Comparison of methods to identify and characterize Post-COVID syndrome using electronic health records and questionnaires

Abstract

Background: Some of those infected with coronavirus suffer from post-COVID syndrome (PCS). However, an uniform definition of PCS is lacking, causing uncertainty about the prevalence and nature of this syndrome. We aim to improve understanding by operationalizing different definitions of PCS in different data sources and describing features and clinical subtypes. Methods: We use different methods and data sources. First, a cohort with electronic health records (EHR) from general practices (GPs) and GP out-of-hours-services combined with sociodemographic data for n≈1.000.000 individuals. Second, questionnaires among n=276 individuals who had been infected with coronavirus. Using both data sources, we operationalized definitions of PCS to calculate frequency and characteristics. In a subgroup of the EHR data we conducted community detection analyses to explore possible clinical subtypes of PCS. Results: The frequency of PCS ranged from 15-33%, depending on the method and data source. Across all methods and definitions, the mean age of individuals with PCS was around 53 years and they were more often female. There were small sex differences in the type of symptoms and overall symptoms were persistent for 6 months. Exploratory network analysis revealed three possible clinical subtypes. Discussion: We showed that frequency rates of post-COVID syndrome differ between methods and data sources, but characteristics of the affected individuals are quite stable. Overall, PCS is a heterogeneous syndrome affecting a significant group of individuals who need adequate care. Future studies should focus on care trajectories and qualitative measures such as experiences and quality of life of individuals living with PCS.

The impact of data extraction and processing on outcomes of research based on routine healthcare data from general practices: an observational study (Preprint)

“…This observational study was conducted in the context of the FAIR work packages of three larger COVID-19 related collaborations: COVID-GP [2,3,24], Long COVID Mixed Methods [25] and GRIP-3. [26,27] For these projects pseudonymized EHR data from general practices were stored and combined on the data platform of Statistics Netherlands.…”