skip to main content

Explore Research Products in the NSF-PAR

Title: Rx-refill Graph Neural Network to Reduce Drug Overprescribing Risks

Prescription (aka Rx) drugs can be easily overprescribed and lead to drug abuse or opioid overdose. Accordingly, a state-run prescription drug monitoring program (PDMP) in the United States has been developed to reduce overprescribing. However, PDMP has limited capability in detecting patients' potential overprescribing behaviors, impairing its effectiveness in preventing drug abuse and overdose in patients. In this paper, we propose a novel model RxNet, which builds 1) a dynamic heterogeneous graph to model Rx refills that are essentially prescribing and dispensing (P&D) relationships among various patients, 2) an RxLSTM network to explore the dynamic Rx-refill behavior and medical condition variation of patients, and 3) a dosing-adaptive network to extract and recalibrate dosing patterns and obtain the refined patient representations which are finally utilized for overprescribing detection. The extensive experimental results on a one-year state-wide PDMP data demonstrate that RxNet consistently outperforms state-of-the-art methods in predicting patients at high risk of opioid overdose and drug abuse.

  more » « less
Award ID(s):
2214376 2203262 2217239 2218762
NSF-PAR ID:
10358124
Author(s) / Creator(s):
; ; ; ; ; ;
Publisher / Repository:
International Joint Conference on Artificial Intelligence
Date Published:
Journal Name:
The 31st International Joint Conference on Artificial Intelligence (IJCAI)
Page Range / eLocation ID:
5379 to 5383
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ( , Mathematical Medicine and Biology: A Journal of the IMA)
    null (Ed.)
    Abstract In 2016, more than 11 million Americans abused prescription opioids. The National Institute on Drug Abuse considers the opioid crisis a national addiction epidemic, as an increasing number of people are affected each year. Using the framework developed in mathematical modelling of infectious diseases, we create and analyse a compartmental opioid-abuse model consisting of a system of ordinary differential equations. Since $40\%$ of opioid overdoses are caused by prescription opioids, our model includes prescription compartments for the four most commonly prescribed opioids, as well as for the susceptible, addicted and recovered populations. While existing research has focused on drug abuse models in general and opioid models with one prescription compartment, no previous work has been done comparing the roles that the most commonly prescribed opioids have had on the crisis. By combining data from the Substance Abuse and Mental Health Services Administration (which tracked the proportion of people who used or misused one of the four individual opioids) with data from the Centers of Disease Control and Prevention (which counted the total number of prescriptions), we estimate prescription rates and probabilities of addiction for the four most commonly prescribed opioids. Additionally, we perform a sensitivity analysis and reallocate prescriptions to determine which opioid has the largest impact on the epidemic. Our results indicate that oxycodone prescriptions are both the most likely to lead to addiction and have the largest impact on the size of the epidemic, while hydrocodone prescriptions had the smallest impact. 
    more » « less
  2. ; ; ; ( , Proceedings American Academy of Forensic Sciences)
    null (Ed.)
    In this century, drug abuse continues to be a national crisis. Since 1999, the number of opioid-induced overdoses has increased four-fold to more than 500,000 deaths. The microbiota gut-brain-axis is a bidirectional circuit that links the neural, endocrine, and immunological systems with gut microbial communities. Gut microbiota play significant roles in human mind and behavior, specifically pain perception, learning capacity and memory, mood, and emotion, and anxiolytic effects and temperament. Also, disruptions in the gut microbiome have been associated with substance use disorders. While much research still needs to be performed, elucidating the interplay of gut microbiota in substance abuse disorders may produce promising avenues for future forensic development. The goal of the current study was to determine gut microbiome composition in substance abuse disorder cases using transverse colon tissues of 21 overdose criminal cases versus 19 non-overdose-related cases. The hypothesis was that postmortem samples of the same origin will reveal similar taxonomic relationships. Using weighted UniFrac analysis, drug abuse was found to be a significant factor in determining microbiome similarity (F = 1.93; df = 1, 35; p < 0.048; R2 = 0.05) indicating that there are detectable differences in composition that are attributable to substance abuse. Using unweighted UniFrac, however, sex was instead found to be a significant predictor of microbiome similarity (F = 1.88; df = 1, 30; p = 0.028; R2 = 0.05). A heatmap was generated of the relative abundances of the 30 most prevalent bacteria per case and their associated substance profile. The results revealed that samples of the same origin cluster together, showing a high degree of similarity between samples and a low degree of similarity among samples of different origin. This examination of human transverse colon microflora in decomposing cadavers expands the emerging literature on postmortem microbial communities, which will ultimately contribute to advanced knowledge of putrefaction. 
    more » « less
  3. ; ; ; ; ; ; ; ; ; ( , PloS one)
    This paper examines network prominence in a co-prescription network as an indicator of opioid doctor shopping (i.e., fraudulent solicitation of opioids from multiple prescribers). Using longitudinal data from a large commercially insured population, we construct a network where a tie between patients is weighted by the number of shared opioid prescribers. Given prior research suggesting that doctor shopping may be a social process, we hypothesize that active doctor shoppers will occupy central structural positions in this network. We show that network prominence, operationalized using PageRank, is associated with more opioid prescriptions, higher predicted risk for dangerous morphine dosage, opioid overdose, and opioid use disorder, controlling for number of prescribers and other variables. Moreover, as a patient’s prominence increases over time, so does their risk for these outcomes, compared to their own average level of risk. Results highlight the importance of co-prescription networks in characterizing high-risk social dynamics. 
    more » « less
  4. ; ; ; ; ; ( , Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies)
    Opioid use disorder is a medical condition with major social and economic consequences. While ubiquitous physiological sensing technologies have been widely adopted and extensively used to monitor day-to-day activities and deliver targeted interventions to improve human health, the use of these technologies to detect drug use in natural environments has been largely underexplored. The long-term goal of our work is to develop a mobile technology system that can identify high-risk opioid-related events (i.e., development of tolerance in the setting of prescription opioid use, return-to-use events in the setting of opioid use disorder) and deploy just-in-time interventions to mitigate the risk of overdose morbidity and mortality. In the current paper, we take an initial step by asking a crucial question: Can opioid use be detected using physiological signals obtained from a wrist-mounted sensor? Thirty-six individuals who were admitted to the hospital for an acute painful condition and received opioid analgesics as part of their clinical care were enrolled. Subjects wore a noninvasive wrist sensor during this time (1-14 days) that continuously measured physiological signals (heart rate, skin temperature, accelerometry, electrodermal activity, and interbeat interval). We collected a total of 2070 hours (≈ 86 days) of physiological data and observed a total of 339 opioid administrations. Our results are encouraging and show that using a Channel-Temporal Attention TCN (CTA-TCN) model, we can detect an opioid administration in a time-window with an F1-score of 0.80, a specificity of 0.77, sensitivity of 0.80, and an AUC of 0.77. We also predict the exact moment of administration in this time-window with a normalized mean absolute error of 8.6% and R2 coefficient of 0.85. 
    more » « less
  5. ; ; ; ; ; ( , Medical Anthropology Quarterly)
    Abstract

    Public narratives often attribute the opioid overdose epidemic in the United States to liberal prescribing practices by health care providers. Consequently, new monitoring guidelines for the management of opioid prescriptions in patients with chronic pain have become recognized as key strategies for slowing this tide of overdose deaths. This article examines the social and ontological terrain of opioid‐based pain management in an HIV clinic in the context of today's opioid overdose epidemic. We engage with anthropological analyses of contemporary drug policy and the nonverbal/performative ways patients and clinicians communicate to theorize the social context of the opioid overdose epidemic as a “situation,” arguing that the establishment of new monitoring strategies (essentially biomedical audit strategies) trouble patient subjectivity in the HIV clinic—a place where that subjectivity has historically been protected and prioritized in the establishment of care.

     
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email