Summary

A surveillance-based modeling study estimates that a 2020 mass oral cholera vaccination campaign in Uvira, Democratic Republic of the Congo, averted 56% of infections and deaths over the following three years. The preprint estimates that 2,350 facility-attended cases and 44 facility and community deaths were prevented.

A 2020 mass oral cholera vaccination campaign in Uvira, Democratic Republic of the Congo, is estimated to have averted 56% of the infections and deaths that would otherwise have occurred over the next three years. The estimate comes from a surveillance-based counterfactual modeling analysis posted as a medRxiv preprint.

The campaign reached 66% of Uvira's population with at least one dose of the oral cholera vaccine Euvichol Plus. The model translates the estimated impact into a median of 2,350 facility-attended cases and 44 facility and community deaths averted. The reported 95% credible interval for the overall reduction was 34% to 81%.

Contents

What the analysis combined

The researchers collected enhanced cholera surveillance data from Uvira's two official cholera treatment facilities between January 2017 and December 2023. The surveillance included systematic rapid diagnostic testing of suspected cases.

The team also conducted repeated, representative population surveys. These recorded healthcare-seeking behaviour, vaccination status, population mobility and antibody profiles. The antibody data were used to estimate seroincidence, an antibody-based measure of new infections in a population.

These data were integrated in a Bayesian framework with an ensemble of mechanistic cholera transmission models. The models incorporated changes in transmission over time, the development and loss of immunity, and the effect of population turnover on vaccination coverage.

To estimate the campaign's effect, the researchers compared the observed setting with modeled counterfactual vaccination scenarios. A counterfactual is an analytically constructed scenario used to estimate what might have happened under a different intervention level.

Estimated impact of the 2020 campaign

The analysis estimated that the campaign prevented 56% of infections and deaths over the three years after vaccination. In absolute terms, the median estimate was 2,350 fewer cases treated at facilities, with a 95% credible interval of 890 to 8,960 cases. The estimated median reduction in facility and community deaths was 44, with a 95% credible interval of 16 to 150.

The researchers also modeled a scenario in which the entire eligible population of Uvira was vaccinated. That scenario produced a median reduction of 64% in cases and deaths, with a 95% credible interval of 43% to 87%.

The credible intervals show the range of model estimates supported by the data and assumptions used in the Bayesian analysis. They are not a direct count of prevented infections or deaths, because the study estimated the campaign's effect through surveillance data and transmission modeling rather than by assigning vaccination in a randomized trial.

Why coverage and infrastructure mattered

The model indicates that imperfect campaign coverage, population turnover and high transmission rates reduced the larger potential benefit of oral vaccination. When the model used half the population turnover, the estimated median reduction increased to 71%, with a 95% credible interval of 52% to 90%.

Population turnover can reduce coverage over time as people enter or move through the population without receiving the campaign dose. This helps explain why the modeled benefit of vaccinating the entire eligible population was higher than the estimated impact of the campaign that reached 66% of residents.

The authors conclude that mass vaccination prevented a substantial share of cholera cases in this endemic setting, but that sustained control in Uvira requires a multisectoral approach. In particular, the analysis identifies safe water and sanitation alongside vaccination as important components of cholera control.

The evidence is specific to Uvira and comes from a preprint reporting model-based estimates. Its main uncertainty is reflected in the wide credible intervals and in the results' dependence on assumptions about transmission, immunity, healthcare use and population movement.

Sources