Abstract
BackgroundThe opioid epidemic in the United States is averaging over 100 deaths per day due to overdose. The effectiveness of opioids as pain treatments, and the drug-seeking behavior of opioid addicts, leads physicians in the United States to issue over 200 million opioid prescriptions every year. To better understand the biomedical profile of opioid-dependent patients, we analyzed information from electronic health records (EHR) including lab tests, vital signs, medical procedures, prescriptions, and other data from millions of patients to predict opioid substance dependence.ResultsWe trained a machine learning model to classify patients by likelihood of having a diagnosis of substance dependence using EHR data from patients diagnosed with substance dependence, along with control patients with no history of substance-related conditions, matched by age, gender, and status of HIV, hepatitis C, and sickle cell disease. The top machine learning classifier using all features achieved a mean area under the receiver operating characteristic (AUROC) curve of ~ 92%, and analysis of the model uncovered associations between basic clinical factors and substance dependence. Additionally, diagnoses, prescriptions, and procedures prior to the diagnoses of substance dependence were analyzed to elucidate the clinical profile of substance-dependent patients, relative to controls.ConclusionsThe predictive model may hold utility for identifying patients at risk of developing dependence, risk of overdose, and opioid-seeking patients that report other symptoms in their visits to the emergency room.
Highlights
In a highly visible report it was described how drug overdose deaths have substantially increased in the United States from 2010 to 2015 [1]
The challenges for physicians combating the opioid epidemic include: 1) Determining which patients are at risk of developing opioid dependence when prescribed these medications for conventional pain treatment; 2) Determining which patients known to be addicted to opioids are most at risk of opioid overdose; and 3) Identifying drug-seeking patients who visit the Emergency Department (ED) for
We describe the application of a machine learning classifier to predict substance dependence based on lab tests and vital signs using patient data derived from the Mount Sinai Medical Center (MSMC) electronic health records (EHR) system
Summary
In a highly visible report it was described how drug overdose deaths have substantially increased in the United States from 2010 to 2015 [1]. Previous studies of biomedical variables predictive of opioid misuse and abuse have unraveled several salient factors, including chronic opioid prescriptions, history of psychiatric illness, non-opioid substance disorders, having a family member diagnosed with an opioid use disorder, the use of multiple pharmacies to fill prescriptions, having hepatitis C, and tobacco addiction [7,8,9,10] These studies are based on various types of data, including pharmacy prescriptions, insurance claims, vital signs, and medical notes from electronic health records (EHR). To determine and predict opioid misuse, Hylan et al utilized natural language processing to analyze clinicians’ notes All these past studies point to few common clinical factors that contribute to opioid pathology. To better understand the biomedical profile of opioid-dependent patients, we analyzed information from electronic health records (EHR) including lab tests, vital signs, medical procedures, prescriptions, and other data from millions of patients to predict opioid substance dependence
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
