Module 1 · 45 min
From AI Idea to Healthcare Impact
The bridge from Essentials: where healthcare AI value actually appears, how to read any proposal along the chain from problem to output to decision to action to outcome to value, four questions that screen a proposal out before a business case, actionability under real capacity, and stating a proposal as a falsifiable value hypothesis.
- Name which surface of value a proposal is actually pursuing, and what evidence that surface demands.
- Trace a proposal along the impact chain: problem, output, decision, action, outcome, value.
- Find the weakest link in that chain before any model, vendor or metric is discussed.
- Screen a proposal with four questions, including whether something simpler would do.
- Test actionability: the action, the owner, the moment, and the capacity to act at that moment.
- State a proposal as a value hypothesis — for whom, what changes, what value, how measured, what would falsify it.
- Say what an external performance result does and does not license for your own population and setting.
Why it matters
Healthcare systems are being asked to do more with the same workforce and estate. AI matters here because it can act on exactly those pressure points — reading a queue of studies, drafting the paperwork, finding the patients who need a scarce clinic slot. That is the opportunity, and it is large enough to be worth doing properly.
Most AI proposals that fail in healthcare do not fail because the model was inaccurate. They fail because no decision changed, nobody had the capacity to act, or the value was never stated in a form that could be shown to be wrong.
Earlier detection, more consistent decisions, fewer missed findings, better-targeted treatment.
- Triage of imaging worklists so time-critical studies are read sooner
- Risk stratification that directs a limited follow-up clinic to the patients most likely to benefit
Example metrics to test: Time from study to report for time-critical findings; Proportion of flagged patients receiving the intended intervention.
The same staff and estate delivering more care, or the same care with less rework and waiting.
- Drafting discharge summaries and referral letters for clinician review
- Automated coding suggestions and scheduling optimisation
Example metrics to test: Documentation minutes per encounter; Turnaround time from request to completed referral.
Care that is easier to reach and understand, and work that is less draining to do.
- Plain-language and multilingual explanations of results and instructions
- Ambient documentation that returns attention to the consultation
Example metrics to test: After-hours documentation time per clinician; Patient-reported comprehension of discharge instructions.
Faster learning from the organisation's own data and from the wider evidence base.
- Screening and extraction support for evidence reviews
- Cohort discovery for trials and service evaluation
Example metrics to test: Time to assemble an eligible cohort; Screening throughput per reviewer at a maintained recall level.
Essentials gave you the vocabulary to follow the conversation. This module gives you the first practitioner instrument: a way to read any proposal — clinical, administrative or generative — from the problem it claims to solve through to the value it claims to produce, and to say out loud which link is weakest.
That instrument is what the rest of the course refines. Module 2 examines whether the data can support the claim, Module 3 whether the prediction reaches a clinical decision, Module 6 what evidence would settle it, and Modules 8 to 10 how it is funded, governed and actually implemented.
Sources & evidence · 4 sources
This module cites public or consensus guidance, scholarly literature.
Content reviewed: September 2026. Publication dates of the individual sources are shown in each citation.
Collins GS, Moons KGM, Dhiman P, et al. TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods. BMJ. 2024;385:e078378. doi:10.1136/bmj-2023-078378
Sets out what a prediction-model report should contain, including how predictors and outcomes are defined and timed. It is a reporting standard: following it makes evidence legible, but does not itself establish that a model is safe or effective.
Open sourceVasey B, Nagendran M, Campbell B, et al. Reporting guideline for the early-stage clinical evaluation of decision support systems driven by artificial intelligence: DECIDE-AI. Nature Medicine. 2022;28:924–933. doi:10.1038/s41591-022-01772-9
Supports the distinction between model performance and clinical evaluation of a decision-support system in live use. It covers early-stage clinical evaluation and does not replace comparative effectiveness evidence.
Open sourceWorld Health Organization. Ethics and governance of artificial intelligence for health. WHO guidance. 2021. ISBN 978-92-4-002920-0
Supports the governance and 'does this need AI at all' framing at policy level. It is guidance rather than binding regulation, and does not determine any specific jurisdiction's legal requirements.
Open sourceNational Institute of Standards and Technology. Generative artificial intelligence. NIST Computer Security Resource Center Glossary; definition sourced by NIST to NIST SP 800-218A, Secure Software Development Practices for Generative AI and Dual-Use Foundation Models (July 2024). doi:10.6028/NIST.SP.800-218A. Entry status checked 25 August 2026.
Used only for a stable, citable definition of generative AI. A glossary entry standardises terminology; it supports no claim about healthcare performance, safety or risk.
Open source
Capstone Board Pack
Add what you just learned to your own strategy document while it is fresh.