Chapter 9 288 Outcome definition First and foremost, no case definition for long COVID that is informative enough for both research and clinical practice has yet been agreed upon. This is perhaps not surprising given that it is difficult to strike a balance between sensitivity (capturing all possible phenotypes, mainly for exploratory research) and specificity (allowing for in-depth evaluation of specific sub-types) when researching an emerging condition. Existing definitions focus mainly on the duration of a series of diverse symptoms[50, 112, 113], and generally do not differentiate in detail between the type, course (fluctuating or constant), severity and objective measurement of symptoms. Going forward, it is likely that definitions of various subtypes of long COVID (for instance, based on organ system) may arise to allow for greater focus on specific pathologies and clinical clusters. Unifying features under the umbrella term of long COVID may include a list of “core symptoms”[55, 114] to act as an overarching criterion. In the process of improving the case definition, it is crucial to take a patientcentred approach to long COVID research[115, 116] through qualitative or mixed-methods studies[117], to ensure that future definitions reflect the lived experience of long COVID. Study design and data analysis In Section 9.2.2., the challenges in drawing conclusions from studies without an appropriate control group have been extensively discussed. Below, we briefly highlight additional biases in study design and data analysis that contribute to why the knowledge gaps presented in Section 9.2.2 exist. A major challenge in long COVID study design is ensuring that different sub-populations are adequately well represented. This is of vital importance to determine the external validity of study findings. Individuals living in remote rural communities, with lower SES status or those with occupational or care-related obligations may be less likely to enrol in a time-consuming, intensive prospective cohort studies on COVID-19. As such, the vast majority of long COVID research has been disproportionately representative of highly educated and non-migrant populations in urban, high-income settings[118, 119]. Identification of eligible study participants from notification data or population registries also introduces selection bias based on uptake of testing, possibly with opposite effects in different settings[120]. It therefore is unknown whether the clinical presentation of long COVID varies among different populations with diversity in COVID-19 burden, underlying population health, culture, health beliefs and levels of engagement with healthcare services. Both funders of research and researchers themselves should ensure that inclusivity is prioritised in future studies.
RkJQdWJsaXNoZXIy MTk4NDMw