Uncovering Problem Opioid Use In Chronic Opioid Therapy With Patient Medical Records
In the NLP search, researchers screened the clinical notes of COT patients for keywords including abuse, misuse, overuse, addiction, and diversion.
At the end of the last century, there was a movement in pain management toward opioid use for chronic pain. The rationale was that people with chronic pain were suffering needlessly because of concerns that chronic opioid therapy (COT) would lead to addiction. From 1999 to 2010, opioid prescriptions increased dramatically. During that same period, we also have seen dramatic increases in opioid misuse, and opioid-related deaths increased threefold.1
Studies and surveys of opioid misuse and abuse in COT have ranged widely, from 3% to 40%. A 2015 study published in PAIN used a novel technique to uncover problem opioid use in COT and identified 14 risk indicators that may predict problematic opioid use during COT. The novel technique is natural language processing (NLP), which is a computer-generated search of medical records to look for keywords that may indicate a problem or concern.1
"NLP is like an advanced Google search. We searched the free text in medical notes for keywords and keywords in association with other keywords to uncover problematic use of opioids and behaviors around opioids,” said Roy E. Palmer, PhD, field medical director at Pfizer in Seattle, Washington, and lead author of the study.
In the NLP search, researchers screened the clinical notes of COT patients for keywords including abuse, misuse, overuse, addiction, and diversion. Researchers developed an algorithm of 1,248 word combinations that could indicate problem use.1