Paper
Discordant reporting of nonmedical opioid use: reply to letter to the editor
Published Jan 2, 2017 · J. Palamar, J. Shearston, C. Cleland
The American Journal of Drug and Alcohol Abuse
2
Citations
0
Influential Citations
Abstract
Below is our brief invited response to Ruan, Chiravuri, and Kaye who submitted a Letter to the Editor (1) regarding our recent manuscript “Discordant reporting of nonmedical opioid use in US high school seniors,” published here in The American Journal of Drug and Alcohol Abuse (2). The published prevalence estimates of nonmedical use of specific opioids such asVicodin reported by theNational Survey onDrugUse andHealth (NSDUH) are presented in Table 1.89B of the 2013 annual report. This table is found in the “Detailed Table” section of their annual report. While the most formal method of citing NSDUH annual results is to cite the main report (to appropriately provide the investigators with credit), here (3), we provide a citation to the direct online tables from the full report.We, too, wish these tables were easier to locate online as not all new versions of NSDUH reports contain links to these detailed tables. Similarly, these prevalence estimates can be computed via the publicly available NSDUH datasets made available by The Interuniversity Consortium for Political and Social Research (ICPSR). We are aware of the study by McCabe et al. (4) described by the authors and feel their results are a great contribution to the literature. However, as described in their report, they only utilized Survey Form 1 of the Monitoring the Future (MTF) dataset. This form contains variable V1564, which asks students about medical or prescribed use of opioids. This is something we ourselves wanted to consider in our analyses; however, as we describe in our report, only three (of six) MTF forms included variables assessing self-reported nonmedical use of (1) opioids as a general category, (2) Vicodin, and (3) OxyContin, and these three forms were Form 3, 5, and 6. Thus, we were not able to consider medical or prescribed use as McCabe et al. analyzed in Form 1 because it does not include the standard questions about Vicodin or OxyContin that are included in the other three forms. We do appreciate the authors’ insight and regret omitting a mention in our limitations section that it is possible that some students may have not closely read the MTF definition for nonmedical use which is accompanied with all nonmedical opioid use questions. Similar to our point above, since most variables (other than some core demographic and drug use variables) are split up between six different survey forms, we are unable to determine how the psychological correlates of use relate to reported nonmedical opioid use in our analytic sample. Reasons for use of various drugs, for example, are only queried in Form 1, so similar to what we discus above, in our analytic sample we are unable to determine the potential associations between reasons for use and nonmedical use of Vicodin and OxyContin, or regarding potential discordant reporting. Finally, we are aware that only a small portion of opioid users initiate heroin and we made no such claim implying that a large proportion of opioid users move on to heroin. In the introduction of this report and in the results of our other report published in Drug and Alcohol Dependence (5), we describe the link between nonmedical opioid use and heroin use in a correlational manner. Results did suggest that almost a quarter of students reporting nonmedical opioid use more than 40 times also reported heroin use, and over three quarters of heroin users also reported lifetime nonmedical use of opioids. We do fluctuate terminology with regard to heroin “use” and “initiation” at times as most students in the MTF sample who have reported heroin use reported using infrequently (hence “initiation” and not necessarily “continued use”). Wewere very cautious in both reports not to infer temporality as MTF is cross-sectional. We agree that if a student surveyed did in fact misread or misinterpret the questions and over-report nonmedical use (when use was only used medically) that this is problematic. As discussed above, we had no way to determine medical use within this analytic sample. As we continue to examine validity of survey
Full text analysis coming soon...