Covid-19 risk factors: Statistical learning from German healthcare claims data

We analyse prior risk factors for severe, critical or fatal courses of Covid-19 based on a retrospective cohort study using claims data of the AOK Bayern. As a methodological contribution, we avoid prior grouping and pre-selection of candidate risk factors and use fine-grained hierarchical information from medical classification systems for diagnoses, pharmaceuticals and procedures, using more than 33,000 covariates. Our approach is competitive to formal analyses using well-specified morbidity groups without needing prior subject-matter knowledge. The methodology and our published coefficients may be of interest for decision makers when prioritizing protective measures towards vulnerable subpopulations as well as for researchers aiming to adjust for confounders in studies of individual risk factors also for smaller cohorts.