Quirin Werthner (1), Jürgen Rissland (2), Dominik Selzer (1), Barbara Berko (2), Stefan Wagenpfeil (3), Sigrun Smola (2) and Thorsten Lehr (1)
(1) Clinical Pharmacy, Saarland University, Saarbruecken, Germany. (2) Institute for Virology, Saarland University Hospital, 66421 Homburg, Germany. (3) Institute for Medical Biometry, Epidemiology and Medical Informatics, Saarland University Hospital, 66421 Homburg, Germany
Introduction: Worldwide, there are an estimated number of 569,800 cervical cancer cases per year due to an infection with human papilloma viruses (HPV) which is associated with about 311,400 deaths every year [1]. Despite recommendations for screening programs and vaccinations by the German Standing Committee on Vaccination (STIKO) as well as reimbursement for these measures, the corresponding German numbers stand at about 4,600 cervical cancer cases and 1,550 deaths per year [2]. Considering Germany’s low vaccination rate of 44.6% in female teenagers [3] compared to other western countries like Denmark (81%) or Sweden (78.5%) [4], and the newly arisen recommendation for the vaccination of boys, it is of ever more importance to analyze the current situation including potential barriers to as well as fostering possibilities for the reduction of incidence numbers. Therefore, a network model holds several advantages over compartment models. Namely, the modelling of individuals as separate nodes, together with the added stochasticity, which in combination leads to a more precise modelling of individual interactions and impacts of interventions [5].
Objectives:
- Development of a network-based mathematical model, describing the transmission and natural history of HPV and cervical cancer in Germany.
- Evaluate the potential implications of policy and behavioral changes on the spreading and prevalence of infections and mortality due to cervical cancer.
Methods: The model development was done using R (version 3.4.3) and the package suits statnet and EpiModel [6,7]. For the model development an approximate Bayesian computation with sequential Monte Carlo (ABC-SMC) was used in an open population modelling approach. The population was set up to consist of 1,000,000 individuals split between men and women from the ages of 12 to 100 to closely reflect the circumstances given by the Saarland region of Germany. Data for this was taken from the Federal Statistical Office of Germany [8].
Each individual was assigned fixed (e.g. gender) and dynamic (e.g. age, infection status) attributes, with infected individuals possessing additional attributes (e.g. diagnosis status, stage). Furthermore, different behavioral attributes were incorporated to fully describe interaction and thus spreading of HPV. Condom use was set to 20% [9] of all sexual intercourses with an assumed 3% rate of faulty condoms. Additionally, different sexual activity levels were taken into account with 21% of men and 15 % of women being categorized as increased sexually active and 8% of men and 6% of women being considered as highly active [10]. The vaccination rate was set to 44.6% [3] for women and in an alternative set up assumed to be 20% for men. Multiple transitions per individual on an annual basis were modeled. The necessary data for the model calibration were obtained from the literature.
Results: The chosen model was an adapted susceptible-infected-recovered (SIR) model with individuals passing from a susceptible to an infected state after being infected via the formation of a relationship with an infected individual. Furthermore, individuals could pass from the infectious to a recovered as well as to a cancer state through recovery or exacerbation, respectively. Over the course of the next 100 years, the model predicts a substantial reduction of cervical cancer cases. Depending on the actual turn out this might become even more amplified by the upcoming vaccinations for boys in Germany. However, the level of acceptance for HPV vaccination for boys might prove to be a critical point and needs to be assessed. Furthermore, the model predicts an increasing effectivity of HPV vaccination through the usage of nonavalent vaccines and the resulting higher degree of protection against additional high risk HPV types.
Conclusion: A functional model for the description of the natural history and spreading of HPV has been developed. Furthermore, the model allows analyzing of potential impacts of behavioral parameters and policy changes on the prevalence of HPV and cervical cancer in the Saarland region.
Acknowledgements:
This project is supported through subsidies of the Innovation Fond at the Federal Joint Committee of Germany (G-BA) under the project title Prävention des Zervixkarzinoms und dessen Vorstufen bei Frauen im Saarland/PRÄZIS (Prevention of cervical cancer and its precursors in women of the Saarland, grant number: 01VSF16050).
References:
[1] Ferlay, J. Estimating the global cancer incidence and mortality in 2018 : GLOBOCAN sources and methods. 1–13, 2018.
[2] Zentrum für Krebsregisterdaten Bericht zum Krebsgeschehen in Deutschland 2016; Krebs in Deutschland für 2013/2014, Ed.; 11th ed.; Robert Koch Institut: Berlin, 2016;
[3] Rieck T, Feig M, Siedler A et al. Aktuelles aus der KV-Impfsurveillance. Impfquoten Ausgewählter Schutzimpfungen Deutschland Epid Bull 1:1–14, 2018.
[4] Brotherton, J.M.L.; Bloem, P.J.N. management of HPV and associated cervical lesions, HPV Vaccination : Current Global Status. 220–233, 2015.
[5] Khanna, A. What can mathematical models tell us about the realtionship between circular migration and HIV transmission dynamics. 11, 1065–1090, 2015.
[6] Tools, T.S.; Analysis, S.; Depends, N.D.; Statnet, D.; Gpl-, L.; Url, L.; Handcock, A.M.S.; Hunter, D.R.; Butts, C.T.; Goodreau, S.M.; et al. Package ‘ statnet .’ 1–6, 2018.
[7] Jenness, S.M.; Goodreau, S.M. EpiModel : An R Package for Mathematical Modeling of Infectious Disease over Networks, 2017.
[8] https://www.destatis.de/DE/Startseite.html
[9] Reece M, Herbenick D, Schick V, Sanders SA, Dodge B, and Fortenberry JD, Condom use rates in a national probability sample of males and females ages 14 to 94 in the United States. J Sex Med, 2010.
[10] https://www.aerzteblatt.de/int/archive/article/193180/Sexual-behavior-in-Germany-results-of-a-representative-survey
Reference: PAGE 28 (2019) Abstr 9114 [www.page-meeting.org/?abstract=9114]
Poster: Drug/Disease Modelling - Infection