Hypotheses Tests of Strain-specific Vaccine Efficacy Adjusted for Covariate Effects

Abstract

In the evaluation of efficacy of a vaccine to protect against disease caused by finitely many diverse infectious pathogens, it is often important to assess if vaccine protection depends on variations of the exposing pathogen. This problem can be formulated under a competing risks model where the endpoint event is the infection and the cause of failure is the infecting strain type determined after the infection is diagnosed. The strain-specific vaccine efficacy is defined as one minus the cause-specific hazard ratio (vaccine/placebo). This paper develops some simple procedures for testing if the vaccine affords protection against various strains and if and how the strain-specific vaccine efficacy depends on the type of exposing strain, adjusting for covariate effects. The Cox proportional hazards model is used to relate the cause-specific outcomes to explanatory variables. The finite sample properties of proposed tests are studied through simulations and are shown to have good performances. The tests developed are applied to the data collected from an oral cholera vaccine trial.Competing risks model, cause-specific hazard function, Cox proportional hazards model,

Testing and Estimation of Time‐Varying Cause‐Specific Hazard Ratios with Covariate Adjustment

“…It is more efficient to use the standard Cox model for the assessment of WC vaccine and for the age group < 5 (see Hyun and Sun, 2007). The methods developed in are applicable nonetheless and some results are listed here for ease of comparisons between younger and older age groups.…”

“…If the covariates are time independent, the Cox model stipulates that the cause‐specific hazard rates for different values of the covariates are proportional. Under this model Hyun and Sun (2007) developed some covariate‐adjusted tests for strain‐specific vaccine efficacy, defined as one minus the cause‐specific hazard ratio (vaccine/placebo). However, the proportional hazards assumption is often violated when treatment effects are investigated, for example, due to time‐varying treatment efficacy.…”