Latent class analysis of non-opioid dependent illegal pharmaceutical opioid users in Ohio

Background: Increases in non-medical use of pharmaceutical opioids in the U.S. have resulted in increases in opioid dependence and unintentional overdose deaths. We characterize heterogeneity in opioid use patterns among a community-based sample of 18-23 year-olds who use non-medical pharmaceutical opioids, yet are not opioid dependent. Methods: Respondent-driven sampling recruited 390 participants. Latent class analysis stratified by racial/ethnic group identified subgroups of non-medical opioid users based on: six-month frequency of use; number of opioid disorder criteria; oral vs. non-oral administration; number of types of opioids used; use of CNS depressants while under using opioids; and reason for opioid use. Multinomial logistic regression estimated the significance of covariates in predicting class membership. Results: Within whites and non-white groups, three classes emerged that were, generally, hierarchically ordered with respect to negative characteristics associated with non-medical opioid use. Within each group, the class with the least negative characteristics also had the highest proportion of individuals who use opioids only to self-medicate a health problem. Within each group's three classes, a larger proportion who had >= 2 opioid abuse and dependence disorder criteria always coincided with a larger proportion who use opioids >= 3 days per week, a larger proportion who used CNS depressants while under the influence of opioids, and a smaller proportion who used opioids only to self-medicate. Conclusion: Differences in patterns of opioid use within each racial/ethnic group of young people who are not opioid dependent suggest the need for tailored interventions designed to reduce the risk of transition to opioid dependence. (C) 2013 Elsevier Ireland Ltd. All rights reserved.