Measles epidemic in Yakusu Health Zone, Democratic Republic of Congo, 2018–2019: a retrospective observational study

A large measles outbreak occurred in Yakusu Health Zone, Democratic Republic of the Congo, between January 2018 and June 2019. We aimed to estimate the magnitude, mortality burden, and vaccination coverage of this epidemic, and to evaluate whether the reactive mass vaccination campaign reached previously unvaccinated children. We conducted a retrospective two-stage cluster household survey to estimate the attack rate (AR), case fatality ratio (CFR), mortality rates, and vaccination coverage. We extrapolated survey-weighted estimates to the population of Yakusu to estimate true outbreak size. Kaplan-Meier survival analysis characterised the temporal distribution of post-measles mortality. Geospatial analyses described the spatial distribution of vaccination coverage and attack rates, and Spearman rank correlation quantified the spatial relationship between routine and campaign coverage at the cluster level. Among 8,968 individuals surveyed, 1,390 (15.5%, 95% CI: 13.2–17.8) reported measles during the recall period. The estimated true outbreak size was 22,068 cases (95% CI: 18,785–25,351), suggesting routine surveillance detected only 27.9% of cases. Among children under 5 years, the AR was 45.7% (95% CI: 39.6–51.9) and the 31-day CFR was 4.9% (95% CI: 3.4–6.3). Among infants under 1 year, survival at 31 days post-onset was only 78.7%, with deaths continuing to accrue beyond this point, reflecting a distribution of onset-to-death intervals that extends beyond the standard attribution cutoff. Routine EPI and campaign vaccination coverage were each approximately two-thirds of the target population (68.2% and 67.7%, respectively). Over half of the target age group (53.5%, 95% CI: 46.8–60.3) had received both EPI and campaign vaccination, while 19.2% had received neither. Among children vaccinated during the campaign, 81.6% had already received routine EPI vaccination, while only 39.8% of previously unvaccinated children were reached. EPI and campaign coverage were strongly spatially correlated at the cluster level (Spearman ρ = 0.59, p < 0.001). This was a severe epidemic with high attack rates and case fatality, particularly among infants. The reactive vaccination campaign, implemented more than one year into the outbreak, overwhelmingly reached children already vaccinated through routine services while failing to reach the majority of unvaccinated children. Reactive campaigns in similar settings should prioritise identifying and reaching unvaccinated populations through improved microplanning and spatial targeting, rather than maximising total doses administered.