Toward Transparency: Implications and Future Directions of Artificial Intelligence Prediction Model Reporting in Healthcare (Preprint)

Abstract

<sec> <title>UNSTRUCTURED</title> The rapid integration of Artificial Intelligence (AI) in healthcare emphasizes the transformative potential it holds for improving patient outcomes through data-driven decision-making. There is a drive toward implementing more complex predictive algorithms for disease diagnosis and prognosis. However, there are unique implications of AI that limit its clinical applicability and validity. To address these challenges, the Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD) group released an extension and update of its 2015 reporting guideline, the TRIPOD+AI, to enhance transparency and methodological rigor in regression and AI prediction model studies. The TRIPOD+AI framework encompasses an expanded scope, incorporating new domains such as fairness, open scientific practices, and patient and public engagement. It is anticipated that these augmented guidelines will facilitate the rigorous evaluation and subsequent adoption of artificial intelligence tools across diverse healthcare settings. </sec>

Ähnliche Arbeiten