201 Summary and general discussion 9 different geographical locations, while safeguarding privacy. This approach could be used both for continuous surveillance of known pathogens, like SARS-CoV-2 and disease syndromes, like SARI, but also in the case of new outbreaks with unknown pathogens (as with the hepatitis outbreak). While systems for continuous surveillance are in place nationally and globally for some pathogens, like SARS-CoV-27, a unified approach would also improve data quality. Robust surveillance systems enable easier outbreak prediction, monitoring and detection, and allow for evaluation of different implementations, such as vaccination strategies4,8. As mentioned above, a distinction can be made between disease surveillance for known pathogens or disease syndromes, like SARS-CoV-2 and SARI, and outbreak detection of unknown pathogens. Of course, pathogens that are now known were at one point unknown. Outbreak detection is a complex and elaborate process9, and giving a full overview is outside the scope of this thesis. Historically, outbreak detection depends on healthcare workers reporting unusual cases, but this has several disadvantages10,11. Most importantly, recognising atypical numbers or new patterns is complex, especially when there are still few cases, while the early stage is vital for disease control12. Alternatively, (automated) collection of non-specific indicators, like fever or the number of emergency room visits, could be used as proxies for outbreak detection10. Moreover, a combination of these methods with big data sources and big data analytics, could further improve outbreak detection, by adding new data sources (e.g., from animal and environmental health sources) and by including data from otherwise hidden populations13. After detection of an outbreak, there is a need for ‘outbreak characterisation’. Some of the main aims of outbreak characterisation are to set a case definition, identify the pathogen, elucidate the route of transmission, incubation and infectious period, and estimate the disease burden across the disease severity pyramid9. The number of subclinical and mild infections in an outbreak is typically much larger than the number of clinical infections, and the number of hospitalised infections is typically larger than the number of deaths5. This distribution is commonly represented as a severity pyramid, where the apex corresponds to the number of deaths and the base reflects the number of asymptomatic infections. While estimating the number of subclinical infections is complex, it can have important consequences for the timing of policy implementations14. Our study attempting to identify the presence of a larger outbreak of mild cases of hepatitis in children (Chapter 3) is an example of outbreak characterisation, using anonymised summary data. After outbreak detection and characterisation, there is a need for longer term surveillance systems. Important aims of disease surveillance, as discussed in the introduction, are to describe the current status and burden of disease, and to monitor trends15. Surveillance can be active (such as sentinel influenza surveillance at multiple general practices) or passive (such as the SARI surveillance study in Chapter 215. Surveillance can be done using aggregated data, as demonstrated in Chapter 2 and Chapter 3 or by collecting information from individual cases15. Both approaches
RkJQdWJsaXNoZXIy MTk4NDMw