Regenerative Medicine Utrecht


Reporting of Artificial Intelligence Prediction Models – contribute to setting up new guidelines!

Technologies are shaping the future, and health care is certainly no exception. A hot topic is digital health, where emerging technologies form the basis of a potential health care revolution. Health apps, e-health technologies, intelligent monitoring and other data-driven technologies could advance innovation in medical research and improve individual care.

Health care digitalization is especially driven by advancements in artificial intelligence, which includes machine learning: computer algorithms learn from data and can help to identify patterns and to make predictions. However, it can be difficult to distinguish value-producing innovations from hype. If not used with proper guidance, knowledge, or expertise, artificial intelligence in clinical medicine has methodological shortcomings and poor reproducibility. The clinical community should not get mesmerized by the artificial intelligence and machine learning revolution. Prediction models must be appropriately developed, evaluated, and tailored to different situations before they can be used in daily medical practice.

There is a massive growth in the number of prediction model studies, in which different models are being developed partly for the same target population and outcomes. Comprehensive and transparent reporting of the key aspects of machine learning prediction model studies is vital and yet often incomplete.

Prof. Gary Collins (Oxford) and Prof. Carl Moons (UMC Utrecht and UU) announce a new initiative to develop a version of the TRIPOD statement, a reporting guidance tool, specific to machine learning (TRIPOD-ML). The aim of TRIPOD-ML will be to focus on the introduction of machine learning prediction algorithms, building on a long and established methodology of prediction research, while harmonizing terminology. To make this new guideline as usable as possible, they invite interested individuals from the machine learning community, particularly those with an interest in machine learning and artificial intelligence applications in health care, to contribute.

Read more here or contact Prof. Carl Moons.