Tuberculosis vaccine non-inferiority trials using an infection primary outcome must be designed and interpreted accounting for the specificity of the tools used to measure infection, otherwise they risk declaring non-inferiority by default.
Grint et al. (Wed,) studied this question.