greenbook
Turning the Green Book from a PDF into a computable, versioned source of truth - and an engine that can actually read it.
The UK childhood immunisation schedule - Green Book Chapter 11 - is published as human-readable PDFs. As of 2026 there is no computable version: every digital system that needs the schedule re-implements it by hand from the PDF, a lossy and error-prone derivative.
greenbook is a proof-of-concept for the other way round: a computable, versioned schedule as the upstream source of truth, and an evaluation engine that takes a patient's FHIR vaccination history and answers - honestly - whether they are up to date for their age.
What the Green Book is · Why a computable version · Getting started
See it run
Pick an example patient and watch the engine evaluate them - the headline verdict, conformance by series, and antigen coverage. This is a pared-back, read-only slice of the full interactive demo.
-
The Green Book today
What the schedule is, the "digital paper" problem, and why answering "is this child fully vaccinated?" is genuinely hard.
-
The concepts
The domain model - Schedule → Series → Dose, product class vs antigen - and the two ideas that make the engine correct: conformance vs coverage and the status model.
-
Getting started
Install and run the engine. A reference implementation in Rust and a peer implementation in JavaScript, validated against one shared conformance suite.
-
Try it live
An interactive demo that evaluates example patients - or build your own and watch the result update live - and the presentation behind the project.
The long-term goal
What if the computable schedule were the source of truth, and the PDF, the website, and clinical tools were all generated from it? Experts would author schedule changes directly in a computable form, and every downstream publication - PDFs, websites, clinical decision support - would be rendered from the same trusted data. That replaces today's "digital paper" workflow, where each vendor reverse-engineers the rules from a PDF and quietly diverges.
This prototype builds the current schedule and the first historical slices in the same format, so historical versioning is additive: older Green Book schedules are added as dated source files, and the evaluator selects the version that applied when each dose first became due.