Using non-targeted HPV infections in studies with risk compensation

Area of Science:

• Epidemiology
• Vaccinology
• Biostatistics

Background:

• Estimating human papillomavirus (HPV) vaccine efficacy typically involves tracking infections from both targeted and non-targeted HPV strains.
• Observational studies assessing vaccine efficacy face challenges like confounding bias and risk compensation, unlike blinded randomized controlled trials.
• A previous method proposed using non-targeted HPV strains to mitigate confounding bias in efficacy estimates for targeted strains.

Purpose of the Study:

• To investigate the causal interpretation of a method using non-targeted HPV strains for vaccine efficacy estimation when unmeasured sexual behavior acts as both a confounder and mediator.
• To determine if this method can accurately isolate the direct immunological effect of the HPV vaccine.

Main Methods:

• The study extends a previously proposed method that utilizes infections with non-targeted HPV strains.
• It incorporates a setting where unmeasured sexual behavior functions as both a confounder and a mediator.
• Mathematical and statistical modeling is employed to analyze causal pathways.

Main Results:

• Under specific assumptions, using non-targeted HPV infections can effectively remove confounding bias.
• The method also accounts for the portion of vaccine effect mediated through behavioral changes.
• The estimated quantity accurately represents the direct immunological effect of the HPV vaccine.

Conclusions:

• The refined method provides a clearer causal interpretation of HPV vaccine efficacy by addressing behavioral confounders and mediators.
• Non-targeted HPV infections serve as a valuable tool for isolating both direct immunological and indirect behavioral effects of the vaccine.
• This approach enhances the reliability of vaccine efficacy studies, particularly in observational settings.