Using data to improve public health: COVID-19 secondment

The fellowship will focus on making the best use of data from multiple sources to better understand the impact of COVID-19 on population health in the UK. The proposed research includes, 1. Identifying risk factors associated with Long COVID. o Use data from multiple sources, such as longitudinal cohort studies and/or electronic health records, to develop a prediction model to classify patients into groups with and without Long COVID. o Use unsupervised learning to further identify and define the subgroups of patients with Long COVID. 2. Evaluating vaccine effectiveness and safety using real-world (out of trial) data. o Evaluate and compare the effectiveness of different methods of vaccine offering: one dose, two doses of the same brand, and mixed brands of vaccine. o Evaluate vaccine safety, such as possible blood clots and heart inflammation, by using regression models for rare adverse events to better identify subgroups of patients at high risk. 3. Quantifying healthcare disruptions during different waves of the pandemic. o Analyse the data related to COVID-19 healthcare activity, to identify how the healthcare burdens change in different waves of the pandemic. o Investigate what factors are associated with healthcare burdens, what are the contributing factors in heterogeneity, and identify good practices in mitigating healthcare disruptions. 4. Assessing the effects of COVID-19 infection on other health outcomes. o Use electronic health records and/or multiple longitudinal cohort studies, to examine how COVID-19 infection impacts on other health outcomes. o Quantify how the effects of COVID-19 infection on health outcomes change with time. o Identify subgroups of the population who are more vulnerable to adverse health outcomes associated with COVID-19 infection. Well documented statistical code will be produced to allow for future updated analysis when new data become available, and for reproducible research.