Longitudinal Studies Assessing the Population-Level Impact of Human Papillomavirus Vaccination on Reducing Cervical Premalignant Lesions
Shing, Jaimie Zhi
The human papillomavirus (HPV) is the most common sexually transmitted infection. The HPV vaccine became available in 2006 and can prevent several adverse health outcomes and their associated morbidity and mortality. Examining trends in HPV-related outcomes, such as cervical premalignant lesions, including cervical intraepithelial neoplasia grades 2 and 3 and adenocarcinoma in situ (collectively referred to as CIN2+), demonstrates population-level vaccine impact. Assessing CIN2+ incidence in the United States is time-consuming and costly, as CIN2+ case confirmation requires cervical biopsy data, which are not included in most cancer registries or surveillance systems. Billing claims, such as international classification of diseases (ICD), can provide surrogate metrics but, have not been validated across the major coding transition from the 9th (ICD-9) to 10th (ICD-10) revision. Using several modeling approaches, CIN2+ claims-based models were built, validated, and compared between the ICD-9 era (January 1, 2008-September 30, 2015) and ICD-10 era (October 1, 2015-December 31, 2017). The model built by least absolute shrinkage and selection operator logistic regression performed well in both eras and did not significantly differ between eras. Using the validated model, age-group-specific CIN2+ incident trends were assessed among Tennessee Medicaid-enrolled women aged 18-39 years and the subset of those who were screened for cervical cancer to account for changes in screening trends. Urbanicity-stratified CIN2+ trends were examined using age-period-cohort models. From 2008-2018, CIN2+ incidence significantly declined in young women aged 18-20, 21-24, and 25-29 years, with overall stable trends in older women aged 30-34 and 35-39 years. Trends in CIN2+ incidence were similar between women living in urban versus rural counties. Trends were driven by age and cohort effects, likely from generational differences in HPV vaccine eligibility, vaccination behaviors, and screening recommendations. Patterns were consistent after restricting to screened women, suggesting declines in CIN2+ incidence were not entirely due to decreases in screening. These results demonstrate the feasibility of claims data to examine HPV vaccine impact on reducing CIN2+ incidence.