TY - JOUR
AU - I. Lopez
AU - S. Fouladvand
AU - S. Kollins
AU - C. A. Chen
AU - J. Bertz
AU - T. Hernandez-Boussard
AU - A. Lembke
AU - K. Humphreys
AU - A. S. Miner
AU - J. H. Chen
A1 - 
AB - Medications such as buprenorphine-naloxone are among the most effective treatments for opioid use disorder, but limited retention in treatment limits long-term outcomes. In this study, we assess the feasibility of a machine learning model to predict retention vs. attrition in medication for opioid use disorder (MOUD) treatment using electronic medical record data including concepts extracted from clinical notes. A logistic regression classifier was trained on 374 MOUD treatments with 68% resulting in potential attrition. On a held-out test set of 157 events, the full model achieved an area under the receiver operating characteristic curve (AUROC) of 0.77 (95% CI: 0.64-0.90) and AUROC of 0.74 (95% CI: 0.62-0.87) with a limited model using only structured EMR data. Risk prediction for opioid MOUD retention vs. attrition is feasible given electronic medical record data, even without necessarily incorporating concepts extracted from clinical notes.
AD - Department of Medicine, Biomedical Informatics Research, Stanford Medicine, Stanford University, CA.; Holmusk, NY.; Primary Care and Population Health, Stanford Medicine, Stanford University, CA.; Center for the Clinical Trials Network, National Institute on Drug Abuse, MD.; Department of Psychiatry and Behavioral Sciences, Stanford Medicine, Stanford University, CA.; Veterans Affairs Palo Alto Health Care System, Palo Alto, CA.
AN - 38222349
BT - AMIA Annu Symp Proc
C5 - Opioids & Substance Use; HIT & Telehealth
DP - NLM
ET - 20240111
JF - AMIA Annu Symp Proc
LA - eng
N2 - Medications such as buprenorphine-naloxone are among the most effective treatments for opioid use disorder, but limited retention in treatment limits long-term outcomes. In this study, we assess the feasibility of a machine learning model to predict retention vs. attrition in medication for opioid use disorder (MOUD) treatment using electronic medical record data including concepts extracted from clinical notes. A logistic regression classifier was trained on 374 MOUD treatments with 68% resulting in potential attrition. On a held-out test set of 157 events, the full model achieved an area under the receiver operating characteristic curve (AUROC) of 0.77 (95% CI: 0.64-0.90) and AUROC of 0.74 (95% CI: 0.62-0.87) with a limited model using only structured EMR data. Risk prediction for opioid MOUD retention vs. attrition is feasible given electronic medical record data, even without necessarily incorporating concepts extracted from clinical notes.
PY - 2023
SN - 1559-4076
SP - 1067
EP - 1076+
ST - Predicting premature discontinuation of medication for opioid use disorder from electronic medical records
T1 - Predicting premature discontinuation of medication for opioid use disorder from electronic medical records
T2 - AMIA Annu Symp Proc
TI - Predicting premature discontinuation of medication for opioid use disorder from electronic medical records
U1 - Opioids & Substance Use; HIT & Telehealth
VL - 2023
VO - 1559-4076
Y1 - 2023
ER -