Gillian D. Sanders, PhD, Duke University, Durham, NC and Scott C. Schmidler, PhD, Duke University, Durham, NC.
PURPOSE: Human papillomavirus (HPV) is the major risk factor for cervical cancer with HPV 16/18 accounting for ~70% of cancers. Initial trials of prophylactic vaccines against HPV 16/18 have had promising results. How potential vaccination implementation strategies will affect the US burden of disease is uncertain. METHODS: We developed a dynamic disease transmission model of HPV and sexual behavior which simulated a population's HPV 16/18 incidence. The model transmission and mixing rates were calibrated to match published prevalence data. Boys and girls entered the model at 12-years old uninfected with HPV. Based on sexual mixing between uninfected and infected individuals, efficacy and duration of the vaccine's protection, and transmission of HPV between sexual contacts, individuals could contract HPV over their lifetime. Individuals in the population who were not vaccinated could be protected through the effects of herd immunity. We assessed the effect on HPV incidence of vaccination programs which varied in vaccine efficacy (70-100%), targeted population (both sexes, girls-only), coverage rates (30-100%), and duration of protection (5-10 years, lifetime). RESULTS: We validated the size and shape of our model's HPV 16/18 predicted prevalence against the published literature. Prior to vaccination, age-specific HPV 16 incidence varied between 0.08-2.6% and HPV 18 incidence varied between 0.12-0.8% with the highest incidence occurring between ages 21-24. Assuming lifetime immunity with vaccination, if the vaccine reached 70% of 12-year old girls and provided 90% efficacy against HPV 16/18 – vaccinated girls experienced up to a 98% reduction in HPV 16/18. Through the benefits of herd immunity, unvaccinated girls also experienced a reduction with ~80% of infections being eliminated. If the vaccine was only able to reach 30% of girls, this reduction was lessened to 92% and 31% in vaccinated and unvaccinated girls respectively. If widespread coverage of girls is possible, inclusion of boys in a HPV 16/18 vaccination program provided small additional reductions in incidence. Inclusion of boys in vaccine programs was more considerable if vaccine immunity or efficacy were reduced or widespread coverage of girls was not feasible. CONCLUSIONS: Dynamic disease transmission models allow estimation of the natural history of HPV infections and evaluation of the effects of vaccination programs on outcomes. As policy makers explore how best to distribute new vaccines, such models can help inform decision making.