Research has shown that opioid misuse has significant financial and clinical implications for both patients and risk-bearing entities. This paper analyzes the risks and costs of opioid misuse and compares machine learning techniques to identify high-risk opioid users against rules-based approaches suggested by CMS.