Contrasting the impact and cost-effectiveness of successive intervention strategies in response to Ebola in the Democratic Republic of the Congo, 2018-2020

Abstract

Introduction: The 10th outbreak of Ebola Virus Disease (EVD) in the Democratic Republic of the Congo (DRC) in 2018-2020 was the largest in DRC's history and the second largest worldwide. Different strategic response plans (SRPs) were implemented, and the outbreak was eventually stopped after a large scale-up of operations with the SRP 4, which benefited from all public health measures deployed during SRPs 1-3, upon which it developed a more holistic approach including community engagement, logistics and security.

Methods: We used modelling to characterise EVD transmission and assess the epidemiological impact of the two main response strategies (SRPs 1-3 vs SRP 4). We simulated potential future epidemics with different intervention scenarios, combined with a costing model to evaluate the incremental cost-effectiveness of different strategies.

Results: We estimated a mean effective reproduction number R of 1.19 (credible interval (95% CrI) = (1.13 ; 1.25)). The spatial spread was moderate with an average 4.4% (95% CrI = (3.5%; 5.4%)) of transmissions moving to different health zones. The scale-up of operations in SRP 4 coincided with a threefold reduction in transmission, and 30% faster control of EVD waves. In simulations, SRP 4 appears cost-saving, although most simulated outbreaks remain small even with SRPs 1-3.

Conclusion: Most EVD outbreaks are expected to be small and can be contained with SRPs 1-3. In outbreaks with increased transmissibility or in the presence of insecurity, rapid scale-up to SRP 4 is likely to save lives and be cost-effective.

Keywords: Health systems evaluation; Mathematical modelling; Viral haemorrhagic fevers.

Figure 1. Epidemic curves of the five most affected health zones. This graph shows the weekly incidence by date of symptom onset of the five health zones with most reported cases. Colours indicate the different waves in each health zone. As waves were identified from daily incidence, some weeks may include overlapping waves. The vertical dashed line marks the change of strategy from SRPs 1–3 to SRP 4, on 15 July 2019.

Figure 2. Estimation of transmission parameters of Ebola Virus Disease in North Kivu/Ituri. This figure summarises the estimation of the model parameters using MCMCs, using 1960 posterior samples after removing burn-in. (A) Posterior density of the joint distribution of the reproduction number (R) during growth phases and the dispersal parameter (ρ) expressed as the percentage of infections moving to different health zones; warmer colours indicate higher density. (B) Marginal posterior distribution of the reproduction number R. (C) Marginal posterior distribution of the dispersal parameter ρ. (D) Marginal distributions of the strategy efficacies for SRPs 1–3 and SRP 4, expressed as per cent reduction of R during the control phases.

Figure 3. Results of simulated Ebola Virus Disease (EVD) outbreaks. Summary of simulated EVD outbreaks under two scenarios, first contrasting strategic response plan (SRP) 1–3 and SRP 4 strategy (A, B), then studying the impact of different delays to implementing SRP 4 after SRPs 1–3 (C, D). Results are reported as probabilities of simulated outbreaks being at least certain sizes, expressed as total number of patches affected out of a maximum of 50 (A, C), and as total number of cases (B, D). Each curve was derived from 10 000 simulations.