Effects of out-of-pocket health expenditure on universal health service coverage: evidence from panel data from low-, middle-, and high-income countries

Abstract Background Achieving Universal Health Coverage (UHC) by 2030, a key SDG target, requires progress in both service coverage and financial protection. Out-of-pocket (OOP) financing remains an important component of health-financing systems, but its association with UHC service coverage may differ across country income groups. Objective To investigate how OOP health expenditure as a share of current health expenditure is associated with UHC service coverage across low-income countries (LICs), middle-income countries (MICs), and high-income countries (HICs), while accounting for demographic, economic, governance and temporal factors. Methods We analyzed a balanced panel of 147 countries from 2015 to 2023. The outcome was the UHC service coverage index (UHCSCI) and the main exposure was OOP health expenditure as a percentage of current health expenditure, representing the composition of national health financing rather than household-level OOP burden or per-capita OOP expenditure. GDP per capita, population size, elderly population (65+), external health expenditure, and the Corruption Perception Index (CPI) were also used as covariates. The fractional logit generalized estimating equation (GEE) model with an AR(1) correlation structure was employed to estimate the population-averaged associations between OOP health spending and UHCSCI across LICs, MICs, and HICs. Results Descriptive findings revealed that UHCSCI achievements in 147 countries ranged from 23% to 92%, with an average of 65.28%. The main exposure OOP expenditure ranged from 2.98% to 84.79%, with an average of 34.02%. The GEE estimates showed that the association between OOP expenditure and UHCSCI varied across country income groups. GDP per capita, the elderly population, external health expenditure, and governance were also associated with UHCSCI. The average marginal association of OOP was progressive, negative in LICs, nearly zero in MICs, and a small positive in HICs. Conclusions The association between OOP financing and UHC service coverage underscores the need to reduce dependence on OOP payments to improve UHC service coverage, particularly in LICs. The heterogeneous association across country income groups highlights the need to consider both financing composition and country context when assessing progress toward UHC. Considering the observational design, the estimates represent population-averaged associations rather than causal effects and do not identify the impacts of specific health financing reforms.