The feasibility of pragmatic influenza vaccine randomized controlled real-world trials in Denmark and England
Joshua Nealon, Daniel Modin, Rebecca Ghosh, Deborah Rudin, Gunnar Gislason, Helen Booth, Jens‐Ulrik Stæhr Jensen, Rachael Williams, Hilary Shepherd, Eleanor Yelland, Hélène Bricout, Sandra S. Chaves, Tor Biering‐Sørensen
Abstract
We estimated the frequency of non-specific influenza-associated clinical endpoints to inform the feasibility of pragmatic randomized controlled trials (RCT) assessing relative vaccine effectiveness (rVE). Hospitalization rates of respiratory, cardiovascular and diabetic events were estimated from Denmark and England's electronic databases and stratified by age, comorbidity and influenza vaccination status. We included a seasonal average of 4.5 million Danish and 7.2 million English individuals, 17 and 32% with comorbidities. Annually, approximately 1% of Danish and 0.5% of English individuals were hospitalized for selected events, ~50% of them respiratory. Hospitalization rates were 40-50-fold and 2-10-fold higher in those >50 years and with comorbidities, respectively. Our findings suggest that a pragmatic RCT using non-specific endpoints is feasible. However, for outcomes with rates <2.5%, it would require randomization of ~100,000 participants to have the power to detect a rVE difference of ~13%. Targeting selected groups (older adults, those with comorbidities) where frequency of events is high would improve trial efficiency.