CHDS’s Emily Burger, Jane Kim, and colleagues compared four cervical cancer simulation models from the Cancer Intervention and Surveillance Modeling Network (CISNET) to understand why estimates of quality-adjusted life-years (QALYs) vary for cervical cancer screening. The open access article was published in Medical Decision Making.
Each model drew on common US data and identical quality-of-life weights to evaluate five recommended US screening strategies against no screening. Life-years gained differed by only 1% to 12% between the least and most effective strategies within each model, but QALYs gained differed by 47% to 120%. Quality-of-life losses from repeat tests were the main driver of this variation, reflecting differences in how many repeat tests each model’s natural-history and test-performance assumptions generated. Sensitivity analyses showed results shifted substantially with the assumed duration of those effects and with discounting.
In the authors’ view, QALY estimates stayed highly dependent on each model’s underlying assumptions, even after key inputs were aligned. They noted that comparative modeling can help identify sources of this variability, and they recommended reporting disaggregated health and resource use outcomes, such as the number of false positives and diagnostic interventions, alongside QALYs.
Learn more: Read the full publication, Why QALY Estimates Differ in Cervical Cancer Screening: An In-Depth Exploration Using CISNET Comparative Modeling
Related news: Improving HPV Transmission Models
Related news: HPV Vaccine May Reduce Cervical Cancer Screenings