Prescription Opioid Misuse in a Large HMO

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): Licit and illicit use of prescription opioids, as well as associated morbidity and mortality, have increased significantly in the past six years in Washington State and nationally. Prescription opioid misuse is a constellation of behaviors characterized by unstable or chaotic patterns of medication use and aberrant patterns of medical care seeking. The biopsychosocial model will be used to examine the transition from opioid use to opioid misuse in the context of chronic pain treatment with opioids from the perspectives of health systems and clinicians. The specific aims of this application are to: 1. Determine the prevalence of prescription opioid misuse, abuse and dependence among those prescribed chronic opioids for non-malignant pain, and to determine the correlation between opioid misuse and the commonly used diagnostic categories of opioid abuse and dependence. 2. Model the utility of automated medical record data in differentiating between opioid misusers and non- misusers and determine the screening properties of the derived model (e.g. sensitivity, specificity, positive and negative predictive values, likelihood ratio). 3. Assess the utility of the automated data model obtained in Aim #2, enhanced with brief interview items related to pain severity, substance abuse, and mental health symptoms to identify potential opioid misusers. To meet these aims, we propose a retrospective cohort study of moderate to long-term opioid users in a large HMO in Washington State. Eligible subjects include all HMO enrollees, including those receiving opioids from primary and specialty care physicians. This large (n=1,000) and diverse sample will allow for greater generalizability of the findings than most of the preliminary studies to date, which were often conducted in pain clinics or the V.A. with its primarily male clientele. Data on medication use, medical conditions, and other risk factors for misuse will be obtained from automated data files. A structured phone interview will be used to obtain the predictor variables for Aim 3, the outcome variables related to opioid misuse, and DSM IV diagnoses of opioid abuse and dependence. Models to identify predictors of opioid misuse will be developed with half of the sample and validated with the other half. De-identified data on non- respondents will be utilized to adjust analyses for any non-response bias.