58 Chapter 3 Limitations This approach also comes with limitations. First, AST and ALT are used as a proxy for hepatitis. In times when the causative agent is incompletely understood, syndromic surveillance for a proxy has utility, but when the cause is determined, surveillance of more specific test results is typically more useful. However, surveillance for AAV2 would be complicated given that it is not a routine laboratory test. Secondly, there are causes other than AAV2/ HAdV infection that could contribute to an increase in the proportion of elevated AST and ALT tests. Lockdowns and a sharp decrease in social interactions delayed exposure to many respiratory viruses in young children, and since many respiratory viral infections can lead to an increase in transaminases20, 21, the increase could also be caused by an increase in influenza virus, Respiratory Syncytial Virus (RSV) or Epstein-Barr Virus (EBV) infections. Using our data, it is not possible to differentiate between different causes of an increase in AST and ALT, although given the increase in HAdV in stool samples found in several countries, this is a plausible explanation22. Moreover, peaks in positive RSV or influenza virus tests in young children were only reported during a short period of the non A-E hepatitis outbreak, and not during the entire period23. Thirdly, our results only apply to the population of hospitalised children, which is likely to be different from the general population of 3-week to 5-year-olds. Fourthly, we decided to only include measurements from individuals without previously elevated ALT or AST measurements, which is a strict inclusion criterion. However, the number of children with multiple elevated ALT or AST measurements is likely to be limited, and therefore unlikely to influence our results or conclusion. Finally, the cutoff used in our study was 200 U/L, reflecting a mild to moderate elevation, but a more subtle increase in AST and ALT elevations could possibly have been picked up with a lower threshold but comes with a higher risk of false-positive results. Policy implications and interpretation We demonstrated that it is feasible to use anonymous, aggregated routinely collected clinical laboratory data for syndromic disease surveillance, and this approach could be used for new outbreaks of diseases with unknown aetiology, widespread poisoning incidents, or if there are few or no diagnostic tests available. Hospital laboratories collect large amounts of data on a daily basis that can potentially be of use for disease surveillance, but these are currently not optimally utilised. Completely anonymous data can be shared lawfully and safely between institutions, without the need for extensive contracting or additional information governance; although organisations must still have an appropriate legal basis, and in the UK additionally meet the common law duty of confidence, to process the routinely collected identifiable data into an anonymous ready format. Collecting summary data from different locations is an efficient way to aggregate large amounts of data and identify changes over time and in different geographical locations. Additionally, the emergence of large electronic databases containing routinely collected health data has made syndromic surveillance of communicable diseases significantly easier24. Moreover, because the data is routinely collected, there is little to no administrative burden for staff, which is especially relevant during outbreaks of infectious diseases. Finally,
RkJQdWJsaXNoZXIy MTk4NDMw