Imputation des Erkrankungs-Datums bei unvollständigen Daten im Rahmen der COVID19 Epidemie, Österreich

The date of disease onset in case reports of notifiable diseases is often missing or reported with delay.Then, a surrogate date, such as the date of the diagnostic laboratory report, has to be used for time seriesanalyses. We introduce a statistical model, which allows the imputation of the missing onset dates based onthe known difference between date of lab report and of disease onset, obtained from complete case reports.The model was applied on the Austrian COVID19 surveillance data. The mean difference between lab reportdate and disease onset was 5.4 days with variation by calendar week of the report: the more case reportsper week, the higher the difference. We found the Austrian COVID19 epidemic peaked on March 26, 2020,using lab report dates, compared to an epidemic peak on March 16, 2020, when using known and imputeddisease onset dates