Estimating incidence rates of grouped HPV types: A systematic review and comparison of the impact of different epidemiological assumptions

Abstract Background Some studies on human papillomavirus (HPV) provide not only type-specific incidence rates (IR), but also IRs of HPV groupings (e.g. the nonavalent grouping). We made an inventory of the different approaches used to calculate such IRs and assessed their impact on the estimated IRs of HPV groupings. Methods We performed a systematic review assessing all approaches used in literature to estimate IRs. Subsequently we applied these approaches to data of a Dutch cohort study on HPV in men who have sex with men (H2M). IRs were estimated for six different HPV groupings. Results The systematic review yielded six different approaches (A-F) for estimating the IRs, varying in exclusion criteria at baseline, the definition of an incident event and person-time. Applying these approaches to the H2M dataset (n = 749), we found differences in the number of participants at risk, number of incidents events, person-time, and IR. For example, for the nonavalent grouping, depending on the approach chosen, the IR varied between 3.09 and 6.54 per 100 person-months. Conclusion In published studies different epidemiological assumptions are used to estimate IRs of grouped HPV types, leading to widely differing estimates of IRs. IRs between different studies may therefore not be comparable.