Key data for outbreak evaluation: building on the Ebola experience

Cori, Anne; Donnelly, Christl A.; Dorigatti, Ilaria; Ferguson, Neil M.; Fraser, Christophe; Garske, Tini; Jombart, Thibaut; Nedjati-Gilani, Gemma; Nouvellet, Pierre; Riley, Steven; Van Kerkhove, Maria D.; Mills, Harriet L.; Blake, Isobel M.

For instance, case line-lists are likely to contain information on only a proportion of all infected individuals: typically those with symptoms, or those who sought care. The level of reporting may also be influenced by the capacity of the local health systems, which can vary over time and space. During the West African Ebola epidemic, less than a third of cases were estimated to be reported [122] and severe cases were probably over-represented compared to mild cases. At the end of 2014, health-care capacity was exceeded in many parts of Guinea, Liberia and Sierra Leone [14] , but new health-care facilities were subsequently built; hence the line-list of cases is likely to be more complete towards the end of the outbreak. Underreporting might also have been higher in this compared to previous Ebola outbreaks, during which the health-care systems were less overburdened. Systematic evaluation of the surveillance system [123] over different spatial units and time periods could help inform the level of underreporting. In addition, joint analysis of genetic sequence and surveillance data can provide insight into the degree of underreporting [124] . Quantifying completeness of, and potential biases in the line-list is important, e.g. to adequately quantify the CFR [28] . Although differences in the CFR were observed across rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 372: 20160371 different health-care facilities in the West African Ebola epidemic, it was not possible to determine whether these were due to reporting differences or underlying differences between settings [125] .