We were delighted to host Prof Ben Van Calster, Professor of Medical Statistics at KU Leuven in Belgium, in the Institute today and gave a truly thought-provoking seminar as part of the Royal Statistical Society Highlands Local Group series.
Clinical prediction models, now often referred to as “clinical AI,” are ubiquitous in healthcare. These models estimate the risk of adverse clinical outcomes based on predictor variables derived from clinical measurements and/or medical images.
Ben discussed issues that may undermine the trustworthiness of the predictions generated by these models, including challenges related to model development, validation, heterogeneity, and uncertainty.
Please get in touch if you would like more information about the work we do locally on diagnostic and prognostic models, medical decision-making, artificial intelligence in healthcare, and research methodology: BHDS@abdn.ac.uk