Use of the consort-ai protocol in reporting artificial intelligence applications in clinical trials

Abstract

Aim: Artificial intelligence has begun to be used in many branches of science, including medicine, with the development of technology in the 21st century. However, these studies are reported outside of a certain standard and this situation causes the studies to have a complex structure. Reporting the research results in accordance with international standards will make meaningful contributions to the process of understanding, interpreting and critically evaluating the research in question. CONSORT-AI (Consolidated Standards of Reporting Trials- Artificial Intelligence) is a reporting protocol that has become widespread in medicine and health due to the rapidly increasing number of artificial intelligence applications. In this study, it was aimed to evaluate whether the CONSORT-AI protocol, which was adapted into Turkish, was correctly understood by the researchers. Materials and Methods: 22 articles selected from TR Index journals were evaluated by the authors who were in the evaluator position with the CONSORT-AI protocol, and the evaluation results were tested with interauthor agreement analysis. Results: The obtained findings showed that the agreement between raters was statistically significant. However, in the studies examined, no article was found that fully observes the principles in the CONSORT-AI protocol. Conclusion: It is expected that the study will contribute to information professionals working in the field of medicine and health, researchers related to the subject, and editors and referees who are in the position of evaluating research by drawing attention to the principles to be followed in artificial intelligence-oriented research outputs.

Keywords

• artificial intelligence
• clinical studies
• protocol
• consort-ai