From Health Economic Models to Decision Support Software: A MedSID/PINCER Researcher RSE Collaboration

Abstract

The MedSID/PINCER project is developing decision support software to help understand the benefits, harms and costs of prescribing safety interventions. Building on existing health economic models, the project translated Excel based modelling work into reusable research software and an interactive dashboard for decision makers. This submission showcases two main project outputs: an R package containing tested, reusable model code, and a Shiny dashboard that makes the models easier to explore, communicate and apply. The work involved understanding model assumptions, making implicit spreadsheet logic explicit, comparing software outputs against expected results, structuring the code for reuse, and designing an interface that supports non-technical users. The project also demonstrates the value of close collaboration between researchers and RSEs. Researchers contributed clinical, medication safety and health economic expertise, model assumptions, expected outputs, stakeholder context and scientific validation. RSEs contributed software architecture, testing, package design, dashboard development, usability, documentation, and planning for sustainable use. In addition, the project explicitly aims to upskill jobbing researchers in robust software engineering practices By showing both the software outputs and the collaboration behind them, this submission highlights how RSEs contribute to research quality, reproducibility and impact. It also offers a practical example of how RSE contributions can be evidenced through artefacts such as tests, code review, model comparisons, prototypes, documentation and stakeholder feedback.