverage in decentralized contexts.