A quantitative framework to define the end of an outbreak: application to Ebola Virus Disease

Bimandra A Djaafara; Natsuko Imai; Esther Hamblion; Benido Impouma; Christl A Donnelly; Anne Cori

Based on our simulation results, we recommend a new criterion using the day of onset of the last detected case as the baseline of the waiting time rather than the outcome day. The onset day is usually captured better and is less affected by diagnostic waiting times. Thus, it is also generalisable to other diseases. However, given the sensitivity of the probability of cases arising in the future to the value of , it is difficult to suggest a single criterion for EO irrespective of . Based on our results we therefore make three general recommendations: i) a shift from counting down to EO from 42 days from the day of outcome to 63 days from the day of onset of the last detected case leading to a preliminary EO declaration; ii) emphasis on the importance of properly resourcing the enhanced 90-day surveillance post-EO declaration to ensure that any flareups are quickly detected and controlled before the final EO declaration (63 + 90 days after the onset of the last detected case); and iii) regular estimation and re-estimation of the reproduction number particularly in the decline phase.