Cloud Computing

Theophile Bigirumurame

Hasselt University, Belgium

Ziv Shkedy

Hasselt University, Belgium

The Surrogate Shiny App

Shiny is an R package (available on CRAN) developed by RStudio that allows us to create web-based applications from R-code. The Surrogate Shiny App was developed as an online shiny application for the evaluation of surrogate endpoints in randomized clinical trials and has the same capacity, in terms of the methods implemented in the App, as the Surrogate R package discussed in Chapter 13.

The Surrogate Shiny App can be used on a local computer or online using the shiny cloud platform. Other cloud platforms, such as Amazon Web Service or Google Cloud platform can be used as well. In contrast with the Surrogate R package, the user does not need to install R in order to conduct the analysis. The Surrogate Shiny App is a graphical user interface (GUI) and the user is not exposed to the R code behind the analysis.

The Surrogate Shiny App can be found on the Shiny Cloud at:

https://uhasselt.shinyapps.io/surrogate

FIGURE 14.1

The Age-Related Macular Degeneration Data and variables for the analysis are specified in the left panel. A short summary of the data and a partial print are shown in the right panel.

In addition, the app is available as a stand-alone version in a SurrShiny.zip file that can be downloaded from

http://ibiostat.be/online-resources.

In this chapter, we briefly illustrate the capacity of Surrogate Shiny App for selected methods that were discussed in previous chapters. For each method, we present the GUI screen that can be used to conduct the analysis and the corresponding R code of the Surrogate package to perform an identical analysis. The code is presented only for clarity and it is not needed for the Surrogate Shiny App. The capacity of the Surrogate Shiny App is illustrated using case studies for three surrogacy settings: two continuous endpoints (Section 12.3.1, 12.3.2, and 13.2), two survival endpoints (Section 12.4 and 13.3), and two binary endpoints (Section 12.8 and 13.4).

The first step of the analysis requires uploading the data to the app. Figure 14.1 shows the data loading screen for the age-related macular degeneration data. Similar to Chapter 12, we need to specify the true and surrogate endpoints (Diff52 and Diff24, respectively), the treatment (Treat), the unit for which Rrial will be calculated (center), and the patient’s identification number (Id).

 
Source
< Prev   CONTENTS   Source   Next >