AI-assisted study finds limited code sharing in clinical prediction research
Source check: Source checked. Not independent scientific confirmation.
Finding: An AI-assisted empirical review found code-sharing statements in 482 of 3,967 clinical prediction studies (12.2%) and assessed 380 repositories. GPT-5.2 screened papers, extracted links and characterized repository features; human reviewers checked 500 papers and 35 repositories. The sample covered studies citing TRIPOD or TRIPOD+AI that were accessible through PubMed Central Open Access, so it does not represent all clinical research. Automated errors remain possible, and repository features do not establish executable reproducibility. The journal article was published October 9, before the November 19 MIT announcement.
Technology: GPT-5.2 (2025-12-11)
Family GPT-5: profile · models page
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- LLM performs screening, link extraction and structured feature assessment.
- Test and result
- Retrospective articles/repositories with human-reference checks.
- Limitation / still unknown
- Selection and classification error limit generalization; actual reproducibility was not established.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://pure-oai.bham.ac.uk/ws/portalfiles/portal/322809295/SounackT2026AIassisted.pdf · https://pubmed.ncbi.nlm.nih.gov/42855584/ · https://www.dana-farber.org/newsroom/news-releases/2026/ai-assisted-scoping-review-of-code-sharing-in-clinical-prediction-model-research
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 20 Nov 2026. Review scope: Institution-deposited publisher PDF: bounded results, Methods and limitations; publisher indexed text and PubMed/institutional publication-date metadata. Direct publisher route blocked; no code executed..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.