The ISPOR Machine Learning Methods Emerging Good Practices Task Force has developed guidance for HEOR and decision-makers in the use of machine learning (ML) methods. The report considers six applications of ML methods that are important to HEOR: (1) machine learning-assisted cohort selection; (2) feature selection; (3) predictive analytics; (4) causal inference; (5) health economic evaluation; and reflection on (6) ethics and transparency. The task force’s goal is to introduce ML methods and their value in conducting research on health economics, as well as patient- and system-level outcomes research to the ISPOR audience and to describe problems for which ML methods are appropriate with particular attention to four major content areas: the prediction of risk of various health care events; the causal estimation of treatment effects; developing models for economic evaluation; and model/data transparency. Dr. Crown will moderate the session with Dr. Kreif discussing the difference in using ML for prediction / classification versus causal inference. Her examples will illustrate the use of ML for special populations (personalized medicine and the identification of patient clusters) and its value for identifying potentially beneficial treatments for COVID-19, rather than waiting for the results of randomized trials. Dr. Jonsson will discuss the payer perspective on the use of evidence generated by ML with particular focus on issues of transparency--both with respect to the black box nature of some ML algorithms, and clarity in how the analyses were conducted. He will also introduce the PALISADE Checklist. ML is a very powerful tool for exploring data but when does the exploration end and the analysis begin? Dr. Felizzi will discuss a case study of improving the prediction of disease progression with health systems implications and optimal delivery of care. Polling will be used during the presentations and for feedback on the task force’s recommendations.