Advisory Committee Chair
Advisory Committee Members
Date of Award
Degree Name by School
Doctor of Philosophy (PhD) School of Health Professions
A large portion of the American Opioid Crisis is due to opioid naÃ¯ve patients who become new persistent post-surgical opioid users, although the risk factors for the development of this addiction are not well studied. The objective of this study was to analyze multiple layers of pre-operative and procedural risk factors using an ecological perspective theoretical framework in adult patients undergoing invasive surgery. We performed a retrospective analysis of 13,970 opioid naÃ¯ve adults in a mixed surgical cohort with data available at the University of California Los Angeles that was merged with narcotics data for the State of California (IRB#19-000625). Opioid naÃ¯ve patients were defined as those with no opioid prescriptions filled 365 to 31 days before surgery. Univariate analyses of 46 risk factors were performed using various regression and machine learning methods such as stepwise regression, lasso and elastic net models, to develop, train and internally validate the final model. Persistent post-surgical opioid use was defined as a refill or supply of opioid medication between 90-365 post-operative days. The incidence of new persistent post-surgical opioid use was 21.2%. Top risk factors included female sex, ages between 35 to 79 years old, white, black and Pacific Islander races, pulmonary hypertension, bipolar disorder, tobacco use, depression, and Farsi-speaking language. Procedural factors included moderate and major surgical complexity, and providers of ophthalmology, cardiology and podiatry procedures. Spinal fusion, cataract removal, transurethral resection of prostate and spinal nerve transection were procedures with increased risk. Model validation scores indicated backwards stepwise regression was the best overall model. Persistent post-surgical opioid use is a significant and common issue impacting 1 in 5 opioid naÃ¯ve patients undergoing invasive surgery in our cohort. We were able to combine two disparate data sources in a novel way to create a high-quality dataset. We identified pre-surgical referable factors of smoking, hypertension, bipolar disorder, and depression. Future research is needed to better understand the unexpected high incidence of persistent opioid use seen with certain procedures.
Pittet, Gia Marie, "Disrupting the Perioperative Opioid Gateway: Identification of Risk Factors for New Persistent Post-Surgical Opioid Use" (2021). All ETDs from UAB. 891.