Purpose: Disease simulation modeling within a decision-analytic framework provides an opportunity to estimate population-level effects of an intervention and provide insight into the public health strategies best able to achieve the greatest good in terms of health for available resources. We describe a collaborative project to evaluate global cervical cancer prevention strategies and disseminate results to various audiences.
Method: We developed models of cervical carcinogenesis including a dynamic transmission model and an individual-based stochastic model to estimate intermediate and long-term clinical and economic outcomes associated with screening and vaccination. These models simulate the underlying natural history of disease, are calibrated utilizing a likelihood-based approach to country-specific epidemiologic data, and reflect uncertainty in major parameters. Empirical data for calibration targets include age- and type-specific incidence and prevalence of HPV, cervical lesions, and cervical cancer, drawn from primary data in longitudinal studies, survey and cancer registry data from collaborators (e.g., IARC, ICO), economic databases, health and demographic surveys, and published studies. We also developed an Excel-based population model to provide approximations of avertable burden and cost-effectiveness where data are limited; this model is also used to conduct financial forecasting and assessment of affordability and budget impact over different time horizons.
Result: We have empirically calibrated 16 country-specific simulation models, and completed analyses for individual countries as well as world regions (e.g., Latin America/ Caribbean, Asia/Pacific) and groups of countries (e.g., low-income countries eligible for financial assistance from GAVI for vaccine introduction). Our first phase of work included dissemination of results through scientific publications and presentations at academic research conferences, monographs or technical reports coordinated by international clinical and policy organizations (i.e., WHO, FIGO), and white papers for specific governments and regional organizations (i.e., Ministries of Health, PAHO). Our second phase of work emphasizes two major areas: (1) country contextualization, development of costing tools, and analysis of alternative delivery strategies; and (2) translation, communication, and dissemination of analyses to aid national and regional decision making, targeting a range of audiences and stakeholders.
Conclusion: Our goal in this multidisciplinary effort is to use models iteratively to reduce cervical cancer by addressing critical short- and long-term questions about the feasibility, sustainability, and cost-effectiveness of screening and vaccination, and to disseminate results of analyses to scientists, policy makers and local decision-makers.
Candidate for the Lee B. Lusted Student Prize Competition