Background/Aims: Despite the drastic increases in prescription opioid misuse and abuse, risk assessment for aberrant drug-related behaviors prior to initiating opioid therapy for chronic non-cancer pain management continues to be underutilized in clinical practice. The purpose of the study was to investigate availability of data elements in the electronic health record that could be used to assess risk for aberrant drug-related behaviors with the automated Diagnosis, Intractability, Risk, Efficacy (DIRE) opioid risk assessment tool. Methods: DIRE is a 7-item tool usually administered by a clinician and used to predict efficacy of analgesia and patient compliance with long-term opioid therapy. Each factor is rated from 1 (least favorable case) to 3 (more favorable case for opioid prescribing). The total score is used for risk stratification with scores < 14 being an unsuitable candidate and scores > 14 being possible candidate for opioid therapy. The validation of the automated process versus clinician-administered rating was conducted using kappa analysis and test characteristics (sensitivity, specificity, positive and negative predictive values). Results: We developed structured data queries, natural language processing (NLP) algorithms for unstructured data, and data mapping strategies to populate the DIRE for a cohort of chronic non-cancer pain patients who were on long-term opioid therapy and who had a clinician-administered DIRE documented in the electronic health record prior to signing the most recent opioid agreement. We used ICD-9 diagnosis codes and NLP to populate diagnosis, psychological and chemical risk items. Encounter data and NLP were used for the reliability item. Intractability and social support items were populated using NLP only. Information on oral morphine equivalents, length of treatment, changes in pain scores and NLP were used to populate the efficacy item. If no information was found, most items were scored as 3 and efficacy as 2. The results of the NLP versus clinician-administered validation kappa analysis and test characteristics are pending. Conclusion: Among major barriers to appropriate management of chronic non-cancer pain with opioids are inadequate time and resources available to clinicians at a point of care for risk assessment. Novel approaches, such as NLP, may support clinical decision-making by automating the process of data extraction.