Patient- and health system-level impacts of home-based HPV screening for cervical cancer

Abstract

Cervical cancer screening is associated with substantial global reductions in cervical cancer incidence and mortality. However, in the U.S., screening adherence among individuals 21-65 years of age declined from 86% to 72% between 2005 and 2021. HPV self-sampling (HPV-SS), or home-based primary HPV screening, addresses well-documented screening barriers and could play a critical role in the U.S. effort to increase screening coverage. However, little is known about HPV-SS in a U.S. setting. This dissertation addresses the impacts on both U.S. patients and health systems of offering home-based HPV screening interventions. The first chapter examined adherence to clinical follow-up among individuals overdue for cervical cancer screening who were directly mailed HPV-SS kits. We reviewed the 1) recommendation of and 2) adherence to follow-up care among individuals enrolled in the HOME trial intervention (mailed HPV-SS kit) and control (usual care reminders for in-clinic screening) whose screening results (whether from HPV-SS or in-clinic screening) were abnormal and thus warranted in-clinic follow-up. We determined whether recommendation of and adherence to follow-up care varied by screening modality; kit returners were less likely than both kit recipients who screened in-clinic and control participants to receive correct follow-up care recommendations. Among the 228 HOME patients for whom follow-up was indicated, 62% of kit recipients with HPV16/18+ findings received correct follow-up recommendations for diagnostic colposcopy, with 35% ultimately completing colposcopy (vs. 81% of control). Clinician electronic health record (EHR) notes revealed uncertainty and confusion over how to correctly manage follow-up after screening via HPV-SS. In the era of primary HPV screening, this study emphasized the importance of patient and clinician education and centralized management of follow-up of abnormal results. In Chapter 2, we evaluated the role of patient education in motivating cervical cancer screening behavior. First, we conducted 8 patient focus groups to understand 1) knowledge gaps surrounding HPV and cervical cancer screening and 2) patient risk perception surrounding cervical cancer. We then embedded a trial within a randomized clinical trial evaluating strategies to increase cervical cancer screening (the STEP trial) to understand the impact of patient education materials (PEMs) on screening completion and determine whether age-targeting of materials enhanced their effectiveness. Partway through the trial, the original PEMs were replaced with three alternatives, informed by focus group feedback: 1) a non-age-specific “optimized” version, 2) an age-targeted version for ages ≤45 years and 3) an age-targeted version for ages >45 years. We evaluated the impact of optimized and age-targeted PEMs on screening uptake across by screening history and age group. In the embedded trial, optimized and age-targeted PEMs significantly increased screening completion among individuals >45 years and those who were previously screening adherent. In the first part of the trial, the association between older age and screening completion was also significant among PEMs recipients. Across the full trial, PEMs were associated with higher screening completion rates among both younger and older individuals. Finally, in the third chapter, we expanded our understanding of the impact of HPV-SS on health systems to include an economic analysis. We extended the findings of the STEP cost-effectiveness analysis by conducting a Value of Information (VOI) analysis. Specifically, we applied a methodology of subgrouping to the VOI analytic framework that enabled us to determine whether future research around at-home cervical cancer screening interventions should focus on specific subgroups that drive heterogeneity in a patient population and, thus, render population-wide cost-effectiveness estimates suboptimal. We created subgroups according to patient age, travel time from an individual’s home to a clinic, and engagement with the healthcare system in the past year. Among screening adherent and overdue individuals, VOI estimates were largely zero, indicating additional research into home-based screening programs is not warranted to reduce decision uncertainty, and private health systems can confidently deploy direct mailing of home-based HPV-SS kits (as indicated by the STEP cost-effectiveness analysis). Among participants with unknown screening histories, VOI estimates were non-zero, but low, suggesting there is additional utility in understanding more about these participants before implementing screening outreach; each health system should consider the size of their “unknown” population before investing in additional data collection or research. Across these three chapters and leveraging data from two randomized controlled trials, we generated a robust knowledge base around the implications of at-home cervical cancer screening on 1) patient follow-up adherence and clinician management of abnormal screening results, 2) patient screening behavior and knowledge of HPV and cervical cancer, and 3) health system implementation of cost-effective, targeted screening approaches.