Moriña et al. BMC Medical Informatics and Decision Making (2020) 20:211 https://doi.org/10.1186/s12911-020-01232-9 SOFTWARE Open Access Online Cost-Effectiveness ANalysis (OCEAN): a user-friendly interface to conduct cost- effectiveness analyses for cervical cancer David Moriña1,2,3* , José Ignacio Martí4, Pedro Puig1,2 and Mireia Diaz4,5 Abstract Background: Most cost-effectiveness analyses in the context of cervical cancer prevention involve the use of mathematical models to simulate HPV infection, cervical disease and prevention strategies. However, it is common for professionals who would need to perform these analyses to not be familiar with the models. This work introduces the Online Cost-Effectiveness ANalysis tool, featuring an easy-to-use web interface providing health professionals, researchers and decision makers involved in cervical cancer prevention programmes with a useful instrument to conduct complex cost-effectiveness analyses, which are becoming an essential tool as an approach for supporting decision-making that involves important trade-offs. Results: The users can run cost-effectiveness evaluations of cervical cancer prevention strategies without deep knowledge of the underlying mathematical model or any programming language, obtaining the most relevant costs and health outcomes in a user-friendly format. The results provided by the tool are consistent with the existing literature. Conclusions: Having such a tool will be an asset to the cervical cancer prevention community, providing researchers with an easy-to-use instrument to conduct cost-effectiveness analyses. Keywords: Cervical cancer screening, Markov chain model, Cost-effectiveness, Online tool, cancer prevention, Decision making Background Cervical cancer and other cancer and noncancer dis- Health care resources are scarce and therefore, their effi- eases are caused by or related to human papillomavirus cient allocation is a priority for policy-makers. To guide (HPV), a common sexually transmitted infection [2]. In the process of decision making in this context, cost- fact, it is estimated that more than 80% of sexually active effectiveness analysis is an essential tool. When the avail- men and women will acquire an HPV infection by age able prevention strategies are multiple and potentially 45 years [3]. HPV infections are asymptomatic in most synergic, as occurs in the context of cervical cancer, cases, although some can lead to the formation of cer- cost-effectiveness analysis becomes critical to ensure the vical abnormalities called cervical intraepithelial neopla- optimal allocation of resources [1]. sia (CIN), which can lead to cervical cancer. This cancer is one of the most prevalent cancers among women worldwide, especially in less developed countries [4]. * Correspondence: dmorina@mat.uab.cat Cervical cancer can be prevented by means of screening 1 Barcelona Graduate School of Mathematics (BGSMath), Barcelona, Spain to find precancerous lesions (secondary prevention) -so 2 Departament de Matemàtiques, Universitat Autònoma de Barcelona (UAB), Cerdanyola del Vallès, 08193 Barcelona, Spain they can be treated- or by HPV vaccination (primary Full list of author information is available at the end of the article prevention) to prevent infection from some of the most © The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Moriña et al. BMC Medical Informatics and Decision Making (2020) 20:211 Page 2 of 7 frequent high-risk types. Several countries are imple- be used in the cost-effectiveness analyses. The calibra- menting a combined prevention strategy including both tion panel requires the following inputs: vaccination and screening, which is recommended on the basis of current evidence, although details may vary.  Number of simulations: Number of simulated The World Health Organization (WHO) recommends cohorts. that countries should implement HPV vaccination in  Number of simulations to keep: Number of preadolescent girls if it is affordable, cost-effective, and simulated cohorts that will be kept and used in the sustainable and that achieve the highest possible cover- cost-effectiveness analyses. Among all simulated co- age [5]. However, there are still many challenges faced horts, the best ones (the matrices producing the out- by cervical cancer prevention programmes [6]. Cur- comes that minimize the differences with respect to rently, most cost-effectiveness analyses focused on the target HPV infection prevalence and CC incidence). evaluation of cervical cancer prevention strategies are All these simulated cohorts will be used to obtain conducted by means of mathematical models that simu- the cost-effectiveness outputs if the Uncertainty late the natural history of HPV and cervical cancer. Level (see the inputs for the cost-effectiveness ana- These complex models must integrate data from various lysis in the next subsection) is not zero, and stored sources and should be effectively calibrated to health tar- each in one sheet on the same file. This allows the gets. The tool presented in this work is based on a Mar- tool to generate several estimates for each cost- kov model that simulates the natural history of HPV effectiveness output (one per matrix) and therefore infection and subsequent cervical disease. The model to generate and display confidence intervals. computes relevant health and economic outcomes as  Percentage of change: Maximum allowed difference cases averted, life expectancy (LE) from 11 years, reduc- between two equivalent transition probability tion in the lifetime risk of CC, life years saved, quality- matrices. adjusted life years (QALYs), net health benefits, and life-  Mortality: Checkbox that should be ticked if time costs. In this paper we present OCEAN, a new easy mortality data are available and it is wanted to be to use web application that allows for cost-effectiveness used in the calibration process. analyses based on a prespecified mathematical model to  Transition probability matrix: Starting values for the be run. This model is described in detail in [7]. The tool transition probability matrix. A general transition and a video tutorial illustrating its usage are available at probability matrix will be used if no user-specific https://iconcologia.shinyapps.io/HECR-OCEAN/. Despite matrix is provided. the tool’s ease of use and design, thought to be usable for  Incidence file: Registered cervical cancer incidence inexperienced users, the authors are available to provide for the specific setting, without any medical guidance in case it is needed. intervention (rate × 100,000 women). It is uploaded in a spreadsheet file and will be used as a target in Implementation the calibration process. The tool is based on a previously described Markov  Prevalence file: Registered HPV16/18 infection model [7]. The details of this model are also available as prevalence for the specific setting. It is uploaded in supplementary material. an Excel file and will be used as a target in the calibration process.  Mortality file: Registered mortality due to cervical The OCEAN tool cancer for the specific setting (rate × 100,000 The tool provides an easy-to-use web interface to con- women). It is uploaded in an Excel file and will be duct cost- effectiveness analyses. The sample data avail- used as a target in the calibration process if able from the OCEAN tool, and in particular the yearly mortality is checked. regression and progression transition probabilities be- tween health states were extracted from a literature re- An Excel file including all calibrated matrices can be view [8–12]. The web interface is based on the shiny downloaded and used for the cost-effectiveness analyses. package [13] for R [14]. Its main screen is divided into The default transition probability matrix is calibrated to two panels, one focused on the calibration process and Spanish data, but the users could upload their own HPV the second on the cost-effectiveness analyses. prevalence, incidence and mortality data and calibrate the transition probabilities matrix to their own settings. Calibration process The goal of the calibration part is to provide the user Cost-effectiveness analysis with a reliable transition probability matrix adjusted to The cost-effectiveness panel of the tool requires the fol- the specific setting, to feed the Markov model that will lowing inputs: Moriña et al. BMC Medical Informatics and Decision Making (2020) 20:211 Page 3 of 7  Number of different scenarios: The number of additional Excel file specifying the proportion of scenarios to be analyzed in the current session. Note women screened each period is required. that each scenario may include none (no  Vaccination: Preadolescent girls are successfully intervention or natural history), one or more vaccinated at the age of 11 years with one, two or prevention strategies (screening alone, vaccination three doses of the vaccine against HPV types 16 and alone or vaccination followed by screening). 18. Efficacy and coverage are set by the user and  Matrix of transition probabilities between different vaccination costs are read from the Excel file loaded health states without medical intervention: as “Scenario-specific values”. The structure of this Transition probability matrix obtained in the file can be explored by downloading the example file previous panel or uploaded by the user. Additionally, from the tool. Currently, only bivalent vaccine is the included default file can be used. considered but it is planned that quadrivalent and  Medical costs and utility coefficients: An Excel file nonavalent vaccines will be available soon as well. containing the treatment direct medical and nonmedical costs, indirect costs and utility An Excel file with all generated results can be down- coefficients. An example can be downloaded from loaded, including information about the inputs used to the tool. generate those particular results.  Discount rate: Discount rate to be applied to health and costs (undiscounted results can be obtained by Results setting the discount rate to 0). This section reproduces the cost-effectiveness analysis  Uncertainty level: It is known that the results of reported in [7] using the OCEAN tool. The goal is to cost-effectiveness analyses are affected by a certain mimic some of the cervical cancer prevention strategies degree of uncertainty at different levels. They can be available in Spain, particularly comparing conventional reflected in the OCEAN tool by using more than cytology to HPV testing, with and without vaccination. one transition probability matrix to feed the Markov All the input files used in this section are available as model. If a file with several sheets is used and the supplementary materials. First, we can use the calibra- uncertainty level is set to 0, an averaged matrix will tion part of the tool to check whether the input matrix be used, and only point estimates will be reported. If we use fits the Spanish registered data in a reasonable the uncertainty level is set to a value α between 0 manner (Tables S2 and S3 show the considered HPV in- and 50, all the matrices are used, and the outcomes fection prevalence and cervical cancer incidence). obtained from each one are recorded. Then, the Through the graphs provided by the calibration part of mean and percentiles α2 and 1 − α2 for each outcome the OCEAN tool, it can be seen whether the input are calculated and reported. matrix fits the targeted values. The output calibrated matrices are also useful to introduce some random noise The considered prevention strategies that can be se- (determined by the percentage of change and the num- lected for each scenario (alone or combined) include the ber of simulations to keep), allowing us to incorporate following: some uncertainty into the final outcomes, providing the researcher with a more realistic picture of the cost-  Natural history: No prevention strategy is effectiveness results. The output matrices from the cali- considered in this scenario, which reproduces the bration part can be used as input for the cost- natural history of HPV infection and cervical cancer. effectiveness analysis part, where the details on the dif- If this option is chosen, the other options disappear. ferent prevention strategies considered are set. The out-  Screening: Several screening scenarios can be put from the tool shows the user that the input matrix is defined, depending on test (cytology, HPV DNA well calibrated (Fig. 1). test or visual inspection), frequency (every 1–10 Once a -or multiple- calibrated transition probability years), targeted ages, and switch age from matrix that feeds the Markov model has been generated cytology to HPV testing. The screening coverage, in the calibration part, it is time for the cost- positive predictive value, sensitivity and costs are effectiveness analysis. In this example, we will consider read from the Excel file loaded as “Scenario- the following prevention strategies: specific values”. The structure of this file can be explored by downloading the example file from  Natural history: The first scenario considers no the tool. Screening may be organized (all women prevention strategy. are screened with the selected frequency) or  Vaccination alone: In this scenario, preadolescent opportunistic (the screening period is variable). girls are successfully vaccinated at the age of 11 To set an opportunistic screening scenario, an years with three doses of the vaccine against HPV Moriña et al. BMC Medical Informatics and Decision Making (2020) 20:211 Page 4 of 7 Fig. 1 Example of the calibration panel and outputs for the age-specific cervical cancer incidence and the age-specific HPV16/18 prevalence compared to the observed data. The output is based on the 5 best-fitting (the matrices producing the outcomes that minimize the differences with respect to target HPV infection prevalence and CC incidence) simulations out of 10 demanded with a 10% change types 16 and 18. The analysis was carried out specificity of cytology to detect CIN2+ are 38.2 and assuming favorable vaccine with 70% efficacy and a 97.8%, respectively [15]. lifelong duration of vaccine immunity to prevent  Combined vaccination and screening: In this cervical lesions caused by HPV 16 and 18 among scenario, we implement vaccination in girls aged 11 uninfected women. No cross-protection against years, followed by screening according to the other high-risk HPV types was assumed. The uptake parameters and assumptions described previously is set to 70%. for vaccination and screening alone.  Screening alone: Assuming that screening with conventional cytology starts at 25 years of age and For all the scenarios, a 3% discount rate for health out- continues until age 65, this strategy considers that comes and costs and a 5% uncertainty level are applied. primary organized HPV DNA testing is performed The unitary costs used in this example are reproduced in women older than 30 years of age with cytology in Table S4. Figure 2 is a screenshot from the tool with triage for positive women. Women are screened in a the proper inputs corresponding to each described 5-year period. For women younger than 30 years of scenario. age, cytology is the reference test. The sensitivity The output tables reported by the tool can be seen in and specificity of HPV DNA testing to detect CIN2+ Fig. 3 and are available as supplementary material in the are 90.5 and 93.0%, respectively, and 90.5 and 91.9% same format in which can be downloaded from the tool for cytology after a positive HPV test [15]. Screening (Table S5). coverage is 70%. On the basis of a study carried out The results tables from Fig. 3 show that the health and in Spain, we assume that the sensitivity and cost outcomes are very similar to those reported in [7]. Moriña et al. BMC Medical Informatics and Decision Making (2020) 20:211 Page 5 of 7 Fig. 2 Example of inputs and parameters for four different scenarios: no intervention or natural history, vaccination alone, screening alone and combined vaccination and screening Fig. 3 Example of output tables as reported by the tool for four different scenarios Moriña et al. BMC Medical Informatics and Decision Making (2020) 20:211 Page 6 of 7 A figure reproducing Fig. 3 from [7] is available as sup- Availability and requirements plementary material (Fig. S2), also leading to the ex- The OCEAN tool is accessible with only a working pected results. internet connection. Sample input files are available from the tool website and are ready to be used. Conclusions Carrying out complex cost-effectiveness analyses can be Supplementary information challenging for many researchers involved in cervical Supplementary information accompanies this paper at https://doi.org/10. 1186/s12911-020-01232-9. cancer prevention and for public health decision makers. The OCEAN tool provides an easy-to-use interface, re- Additional file 1. quiring only a working internet connection and fairly available input information to allow its users to compare Acknowledgments the most common cervical cancer prevention strategies Not applicable. worldwide using their specific parameters. There is a lack of robust, rigorous, accessible tools allowing users Authors’ contributions DM, PP and MD developed the underlying model. DM, PP and JIMG to conduct reliable cost-effectiveness analyses, as these designed and built the web interface. All authors discussed the results and analyses may play an important role in deciding public contributed to the final manuscript. The author(s) read and approved the health policies in cervical cancer. We consider that the final manuscript. outputs of the OCEAN tool are of potential usefulness to Funding the public health and cervical cancer prevention com- This work was supported by RecerCaixa (2015ACUP00129) and by Instituto munities, but with some constraints. de Salud Carlos III FIS - PI16/01254 (CECO: 17FIS012). The research leading to these results received funding from the Instituto de Salud Carlos III-ISCIII In general, it is known that the mathematical models (Spanish Government) co-funded by FEDER funds/European Regional Devel- used in cost-effectiveness analyses are subject to uncer- opment Fund (ERDF) - a way to build Europe (RD12/0036/0056, PIE16/00049, tainty at different levels [16], and therefore, the results PI14/01219, PI11/02090 and PI16/01254) and received funding from the Agència de Gestió d’Ajuts Universitaris i de Recerca (2017SGR1718) and the provided by the tool (even setting the uncertainty level European Commission 7th Framework Programme COHEAHR (Health-F3– over zero) must be taken with caution. In particular, the 2013-603019). David Moriña acknowledges financial support from the Span- calibration process of the OCEAN tool does not include ish Ministry of Economy and Competitiveness, through the María de Maeztu programme for Units of Excellence in R&D (MDM-2014-0445) and from Fun- optimization routines, although they have been recently dación Santander Universidades. This work was partially funded by grant recommended in the literature [1] due to time limita- RTI2018–096072-B-I00 from the Spanish Ministry of Science, Innovation and tions on the on-line interface. Universities. Assessing the impact of uncertainty in the results, ei- Availability of data and materials ther through statistical analysis or through sensitivity The example used in this article can be reproduced by using sample input analysis, is recommended by recent guidelines on health files from the tool. Additional tables and figures are available as economic evaluation [17, 18]. The OCEAN tool facili- supplementary material. tates the task of handling uncertainty through determin- Ethics approval and consent to participate istic and probabilistic sensitivity analyses [19–21], so the Not applicable. implications on the cost-effectiveness analysis can be Consent for publication easily examined. Not applicable. Several extensions of the software are currently planned. For instance, other vaccines (quadrivalent and Competing interests None of the authors have any conflicts of interest. nonavalent) will be included, and the vaccination admin- istration period will be extended to other ages, including Author details 1 catch-up vaccination and male vaccination. Using a 2 Barcelona Graduate School of Mathematics (BGSMath), Barcelona, Spain. Departament de Matemàtiques, Universitat Autònoma de Barcelona (UAB), static underlying model also has some relevant limita- Cerdanyola del Vallès, 08193 Barcelona, Spain. 3Department of Econometrics, tions such as not capturing herd immunity benefits due Statistics and Applied Economics, Riskcenter-IREA, Universitat de Barcelona to HPV vaccination, although this Markov model can (UB), Barcelona, Spain. 4Unit of Infections and Cancer - Information and Interventions (UNIC-I&I), Cancer Epidemiology Research Program (CERP), handle complex screening strategies and it is known that Catalan Institute of Oncology (ICO)-IDIBELL, Barcelona, Spain. 5Centro de static models improve the transparency and robustness Investigación Biomédica en Red (CIBERONC), Barcelona, Spain. of the results compared to dynamic models [22]. To Received: 15 April 2019 Accepted: 26 August 2020 overcome this issue, more sophisticated models based on microsimulation will be included in the near future. Other HPV-related diseases (genital warts, recurrent References respiratory papillomatosis and other cancer locations 1. Moriña D, de Sanjosé S, Diaz M. Impact of model calibration on cost- effectiveness analysis of cervical cancer prevention. Sci Rep [Internet]. 2017; such as the vulva, vagina, penis, anus and oropharynx) 7:17208 Available from: http://www.ncbi.nlm.nih.gov/pubmed/29222509, will also be considered in upcoming versions of the tool. [cited 2018 Feb 27]. Moriña et al. BMC Medical Informatics and Decision Making (2020) 20:211 Page 7 of 7 2. Bosch FX, Lorincz A, Muñoz N, Meijer CJLM, Shah KV. The causal relation 19. Stout NK, Knudsen AB, Kong CY, McMahon PM, Gazelle GS. Calibration between human papillomavirus and cervical cancer. J Clin Pathol [Internet]. methods used in cancer simulation models and suggested reporting 2002;55:244–65 Available from: http://www.ncbi.nlm.nih.gov/pubmed/1191 guidelines. Pharm Int. 2009;27:–533, 45 Available from: http://link.springer. 9208, [cited 2018 Feb 27]. com/10.2165/11314830-000000000-00000, [cited 2018 Feb 27]. 3. Chesson HW, Dunne EF, Hariri S, Markowitz LE. The Estimated Lifetime 20. Kim JJ, Kuntz KM, Stout NK, Mahmud S, Villa LL, Franco EL, et al. Probability of Acquiring Human Papillomavirus in the United States. Sex Multiparameter calibration of a natural history model of cervical cancer. Am Transm Dis [Internet]. 2014;41:660–664. Available from: http://www.ncbi.nlm. J Epidemiol. 2007;166:137–50. nih.gov/pubmed/25299412, [cited 2018 mar 13]. 21. Van de Velde N, Brisson M, Boily M-C. Modeling human papillomavirus 4. Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, et al. vaccine effectiveness: quantifying the impact of parameter uncertainty. Am Cancer incidence and mortality worldwide: Sources, methods and major J Epidemiol [Internet]. 2007;165:762–75 Available from: http://www.ncbi.nlm. patterns in GLOBOCAN 2012. Int J Cancer [Internet]. 2015;136:E359–86 nih.gov/pubmed/17276976. Available from: http://www.ncbi.nlm.nih.gov/pubmed/25220842, [cited 2018 22. Westra TA, Rozenbaum MH, Rogoza RM, Nijman HW, Daemen T, Postma MJ, Feb 27]. et al. Until which age should women be vaccinated against HPV infection? 5. WHO. Human papillomavirus vaccines: WHO position paper. Biol Int. 2009; Recommendation based on cost-effectiveness analyses. J Infect Dis. 2011;204: 37:338–44 Available from: http://www.ncbi.nlm.nih.gov/pubmed/19525124, 377–84 Available from: https://www.ncbi.nlm.nih.gov/pubmed/21742836. [cited 2018 Feb 27]. 6. Ekwunife OI, O’Mahony JF, Gerber Grote A, Mosch C, Paeck T, Lhachimi SK. Challenges in Cost-Effectiveness Analysis Modelling of HPV Vaccines in Low- Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in and Middle-Income Countries: A Systematic Review and Practice published maps and institutional affiliations. Recommendations. Pharm Int. 2017;35:65–82 Available from: http://link. springer.com/10.1007/s40273-016-0451-7, [cited 2018 Feb 27]. 7. Georgalis L, De Sanjosé S, Esnaola M, Bosch FX, Diaz M, To C. Present and future of cervical cancer prevention in Spain: a cost-effectiveness analysis. Eur J Cancer Prev. 2015;00:0–0. 8. Myers ER, McCrory DC, Nanda K, Bastian L, Matchar DB. Mathematical model for the natural history of human papillomavirus infection and cervical carcinogenesis. Am J Epidemiol [Internet]. 2000;151:1158–71 Available from: http://www.ncbi.nlm.nih.gov/pubmed/10905528, [cited 2018 Feb 27]. 9. Canfell K, Barnabas R, Patnick J, Beral V. The predicted effect of changes in cervical screening practice in the UK: results from a modelling study. Br J Cancer [Internet]. 2004;91:530–6 Available from: http://www.pubmedcentral. nih.gov/articlerender.fcgi?artid=2409838&tool=pmcentrez&rendertype= abstract%5Cnhttp://www.ncbi.nlm.nih.gov/pubmed/15266332%5Cnhttp:// www.pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC2409838. 10. Insinga RP, Glass AG, Myers ER, Rush BB. Abnormal Outcomes Following Cervical Cancer Screening: Event Duration and Health Utility Loss. Med. Decis. Mak. [Internet], vol. 27. Los Angeles: Sage PublicationsSage CA; 2007. p. 414–22. Available from: http://mdm.sagepub.com/cgi/doi/10.1177/0272 989X07302128, [cited 2017 Feb 3]. 11. Kulasingam SL, Benard S, Barnabas RV, Largeron N, Myers ER. Adding a quadrivalent human papillomavirus vaccine to the UK cervical cancer screening programme: a cost-effectiveness analysis. Cost Eff Resour Alloc [Internet]. 2008;6:4 Available from: http://www.ncbi.nlm.nih.gov/pubmed/1 8279515, [cited 2018 Feb 27]. 12. Kohli M, Lawrence D, Haig J, Anonychuk A, Demarteau N. Modeling the impact of the difference in cross-protection data between a human papillomavirus (HPV)-16/18 AS04-adjuvanted vaccine and a human papillomavirus (HPV)-6/11/16/18 vaccine in Canada. BMC Public Health [Internet]. BMC Public Health. 2012;12:872 Available from: http://www. biomedcentral.com/1471-2458/12/872. 13. Chang W, Cheng J, Allaire JJ, Xie Y, McPherson J. shiny: Web Application Framework for R [Internet]. 2017. Available from: https://cran.r-project.org/ package=shiny. 14. R Core Team. R. A Language and Environment for Statistical Computing [Internet]. Vienna; 2017. Available from: https://www.r-project.org/. 15. Ibáñez R, Autonell J, Sardà M, Crespo N, Pique P, Pascual A, et al. Protecting the underscreened women in developed countries: the value of HPV test. BMC Cancer [Internet]. 2014;14:574 Available from: http://bmccancer. biomedcentral.com/articles/10.1186/1471-2407-14-574, [cited 2018 Feb 27]. 16. Drummond M, Stoddart G. Methods for the economic evaluation of health care programmes [Internet]: Oxford University Press; 1997. Available from: https://books.google.es/books/about/Methods_for_the_Economic_ Evaluation_of_H.html?id=sz1RnwEACAAJ&redir_esc=y, [cited 2018 Feb 27]. 17. Kim S-Y, Russell LB, Sinha A. Handling Parameter Uncertainty in Cost- Effectiveness Models Simply and Responsibly. Med Decis Mak [Internet]. 2015;35:567–9 Available from: http://www.ncbi.nlm.nih.gov/pubmed/262 80060, [cited 2018 Feb 27]. 18. Brisson M, Edmunds WJ. Impact of model, methodological, and parameter uncertainty in the economic analysis of vaccination programs. Med Decis Making [Internet]. 2006;26:434–46 Available from: http://journals.sagepub. com/doi/10.1177/0272989X06290485, [cited 2018 Feb 27].