BACKGROUND The COVID-19 pandemic underscored the need and value of a standardised and timely surveillance system for severe acute respiratory infections (SARI) to inform epidemic preparedness and response. AIM We aimed to develop an automated SARI surveillance system using electronic health records retrieved from pre-existing national health registers in Denmark. METHODS We used the Danish Civil Register, the Danish National Patient Register and the Danish Microbiology Database to set up the system. First, we determined a SARI case definition for surveillance, choosing among six different potentially usable combinations of ICD-10 diagnosis codes by exploring how each combination captured patient characteristics (age, hospital admission length, mortality, laboratory tests and seasonality). Second, using this case definition, we evaluated the surveillance system’s timeliness and completeness by comparing weekly data reported with a delay of 1, 8, 15, 22 and 29 days, respectively, against a complete set of data extracted after 120 days. RESULTS The selected case definition combined ICD-10 codes for influenza (J09-J11), acute lower viral and bacterial respiratory tract infections and bronchiolitis (J12-J22) and COVID-19 (B342A and B972A). With regards to timeliness and completeness of this definition, weekly data reported with a delay of 8 days was 89–93% complete and showed very similar patterns in weekly changes in SARI cases as data reported after 120 days. CONCLUSION Our SARI surveillance system detected fluctuations in weekly SARI cases in a consistent and timely manner. We recommend countries to explore using electronic health registers as a resource-efficient alternative to standard SARI sentinel surveillance.
Lomholt et al. (Thu,) studied this question.