Estimating the prevalence of and characteristics associated with prescription opioid diversion among a clinic population living with HIV: Indirect and direct questioning techniques.

Abstract

Evidence suggests that over 50 % of individuals who used prescription opioids non-medically obtained their prescription from friends or family. Despite its high consequence, reliable opioid diversion prevalence estimates are lacking due to social desirability bias. We used indirect questioning, a technique designed to measure sensitive behaviors, to assess the prevalence of prescription opioid diversion among a cohort of individuals with HIV. We randomized 581 participants from a large urban HIV clinical cohort to answer either a direct or indirect question about opioid diversion between October 2016-July 2018. We estimated the prevalence of diversion under each method. We also estimated diversion prevalence in subsets of the sample by age, sex, race, HIV risk group, substance use, and mental health co-morbidities. Of 1,285 patients screened, 581 (45.2 %) reported ever having received an opioid prescription. Of these, 252 (43.4 %) directly answered whether they had ever diverted opioids and 313 (53.9 %) answered the indirect question. The prevalence of opioid diversion under direct and indirect questioning was 6.3 % (95 % CI 3.7 %-10.1 %) and 15.3 % (95 % CI 10.4 %-20.3 %), respectively. In unadjusted analyses, males, non-African Americans, and patients with a history of illicit drug use had a higher diversion prevalence. In adjusted analyses, ever having used cocaine was most associated with diversion (OR 15.67, 95 % CI 0.93-263.17). Opioid diversion was common among this population, with the estimated prevalence more than doubling under the indirect questioning method designed to elicit less biased responses.