Uncovering Problem Opioid Use In Chronic Opioid Therapy With Patient Medical Records

What Does the Study Mean for Clinicians?

The electronic medical records of the patients in the study contained close to 8 million clinical notes.1 “Our study showed that there is a large amount of clinical data painstakingly recorded but not being appreciated. Clinicians can benefit from the emerging technology of NLP to help them make better decisions. As this technology develops and becomes more available, it could be very useful,” Palmer said. 

In a commentary that accompanied the study in PAIN, Jane C. Ballantyne, MD, professor of anesthesiology and pain medicine at the University of Washington, Harborview Medical Center in Seattle, draws attention to the disconnect between the ICD-9 coding and the NLP findings.2 

“Clinicians find it difficult to label their patients with the stigma of addiction. There is stigma for the patient but also risk for the clinician. If you note or diagnose a patient as having problems with opioid abuse, you need to answer the question of why you are continuing to prescribe opiates,” Ballantyne said. 

The authors of the study noted that both ICD-9 and NLP may underestimate the risk of problem opiate use.1 “We feel that an abuse rate of 9 to 10% probably underappreciates the problem. It is likely that some problem use is either not documented or not recognized,” Palmer noted. 

“The risk indicators developed for the study could be very important for clinicians. We have scoring systems and risk stratification scores for other disorders. Clinicians may use these risk indicators to identify patients at risk. Then the question becomes what to do. The best course may be not to start those patients on opioids therapy or to take them off therapy. Most pain management experts agree that there are better options for many patients with chronic pain,” Ballantyne said.

“Future studies could build on our risk indicator findings to develop an opioid use risk stratification scale. This could help identify patients at risk for overuse, misuse, or abuse. The scale would need to be validated and tested but could be a valuable tool for the future,” Palmer said. 


1. Palmer RE, et al. Pain. 2015; doi: 10.1001/jamaoncol.2015.2316.

2. Ballantyne, J, et al. Pain. 2015; doi: 10.1001/jamaoncol.2015.2316.