Machine learning in practice-Evaluation of clinical value, guidelines

Abstract

Machine learning research in health care literature has grown at an unprecedented pace. This development has generated a clear disparity between the number of first publications involving machine learning implementations and that of orienting guidelines and recommendation statements to promote quality and report standardization. In turn, this hinders the much-needed evaluation of the clinical value of machine learning studies and applications. This appraisal should constitute a continuous process that allows performance evaluation, facilitates repeatability, leads optimization and boost clinical value while minimizing research waste. The present chapter outlines the need for machine learning frameworks in healthcare research to guide efforts in reporting and evaluating clinical value these novel implementations, and it discusses the emerging recommendations and guidelines in the area.