Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Mar 27.
Published in final edited form as: Drug Alcohol Depend. 2022 Apr 6;235:109442. doi: 10.1016/j.drugalcdep.2022.109442

Young adult opioid misuse indicates a general tendency toward substance use and is strongly predicted by general substance use risk

Danielle Pandika 1, Jennifer A Bailey 1, Sabrina Oesterle 2, Margaret R Kuklinski 1
PMCID: PMC10967660  NIHMSID: NIHMS1976973  PMID: 35461085

Abstract

OBJECTIVE:

To examine whether young adult opioid misuse reflects a general tendency toward substance use and is influenced by general substance use risk or whether it is a different phenomenon from other drug use.

METHODS:

At ages 23 (2016) and 26 (2019), a panel of young adults (n = 3,794 to 3,833) in the United States self-reported their past-month substance use (opioid misuse, heavy drinking, cigarettes, cannabis) and substance-specific risk factors (perceptions of harm; approval of use; and use of each substance by friends and romantic partners). Structural equation models examined non-opioid and opioid-specific associations between latent risk and substance use factors.

RESULTS:

Opioid misuse and opioid-specific risk factors shared significant variance with latent substance use and latent substance use risk, respectively, which were strongly associated. A statistically significant residual correlation between opioid-specific risk and opioid misuse remained.

CONCLUSION:

Young adult opioid misuse reflects a general tendency toward substance use and is strongly predicted by risk for substance use. Opioid-specific risk factors play only a small independent role. Existing evidence-based substance use interventions may be effective in preventing opioid misuse among young adults.

Keywords: young adult opioid misuse, prevention of opioid misuse, risk for opioid misuse, general substance use, opioid and substance

1. Introduction

The opioid epidemic, including use of heroin and non-medical use of prescription opioids (i.e., using without a prescription or for other purposes or in greater quantities than prescribed), continues to gravely impact communities in the United States. Over 90,000 individuals died by opioid overdose between 2019 and 2020, and synthetic opioid-related deaths increased by 1,040% between 2013 and 2019 (Ahmad et al., 2021; Mattson et al., 2021). Young adults (YAs) between the ages of 18 and 25 continue to be disproportionally affected by the opioid epidemic. Relative to older adults, YAs had greater odds of being diagnosed with opioid use disorder between 2015 and 2018 (Haider et al., 2020). About 20% of YA deaths in 2016 involved opioids (Gomes et al., 2018). Relative to older individuals, YAs were more likely to relapse during a medication trial for opioid use disorder (Fishman et al., 2020).

Although changes in prescribing practices have had some success in reducing YA opioid misuse by curtailing supply, these approaches do not target other risk factors that may be central to YA opioid misuse (Compton and Wargo, 2018). Primary prevention approaches that address a broad set of risk factors are needed and have potential to reduce YA opioid misuse at the population level. Elucidating whether these factors are opioid specific or reflect more general risk for substance use has important prevention implications. If risk factors for YA opioid misuse are specific, new programming addressing opioid-specific risk factors (e.g., perceptions of harm or favorable attitudes toward opioid misuse) or adaptations to existing substance use prevention programming may be needed. If, however, YA opioid misuse is most strongly predicted by non-substance specific risk for substance use (e.g., substance-using peers) and is strongly associated with substance use in general, existing evidence-based substance use prevention programming (e.g., Project Towards No Drug Abuse) should also be effective in reducing opioid misuse (Mihalic and Elliott, 2015). Expanded implementation of these programs would be supported as a means of reducing YA opioid misuse.

The association between opioid misuse and other substance misuse among young people is well supported by empirical studies in general populations. A large majority of adolescents and YAs who misuse opioids report using other substances (Catalano et al., 2011; McCabe et al., 2012; Monnat and Rigg, 2016), and co-ingestion of opioids with other drugs is highly prevalent (Cicero et al., 2020; McCabe et al., 2012; Monnat and Rigg, 2016). One study of YAs who had misused prescription opioids by age 21 found that 99% had used alcohol, about 90% had used tobacco, and 93% had used marijuana ever in their life. Over 95% had ever used an illicit drug (e.g. marijuana, cocaine, heroin) in addition to opioids (Catalano et al., 2011). Altogether, these studies suggest that YA opioid misuse likely reflects a tendency to use substances in general. While this general use can include the use of multiple types of substances within an acute use-session or specific period of time, or polysubstance use, for this study “general substance use” refers to a broader tendency to use any substance after initiating use—including polysubstance use or using different substances separately during different periods of time (Connor et al., 2014).

Less studied is whether risk factors for YA opioid misuse are unique or shared with other substances, although emerging research in adolescents and YAs supports both possibilities. Opioid-specific risk factors (e.g., parent use of prescription opioids for medical and nonmedical purposes) predict youth alcohol, tobacco and marijuana use, not just opioid misuse (Kerr et al., 2020). General family and peer risk factors (i.e. family conflict); risk factors specific to other substances (i.e. peer use of cannabis; positive drinking expectancies); and adolescent and YA alcohol, tobacco, and cannabis use are associated with YA opioid misuse (Griesler et al., 2019; Reboussin et al., 2020; Thrul et al., 2021; Tucker et al., 2020). However, some studies also have found unique associations between opioid-specific risk and nonmedical prescription opioid use. For example, Griesler et al (2019) found that parental nonmedical prescription opioid use was positively associated with adolescent nonmedical prescription opioid use independent of other parent and adolescent substance use, psychosocial risk factors, and social demographics. Likely, both opioid-specific risk factors and those linked to general substance use contribute to YA opioid misuse. However, few studies have tested both simultaneously and even fewer with YA risk factors. Identifying the degree to which general substance-use versus opioid-specific risk predicts opioid misuse in YAs would inform whether and in what way existing prevention programs need to be adapted or new programs developed to effectively reduce YA opioid misuse.

Within a community sample, the present study addressed this knowledge gap by using substance-specific risk factors relevant to YAs (e.g. friend and romantic partner use versus parent use) and measures that captured fairly common and likely harmful substance use in YAs (e.g. heavy drinking versus generally drinking alcohol). The study used SEM to test the hypotheses that 1) opioid misuse, along with heavy alcohol, tobacco, and cannabis use, is an indicator of general substance use and that 2) YA opioid-specific risk, along with alcohol-, tobacco, and cannabis-specific risk, contributes to general substance use risk in YAs. Additionally, the study tested the hypothesis that 3) opioid-specific risk would have a small but significant association with opioid misuse once the overlap with general substance use risk and use of other substances and their association was taken into account. Hypotheses were tested at ages 23 and 26 years to determine the stability of associations through young adulthood.

2. Methods

2.1. Participants and Procedures

Data were drawn from the Community Youth Development Study (CYDS), a community-randomized trial testing Communities That Care (CTC). CTC is a coalition-driven prevention planning and implementation system that supports the high-quality and sustainable adoption of evidence-based programs, policies, and practices across a community to promote positive youth development. CTC and the CYDS trial have been described in detail elsewhere (Brown et al., 2009; Hawkins et al., 2008; Oesterle et al., 2018). Briefly, twenty-four rural communities (average population = 14,646; range=1,578 – 40,787 residents per 2000 census data) in 7 states (Colorado, Illinois, Kansas Maine, Oregon, Utah, and Washington) participated in the trial. Communities were paired within state based on sociodemographic characteristics; one community from each pair was randomly assigned to implement CTC and the other served as a control community. All students in 5th grade (age 10) in 2003–2004 in participating communities were eligible to join a longitudinal panel; parental consent for participation was obtained for 4,420 students (76% of those eligible). The baseline sample of 4,407 students included youth with parental consent who completed a survey in either 2003 (grade 5) or 2004 (grade 6) and who remained in their original community for at least 1 semester. About 53% of panel members were female and 20% were Latinx. Participants could select multiple race categories: 71% were White, 4% were Black, 2% were Asian, 6% were Native American/Alaska Native, 0.8% were Native Hawaiian/Pacific Islander, and 21% were Other (of which 76% were Latinx). Participants were blind to study assignment and were followed even if they subsequently moved from their original study community. Data used in the current study were collected in 2016 (n = 3,833; mean age = 23.2, SD = 0.5) and 2019 (n = 3,765; mean age = 26.2, SD = 0.5); 88% and 87% of the still living sample were retained in these years, respectively. The majority of participants completed the survey online (88% online, 12% paper at Age 23; 92% online, 8% paper at Age 26) and received a $50 incentive. The study protocol was approved by the University of Washington Institutional Review Board.

2.2. Measures

2.2.1. Past month opioid misuse and other substance use.

Participants were asked on how many occasions in the past 30 days they used heroin, or used prescription opioids not prescribed to them by a doctor, smoked cigarettes, or used marijuana. Heavy alcohol use was assessed by asking on how many occasions respondents had 5 or more drinks in a row during the two weeks prior to the survey. Responses were dichotomized into any use (1) or no use (0). Any heroin use and any prescription opioid misuse were combined to create a past-month opioid misuse variable (1 yes to either, 0 no to both).

2.2.2. Opioid- and other substance use-specific risk factors.

Four risk factors were measured, including perceived partner use, perceived peer use, perceived harm, and personal approval of use, referencing each of the four substances - prescription opioids, heavy alcohol, cigarettes, and cannabis. Partner substance use was assessed by asking respondents whether, in the past year, their romantic partner or spouse used prescription opioids not prescribed for them by a doctor, consumed 5 or more drinks per occasion at least once or twice each week, smoked one or more packs of cigarettes per day, or regularly used marijuana (each scored 1 yes, 0 no). Peer substance use was assessed by asking respondents how many of their best friends used prescription opioids not prescribed for them by a doctor in the past year, consumed 5 or more drinks per occasion at least once or twice each week, smoked one or more packs of cigarettes per day, and used any marijuana (1 none to 4 all friends). Perceived harm measures assessed how much participants felt people risk harming themselves by using prescription opioids not prescribed for them by a doctor, having 5 or more drinks per occasion at least once or twice a week, smoking one or more packs of cigarettes per day, and using marijuana regularly (1 great risk to 4 no risk). Approval measures used the same wording to assess how wrong respondents felt it was for someone their age to participate in these substance use behaviors (1 very wrong to 4 not wrong at all). Responses were recoded as necessary so that higher scores indicated greater risk. Substance-specific mean scores were created by first standardizing the risk items (because response scales differed) and then taking their average. Mean scores were used as indicators of a latent substance use risk factor.

2.2.3. Covariates.

To adjust for potential confounds and variability in the socioeconomic and demographic composition and life course status of the sample in young adulthood, analyses included several covariates. Sex (male/female) was based on self-report at Grade 7. Ethnicity was coded as Latinx (1 yes, 0 no); race was coded using mutually exclusive dummy codes for Black and Other race (referent = White). Both ethnicity and race were self-reported in Grades 6, 7, or 8. Past-year income and marital or cohabitation status (1 married/cohabiting, 0 not married/not cohabiting) were self-reported at Age 23 and 26.

2.3. Statistical Analysis

Analyses used structural equation models (SEM) estimated in Mplus v8.4 to test study hypotheses separately at ages 23 and 26 to examine the developmental stability of the associations (Muthen and Muthen, (1998–2017)). Full information maximum likelihood estimation (FIML) was used to adjust for missing data. Analyses accounted for clustering of individuals in the 24 original study communities by using a sandwich variance estimator to adjust fit statistics and standard errors, using the TYPE = COMPLEX specification in Mplus (Stapleton et al., 2016a; Stapleton et al., 2016b). Because 24 clusters are just above the minimum recommended number of clusters needed to accurately estimate standard errors with this estimator, this approach could lead to slightly inflated Type-I error rates (Bell and McCaffrey, 2002; McNeish and Harring, 2016; Muthen and Muthen, (1998–2017)). To address this potential bias, we manually calculated p-values from a t-distribution with degrees of freedom equal to the number of clusters minus 1 rather than Z tests provided by Mplus. This has been shown to provide a simple but reasonably effective improvement for between 20 and 25 clusters (Cameron et al., 2008). To accommodate the estimation of latent factors with categorical indicators, we used the weighted least squares means and variances adjusted (WLSMV) estimator; theta parameterization enabled testing of residual correlations among categorical variables.

Although the CTC intervention was not hypothesized to have changed the relationships under investigation, we first tested whether measurement models and associations of interest were the same in the experimental and control groups before pooling data. Tests of factorial invariance in the latent variables across experimental and control groups showed significant evidence for scalar invariance, supporting pooling (results available from first author). We also estimated models separately for the experimental and control communities. Parameter estimates were very similar across conditions (results available from first author). Thus, data from experimental and control community participants were pooled for subsequent analyses and included a main effect of intervention group in addition to the other covariates.

The models’ goodness-of-fit was evaluated with the chi-square test of model fit, root mean square error of approximation (RMSEA), comparative fit index (CFI) and the Tucker-Lewis index (TLI). A statistically significant chi-square test, RMSEA values less than .06 and CFI and TLI values greater than .95 indicate good fit (Hu and Bentler, 1999). Conventionally, models with acceptable fit have all or most fit indices that meet these cut-off values.

In a single model, we created two latent variables: a general substance use risk factor with four indicators (opioid-specific, alcohol-specific, cigarette-specific, and cannabis-specific risk scores) and a general substance use factor with four indicators (opioid use, heavy alcohol use, cigarette use, and cannabis use). The latent substance use factor was regressed on the latent general substance use risk factor. We hypothesized that 1) opioid-specific risk scores would significantly load with other substance-specific risk scores onto a single latent factor of general substance use risk and 2) that opioid misuse would significantly load with other substance use onto a single latent factor of substance use. In the same model, we examined the association of opioid-specific risk with opioid misuse beyond risk for general substance use and use of other substances by estimating substance-specific residual correlations between each substance-specific risk indicator and its corresponding substance use indicator (e.g. residual of opioid-specific risk correlated with opioid misuse residual; see Figure 1). We hypothesized that 3) after accounting for latent general substance use risk and latent substance use, the residual association between opioid-specific risk and opioid misuse would be statistically significant.

Figure 1.

Figure 1.

Open in a new tab

Age 23 general risk for substance use and general substance use latent factor model (n = 3,760). Latent past month substance use regressed on latent general risk for substance use, with residual correlations between substance-specific risk scores and corresponding substance use indicators. Each latent factor indicator was regressed on covariates (race/ethnicity, sex, income, marital status, experimental/control group). Model fit indices: χ2(15) = 51.963, p <.001; RMSEA = .026, 90% CI .018–.033; CFI = .967, TLI = .815. Parameter estimates are standardized; all listed parameter estimates are statistically significant.

A secondary model examined which opioid-specific risk factors were associated with opioid misuse beyond use of other substances. Figure 2 shows the model specifying an opioid-specific latent risk factor using the four opioid-specific risk items as indicators, along with the general substance use latent factor described above. The latent substance use factor and residual of opioid misuse were regressed on the latent opioid risk factor. To identify specific risk factors that may uniquely contribute to opioid misuse, residual correlations of specific risk items (e.g. romantic partner opioid misuse, positive attitudes towards opioid misuse) with opioid misuse were tested one at a time to facilitate model identification.

Figure 2.

Figure 2.

Open in a new tab

Age 23 opioid-specific risk and general substance use latent factor model (n = 3,760). Latent past month substance use and past month opioid misuse regressed on latent opioid-specific risk, with residual correlations between each specific type of risk and opioid misuse, tested one at a time for model identification. Each latent factor indicator was regressed on covariates (race/ethnicity, sex, income, marital status, experimental/control group). Model fit indices: χ2(16) = 119.436, p < .001; RMSEA = .041, 90% CI .035–.049; CFI = .95, TLI = .74. Parameter estimates are standardized; all listed parameter estimates are statistically significant.

3. Results

Table 1 shows sample demographic data, prevalence of substance use, and mean levels of substance-specific risk factors at each age. Compared to heavy alcohol, cigarette, and cannabis use, past-month opioid misuse in this sample was relatively rare (3% at age 23 and 2% at age 26) and less prevalent compared to national levels of past year opioid misuse in young adults (Hudgins et al., 2019).

Table 1.

Sample demographics, substance use, and substance use risk at ages 23 and 26.

Age 23 Age 26
Demographics N (%) N (%)
Female 2009 (53%) 1990 (53%)
Male 1811 (47%) 1769 (47%)
Race/Ethnicity
Black/African American 144 (4%) 137 (4%)
White (referent) 2576 (68%) 2532 (68%)
Other 1082 (29%) 1069 (28%)
Asian/Pacific Islander 60 (2%) 61 (2%)
Native American 132 (3%) 120 (3%)
Multiracial 162 (4%) 161 (4%%)
Unspecifieda 725 (19%) 724 (19%)
Latinx 769 (20%) 766 (20%)
Income
Under $5,000 837 (22.1%) 513 (13.7%)
$5,000 – $9,999 539 (14.2%) 226 (6.1%)
$10,000 – $14,999 616 (16.3%) 307 (8.2%)
$15,000 – $19,999 442 (11.7%) 308 (8.2%)
$20,000 – $29,999 644 (17.0%) 725 (19.4%)
$30,000 – $39,999 373 (9.9%) 672 (18.0%)
$40,000 – $49,999 162 (4.3%) 389 (10.4%)
$50,000 – $59,999 83 (2.2%) 251 (6.7%)
$60,000 – $69,999 48 (1.3%) 134 (3.6%)
$70,000 – $79,999 17 (0.4%) 90 (2.4%)
$80,000 – $89,999 11 (0.3%) 45 (1.2%)
$90,000 – $99,999 5 (0.1%) 29 (0.8%)
$100,000 or more 8 (0.2%) 50 (1.3%)
Married or Cohabiting (vs. not) 1788 (47%) 2213 (59%)
Substance Use – Past Month
Opioid Misuse 108 (2.8%) 72 (1.9%)
Heavy Alcohol Use 1349 (35%) 1234 (33%)
Cigarette Use 939 (25%) 785 (21%)
Cannabis Use 933 (24%) 1000 (27%)
Open in a new tab

NOTE:

a

About 75% of individuals who did not specify a race endorsed Latinx ethnicity.

3.1. Primary Model: Latent General Substance Use Risk and Substance Use

Results from analyses of age 23 data (Figure 1) indicated that opioid misuse shared significant variance with the other substance use indicators and all loaded significantly on one latent substance use factor (factor loadings ranged between .49 and .76). Similarly, opioid-specific risk shared significant variance with the other substance-specific risk scores and all loaded significantly on one latent factor indicating risk for general substance use (loadings ranged between .59 and .76). As would be expected, latent risk for substance use was strongly and positively associated with the latent substance use factor (β = .74, p < .001). Model fit was acceptable, with chi-square, RMSEA and CFI indicating good fit but the TLI estimate less than .95 (see Figure 1). The model explained 47% of the variance in past-month opioid misuse and 36% of the variance in opioid-specific risk. Appendix A presents Mplus output for this model.

The correlations between residual variances of substance-specific risk indicators and their corresponding substance use indicators were all significant (Figure 1). Supplementary testing (results not shown) indicated that these residual correlations were substance specific. That is, residual variance in opioid-specific risk, for example, was correlated with residual variance in opioid misuse but not residual variance in heavy drinking, cigarette use, or cannabis use. Although there were residual correlations for all substance-specific indicators, the correlation was weakest for opioid misuse (e.g. ropioid = .23 versus ralcohol = .51). Age 26 data showed an identical pattern of significant associations and similar parameter estimates (see Appendix B).

3.2. Secondary Model: Latent Opioid-Specific Risk and Substance Use

The model had acceptable fit, with chi-square, RMSEA, and CFI indices indicating good fit but a TLI estimate less than .95 (see Figure 2), and explained a significant portion of the variance in past month opioid use (r2 = .57). All opioid risk indicators loaded significantly on a single latent opioid risk factor. Similar to the prior model, the regression of the latent general substance use factor on the opioid risk factor was significant, such that greater opioid risk was associated with greater general substance use (β = .668, p < .001). In addition, overall latent opioid risk was related to past month opioid use directly (β = .453, p < .001). Of the four residual correlations modeled between specific types of risk and past month opioid misuse, only partner and peer opioid use residuals were significantly associated with the opioid misuse residual (rpartner = .50, rpeer = .27; Figure 2). Appendix C provides Mplus output for this model. Age 26 data showed an identical pattern of significant associations and similar parameter estimates, with the exception that past month opioid use was not directly predicted by the latent opioid risk factor (Appendix D).

4. Discussion

Opioid misuse is often portrayed as a unique issue that requires new solutions. However, existing evidence from both adult treatment and general YA populations suggests that YAs who misuse opioids also tend to use other substances (Catalano et al., 2011; Cicero et al., 2020; Griesler et al., 2019). YA opioid misuse is thus indicative of substance use in general rather than a unique substance use behavior. This study supports the idea that existing youth and YA programs shown to prevent use and misuse of other substances could also be effective for the prevention of opioid misuse. It additionally suggests that risk pathways to YA opioid misuse are mostly shared with those for other substances, including alcohol, tobacco, and cannabis. Although previous studies have examined how general substance use-related and opioid-specific risk factors relate to youth opioid and other substance use, this study added new knowledge by testing the overlap between opioid-specific and other substance-related risk factors in a community-based YA sample (Griesler et al., 2019; Kerr et al., 2020). Although not the focus of the current analyses, this study did not find significant differences by intervention condition in the associations between risk with substance use, which is consistent with the intervention aim of reducing overall levels of risk and use rather than reducing the association between risk and use outcomes.

Results strongly suggest that much of the variance in YA opioid misuse was explained by a general tendency towards substance use; however a small, but statistically significant opioid-specific correlation between risk and misuse was not explained by these shared factors. Some of this unexplained unique association may be due to opioid use related to pain and other medical reasons that were not examined in the present study (Hosier et al., 2020; Hudgins et al., 2019; Papp et al., 2020). However, a closer look at the specific opioid-related risk factors measured in this study suggested that partner and peer misuse of opioids may pose a unique independent risk for YA opioid misuse. These findings align with a recent scoping review that found that youth who misused opioids most frequently cited a family member or friend (versus a doctor) as their source of prescription opioids (Bonar et al., 2020). Moreover, YAs with romantic partners or close friends who misuse opioids may not only have easier access to opioids, but may also experience more frequent reinforcement of the behavior through social approval and shared experiences with their close friends/partners (Hawkins and Weis, 1985). This finding suggests that opioid-specific social norms, created by the behaviors of YAs’ proximal social networks, should be targeted by preventive efforts alongside substance-use risk like perceived harm of and favorable attitudes towards heavy alcohol, tobacco, and cannabis use. The findings of the present study are critical because there are a number of tested-effective prevention programs– universal, selected, and indicated – that target substance use broadly and can be implemented across childhood, adolescence, and young adulthood to reduce substance use (Mihalic and Elliott, 2015). The substantial overlap between YA opioid misuse and general substance misuse and risk suggests that these prevention programs should also work to reduce opioid misuse.

This study has some limitations. First, there were a small number of people reporting opioid misuse in this community sample. Second, study participants were originally from 24 rural communities in 7 states and may not be representative of the U.S. population. However, a substantial proportion had moved to other parts of the country by young adulthood. The contexts within which opioid misuse was measured, therefore, represent a broader set of YA experiences and locales. Analyses adjusted for sociodemographic and economic factors, including sex, race/ethnicity, income, and marital status, as well as intervention/control group status, which strengthens conclusions. Third, the study sample was largely White (68%), with about 20% of the sample identifying as Hispanic/Latinx, which may limit the generalizability of study findings to other racial and ethnic groups. Fourth, the study focused on substance-specific risks and opioid misuse as they relate to commonly used substances. Although an important first step in elucidating whether opioid misuse is a unique substance use phenomenon, other substances (benzodiazepines, stimulants) have been linked to opioid misuse but were not included here and merit future study (Strickland et al., 2019; Votaw et al., 2019). Fifth, educational attainment and other risk factors (e.g. mental health) with strong associations to young adult substance use were not included in study analyses. Finally, analyses were cross-sectional in nature and cannot speak to the causality of examined relationships. However, findings were bolstered by replication at a second time point three years later. In order to strengthen causal conclusions and identify potential early indicators of risk for opioid misuse as targets for intervention, future research should consider the role of adolescent risk factors in longitudinally predicting YA opioid misuse. Although beyond the scope of the current study, future research would benefit from examining psychosocial risk factors with pain, physical health, and medical procedures as they relate to risk and YA opioid misuse, particularly given the central role of prescription opioids in treating pain and the ongoing opioid epidemic.

5. Conclusion

Results from this study suggest that young adult opioid misuse must be understood in the broader context of general substance use and risk for substance-using behavior. This finding argues against the widespread focus on opioid misuse as a unique phenomenon requiring new strategies for prevention. Changes in prescribing practices helped to fuel the opioid epidemic, and are critical to resolving it. However, reducing the supply of available opioids is only one part of the solution; reducing demand is also essential. The current study suggests that expanded implementation of existing, tested-effective substance use prevention programs may be an efficient and effective strategy for combatting the opioid crisis in YA populations.

Supplementary Material

1
2
3
4

HIGHLIGHTS.

  • Young adult opioid misuse likely reflects a general tendency to use substances

  • Risk for general substance use strongly predicts young adult opioid misuse

  • Existing, tested-effective substance use prevention programs may be an efficient and effective strategy for combatting the opioid crisis in young adult populations

Acknowledgements:

The authors gratefully acknowledge CYDS panel participants for their continued contribution to the longitudinal study. We also acknowledge the Social Development Research Group (SDRG) Survey Research Division for their hard work maintaining high panel retention.

Funding Information:

Research reported in this publication was supported by the National Institute on Drug Abuse of the National Institutes of Health under Award Numbers R01DA015183, R56DA044522, and R01DA044522. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The authors have no conflict of interest. ClinicalTrials.gov identifier: NCT01088542.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Conflict of Interest Statement

The authors have no conflict declared

REFERENCES

  1. Ahmad FB, Rossen LM, Sutton P, 2021. Provisional drug overdose death counts. . National Center for Health Statistics. [Google Scholar]
  2. Bell RM, McCaffrey DF, 2002. Bias reduction in standard errors for linear regression with multi-stage samples. Survey Methodology 28(2), 169–182. [Google Scholar]
  3. Bonar EE, Coughlin L, Roche JS, Philyaw-Kotov ML, Bixler EA, Sinelnikov S, Kolosh A, Cihak MJ, Cunningham RM, Walton MA, 2020. Prescription opioid misuse among adolescents and emerging adults in the United States: A scoping review. Prev Med 132, 105972. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Brown EC, Graham JW, Hawkins JD, Arthur MW, Baldwin MM, Oesterle S, Briney JS, Catalano RF, Abbott RD, 2009. Design and Analysis of the Community Youth Development Study Longitudinal Cohort Sample. Evaluation Review 33(4), 311–334. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Cameron AC, Gelbach JB, Miller DL, 2008. Bootstrap based improvements for inference with clustered errors. The Review of Economics and Statistics 90(3), 414–427. [Google Scholar]
  6. Catalano RF, White HR, Fleming CB, Haggerty KP, 2011. Is nonmedical prescription opiate use a unique form of illicit drug use? Addict Behav 36(1–2), 79–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Cicero TJ, Ellis MS, Kasper ZA, 2020. Polysubstance Use: A Broader Understanding of Substance Use During the Opioid Crisis. Am J Public Health 110(2), 244–250. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Compton WM, Wargo EM, 2018. Prescription Drug Monitoring Programs: Promising Practices in Need of Refinement. Ann Intern Med 168(11), 826–827. [DOI] [PubMed] [Google Scholar]
  9. Connor JP, Gullo MJ, White A, Kelly AB, 2014. Polysubstance use: diagnostic challenges, patterns of use and health. Curr Opin Psychiatry 27(4), 269–275. [DOI] [PubMed] [Google Scholar]
  10. Fishman M, Wenzel K, Scodes J, Pavlicova M, Lee JD, Rotrosen J, Nunes E, 2020. Young Adults Have Worse Outcomes Than Older Adults: Secondary Analysis of a Medication Trial for Opioid Use Disorder. J Adolesc Health 67(6), 778–785. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Gomes T, Tadrous M, Mamdani MM, Paterson JM, Juurlink DN, 2018. The Burden of Opioid-Related Mortality in the United States. JAMA Netw Open 1(2), e180217. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Griesler PC, Hu MC, Wall MM, Kandel DB, 2019. Nonmedical Prescription Opioid Use by Parents and Adolescents in the US. Pediatrics 143(3). [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Haider MR, Brown MJ, Gupta RD, Karim S, Olatosi B, Li X, 2020. Psycho-Social Correlates of Opioid Use Disorder among the US Adult Population: Evidence from the National Survey on Drug Use and Health, 2015–2018. Subst Use Misuse 55(12), 2002–2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Hawkins JD, Catalano RF, Arthur MW, Egan E, Brown EC, Abbott RD, Murray DM, 2008. Testing communities that care: the rationale, design and behavioral baseline equivalence of the community youth development study. Prev Sci 9(3), 178–190. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Hawkins JD, Weis JG, 1985. The social developmental model: An integrated approach to delinquency prevention. Journal of primary prevention 6(2), 73–97. [DOI] [PubMed] [Google Scholar]
  16. Hosier GW, McGregor T, Beiko D, Tasian GE, Booth C, Whitehead M, Siemens DR, 2020. Increased risk of new persistent opioid use in pediatric and young adult patients with kidney stones. Can Urol Assoc J 14(8), 237–244. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Hu L.t., Bentler PM, 1999. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal 6(1), 1–55. [Google Scholar]
  18. Hudgins JD, Porter JJ, Monuteaux MC, Bourgeois FT, 2019. Prescription opioid use and misuse among adolescents and young adults in the United States: A national survey study. PLoS Med 16(11), e1002922. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kerr DCR, Tiberio SS, Capaldi DM, Owen LD, 2020. Paternal and maternal prescription opioid use and misuse: General and specific risks for early adolescents’ substance use. Addict Behav 103, 106248. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Mattson CL, Tanz LJ, Quinn K, Kariisa M, Patel P, Davis NL, 2021. Trends and geographic patterns in drug and sythetic opioid overdose deaths -- United States, 2013–2019. Morbidity and Mortality Weekly Report 70(6). [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. McCabe SE, West BT, Teter CJ, Boyd CJ, 2012. Co-ingestion of prescription opioids and other drugs among high school seniors: results from a national study. Drug Alcohol Depend 126(1–2), 65–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. McNeish DM, Harring JR, 2016. Clustered data with small sample sizes: Comparing the performance of model-based and design-based approaches. Communications in Statistics - Simulation and Computation 46(2), 855–869. [Google Scholar]
  23. Mihalic SF, Elliott DS, 2015. Evidence-based programs registry: blueprints for Healthy Youth Development. Eval Program Plann 48, 124–131. [DOI] [PubMed] [Google Scholar]
  24. Monnat SM, Rigg KK, 2016. Examining Rural/Urban Differences in Prescription Opioid Misuse Among US Adolescents. J Rural Health 32(2), 204–218. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Muthen LK, Muthen BO, (1998–2017). Mplus User’s Guide. Muthen & Muthen, Los Angeles, CA. [Google Scholar]
  26. Oesterle S, Kuklinski MR, Hawkins JD, Skinner ML, Guttmannova K, Rhew IC, 2018. Long-Term Effects of the Communities That Care Trial on Substance Use, Antisocial Behavior, and Violence Through Age 21 Years. Am J Public Health 108(5), 659–665. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Papp LM, Kouros CD, Curtin JJ, 2020. Real-time associations between young adults’ momentary pain and prescription opioid misuse intentions in daily life. Am Psychol 75(6), 761–771. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Reboussin BA, Rabinowitz JA, Thrul J, Maher B, Green KM, Ialongo NS, 2020. Trajectories of cannabis use and risk for opioid misuse in a young adult urban cohort. Drug Alcohol Depend 215, 108182. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Stapleton LM, McNeish DM, Yang JS, 2016a. Multilevel and Single-Level Models for Measured and Latent Variables When Data Are Clustered. Educational Psychologist 51(3–4), 317–330. [Google Scholar]
  30. Stapleton LM, Yang JS, Hancock GR, 2016b. Construct Meaning in Multilevel Settings. Journal of Educational and Behavioral Statistics 41(5), 481–520. [Google Scholar]
  31. Strickland JC, Havens JR, Stoops WW, 2019. A nationally representative analysis of “twin epidemics”: Rising rates of methamphetamine use among persons who use opioids. Drug Alcohol Depend 204, 107592. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Thrul J, Reboussin BA, Rabinowitz JA, Maher B, Ialongo NS, 2021. Alcohol trajectories and subsequent risk for opioid misusein a cohort of urban adolescents. Substance abuse, 1–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Tucker JS, Davis JP, Seelam R, Stein BD, D’Amico EJ, 2020. Predictors of Opioid Misuse During Emerging Adulthood: An Examination of Adolescent Individual, Family and Peer Factors. Drug Alcohol Depend 214, 108188. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Votaw VR, Geyer R, Rieselbach MM, McHugh RK, 2019. The epidemiology of benzodiazepine misuse: A systematic review. Drug Alcohol Depend 200, 95–114. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1976973-supplement-1.docx (19.1KB, docx)
2
NIHMS1976973-supplement-2.docx (50.8KB, docx)
3
NIHMS1976973-supplement-3.docx (19.3KB, docx)
4
NIHMS1976973-supplement-4.docx (47.3KB, docx)

RESOURCES