AI for clinical research has to work with the field's own sources and paperwork, not only with papers. AI for Research searches PubMed and Europe PMC, looks up ClinicalTrials.gov registry records and openFDA data, drafts the checklists a trial team prepares, and checks a manuscript against the reporting guideline it will be judged by. It is a research tool: nothing here is treatment advice, and results need a professional's review.
If you want to know how clinicians and health researchers use the product day to day, and which plan fits them, read AI for medical research. This page is about the field: the sources, the checklists and the documents.
Sources for medicine and health
| Source | What it holds | Where in the product | Credits |
|---|---|---|---|
| PubMed and Europe PMC | Biomedical journal articles, with study type labels | Medical source in Deep Research and Find Papers | Search 0; Summary 1, Deep 15, Detailed 45 (paid) |
| medRxiv and bioRxiv | Preprints, flagged as not peer reviewed | Preprints source | As the run's depth |
| OpenAlex and Semantic Scholar | 250M+ scholarly works across fields | All papers source | As the run's depth |
| ClinicalTrials.gov | Registry records: status, design, outcomes, posted results | Official Sources | 0 |
| openFDA | FDA adverse event reports, drug labels and recalls | Official Sources | 0 |
| My Library | Your own PDFs and imported references (Zotero, Mendeley) | My Library source | As the run's depth |
Embase, CINAHL and the Cochrane Library are not searched. For a systematic review, import their export files into the Systematic Review workspace.
Tools for medicine and health
| Tool | What it does | Credits |
|---|---|---|
| Deep Research | Cited answers from medical sources, filtered by study type, with retracted papers left out | Summary 1, Deep 15, Detailed 45 (paid) |
| Clinical research checklists | IRB submission, site initiation visit, ClinicalTrials.gov results posting, consent translation accuracy, genomics incidental-findings plan | Blank 0, tailored to your study 3 |
| Reporting guideline check | CONSORT 2010, STROBE 2007, PRISMA 2020, CARE 2013, STARD 2015, COREQ 2007, SRQR 2014, SQUIRE 2.0, SAGER and SAMPL, item by item | 5 per 10,000 words |
| Ethics statement | The approval statement for a manuscript: committee, reference number, consent or waiver, in the journal's wording | 1 |
| Case report (CARE) | A CARE-structured draft from de-identified case notes, with a timeline table | 15 |
| Clinical audit | An audit plan with criteria and targets, or the report of a completed cycle | Plan 3, report 15 |
| Data Lab | Survival (Kaplan-Meier, log-rank), diagnostic accuracy, ROC curves, logistic regression, calculated in your browser | 0; AI write-up 2 |
| Systematic Review workspace | Protocol, screening, extraction and the PRISMA 2020 diagram | AI screening 1 per 25 records |
The checklists sit in the Proposal Studio under Checklists, and the ethics statement in the Proposal Studio too. Case reports and clinical audits are Writing Studio modules. SAGER (sex and gender reporting) and SAMPL (statistical reporting) are new to the guideline check; both are paraphrased and numbered as published. The ethics approval statement template and the reporting checklists are free to copy.
Case reports and audits need de-identified data and the patient's or the organization's consent where it applies. Do not upload identifiable patient information.
A worked example
Example (invented for illustration): a research coordinator is preparing a two-site trial of a home exercise program after knee replacement.
- In Deep Research, Medical source, study types randomized controlled trial and systematic review, she asks what is known about supervised versus home exercise after knee replacement. Summary answer: 1 credit.
- In Official Sources she looks up registered trials on the same question on ClinicalTrials.gov, to see which outcomes other teams registered. 0 credits.
- She checks openFDA for adverse event reports on the anticoagulant most patients take after surgery, and notes in her protocol that report counts do not show causation. 0 credits.
- She opens the site initiation visit checklist, tailored to her two sites: 3 credits.
- Eighteen months later she checks the 8,000-word manuscript against CONSORT 2010 and SAGER: 5 credits each, 10 credits.
- She writes the ethics statement for the journal: 1 credit. The Kaplan-Meier analysis in Data Lab is free; the AI results paragraph costs 2.
Total: 17 credits, inside the Free plan's 45 a month.
Citation styles and reporting standards
Medical journals mostly use Vancouver or AMA references; both are among the 10,000+ styles in the editor, and Vancouver is one of the journal formatter's presets, with a blank Vancouver template in the editor. The reporting standards in the guideline check follow the study design: CONSORT 2010 for randomized trials (CONSORT 2025 has since been published ), STROBE 2007 for cohort, case-control and cross-sectional studies, STARD 2015 for diagnostic accuracy, CARE 2013 for case reports, SQUIRE 2.0 for quality improvement, PRISMA 2020 for systematic reviews, and COREQ 2007 or SRQR 2014 for qualitative work. SAGER and SAMPL apply across designs.
What it does not do
- It gives no diagnosis, treatment or dosing advice and is not a medical device or clinical decision support.
- openFDA reports are not proof that a drug caused an event, and they do not give incidence.
- It does not register a trial or post results for you. The results posting checklist helps you prepare; the sponsor or investigator submits on ClinicalTrials.gov.
- The guideline check drafts the checklist; the author confirms it. It does not guarantee compliance or acceptance.
- It does not search Embase, CINAHL, the Cochrane Library or registries other than ClinicalTrials.gov.
- Data Lab computes statistics in the browser; it does not choose your analysis plan or replace a trial statistician.
Who uses it
- Clinicians and health researchers: see AI for medical research.
- PhD candidates in health sciences: see AI for PhD students.
- Research teams in pharma, devices and contract research: see AI for industry.
For genes, proteins and variants behind a clinical question, see life sciences.
Learn more
- Site initiation visit checklist
- ClinicalTrials.gov results reporting
- SAGER guidelines
- openFDA for researchers
- How to search PubMed
- Levels of evidence and study types
Last updated 2026-11-13