Artificial Intelligence and Machine Learning in Clinical Trials

What have AI and machine learning to do with medical trials and decision making? Currently, data gathered in trials is the established way to determine the effectiveness of new treatments. Meanwhile, data is also routinely gathered to administer the everyday treatment of patients: advocates of machine learning, including large technology companies, are suggesting that these medical records can be used to achieve precision medicine. We outline the idea of machine learning (some forms of which are familiar) and, based on experience developing decision-support systems, explain the challenges of this vision and suggest how the current approach to trials could be developed into continuous learning to improve treatment.