Who should bear the cost of COVID-19 non-vaccination? An equity analysis using a luck egalitarian framework in a South African insured population
Year: 2026
Type: preprint
Study design: Observational Study
Abstract: Background Personal choice in health behaviours raises difficult questions of equity: when individuals freely decline effective health interventions, who should bear the resulting costs? This question sits at the heart of debates about personal responsibility in health financing, yet it has rarely been examined empirically. Luck egalitarianism offers a framework for assessing these questions: inequalities arising from brute luck (circumstance) should be compensated, while those from option luck (genuine choice) may legitimately be borne by individuals. Applying this framework requires distinguishing empirically between non-vaccination due to choice and non-vaccination due to constrained circumstance — a distinction that has rarely been operationalised in real-world health financing contexts. COVID-19 vaccination in South Africa's privately insured population, where vaccines were freely available and access barriers minimal, offers a favourable setting for such an inquiry. Methods Using de-identified claims data for approximately 550,000 members of two South African medical schemes (March 2020–December 2022), we estimated each individual's predicted probability of vaccination based on demographic and clinical characteristics using logistic regression. We compared observed and predicted uptake across probability strata to infer whether non-vaccination reflected personal choice or constrained circumstance. We then modelled excess costs attributable to non-vaccination using a zero-inflated negative binomial model, comparing observed expenditure with a counterfactual in which all members were vaccinated. Excess costs were calculated across predicted probability strata. Results Predicted and observed vaccination rates aligned closely, suggesting that residual non-vaccination in higher-probability groups reflected personal choice rather than constrained circumstance. Observed costs exceeded predicted costs by an average of 22% under a universal vaccination scenario, with the excess concentrated among older adults and those with comorbidities. Among younger, low-risk members, predicted costs slightly exceeded observed expenditure, as the fixed cost of vaccination was not offset by avoided hospitalisation. Conclusions The study empirically distinguished choice-based from circumstance-based non-vaccination and quantified the financial consequences of that distinction in the pseudo-laboratory setting of the study. This offers an empirical basis for assessing the equity of responsibility-based financing arrangements, and for deliberating when personal responsibility should inform health financing design.
Source: Research Square
Institution: University of Cape Town
Language: English
Subfield: Health
Field: Social Sciences
Domain: Social Sciences
Open Access status: green
Funder: University of Cape Town
OpenAlex ID: W7226000706
