Is Machine Learning Changing Evidence-Based Medicine?

Abstract

Abstract The paradigm of evidence-based medicine has become a fundamental part of health care and health professions education since the 1990s, even as it has drawn criticism. Meanwhile, in the past decade, increased data collection and computational power have brought about advances in machine learning, such as diagnostic algorithms for pattern recognition in medical images. Machine learning has potential to address some of the limitations of evidence-based medicine, but also presents new questions about what counts as evidence in medicine. Critical appraisal of machine learning can become part of the teaching of evidence-based medicine.