Application of Truncated Sequential Probability Ratio Test in Evaluating COVID-19 Vaccine Efficacy

• Activity Number: 123 - The Early Bird Catches the Worm: Sequential Decision-Making as a Framework in Time of COVID-19 Pandemic
• Type: Invited
• Date/Time: Monday, August 8, 2022 : 10:30 AM to 12:20 PM
• Sponsor: Caucus for Women in Statistics
• Author(s): Vladimir Dragalin*
• Companies: Janssen Pharmaceutical Companies of Johnson & Johnson
• Keywords: sequential probability ratio test; interim analysis; COVID-19 vaccine efficacy; truncated test; early stopping; futility

Abstract

The testing hypotheses for establishing vaccine efficacy can be reduced to the problem of testing hypotheses for the parameter of the binomial distribution, especially in the case of rare events. The number of experiments in the binomial distribution corresponds to the number of infections in the two groups combined, the number of infections in the vaccine group corresponds to the number of successes and p=(1-V)/(2-V) is the probability of success, where V is the vaccine efficacy. This allows to apply sequential designs on the scale of total number of accumulating infections, rather than on the total number of subjects enrolled. Phase 3 trial of Johnson & Johnson COVID-19 vaccine used the truncated sequential probability ratio test (TSPRT) allowing early stopping for efficacy or futility after every infection. In this presentation, different options for setting the early stopping boundaries and details for determining the truncation point will be considered. Comparison of the operating characteristics of the proposed TSPRT design with the other designs will be also presented.