Thomas Monks

1. Quantitative models are widely used to support decision-making under multiple and often competing objectives, relying on indicators to represent the system under study. These models enable strat...

1. President Barack Obama introduced a series of grand challenges during his time in office, describing them as “ambitious but achievable goals that harness science, technology, and innovation to s...

Discrete-event simulation (DES) is a popular technique for exploring problems in healthcare. For these models to be reused and have lasting impact, they need to be reproducible, transparent and well structured. Reproducible analytical pipeline (RAP) approaches are structured robust workflows that ensure analyses can be reproduced. They have emerged as best practice, but modellers struggle to implement them due to gaps in accessible guidance, skills, and time. This paper presents an integrated set of resources designed to help modellers bridge this implementation gap: an open-access e-book and four worked example repositories demonstrating complete RAP workflows for DES in both Python and R. The online book provides step-by-step guidance through nine major sections covering introductory material, project set-up, model inputs, model building, output analysis, experimentation, verification and validation, style and documentation, and collaboration and sharing. The case studies demonstrate varying complexity: a classic M/M/s queueing model, and a replication of a stroke care pathway model. The worked examples serve dual purposes: they demonstrate that RAP principles are achievable for healthcare DES models (from canonical queueing systems to real-world clinical pathways), and they provide templates that modelling teams can adopt and adapt within routine decision-support projects.

Objectives: This study assessed whether a previously developed Monte Carlo simulation model can be reused for evaluating various strategies to minimise time-to-treatment in southwest Netherlands for endovascular thrombectomy (EVT) in patients who had an ischaemic stroke. Design: Reuse of a previously developed simulation model to simulate various strategies in another region, using prospectively collected data from stroke centres and retrospective data from emergency medical services. Participants: Data from 509 patients who had an ischaemic stroke (≥18 years) treated with EVT (2014–2018) were used. Interventions: Input for the simulation model reuse included distributions of observed time delays along the acute stroke pathway. Validation of the baseline models was based on face validity and statistical measures (patient data vs model output) using the Assessment of the Validation Status of Health Economic decision models tool. We simulated strategies for a subregion: interhospital patient transfer by helicopter, transport of the neurointerventionalist to the primary stroke centre (‘drive-the-doctor’), interhospital patient transfer to a thrombectomy-capable stroke centre (TSC) outside the region and prehospital triage using the Rapid Arterial Occlusion Evaluation (RACE) scale. Primary outcome: Onset-to-groin time was the outcome. Results: Reuse of the original simulation model was obtained by minimal effort, implying limited adaptation. Compared with the baseline model, interhospital patient transfer by helicopter or to a TSC outside the region and prehospital routing using the RACE scale reduced mean onset-to-groin time by 16, 13 and 39 min, respectively (95% CrI for all: equal to the point estimate). ‘Drive the doctor’ reduced mean onset-to-groin time by 27 (car), 49 (ambulance) or 58 min (helicopter), each with a 95% CrI equal to the point estimate. Conclusions: The original simulation model can be applied to different regions in the Netherlands. Strategies tested within the subregion resulted in promising results of ‘drive the doctor’ and prehospital patient routing using the RACE scale.

Discrete-event simulation (DES) is widely used in healthcare Operations Research, but the models themselves are rarely shared. This limits their potential for reuse and long-term impact in the modelling and healthcare communities. This study explores the feasibility of using generative artificial intelligence (AI) to recreate published models using Free and Open Source Software (FOSS), based on the descriptions provided in an academic journal. Using a structured methodology, we successfully generated, tested and internally reproduced two DES models, including user interfaces. The reported results were replicated for one model, but not the other, likely due to missing information on distributions. These models are substantially more complex than AI-generated DES models published to date. Given the challenges we faced in prompt engineering, code generation, and model testing, we conclude that our iterative approach to model development, systematic comparison and testing, and the expertise of our team were necessary to the success of our recreated simulation models.