What made primary health care resilient against COVID-19? A mixed-methods positive deviance study in Nigeria

Abstract

Introduction The SARS-CoV-2 (COVID-19) pandemic overwhelmed some primary health care (PHC) systems, while others adapted and recovered. In Nigeria, large, within-state variations existed in the ability to maintain PHC service volumes. Identifying characteristics of high-performing local government areas (LGAs) can improve understanding of subnational health systems resilience.Methods Employing a sequential explanatory mixed-methods design, we quantitatively identified ‘positive deviant’ LGAs based on their speed of recovery of outpatient and antenatal care services to prepandemic levels using service volume data from Nigeria’s health management information system and matched them to comparators with similar baseline characteristics and slower recoveries. 70 semistructured interviews were conducted with LGA officials, facility officers and community leaders in sampled LGAs to analyse comparisons based on Kruk’s resilience framework.Results A total of 57 LGAs were identified as positive deviants out of 490 eligible LGAs that experienced a temporary decrease in PHC-level outpatient and antenatal care service volumes. Positive deviants had an average of 8.6% higher outpatient service volume than expected, and comparators had 27.1% lower outpatient volume than expected after the initial disruption to services. Informants in 12 positive deviants described health systems that were more integrated, aware and self-regulating than comparator LGAs. Positive deviants were more likely to employ demand-side adaptations, whereas comparators primarily focused on supply-side adaptations. Barriers included long-standing financing and PHC workforce gaps.Conclusion Sufficient flexible financing, adequate PHC staffing and local leadership enabled health systems to recover service volumes during COVID-19. Resilient PHC requires simultaneous attention to bottom-up and top-down capabilities connected by strong leadership.