Medical specialties
AI for Clinical Research and Practice Diploma
Register interest
A practical diploma in using AI for clinical research and practice: how to work with large language models safely and reproducibly, from the first research question to peer review, and how to judge the AI tools arriving in your clinic.
Ten weekly sessions across two modules follow one research project from start to finish, taught live by practising specialists you can question. Each session pairs teaching with a live demonstration on public or synthetic data and ends with what not to do. You finish with a capstone on your own research.
- Course director
- To be announced
- Format
- 10 live weekly sessions on Zoom, in 2 modules
- Fee
- £600; £1,200 for industry (scholarships available)
- Dates
- Confirmed when enrolment opens
- CPD
- Certificate of Completion for your CPD portfolio. CME/CPD accreditation to be applied for.

Who it is for
Doctors in any specialty who write, review or read research, clinical researchers and allied health researchers, and medical affairs and evidence teams in industry, worldwide.
The programme
10 one-hour sessions in 2 modules, each with its learning objectives. This is the draft programme: the course director reviews and refines it before enrolment opens, so session titles may change.
Module 1 · Sessions 1 to 5From first question to a clean reference list
1. How large language models work, and where they fail on clinical questions
- Explain how a large language model produces text by prediction, not retrieval
- Predict where models fail: fabricated facts, agreeing with the user, knowledge cut-offs and run-to-run variation
- Distinguish general assistants, research tools and agents, and choose between them for a task
- Test a model on questions in your own field and record what it got wrong
2. Safe use: patients, images and unpublished work
- Compare consumer and enterprise tiers, their data retention and their model-training settings
- De-identify clinic letters, operation notes and imaging, including burned-in text and file metadata
- Apply data protection law and your institution's policy to AI tools
- Write a one-page AI use policy for yourself or your department
3. Prompting for consistent, reproducible results
- Structure a prompt with a role, context, task, constraints and output format
- Build reusable templates and worked examples for data extraction and classification
- Measure variation by running the same prompt repeatedly
- Version prompts as you would a study protocol
4. Finding and synthesising the evidence
- Use AI search and citation-mapping tools alongside the bibliographic databases, which remain the record
- Screen and extract data with AI assistance under dual human checks
- Report AI use in an evidence synthesis transparently
- Keep a search log that another researcher could repeat
5. Reference integrity: what editors see
- Explain why models fabricate and misattribute references
- Check identifiers, authors and years for a whole reference list at scale
- Confirm that each cited paper supports the claim made, and identify retracted work
- Audit the reference list of your own last paper
Module 2 · Sessions 6 to 10From analysis to peer review and practice
6. Study design, analysis and code
- Sharpen a research question and draft a protocol and statistical analysis plan with AI support
- Read, check and run AI-written analysis code in R or Python
- Reproduce a published analysis from open or synthetic data
- Recognise when to bring in a statistician
7. Writing papers and grants with AI, and disclosing it
- Use AI as an editor of your own outline and draft, not as the author
- Check a manuscript against the relevant reporting guideline
- Write a disclosure statement that meets journal and authorship rules
- Explain why AI text detectors are unreliable
8. Peer review in the AI era
- State what journals and funders allow reviewers to do with AI
- Protect the confidentiality of manuscripts and grant applications under review
- Recognise fabricated data, references and images, and undisclosed AI text
- Write a structured review of a supplied manuscript using only permitted steps
9. Appraising and adopting AI in clinical practice
- Appraise a study of a clinical AI tool against the reporting guidelines for prediction, diagnostic accuracy and trials
- Judge external validation, bias and generalisability
- Weigh a vendor's claims against published evidence
- Outline how clinical AI is regulated in the UK, Europe and the United States
10. Agentic AI, your own workflow and the capstone showcase
- Describe what deep-research and agentic tools can and cannot be trusted to do
- Set up automated literature surveillance with the checks a human must still make
- Design a personal workflow that survives the next model release
- Present a capstone with its prompt log and verification log
What the diploma includes
- 10 live sessions on Zoom, one evening a week, with recordings for the weeks you miss.
- Practising specialists teaching live, so you can put your questions to them.
- A weekly task and a capstone on your own research, with a prompt log and a verification log.
- A digital certificate carrying your final classification.
Lead or teach the AI for Clinical Research and Practice Diploma
We are looking for a course director and faculty from the specialty. Sessions are one hour in the UK evening, live on Zoom, and paid.