Abstract
Clinical trial designs that address multiple questions more efficiently are desirable. When we designed a hepatitis B vaccine trial to assess seroprotective outcomes in persons with HIV (PWH), we aimed for an efficient design that addressed three primary objectives in two study populations. The study focused on: PWH who did not respond to prior HBV vaccination, and PWH with no known history of HBV vaccination. Whereas one vaccine regimen was studied in the vaccine-naïve participants, multiple interventions were considered for those with prior nonresponse, with two different vaccines and two dosing schedules. Several features of the trial design required statistical considerations related to multiple testing: (1) assessment of vaccine response in two study populations under one trial, (2) comparisons among multiple treatment arms, and (3) sequential repeated significance tests in interim data monitoring. We describe the features aimed to gain statistical and administrative efficiencies, including reduction in the study sample size of 12 %. We also describe how we controlled type I error and planned interim data monitoring, and highlight the time lag issue due to the laboratory-based immunogenicity endpoint in this international, multi-center trial (NCT04193189).