Ann N. Burchell

Background: Self-report of human papillomavirus (HPV) vaccination has ~80–90% sensitivity and ~75–85% specificity. We measured the effect of nondifferential exposure misclassification associated with self-reported vaccination on vaccine effectiveness (VE) estimates. Methods: Between 2017–2019, we recruited sexually active gay, bisexual, and other men who have sex with men aged 16–30 years in Canada. VE was derived as 1−prevalence ratio × 100% for prevalent anal HPV infection comparing vaccinated (≥1 dose) to unvaccinated men using a multivariable modified Poisson regression. We conducted a multidimensional and probabilistic quantitative bias analysis to correct VE estimates. Results: Bias-corrected VE estimates were relatively stable across sensitivity values but differed from the uncorrected estimate at lower values of specificity. The median adjusted VE was 27% (2.5–97.5th simulation interval = −5–49%) in the uncorrected analysis, increasing to 39% (2.5–97.5th simulation interval = 2–65%) in the bias-corrected analysis. Conclusion: A large proportion of participants erroneously reporting HPV vaccination would be required to meaningfully change VE estimates.