A Quantitative Framework to Define the End of an Outbreak: Application to Ebola Virus Disease
Abstract The end-of-outbreak declaration is an important step in controlling infectious disease outbreaks. An objective estimation of the confidence level that an outbreak is over is important to reduce the risk of post-declaration flare-ups. We developed a simulation-based model to quantify that confidence. We tested it on simulated Ebola Virus Disease data. We found these confidence estimates were most sensitive to the instantaneous reproduction number, the reporting rate, and the time between the symptom onset to death or recovery of the last detected case. For Ebola Virus Disease, our results suggest that the current World Health Organization criterion of 42 days since the recovery or death of the last detected case is too short and sensitive to underreporting. Therefore, we suggest a shift to a preliminary end-of-outbreak declaration after 63 days from the symptom onset day of the last detected case. This preliminary declaration should still be followed by 90 days of enhanced surveillance to capture potential flare-ups of cases, after which the official end-of-outbreak can be declared. This sequence corresponds to more than 95% confidence that an outbreak is over in most of the scenarios examined. Our framework is generic, and therefore could be adapted to estimate end-of-outbreak confidence for other infectious diseases.