Abstract

Clinical research in respiratory tract infections (RTIs) is profoundly affected by non-pharmaceutical interventions (NPIs) against COVID-19. Clinical trials based on pre-pandemic data thus need to be put in question. We combined location-specific epidemiological RTI (seasonality) models and a mechanistic immunological model of individual viral RTI episode resolution. Adopting viral transmission so that the RTI disease burden during 2020 lockdown in the UK and China is reproduced, we simulated several NPI scenarios (none, mild, medium, strong=lockdown) and conducted in silico clinical trials in children with recurrent RTIs (RRTI) quantifying annual RTI rate distributions. As intervention, virtual patients received the bacterial lysate OM-85 described by a PBPK/PD approach. The model predicts that sample size (based on the RTI rates ratio) or the post-hoc power of fixed sample size trials are not majorly impacted under NPIs except for lockdown, while absolute benefit is impacted. Moreover, the mild NPI scenario already affected the simulated time to recruit for given eligibility criteria to select RRTI patients. Our model can help forecasting post-COVID-19 trial feasibility and advocates for gauging multiple different benefit metrics and for embracing innovative trial design as opposed to one based on pre-pandemic hypotheses.