Chapter 1 22 biomarkers or distinct histopathological properties for the condition. Finally, there is a striking lack of long COVID studies originating from low-resource settings, and research conducted in high-resource settings is not always representative of marginalised communities. Given these drawbacks, it is likely that we have only scratched the surface of our knowledge on long COVID. Further investigation of this ‘silent pandemic’[68] will require years of dedicated and collaborated efforts, during which time an urgent need exists for support and information for the individuals currently living with long COVID. 1.4 DATA SOURCES COVID-19 surveillance data of notified cases In Chapter 2 and Chapter 3 of this thesis, routinely collected COVID-19 notification data linked to municipality records were used. Notification data were collected as laboratories and hospitals alerted the regional Public Health Service of Amsterdam (PHSA, or “GGD Amsterdam” in Dutch) of any positive SARS-CoV-2 diagnosis, in accordance with Dutch Public Health Law (see Section 1.2). Following notification, the contact tracing team gathered additional information (such as date of illness onset, hospitalisation status, occupation) by telephone. For the purpose of our research, these notification data were consequently matched to registration data from the municipality records of the City of Amsterdam (BRP) to retrieve postal code of residence and the country of birth of a notified individual and their parents, in order to infer migration background. It is important to note that notification data, similarly to all forms of data collected via passive surveillance, are affected by selection bias. During the time period presented in Chapter 2 and Chapter 3 of this thesis (February 2020 to January 2021), for example, the criteria for eligibility for testing varied considerably. During the first COVID-19 wave, described in Chapter 2, tests were restricted to severely-ill individuals and healthcare professionals. Following 1 June 2020 (this thesis, Chapter 3), access to testing services was expanded but remained susceptible to selection bias from access and intention to testing. On the other hand, criteria for admission to hospital are based on more objective parameters and are therefore less prone (although not fully) to changes over time. We therefore utilised rates of COVID-19 hospitalisation instead of SARS-CoV-2 infection as a less biased measure by which to compare COVID-19 burden between groups. Please note that individuals with a migration background are categorised into ‘Western’ and ‘non-Western’ in Chapter 2 of this thesis, but as ‘European background’ and ‘nonEuropean background’ in Chapter 3. Around the time of data analysis, the Dutch Central Bureau for Statistics (CBS) was in the process of reclassifying migration background in order to make it more objective, focussed on the individual rather than their parents,
RkJQdWJsaXNoZXIy MTk4NDMw