35 Surveillance of Severe Acute Respiratory Infections 2 codes especially, there is often a delay in registration, which could have a significant impact in a prospective setting. Thirdly, ICD-10 codes were only included if the same ICD-10 code was not registered within the previous year, to avoid the inclusion of patients with chronic disease. This could lead to a small underestimation of the number of SARI cases based on ICD-10 registrations. Finally, analysing data from multiple hospitals could confirm whether these indicators can be used for surveillance of SARI. As hospitals protocols may could change over time, regular validation of proxy indicators is essential. While the availability of contact and droplet precaution labels that can be extracted from EMRs may differ between countries and hospitals, our approach demonstrates that there are suitable proxies for SARI surveillance. Conclusion The number contact and droplet precaution labels and ICD-10 codes are suitable for (automated) SARI surveillance in our cohort. PCR test results provide valuable additional information. Because of significant changes in public awareness and hospital testing policy, the number of RT-PCR tests is not considered a reliable indicator. As contact and droplet precautions best reflect pressure on hospital capacity, do not have a delay in reporting and registration policies are less likely to change over time, this parameter may be the most suitable indicator. Validating our results in different hospitals and in a prospective setting could confirm the feasibility of this indicator for SARI surveillance. An important future objective would be set up a national SARI surveillance system, collecting data from various hospitals across the country.
RkJQdWJsaXNoZXIy MTk4NDMw