useR! 2024

In the pharmaceutical industry, we aim to create standard R packages for statisticians, economic modelers, and statistical programmers. The challenge is to have both flexible functions and a systematic structure to support automatic reporting systems. In this presentation, we introduce a layered design for R package development that addresses the requirements of these diverse users. Then, through a case study in Health Technology Assessment (HTA) analysis, we show how our design provides solutions while adhering to good programming and documentation practices. The design consists of two layers, first, a verb layer, embracing functional programming and built with pipeable functions to provide flexibility for exploratory analysis. Second, a reporting layer, wrapped around the verb layer, with the ability to generate agreed upon standard analysis with one call to a function, making it easy to repeat analysis for different populations, interventions, comparators, and outcomes (PICOs). Throughout the case study, we illustrate how we leveraged R package development tools, to organize, document, and test R code to ensure quality and maintainability.