Objective: With the emergence of coronavirus disease 2019 (COVID-19) countries had to establish COVID-19 surveillance by adapting existing systems, such as the mandatory notification system and syndromic surveillance systems. We estimated COVID-19 hospitalization and ICU burden from existing severe acute respiratory infections (SARI) surveillance and compared it to COVID-19 notification data.
Methods: Using data from the ICD-10 based hospital sentinel, we estimated SARI incidence and validated our estimations retrospectively using full population data. We estimated age-specific incidences for COVID-19 hospitalization and ICU for the first three COVID-19 waves in Germany, based on information of SARI cases (COVID-SARI). We compared these estimations to hospitalization and ICU burden of COVID-19 from notification data and described potential underreporting of hospitalizations in notification data.
Findings: The estimation of SARI incidence from sentinel data corresponded very well to full population data. The estimated COVID-SARI incidence matched the notified COVID-19 hospitalization incidence in the first wave, but was much higher during the second and the third wave. The proportion of unknown hospitalization status among notified COVID-19 cases was much higher in the later waves compared to the first wave. For all waves, the ICU incidence estimated from COVID-SARI was considerably higher than the results from notification data.
Conclusion: The use of SARI sentinel data adds valid and important information for assessing COVID-19 hospitalization and ICU burden, especially in times with high case numbers where notification data may be more incomplete.