Guidelines for clinical trial protocols for interventions involving artificial intelligence: the SPIRIT-AI extension

AI Generated Summary*

SPIRIT-AI, developed alongside CONSORT-AI for trial reports, extends the SPIRIT 2013 protocol reporting standard to trials of interventions incorporating artificial intelligence. Candidate items came from a literature review and expert input, were rated in a two-round Delphi survey with 103 respondents, debated at a 31-stakeholder consensus meeting using an 80% inclusion threshold, and piloted with 34 people. The checklist adds 15 items (12 extensions, 3 elaborations) addressing intended use and users, prior evidence, site integration needs, input-data eligibility and quality handling, algorithm version, human-AI interaction, output definition, performance error analysis, and code access. Only seven AI trials had been published during development, and continuously learning systems were discussed but judged too early to include.

*This summary was generated by AI and is published unedited. Oku does not alter these summaries. It may contain errors or omissions and is provided for general informational purposes only. Oku does not guarantee its accuracy, completeness, or reliability. For authoritative information, please refer to the original, peer-reviewed article.