The Potential of R Shiny User Interfaces to Support Health Economic Decision Making

Material for the ISPOR Workshop "R Shiny for HTA" - 1 December 2021

Paul Schneider^1,2, Gianluca Baio⁴, Rose Hart³, Robert Smith^1,2,3

¹University of Sheffield; ²Dark Peak Analytics, ³UK Health Security Agency (DHSC); ⁴UCL; ⁵BresMed

---

PURPOSE

To introduce participants to R Shiny user interfaces and demonstrate how they can be used to make health economic decision models more accessible to stakeholders.

DESCRIPTION

As health economic decision models become more computationally burdensome, the shortcomings of Microsoft Excel are becoming increasingly apparent. As a result, high level programming languages such as R are becoming more popular. The script-based nature of these languages may be perceived as lacking an accessible interface, limiting the extent to which stakeholders, who are not modellers, can engage with economic models. The R package ‘shiny’ provides an attractive solution: it allows the creation of easy-to-use web interfaces, which allow decision makers to directly interact with the decision model. Users can easily input their own assumptions and explore completely new scenarios in real-time. The underlying R code can be submitted alongside the interface to be reviewed and transparently debugged. In the future, R Shiny user interfaces could complement HTA submissions and improve health policy decision making.

AGENDA

• Introduction to R and Shiny in HTA: current situation and future potential (Gianluca Baio)
• Interactive coding tutorial: building a simple Shiny app for a Markov model in R (Robert Smith)
• Tips for using R Shiny in practice for efficient scalability and multi-team projects (Rose Hart)
• Open questions and discussion

RESOURCES

Slides:

• Gianluca Baio: Introduction to R and Shiny in HTA
• Robert Smith: Tutorial
• Rose Hart: Tips for using R Shiny in practice

Building a simple shiny app:

• App: https://robertasmith.shinyapps.io/sick_sicker/
• Paper: https://wellcomeopenresearch.org/articles/5-69
• Code: https://github.com/RobertASmith/paper_makeHEshiny

More advanced materials:

• App: https://darkpeakanalytics.shinyapps.io/sadm-mk2/
• Code: https://github.com/bitowaqr/sadm-mk2-demo
• Package: https://github.com/RobertASmith/darkpeak