Model output
Generalized additive models (GAM) fit on 1e+05 simulation runs.
Manchester Centre for Health Economics
MANC-RISK-SCREEN
Manchester Centre for Health Economics
MANC-RISK-SCREEN
A discrete event simulation model to evaluate the costs and outcomes of different approaches to breast cancer screening in the UK.
Welcome to the MANC-RISK-SCREEN Shiny App! This app provides an accessible interface for researchers and decision makers to explore the MANC-RISK-SCREEN model. This page provides a brief description of the underlying model. Click on the help tab for instructions on how use this app.
The MANC-RISK-SCREEN model is a discrete event simulation model which aims to predict the costs, outcomes, and cost-effectiveness of six breast cancer screening strategies in the UK. The model includes three universal screening strategies (No screening, 2-yearly screening for all, and 3-yearly screening for all), and three strategies which use breast cancer risk prediction to alter the screening interval for women at different levels of risk:
PROCAS screening: annual screening for women with 10-year risk higher than 8%, 2 yearly screenings for women with a risk between 5% and 8%, and 3 yearly screenings for all other women.Fully stratified screening: As above, but with 5 yearly screenings for women with a risk below 1.5%.Risk Tertiles: annual screening for women in the highest third of 10 year risk, 2 yearly screenings for women in the middle third, and 3 yearly screenings for women in the bottom third.The risk prediction tool used in the last three cases is the Tyrer-Cuzick questionnaire and Volpara automated breast density measurement.
The model uses a wide range of input parameters to simulate women through each screening programme, producing average outcomes including costs and Quality-Adjusted Life Years (QALYs). In the full analysis, a probabilistic sensitivity analysis (PSA) was also conducted to address parameter uncertainty. Using the PSA draws, a regression model known as a generalised additive model (GAM) was estimated to predict the costs and QALYs for each strategy based on the values of the input parameters. This GAM model forms the basis for this shiny app, allowing the user to choose a selection of input values to explore their impact on the cost-effectiveness results.
Generalized additive models (GAM) fit on 1e+05 simulation runs.
To begin using the app, please select the Main tab at the top of the screen. This will take you to a new page featuring tabs for different input values on the left side and an output table and graph on the right side.
Five tabs on the left hand side (Mortality, Cancer Growth, Screening, Cost, and Utility) allow you to switch between each input value. You can change the value of each input value dragging the corresponding slider. When an input value is changed, the predicted outputs reported in the model output table and graph will automatically update to reflect the new values.
The GAM model was estimated for wide ranges of parameter values. However, you can choose input values outside this range. In this case a warning message will appear, as it is possible that results produced from these values are less reliable.
For some parameters, such as 5-year cancer survival rates, there is also a logical ordering to the parameter values. A red warning message will be shown if you choose values which do not follow the logical ordering.
At any point you can click the reset button in the app to return all parameters to their base-case values. Use the save button to save the current input values into a text file, which can later be used to restore the session.
Ewan Gray, Anna Donten, Nico Karssemeijer, Carla van Gils, D. Gareth Evans, Sue Astley, Katherine Payne, “Evaluation of a Stratified National Breast Screening Program in the United Kingdom: An Early Model-Based Cost-Effectiveness Analysis”. Value in Health, Volume 20, Issue 8, 2017, Pages 1100-1109, ISSN 1098-3015, https://doi.org/10.1016/j.jval.2017.04.012.
Code for the complete model, mantained by Stuart Wright, University of Manchester Centre for Health Economics: https://github.com/stuwrighthealthecon/MANC-RISK-SCREEN
Code for this app, mantained by Martin Herrerias Azcue, University of Manchester Research-IT: https://github.com/UoMResearchIT/MANC-RISK-SCREEN