PURPOSE: To empower healthcare researchers with a greater understanding of predictive analytics, including recent methodological innovations DESCRIPTION: The volume and diversity of studies and applications involving predictive analytics is rapidly growing, from risk stratification tools for major chronic conditions, to predictive algorithms to screen for patients with undiagnosed rare conditions. Many health care researchers are unfamiliar with what is meant by predictive analytics along with recent, relevant methodological innovations from the fields of statistical learning and machine learning. Predictive analytics is an area shrouded in mystery and hence its potential is unrealized. This workshop will help demystify predictive analytics for health care researchers. The workshop will assess what is meant by predictive analytics, drawing on illustrative examples from the health domain. Important predictive analytics concepts will be described, with an emphasis on intuitive and accessible explanations. The importance of clear problem formulation will be emphasized and the interdependence between study objectives, data complexity, and choice of methods discussed. Methods to evaluate and compare model performance, such as cross-validation, will be described. Approaches for deriving predictions, such as Random Forests, Adaptive Boosting and Support Vector Machines, will be explained and contrasted with more familiar techniques, such as multiple regression. Case studies will be used throughout to illustrate concepts. This workshop will contribute to knowledge building in the increasingly important area of predictive analytics. As demands for predictive analytics applications increases and as underlying data becomes ever more voluminous and complex, the need for greater clarity and understanding on modern predictive analytics methods is paramount. The session will be highly interactive, with audience members encouraged to share their experience and ask panel members questions. Attendees will be able to download a ‘fake-data’ example coded in R to continue their learning experience. The panel will comprise experts from academia, industry, and specialist service providers.