Opioid-related harm is an urgent patient safety priority. Identifying patients at higher risk of harm is a critical aspect of opioid safety. This quality improvement team developed a predictive model, based on electronic health record data, to identify high-risk opioid users in order to provide targeted monitoring and intervention via a clinical decision support tool. The model included known risk factors for opioid-related harm, such as type of medication, dose, and coprescribed sedating medications as well as medical and mental health conditions. Investigators developed and validated the model using data from 2010 and tested its ability to predict overdose or suicide attempt in 2011. The model successfully and prospectively identified patients at risk for suicide attempt or overdose. They then used the electronic health record to provide physicians with an overdose or suicide risk estimate and a checklist of risk mitigation strategies at the point of care. The authors suggest that further study of the implementation of this risk mitigation strategy in primary care is needed.