Abuse of prescription drugs is a major public health problem associated with both significant costs to the health care system and consequences to the individual. Identifying patients who misuse these substances is often difficult, as clinicians must disentangle legitimate pain management needs from possible dependency or abuse. Thus, the purpose of this project is: 1) to develop a descriptive understanding of patient characteristics for individuals who receive a diagnosis of opioid dependence or abuse, 2) to develop a predictive model based on patient health care utilization characteristics to predict the development of opioid dependence or abuse among individuals who receive an opiate prescription, and 3) to examine variables, other than an abuse or dependence diagnosis, which may indicate problem use. This proposed study will be one of the first studies to systematically use insurance claims to predict opioid misuse. Data will be obtained from numerous nationally representative commercial health insurers through the MedStat MarketScan database. This proposed study will use a retrospective observational design among pharmacy, medical, and enrollment files, making use of the extensive database of commercially insured patients. The presence of an opioid prescription will be assessed between July 1, 1996 and July 1, 2004. Once patients with at least one prescription claim have been identified, the medical claims will subsequently look for ICD-9CM codes for either opioid abuse or dependence in the two years following this index period. Descriptive and bivariate statistics will be used to describe basic patient characteristics for those who either do or do not receive an opioid misuse diagnosis, and the results will be framed in terms of risk or protective factors for developing opiate misuse. Probit and logit models will be developed to determine the probability of opioid abuse or dependence based on predictor variables. Subsequently, other variables that may indicate misuse will be examined as outcome variables in a series of analyses. The results will provide a model to assist both clinicians and health care insurance administrators to use existing data to predict the potential for future complications. As an initial study examining the characteristics that predict opiate abuse or dependency, the findings will add to the available literature and may be used in primary or secondary prevention of prescription drug misuse. PUBLIC HEALTH SIGNIFICANCE The proposed project seeks to augment our understanding of prescription drug misuse through investigating characteristics of commercially-insured individuals who develop a problem with the misuse of opioid drugs. Data provided from an actuarial database will be used both to describe the characteristics of opiate-misusing individuals and to predict the development of an opiate use disorder following an initial medical prescription for an opiate.
|Cochran, Bryan N; Flentje, Annesa; Heck, Nicholas C et al. (2014) Factors predicting development of opioid use disorders among individuals who receive an initial opioid prescription: mathematical modeling using a database of commercially-insured individuals. Drug Alcohol Depend 138:202-8|