TRIPOD+AI: Updated Reporting Guidelines for Clinical Prediction Models
Sherri Rose joins global network of health experts to improve the transparency and accuracy of prediction algorithms.
Sherri Rose is part of a global consortium of experts who have updated the TRIPOD guidelines for prediction algorithms to include machine learning and AI methods. The new TRIPOD+AI guidelines were recently published in BMJ with the ultimate goal of improving patient care.
The first TRIPOD—or Transparent Reporting of multivariable prediction model for Individual Prognosis or Diagnosis—statement was published in 2015 to provide recommendations for studies developing or evaluating the performance of prediction models. These algorithms are widely used to predict health outcomes and support clinical decision-making.
But it’s been nearly a decade since the TRIPOD guidelines for prediction algorithms were published and there have been many methodological advances using artificial intelligence powered by machine learning methods. Thousands of these predictive models are published each year and there have been longstanding concerns about the transparency and accuracy of these models, giving editors and peer reviewers of medical journal articles incomplete or even inaccurate reporting.
“Poor reporting of a model might also mask flaws in the design, data collection, or conduct of a study that, if the model was implemented in the clinical pathway, could cause harm,” they wrote. “Better reporting can create more trust and influence patient and public acceptability of the use of prediction models in healthcare.”
The global consortium of researchers wrote that the new guidelines supersede the TRIPOD 2015 guidelines. They presented a 27-item checklist with detailed explanations of each reporting recommendation, and the TRIPOD+AI for Abstracts checklist.
