Prescription Opioid Misuse and Risky Adolescent Behavior

Data on participants were drawn from the Youth Risk Behavior Surveillance Survey (YRBSS), a cross-sectional study conducted by the Centers for Disease Control and Prevention (CDC) in 2017 (n = 14 765). This is a nationally representative school-based survey that uses a 3-stage, cluster sample design of US students in grades 9 to 12 in public, Catholic, and other private schools. The CDC obtained samples from all 50 states and the District of Columbia. They first identified large-size counties or adjacent smaller counties from which to select schools geographically (primary sampling units). In the second stage, a list of public and private schools was selected by using the Market Data Retrieval database so as to obtain a wide cross-section of the different types of schools attended by American youth. In the third stage of sampling, the CDC randomly selected 1 or 2 entire classes in each chosen school and in each of the grades, 9 to 12, in which all students of the sampled class were eligible to participate.¹⁷ 

Our independent variable of interest was any misuse of prescription pain medication or prescription opioid misuse. YRBSS queried, “During your life, how many times have you taken a prescription pain medication differently than how a doctor told you to use it (count drugs such as codeine, Vicodin, Oxycontin, hydrocodone, and Percocet)?”; misuse was considered positive for responses indicating use 1 or more times and negative for responses of 0 times.

We identified 22 outcome variables and categorized them into broader domains of risky behavior, including (1) risky driving behavior (4 variables), (2) risky sexual behavior (4 variables), (3) risky substance use behavior (10 variables), (4) suicide attempt (1 variable), and (5) risky violent behavior (3 variables). See Supplemental Table 3 for definitions of these variables.

Potential confounders measured in the 2017 YRBSS that were adjusted for in the analyses included sex, age (in years), and race and/or ethnicity. Race and/or ethnicity categories were defined per the YRBSS manual.¹⁸ 

Data were analyzed by using SAS statistical software version 9.4 (SAS Institute, Inc, Cary, NC). Special sample survey procedures that are appropriate for analyzing complex survey data such as in the YRBSS with appropriate specifications on the STRATA, CLUSTER, and WEIGHT statements were used. The SAS procedure Proc SurveyFreq was used to generate weighted frequencies of each variable overall and by opioid misuse group (yes or no), and Proc SurveyLogistic was used to conduct multiple logistic regression analyses to determine the strength of the association between each outcome and opioid misuse group (yes or no), adjusting for demographic characteristics. To provide adjustment for the number of outcomes (n = 22) studied, we divided the usual significance threshold (0.05) by 22 and considered P <.002 to be statistically significant in all analyses.

Study participant characteristics are provided in Table 1. The study cohort included 14 765 participants; 49.3% were of male sex, and 53.5% were non-Hispanic whites. Overall, 14% of the youth reported ever misusing prescription opioids. There was no statistically significant difference in those who reported opioid misuse compared with those who did not by sex (P = .2) or race and/or ethnicity (P = .005), but significant differences were seen by age (P < .0001). There was an overrepresentation of 17- and 18-year-olds in the opioid misuser category.

In this cross-sectional sample, students who reported ever misusing prescription opioids, compared with those who never misused, were 2.8 times more likely to report never or rarely using a seatbelt (confidence interval [CI] 2.28–3.49; P < .0001), 2.8 times more likely to have ridden with an intoxicated driver (CI 2.40–3.36; P < .0001), 5.8 times more likely to have driven under the influence (CI 4.30–7.70; P < .0001), and 2.3 times more likely to have texted or e-mailed while driving (CI 2.00–2.65; P < .0001; Table 2).

Students who reported ever misusing prescription opioids were 3.9 times more likely to have reported first sexual intercourse before age 13 (CI 2.88–5.40; P < .0001), 4.8 times more likely to have reported sex with 4 or more partners (CI 3.89–5.98; P < .0001), 3.6 times more likely to have used substances before last sexual intercourse (CI 2.87–4.59; P < .0001), and 2.0 times more likely to have not used a condom before last sexual intercourse (CI 1.69–2.35; P < .0001; Table 2).

There were 10 variables representing substance use (Supplemental Table 3). Students who reported ever misusing prescription opioids were significantly more likely to have ever tried other substances (adjusted odds ratios ranged from 5.4 to 22.3; P < .0001 for all associations; Table 2).

Students who reported ever misusing prescription opioids were 4.9 times more likely to have ever attempted suicide (CI 3.89–6.11; P < .0001; Table 2).

Students who reported ever misusing prescription opioids were 3.4 times more likely to have carried a weapon in the past 30 days (CI 2.92–3.91; P < .0001), 5.1 times more likely to have carried a gun in the past 30 days (CI 4.17–6.31; P < .0001), and 4.0 times more likely to have engaged in physical fights in the past year (CI 3.56–4.56; P < .0001; Table 2).

Our aim in this study was to undertake a comprehensive evaluation of the relation between prescription opioid misuse and other risky behaviors in youth as assessed in the 2017 YRBSS. Prescription opioid pain medication misuse was significantly associated with all risky behaviors measured in this study even after adjusting for covariates. These included risky driving behaviors, violent behaviors, risky sexual behaviors, other substance use, and suicide attempts.

Previous studies have examined 1 or 2 risky behaviors or examined only past-year prescription medication misuse.^9–11,15,19  Our study, by comparison, more comprehensively assessed risky behaviors in several categories. Additionally, many of the existing studies were smaller, focusing on 1 school or 1 geographic region.^{9,10,15,19–21}  Given that our study is of a nationally representative sample, we were able to provide a more generalizable assessment of the risky behaviors associated with prescription opioid misuse. Furthermore, much of the existing literature examined prescription drug misuse (including medications such as stimulants, benzodiazepines, steroids, and pain relievers) and did not address opioid use specifically.^10,19–21  By contrast, we were able to more specifically describe the relationship between opioid misuse and risky behavior, thus more precisely characterizing this high-risk population and providing evidence for early targeted interventions.

There are several possible explanations for the causes of the association observed in this study. There is evidence in the literature that those who misuse prescription opioids then engage in other high-risk behaviors. Studies have examined the progression of prescription opioid use to use of other substances, specifically heroin.²²  Alternatively, there is evidence that high-risk–takers are also more likely to engage in other risky behaviors, and thus, prescription opioid misuse may be part of the high-risk–taker profile.²³  However, because of the cross-sectional design of this study, we could not determine the direction of the association.

Our findings have several implications for clinicians prescribing opioid pain medications or providing care for those taking prescription opioid pain medications. Although recognizing that our results cannot determine causality, it seems prudent for physicians to not only discuss with patients and their parents the direct dangers of misuse of prescribed opioid pain medications but also review the observed associations between opioid misuse and other risky behaviors. When physicians and other providers identify youth misusing prescription opioids, they should have a heightened awareness of associated risky behaviors and potential opportunities for education and prevention (eg, advising about riding in a car with an intoxicated driver). The opposite directionality of associations should also be considered. For example, school sexual health educators should consider the greater likelihood of misuse of prescription opioid medications among youth engaging in high-risk sexual behaviors. Additionally, psychiatrists evaluating patients for self-harm risk should take into consideration misuse of prescription opioid medications as a risk factor for suicide attempt.

There were a few limitations in the current study that should be considered when interpreting the results. First, the study was cross-sectional, limiting the ability to determine causality of outcomes. However, this study provides a snapshot of the risky behaviors associated with prescription opioid misuse. Furthermore, additional covariates such as socioeconomic status, mental health disorders, and performance in school were not asked about in the YRBSS but may have implications on the results. Additionally, the motivations to misuse prescription opioid medications and source of prescription and/or access to the opioid mediations were not assessed in this survey, and as such, no conclusions can be made directly via this study. Finally, the current data rely on self-reporting, which could be subject to recall and/or unacceptability bias.

These limitations notwithstanding, we report that adolescents who misuse prescription opioids are more likely to have engaged in a broad range of other risky behaviors, all of which contribute to significant adverse health outcomes. Future efforts should be focused on better understanding the directionality of this association. Additionally, preventing access to prescription opioids will help minimize overall risk in this group. With the ongoing opioid epidemic, pediatricians and child psychiatrists are likely to be more attuned to opioid misuse in their patients. If youth are screening positive for opioid misuse, pediatricians, nurses, social workers, child psychiatrists, and other providers assessing adolescents may have a new, broad range of other risky behaviors for which to screen regardless of the direction of the association.

CDC

Centers for Disease Control and Prevention

CI

confidence interval

YRBSS

Youth Risk Behavior Surveillance Survey