Shiny Applications

Shiny Applications

Introduction

Health economic evaluation models have traditionally been built in Microsoft Excel, but more sophisticated tools are increasingly being used as model complexity and computational requirements increase. Of all the programming languages, R is most popular amongst health economists because it has a plethora of user created packages and is highly flexible. However, even with an integrated development environment such as R Studio, R lacks a simple point and click user interface and therefore requires some programming ability. This might make the switch from Microsoft Excel to R seem daunting, and it might make it difficult to directly communicate results with decisions makers and other stakeholders.

The R package Shiny has the potential to resolve this limitation. It allows programmers to embed health economic models developed in R into interactive web browser based user interfaces. Users can specify their own assumptions about model parameters and run different scenario analyses, which, in the case of regular a Markov model, can be computed within seconds.

When might I use this?

Shiny applications may be particularly useful…

  • to experiment with and to reflect on different scenarios and test the validity/plausibility model of inputs and outputs.
  • in areas where one health economic decision model is applied in range of circumstances (e.g. in public health, models are often built to be used in a number of different countries).
  • when the full model source code can or may not be shared (e.g. for proprietary or privacy reasons).
  • to engage with stakeholders and decision makers, who would otherwise not be able to interact directly with statistical computer models.

Resources

Advanced shiny app and source code for a fast, time-dependant 3-State Markov model

Tutorial: building a simple shiny app for a 4-State Markov model

Bresmed’s ‘IntRface’ application

SHARP CKD-CVD

IVI-RA Value Tool

More papers on the use of shiny for health economic modelling