Update on WP4: Training materials


Amy Heather ORCID iD0000-0002-6596-3479



Part of the “STARS” project:

Stages for developing the training materials

Gather evidence on requirements for reproducible DES

From theory to practice: build working examples

Testing with a real-world stroke model

Knowledge mobilisation: online book

Gather evidence on requirements for reproducible DES



  • Empirical research - Barriers and facilitators identified from reproducing eight published healthcare DES models (Python/R).

Gather evidence on requirements for reproducible DES



  • Empirical research - Barriers and facilitators identified from reproducing eight published healthcare DES models (Python/R).
  • NHS RAP guidelines - Established framework for developing reproducible analytical pipelines (RAP) in the NHS.

Gather evidence on requirements for reproducible DES



  • Empirical research - Barriers and facilitators identified from reproducing eight published healthcare DES models (Python/R).
  • NHS RAP guidelines - Established framework for developing reproducible analytical pipelines (RAP) in the NHS.

These serve different purposes. Distinction between reproducibility and a RAP:

Gather evidence on requirements for reproducible DES



  • Empirical research - Barriers and facilitators identified from reproducing eight published healthcare DES models (Python/R).
  • NHS RAP guidelines - Established framework for developing reproducible analytical pipelines (RAP) in the NHS.

These serve different purposes. Distinction between reproducibility and a RAP:

Gather evidence on requirements for reproducible DES



  • Empirical research - Barriers and facilitators identified from reproducing eight published healthcare DES models (Python/R).
  • NHS RAP guidelines - Established framework for developing reproducible analytical pipelines (RAP) in the NHS.
  • Established DES modeling practices - All the necessary stages of creation of a DES model were identified from domain expertise and textbook recommendations.
    • Why? The RAP should include all the analysis stages of a typical DES modelling process.

From theory to practice: building working examples



Developed simple M/M/s examples as templates in Python and R that meet all requirements we identified.

From theory to practice: building working examples



Developed simple M/M/s examples as templates in Python (SimPy) and R (simmer) that meet all requirements we identified.

  • Meets all requirements (from Heather et al. (2025) and NHS Levels of RAP) - for example:
    • Structured as a package.
    • Unit tests, back tests and functional tests.
  • Complementary structure between Python and R models.
  • Python mostly classes, R mostly functions (as is typical for each language).
  • Provides code for simulation with scenarios and sensitivity analysis as well as choice of model parameters (input modelling, warm-up length, number of replications).

Testing with a real-world stroke model



Applied the templates to a stroke pathway model described in a published paper. This provided a real, more complex example and helped refine the template.

Knowledge mobilisation: online book



The M/M/s and stroke models demonstrate what reproducible DES looks like, they don’t show modellers how to get there.

I am developing an online book which breaks down these models, showing how they met each requirement. The book:

  • Provides actionable steps which guide modellers to create a model which meets all the reproducibility and RAP requirements.
  • Allows people to easily choose particular pages of interest (e.g. already have a model, but want to find out more about citation).

Knowledge mobilisation: online book



Looking forwards




Review: At a good stage for review of some of these materials.

  • Tom has reviewed the M/M/s models.
  • Alison - review of R (mentioning before - timing better now?)
  • Nav - review of book, if happy to?
  • Andy & Chris - at a later stage?


Plan: Continuing to work on these materials. Will describe in a publication.