Connecting data science and translational simulation

Translational simulation is designed to explore and improve healthcare systems and performance. But this potential can only be realised if simulation activities generate actionable insights, using methods that are efficient and cost effective. Robust data strategies are required, embracing established quality‑improvement (QI) frameworks and the recent applications of artificial intelligence. Data collection, analysis and presentation are the primary functions of simulation for quality improvement. This requires translational simulation practitioners to adopt disciplined application of data science, if their diagnostic and interventional simulations are to be translated into tangible gains in healthcare quality and safety.This article is presented in three parts. First, I review contemporary data science principles in QI and emerging capabilities for data capture, synthesis, and tailored dissemination of findings. Second, I illustrate these principles through case vignettes drawn from the literature. Third, I synthesise these lessons to extend Nickson’s Input–Process–Output model, offering guidance for data strategy development for translational simulation initiatives. By integrating a rigorous data orientation into the foundational IPO schema, I argue that translational simulation can better realise its potential.