Keywords: binary, vaccines, equivalence, noninferiority
Clinical immunogenicity data in vaccines are commonly dichotomized into responders and non-responder binary data prior to and following vaccination. Given the `leaning’ of the bio-pharmaceutical industry (Drug Discovery Today 2016; 21: 379-384) increasingly it is the case that multiple antigen vaccines are being developed and positive-controlled trials are applied, resulting in intersection-union tests for inferential non-inferiority or equivalence testing. This lecture will review statistical analysis methods and evaluate coverage probabilities for methods used to analyze clinical immunogenicity data, summarize a commonly applied method used to assess equivalence and non-inferiority, and illustrate this method using anonymized clinical data and simulations. Particular attention will be placed upon the implications and use of imbalanced randomization in study design.