ABSTRACT
Background
Medication sources for prescription opioid misuse in US adolescents and young adults may have changed, particularly related to the COVID‐19 pandemic. We examined how opioid sources vary across age groups and cohorts and whether links between sources and substance use or mental health vary between age groups.
Methods
Cross sectional nationally representative survey data from the 2015–2019 and 2021–2022 National Survey on Drug Use and Health were used. Respondents were interviewed in their homes (2015–19) or via mixed household and web‐based survey methods (2021–2022). Participants were 6045 14–25‐year‐old reporting past‐year prescription opioid misuse. Mutually exclusive opioid sources were physicians, purchases, theft, and friends/family for free. Cross‐tabulations estimated source prevalence by age, and logistic and negative binomial regression models evaluated links between sources (independent variable) and substance use, mental health, and sociodemographics (individual dependent variables).
Results
Theft declined significantly with age in the 2015–2019 cohort (12.4% to 4.7%), with some evidence of opioid source differences between 2015–2019 and 2021–2022. Those using purchases had the most past‐month prescription opioid misuse days and the highest prevalence of substance use disorder from misuse (e.g., 23.4% of 2015–2019 young adults), other substance use, and mental health concerns; those using theft typically had the second highest rates, with the lowest rates often in the physician source group.
Conclusion
We found age‐ and cohort‐based differences in sources, and mechanisms underlying these differences warrant further study. Adolescents and young adults using purchases or theft to obtain opioids are important targets for identification and intervention.
Keywords: adolescents, opioid, prescription drug misuse, source, young adults
1. Introduction
Sources of medications used to treat pain outside of a clinician's supervision, like prescription opioids, can be an important target for prevention and intervention [1, 2]. Adolescents and young adults (AYAs) are key developmental groups for prevention of prescription opioid misuse (POM) initiation and entrenchment, as much of POM initiation occurs by age 25, and initiation is more common in younger generations [3, 4, 5]. Once established, POM is linked to higher rates of other substance use, substance use disorders (SUD), and psychopathology in AYAs [6]. While most individuals with POM do not progress illicit opioid use, POM is a robust risk factor for heroin [7, 8, 9] and illicit fentanyl use [10, 11].
POM is often defined as opioid medication use without a prescription and/or in ways not intended by the prescriber, like at higher doses, more often, or nonorally [12, 13]. Sources for POM include physician sources (i.e., misuse of one's own prescribed medication), purchases from peers, family, or dealers, theft from individuals or medical sources (e.g., pharmacies), and obtaining opioid medication from friends or family for free [2]. In adults, use of physician sources or obtaining opioid medication from friends or family for free is linked to a lower level of concurrent other substance use, less frequent POM or SUD from POM, and fewer mental health concerns than opioid theft or purchases [14, 15, 16, 17]. Research that examines only adolescents [18, 19] or young adults [20, 21] suggests similar relationships.
Purser [22] analyzed data from the 2015–2019 nationally representative National Survey on Drug Use and Health (NSDUH) and found that the odds of receiving a prescription opioid from a friend or relative for free declined over the study period across the US population. Similarly, McCabe and colleagues [23] found that both purchases from friends and friends providing stimulant, opioid, and/or tranquilizer/sedative medication for free declined from 2009/10 to 2021/22 in US high school seniors (modal age 18 years). These papers suggest changes in sources for prescription misuse, but neither specifically examined both AYAs and POM sources at the same time. Given evidence of disruptions in illicit opioid supply networks related to the COVID‐19 pandemic [24], sources for POM may also have been disrupted. Examinations of recent changes in POM sources are needed, and evidence on concurrent substance use and mental health profiles by source used needs updating, as most prior research used data roughly a decade old [18, 19, 21, 25, 26].
To address these knowledge gaps, we first aimed to quantify the prevalence of POM sources by age (i.e., 14/15, 16/17, 18–20, 21–23, 24/25 years) with the 2015–2019 and 2021–2022 NSDUH cohorts. Second, we linked POM sources (independent variable) to prevalence of prescription opioid use disorder and frequency of past‐year prescription opioid misuse (dependent variables). Finally, we examined associations between other substance use, mental health, educational status, housing, and sexual identity (dependent variables) and POM sources (independent variable). All regression analyses occurred separately in adolescents and young adults and separately by cohort.
2. Methods
The NSDUH annually surveys US residents aged 12 and older, oversampling AYAs. The independent multistage area probability sample design allows for the calculation of nationally representative estimates, using survey weights that incorporate US Census population estimates. To maximize complete, valid, and accurate responding, the NSDUH combines skip‐outs, consistency checks, and audio computer‐assisted self‐interviewing (ACASI) methods. For 2015–2019, all data were collected face‐to‐face in the resident's household; beginning with 2020, the NSDUH combined in‐person interviews with web‐based data collection. We exclude the 2020 data here, per NSDUH recommendations, aggregating the 2015–2019 and the 2021–2022 data, as NSDUH staff also recommend against combining the 2015–2019 and 2021–2022 data because of COVID‐related changes in methodology (e.g., shift from exclusively in‐person data collection to mixed in‐person and web‐based data collection) [27]. All NSDUH procedures were approved by the Research Triangle International IRB [28], and this work was exempted by the first author's IRB. Readers are referred to other resources [29, 30, 31, 32, 33] for further methodological details of the NSDUH.
2.1. Participants
In the 2015–2019 cohort, 84.8% of adolescents (unweighted n = 1625) and 94.7% of young adults (unweighted n = 4240) with past‐year POM reported their most recent source. For the 2021–2022 cohort, 252 adolescents (93.4% weighted) and 773 young adults (95.0% weighted) noted their most recent POM source. For the adolescent 2015–2019 analytic sample, 53.9% were female, 52.7% were non‐Hispanic white, and 39.7% had household incomes of $75,000 or greater. In 2015–2019 young adults, 46.9% were female, 62.4% were non‐Hispanic white, and 26.6% had household incomes of $75,000 or greater. Among the 2021–2022 adolescent cohort, 61.7% were female, 46.5% were non‐Hispanic white, and 42.1% had household incomes of $75,000 or greater; 2021–2022 young adults were 49.2% female, 51.5% non‐Hispanic white, and 30.8% had household incomes of $75,000 or greater. Sex and race/ethnicity by age group and by cohort are captured in Table S1, stratified by source category.
2.2. Measures: POM Sources
In the NSDUH, POM is defined as opioid medication use “in any way a doctor did not direct”, including use without a prescription and use with a prescription but in larger doses, more frequently, for a longer period, or in any other nonprescribed way (e.g., nonorally). To promote accurate recall, the POM assessment includes trade (e.g., Oxy‐Contin, Demerol) and generic names (e.g., oxycodone, hydrocodone), with medication pictures.
Those with past‐year POM are asked about the source of medication for their most recent episode. Here, responses are recoded as physician source (“from just one doctor” or “from more than one doctor”), friend/relative for free (“from a friend or relative for free”), purchased (“bought the opioid from a friend or relative” or “bought the opioid from a drug dealer or other stranger”), and theft (“I stole the opioid from a doctor's office, clinic, hospital, or pharmacy”, or “I took the opioid from a friend or relative without asking”). Respondents could select “I got the opioid in some other way”, but we excluded this response because of low rates of endorsement and because of a lack of information about the free‐text response accompanying this “other” response in the NSDUH public‐use data files.
2.3. Measures: Prescription Opioid Use Disorder and POM Frequency
Measures of prescription opioid use disorder relied on the DSM‐IV criteria for the 2015–2019 NSDUH data, with a diagnosis of substance abuse or dependence from POM specifically captured. For the 2021–2022 NSDUH, the DSM‐5 criteria were used. Participants engaged in past‐month POM are also asked about the frequency of POM, with a range of zero to 30 days; those with past‐year POM and source data but not past‐month POM were logically imputed to have zero days of past‐month POM.
2.4. Measures: Other Substance Use
Substance use measures were dichotomous past‐month binge alcohol use, past‐month cannabis use, past‐year prescription stimulant misuse, past‐year prescription benzodiazepine misuse, any past‐year SUD, and past‐year nicotine dependence. The National Institute of Alcohol Abuse and Alcoholism's definition of binge drinking was used (i.e., four or five alcoholic drinks at one occasion for females and males, respectively) [34]. Any SUD relied on the DSM‐IV criteria for the 2015–2019 NSDUH, with any diagnosis of substance abuse or dependence counting as positive for SUD. For the 2021–2022 NSDUH, the DSM‐5 criteria were used, and SUD from use or misuse of alcohol, cannabis, cocaine, heroin, hallucinogens, inhalants, methamphetamine, prescription opioids, tranquilizer/sedatives, or stimulants was used to generate a dichotomous any SUD diagnosis. Finally, nicotine dependence came from the Nicotine Dependence Syndrome Scale (NDSS). Specific criteria and scoring for the SUD diagnoses and the NDSS are captured in the NSDUH codebooks [27, 35].
2.5. Measures: Mental Health and Sociodemographics
DSM‐IV (2015–2016) or DSM‐5 (2017–2019; 2021–2022) past‐year major depressive episode was assessed. In young adults only, serious psychological distress from the six‐item Kessler Psychological Distress Scale [36] (young adults only) and a single‐item assessment of suicidal ideation (“At any time in the past 12 months … did you seriously think about trying to kill yourself?”; young adults only) were captured. Two sociodemographic measures were used: housing instability (i.e., at least two past‐year moves) and educational status (i.e., in school versus not for adolescents and in college/college graduate versus high school graduate or less for young adults). In young adults, membership in a sexually diverse group (i.e., gay, lesbian, or bisexual identity) was also captured. Age was captured as 14/15, 16/17, 18–20, 21–23, and 24/25 years, and race/ethnicity (non‐Hispanic White, non‐Hispanic Black, non‐Hispanic Asian, non‐Hispanic American Indian/Native American, non‐Hispanic Hawaiian/Pacific Islander, non‐Hispanic multiracial, Hispanic), sex, household income, and population density in area of residence were also captured. Twelve‐ and thirteen‐year‐old respondents were excluded due to low counts of respondents engaged in POM.
2.6. Data Analyses
Stata 18.0 was used for analyses. To account for the complex sampling design features of the NSDUH in variance and point estimation, the svy commands were used in Stata, explicitly accounting for NSDUH sampling (i.e., stratification and cluster sampling) and weighting variables. Analyses focused on the subpopulation defined by those with past‐year POM who provided source data and were conducted separately by age group and cohort.
Initial analyses used cross‐tabulations to examine associations of source category to sex, race/ethnicity, and age with POM source category, which included prevalence estimates and 95% confidence intervals (95% CIs) of the prevalence estimate; associations were evaluated using design‐adjusted Rao‐Scott tests [37]. Second, we estimated the prevalence of prescription opioid use disorder and past‐month days of POM (dependent variables) by source category (independent variable). Differences in use disorder and days of POM were evaluated using design‐based logistic and negative binomial regression, respectively.
Further analyses employed design‐based logistic regression models to investigate the relationships between POM source category (independent variable) and the substance use and mental health characteristics; housing instability, educational status, and sexual identity (young adults only) were also dependent variables in regression analyses. For all regression analyses, the physician source group was set as the reference, given evidence of lower rates of other substance use among adolescents [25] and young adults [26] in this POM source group versus other source groups. Odds ratios (ORs) and 95% CIs of the OR were calculated for logistic regressions, and all regression analyses adjusted for sex, age, race/ethnicity, household income, and population density. Both adolescents and young adults were included in regression analyses involving the 2015–2019 NSDUH data, but only young adults were included in regression models using the 2021–2022 data because of the small sample of adolescents engaged in POM in this cohort.
3. Results
3.1. Sex, Racial/Ethnic, Age, and Cohort Patterns of POM Sources
An estimated 3.1% of adolescents, 12–17 years, and an estimated 6.7% of young adults (both weighted), 18–25 years, endorsed past‐year POM in the 2015–2019 data; for 2021–2022, an estimated 1.8% of adolescents and 3.0% of young adults engaged in past‐year POM. Per Table S1, there were significant sex differences in patterns of AYA POM sources in the 2015–2019 cohort. The adolescent physician source (60.9% vs. 39.1%) and friends/family for free groups (55.4% vs. 44.6%) had a higher prevalence of females than males, who were more common in the purchases group (58.2% vs. 41.8%). In young adults, there were more women in the physician sources group (53.9% vs. 46.1%) and more men in the purchases group (66.3% vs. 33.7%). For race/ethnicity, patterns were more complex, with few clear pairwise differences.
Per Table 1, opioid sources changed significantly from age 14/15 to age 24/25 (design‐based Rao‐Scott test: F(8.62, 430.80) = 3.37, p = 0.0006) in the 2015–2019 cohort, while the omnibus chi‐square was nonsignificant for the 2021–2022 cohort (p = 0.50). Theft became less common with aging in both cohorts, though only the 2015–2019 differences were significant (e.g., 14/15 year olds = 12.4%, 95% CI = 9.0–16.7; 24/25 year olds = 4.7, 95% CI = 3.4–6.6). Purchases became more common with aging over the AYA span, though these within‐source differences were nonsignificant. Physician source use was relatively stable by age in the 2015–2019 cohort, though it nonsignificantly decreased from 14/15 years (41.1%) to 24/25 years (27.2%) in 2021–2022. Finally, differences by age group in obtaining medication for POM from family or friends for free were not consistent in either cohort. Use of friends or family to obtain opioids for free was the most common source across ages in the 2015–2019 cohort and was most common among young adults in the 2021–2022 cohort, with physician sources more common in adolescents (Table 1).
TABLE 1.
Prescription opioid misuse sources by age group among those with past‐year misuse.
| 2015–2019 | 2021–2022 | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Sample size | Physician source a | Friend/relative (free) a | Purchased a | Theft a | Sample size | Physician source a | Friend/relative (free) a | Purchased a | Theft a | |
| n | % (95% CI) | % (95% CI) | % (95% CI) | % (95% CI) | n | % (95% CI) | % (95% CI) | % (95% CI) | % (95% CI) | |
| 14/15 years | 530 |
33.8 (27.9–40.4) |
34.5 (29.8–39.5) |
19.3 (15.4–24.0) |
12.4 (9.0–16.7) |
100 | 41.1 (25.4–58.8) | 32.1 (19.5–47.9) | 12.8 (5.6–26.5) | 14.1 (6.8–26.8) |
| 16/17 years | 950 |
26.4 (23.1–30.0) |
44.1 (39.7–48.7) |
19.4 (16.7–22.4) |
10.1 (7.8–13.0) | 125 | 36.4 (24.9–49.7) | 29.0 (18.0–43.2) | 23.5 (14.8–35.3) | 11.1 (5.7–20.5) |
| 18–20 years | 1390 |
28.8 (25.5–32.2) |
40.6 (37.5–43.8) |
21.6 (18.7–25.0) |
9.0 (7.0–11.4) |
221 | 32.3 (23.7–42.3) | 35.9 (26.9–45.9) | 18.8 (11.9–28.5) | 13.1 (7.2–22.7) |
| 21–23 years | 1671 |
25.6 (23.1–28.4) |
46.6 (43.4–49.7) |
21.4 (18.7–24.7) |
6.4 (4.8–8.6) |
319 | 34.2 (26.4–43.0) | 35.9 (27.6–45.2) | 24.4 (17.0–33.7) | 5.5 (2.8–10.6) |
| 24/25 years | 1179 |
29.4 (26.0–33.1) |
42.6 (38.4–46.5) |
23.2 (19.9–27.0) |
4.7 (3.4–6.6) | 233 | 27.2 (19.4–36.7) | 37.8 (27.7–49.1) | 28.5 (18.1–41.7) | 6.6 (2.9–14.4) |
Note: 2015–2019 design‐based chi‐square: F(8.62, 430.80) = 3.37; p = 0.0006; 2021–2022 design‐based chi‐square: F(9.57, 478.74) = 2.74; p = 0.50.
95% CI = 95% confidence interval of the prevalence estimate.
Source Categories are: Obtained opioids from one or more doctors (Physician Source); Obtained from a friend or relative for free (Friend/Relative [free]); Purchased from friend, relative, or stranger/dealer (Purchased); and Took from friend/relative without asking or stole from a medical source (e.g., physician's office, pharmacy) (Theft).
3.2. Prevalence of Prescription Opioid Use Disorder and Past‐Month Days of POM by Source Category in AYAs
Per Table 2, the highest prevalence of prescription opioid use disorder in the 2015–2019 cohort was in AYAs who purchased opioids for misuse (24.4% in adolescents, 23.4% in young adults). This prevalence was significantly higher than that of the physician source reference group (11.3%, OR = 2.36, 95% CI = 1.77–3.16) in young adults, but not significantly higher in adolescents (14.4%, OR = 1.91, 95% CI = 0.99–3.68). Young adults engaged in theft also had a higher prevalence rate of prescription opioid use disorder (16.6%, OR = 1.59, 95% CI = 1.06–2.38) than those using physician sources. Table S2 captures these analyses in the 2021–2022 young adult cohort. As with the 2015–2019 cohort, those using purchases had a significantly higher prevalence of past‐year prescription opioid use disorder (49.0%) than those in the physician source reference group (24.7%, OR = 3.07, 95% CI = 1.56–6.06), with no other significant differences.
TABLE 2.
Prevalence of opioid use disorder by source category in adolescents and young adults, via 2015–2019 NSDUH data.
| Opioid sources a | Adolescents | Young adults | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Prevalence | 95% CI | Odds ratio | 95% CI | p | Prevalence | 95% CI | Odds ratio | 95% CI | p | |
| Physician sources | 14.4% | 10.1–20.1 | Reference group | 11.3% | 9.1–14.0 | Reference group | ||||
| Friend/relative (free) | 11.0% | 8.5–14.1 | 0.67 | 0.40–1.13 | 0.13 | 7.1% | 6.0–8.5 | 0.60 | 0.44–0.80 | 0.001 |
| Purchased | 24.4% | 18.4–31.6 | 1.91 | 0.99–3.68 | 0.053 | 23.4% | 20.0–27.1 | 2.36 | 1.77–3.16 | < 0.001 |
| Theft | 21.2% | 15.2–28.9 | 1.60 | 0.81–3.16 | 0.17 | 16.6% | 12.3–22.1 | 1.59 | 1.06–2.38 | 0.024 |
Note: All regression estimates are adjusted for sex, age, race/ethnicity, income, and population density.
Source categories from most recent episode of misuse are: Obtained opioids from one or more doctors (Physician source); Obtained from a friend or relative for free (Friend/Relative [free]); Purchased from friend, relative, or stranger/dealer (Purchased); and Took from friend/relative without asking or stole from a medical source (e.g., physician's office, pharmacy) (Theft).
Figure 1 captures mean days of past‐month POM by source category in AYAs in the 2015–2019 cohort. Adolescents using theft or purchases each had a greater mean number of days of POM (1.48 and 1.58, respectively) than those using physician sources (0.83, p < 0.05). In young adults, only those who purchased medication for POM had a greater mean number of POM days (2.83) than those using physician sources (0.99, p < 0.001). Figure S1 captures these results in the 2021–2022 cohort. Again, young adults using purchases had a greater mean number of days of POM (2.83) than the physician source group (0.99, p < 0.05); no other significant differences were found.
FIGURE 1.
Mean days of past‐month prescription opioid misuse (POM) by source category in adolescents and young adults, via 2015–2019 NSDUH data. Error bars represent 95% confidence intervals for the mean.
3.3. Other Substance Use, Mental Health, and Sociodemographic Characteristics by POM Source in Adolescents
For most substance use characteristics (Table 3), the highest adjusted odds of use versus the physician source reference group were in those who purchased opioids (e.g., past‐month cannabis use OR of 5.49 in those who purchased opioids). For past‐year SUD and past‐year nicotine dependence, the only group that differed from the reference was adolescents engaged in opioid purchases (SUD OR = 2.73, 95% CI = 1.68–4.42; nicotine dependence OR = 5.17, 95% CI = 1.37–19.55). No significant differences were found by source group in terms of past‐year major depression or housing instability, and only those engaged in theft were more likely to be in school than those using physician sources (OR = 2.24, 95% CI = 1.11–4.51).
TABLE 3.
Logistic regression models of substance use, mental health, and sociodemographic characteristics by prescription opioid source in adolescents with past‐year misuse, via 2015–2019 NSDUH data.
| Opioid sources a | 30‐day binge alcohol use | 30‐day cannabis use | 12‐month Rx stimulant misuse | 12‐month Rx BZD misuse | Any 12‐month SUD b | 12‐month nicotine dependence c | 12‐month MDE | Housing instability d | Currently in school |
|---|---|---|---|---|---|---|---|---|---|
| aOR (95% CI) | aOR (95% CI) | aOR (95% CI) | aOR (95% CI) | aOR (95% CI) | aOR (95% CI) | aOR (95% CI) | aOR (95% CI) | aOR (95% CI) | |
|
Physician source (n = 411) |
Reference Group | Reference Group | Reference Group | Reference Group | Reference Group | Reference Group | Reference Group | Reference Group | Reference Group |
| Friend/Relative (free) | 2.14*** | 2.83*** | 2.69*** | 4.32** | 1.38 | 2.65 | 1.13 | 0.79 | 1.32 |
| (n = 619) | (1.58–2.89) | (1.94–4.13) | 1.70–4.24) | (1.61–11.64) | (0.95–2.00) | (0.78–8.99) | (0.79–1.63) | (0.49–1.29) | (0.86–2.01) |
| Purchased | 2.99*** | 5.49*** | 5.00*** | 16.24*** | 2.73*** | 5.17* | 1.56 | 0.77 | 1.41 |
| (n = 289) | (1.88–4.75) | (3.38–8.92) | (2.85–8.75) | (6.70–39.37) | (1.68–4.42) | (1.37–19.55) | (0.99–2.46) | (0.43–1.39) | (0.80–2.46) |
| Theft | 2.28** | 2.06** | 3.93*** | 5.58** | 1.62 | 3.50 | 1.12 | 1.00 | 2.24* |
| (n = 161) | (1.38–3.77) | (1.23–3.47) | (2.07–7.46) | (1.95–15.98) | (0.96–2.72) | (0.78–15.62) | (0.65–1.93) | (0.51–2.00) | (1.11–4.51) |
| Archer–Lemeshow goodness of fit test |
F (9, 42) = 1.33 p = 0.25 |
F (9, 42) = 0.83 p = 0.59 |
F (9, 42) = 1.80 p = 0.10 |
F (9, 42) = 0.46 p = 0.89 |
F (9, 42) = 0.25 p = 0.98 |
F (9, 42) = 346.72 p < 0.0001 |
F (9, 42) = 0.52 p = 0.85 |
F (9, 42) = 1.76 p = 0.11 |
F (9, 42) = 0.32 p = 0.96 |
Note: All regression estimates are adjusted for sex, age, race/ethnicity, income, and population density.
Abbreviations: 95% CI = 95% confidence interval, aOR = adjusted odds ratio, BZDs = benzodiazepines, MDE = major depressive episode, Rx = Prescription, SUD = substance use disorder.
* denotes p ≤ 0.05; ** denotes p ≤ 0.01; *** denotes p ≤ 0.001.
Source categories from most recent episode of misuse are: Obtained opioids from one or more doctors (Physician source); Obtained from a friend or relative for free (Friend/Relative [free]); Purchased from friend, relative, or stranger/dealer (Purchased); and Took from friend/relative without asking or stole from a medical source (e.g., physician's office, pharmacy) (Theft).
SUD captures any past‐year DSM‐IV Substance Abuse or Dependence from use or misuse or one or more of: alcohol, cannabis, heroin, cocaine, inhalants, hallucinogens, methamphetamine, prescription opioids, prescription stimulants, prescription tranquilizers, or prescription sedatives.
Nicotine Dependence is measured using the Nicotine Dependence Syndrome Scale.
Housing instability is two or more past‐year moves between residences.
3.4. Other Substance Use, Mental Health, and Sociodemographic Characteristics by POM Source in Young Adults
Patterns in the analyses among young adults largely followed those in adolescents, with higher odds of substance use among those who used a friend or relative to obtain opioids for free, purchased them, or used theft to obtain them for POM, versus the reference group using physician sources (see Table 4). Those who purchased opioids generally had higher odds of substance use relative to physician sources (e.g., past‐year prescription benzodiazepine misuse OR = 3.71, 95% CI = 2.31–5.94). With that said, those engaged in theft only had significantly higher odds versus the physician source reference group for three substance use characteristics: past‐month cannabis use, past‐year SUD, and past‐year nicotine dependence. Notably, those who purchased or stole opioids had 2.47 times greater odds of past‐year SUD than those who used physician sources (purchased 95% CI = 1.91–3.19; theft 95% CI = 1.80–3.39).
TABLE 4.
Logistic regression models of substance use characteristics by prescription opioid source in young adults with past‐year misuse, via 2015–2019 NSDUH data.
| Opioid sources a | 30‐day binge alcohol use | 30‐day cannabis use | 12‐month prescription stimulant misuse | 12‐month prescription benzodiazepine misuse | 12‐month substance use disorder (SUD) b | 12‐month nicotine dependence c |
|---|---|---|---|---|---|---|
| aOR (95% CI) | aOR (95% CI) | aOR (95% CI) | aOR (95% CI) | aOR (95% CI) | aOR (95% CI) | |
| Physician source (n = 1168) | Reference group | Reference group | Reference group | Reference group | Reference group | Reference group |
| Friend/relative (free) (n = 1840) |
1.42** (1.14–1.77) |
1.71*** (1.40–2.10) |
1.55*** (1.25–1.91) |
1.75* (1.08–2.83) |
1.29* (1.06–1.58) |
1.79*** (1.34–2.38) |
| Purchased (n = 958) |
1.38** (1.12–1.69) |
2.56*** (2.04–3.22) |
2.31*** (1.75–3.06) |
3.71*** (2.31–5.94) |
2.47*** (1.91–3.19) |
3.57*** (2.82–4.54) |
| Theft (n = 274) |
1.41 (0.99–2.03) |
1.88** (1.33–2.66) |
1.14 (0.79–1.65) |
1.05 (0.56–1.94) |
2.47*** (1.80–3.39) |
1.99** (1.35–2.94) |
| Archer–Lemeshow goodness of fit test |
F (9, 42) = 0.74 p = 0.67 |
F (9, 42) = 0.68 p = 0.73 |
F (9, 42) = 1.72 p = 0.11 |
F (9, 42) = 0.62 p = 0.78 |
F (9, 42) = 0.66 p = 0.74 |
F (9, 42) = 0.79 p = 0.63 |
Note: All regression estimates are adjusted for sex, age, race/ethnicity, income, and population density.
Abbreviations: aOR = adjusted odds ratio, 95% CI = 95% confidence interval.
*denotes p ≤ 0.05; **denotes p ≤ 0.01; ***denotes p ≤ 0.001.
Source categories from most recent episode of misuse are: Obtained opioids from one or more doctors (Physician Source); Obtained from a friend or relative for free (Friend/Relative [free]); Purchased from friend, relative, or stranger/dealer (Purchased); and Took from friend/relative without asking or stole from a medical source (e.g., physician's office, pharmacy) (Theft).
Any SUD captures past‐year DSM‐IV Substance Abuse or Dependence from use or misuse of one or more of: alcohol, cannabis, heroin, cocaine, inhalants, hallucinogens, methamphetamine, prescription opioids, prescription stimulants, prescription tranquilizers, or prescription sedatives.
Nicotine Dependence (Nicotine Dep) is measured using the Nicotine Dependence Syndrome Scale.
All mental health characteristics were more common in those who purchased, stole, or used friends or relatives for free to obtain opioids for POM than the physician source group, except for past‐year major depression in those who purchased opioids (see Table 5). The theft group had the highest odds of the mental health characteristics versus the physician source group, with ORs for major depression, serious psychological distress, and suicidal ideation of 1.70 (95% CI = 1.20–2.42), 2.28 (95% CI = 1.62–3.20), and 2.26 (95% CI = 1.48–3.44), respectively. Only those who used friends or relatives to obtain opioids for free or purchased opioids had lower odds of being in college or having graduated college than the physician source group. The patterns of results were similar in the 2021–2022 cohort, with fewer significant associations. These analyses are captured in Tables S3 and S4.
TABLE 5.
Logistic regression models of mental health and sociodemographic characteristics by prescription opioid source in young adults with past‐year misuse, via 2015–2019 NSDUH data.
| Opioid sources a | Past‐year major depressive episode | Past‐year serious psychological distress | Past‐year suicidal ideation | Housing instability b | In college or college grad | Member of a sexually diverse group c |
|---|---|---|---|---|---|---|
| aOR (95% CI) | aOR (95% CI) | aOR (95% CI) | aOR (95% CI) | aOR (95% CI) | aOR (95% CI) | |
| Physician source (n = 1168) | Reference group | Reference group | Reference group | Reference group | Reference group | Reference group |
| Friend/relative (free) (n = 1840) |
1.42** (1.10–1.83) |
1.28* (1.03–1.59) |
1.25* (1.01–1.54) |
1.12 (0.92–1.36) |
0.79* (0.65–0.97) |
1.28 (0.98–1.67) |
| Purchased (n = 958) |
1.24 (0.92–1.66) |
1.47** (1.11–1.94) |
1.50* (1.10–2.04) |
1.18 (0.94–1.49) |
0.45*** (0.34–0.61) |
1.35 (0.96–1.91) |
| Theft (n = 274) |
1.70** (1.20–2.42) |
2.28*** (1.62–3.20) |
2.26*** (1.48–3.44) |
1.19 (0.76–1.86) |
0.87 (0.61–1.22) |
1.40 (0.88–2.22) |
| Archer–Lemeshow goodness of fit test |
F (9, 42) = 0.54 p = 0.84 |
F (9, 42) = 1.64 p = 0.13 |
F (9, 42) = 2.32 p = 0.03 |
F (9, 42) = 1.07 p = 0.40 |
F (9, 42) = 1.68 p = 0.12 |
F (9, 42) = 2.02 p = 0.06 |
Note: All outcomes are adjusted for sex, age, race/ethnicity, income, and population density.
Abbreviations: aOR = Adjusted odds ratio, 95% CI = 95% confidence interval.
* denotes p ≤ 0.05; ** denotes p ≤ 0.01; *** denotes p ≤ 0.001.
Source categories from most recent episode of misuse are: Obtained opioids from one or more doctors (Physician Source); Obtained from a friend or relative for free (Friend/Relative [free]); Purchased from friend, relative, or stranger/dealer (Purchased); and Took from friend/relative without asking or stole from a medical source (e.g., physician's office, pharmacy) (Theft).
Housing Instability is two or more past‐year moves between residences.
Membership in a Sexually Diverse Group captures those who identify as gay, lesbian, or bisexual.
4. Discussion
This research found initial evidence of age‐ and cohort‐based differences in POM sources. In the 2015–2019 cohort, purchases and obtaining opioids from friends or family for free were slightly more common at older ages in the AYA group, while physician sources and theft became somewhat less common; only theft, though, evidenced a significant decline over the AYA period. Results differed somewhat in the 2021–2022 cohort, with a much greater magnitude of increases from 14/15 to 24/25‐year‐old participants in purchases and a much greater magnitude of decreases in physician source use. While these differences should be interpreted cautiously, given wide confidence intervals in the 2021–2022 results (largely due to smaller sample sizes), they broadly concur with those of McCabe et al. [23], who found significant changes in prescription misuse sources over the 2009–2021 period in 17/18‐year‐old individuals. While our differences between 2015–2019 and 2021–2022 do not mirror those of McCabe et al. [23], our results concur that disruptions seen in illicit opioid procurement networks [24] may have extended to POM.
In line with prior research [14, 15, 16, 17, 18, 19, 20, 21, 25, 26, 38], we found that AYAs who purchased opioids for POM typically had the highest rates of prescription opioid use disorder, rates of other substance use, SUD, and mental health concerns, and a greater mean number of past‐month days of POM. Young adults who purchased opioids were also less likely to be in education or have graduated college, further highlighting the vulnerability of this group. Those engaged in theft or who obtained opioids from friends or family for free also had elevated odds of other substance use and mental health concerns versus the physician source group, though these odds were typically lower than AYAs in the purchases group.
It may be that purchases serve as a marker of greater overall and nonspecific substance use involvement, given their high rates of other substance use and any SUD. At the same time, this greater overall substance use involvement occurs with the highest rates of prescription opioid use disorder and greatest mean days of POM, highlighting that AYAs who purchase opioids for POM likely need both general substance use treatment and interventions specifically focused on POM. Given that those who purchase opioid medication for POM are more likely than other groups to encounter counterfeit pressed pills that contain fentanyl and/or other contaminants like xyazline, harm reduction interventions that include drug checking kits and increased naloxone availability are also potentially important in this group [39, 40]. With that said, any POM is linked to higher rates of other substance use, SUD, and other poor outcomes in AYAs [6]. As such, even the lowest risk group here, AYAs using physician sources, will need identification and intervention.
To better understand POM and POM source groups, research linking POM sources to POM motives could further our understanding of AYA POM and highlight potential treatment targets. To illustrate, if those who purchase opioid medication have high rates of POM to get high, experiment, and alter other drug effects, this would further suggest nonspecific substance involvement and a need for broad substance use treatment. Furthermore, linking sources to POM motives like pain relief, sleep promotion, and anxiolysis would point to nonopioid treatments for these symptoms in those using those sources. A second direction for future research is to evaluate longitudinal trajectories of both POM sources through the AYA period and potential risk factors for use of purchases for POM. Given the differential odds of substance use and mental health concerns by source category, a better understanding of trajectories of POM sources and developmental differences between source groups could help identify those in greatest need of intervention, developmental characteristics of source groups, and key points for intervention. Finally, work to develop prevention programs to limit POM is greatly needed. Currently, the validated Iowa Strengthening Families Program [41, 42, 43] has longitudinal evidence of effectiveness in preventing POM in AYAs, but other interventions are needed both to prevent POM incidence and entrenchment.
4.1. Limitations
The key limitations are those inherent to the NSDUH, which include the cross‐sectional data, potential measurement error associated with self‐reporting of behaviors, selection bias that is not fully corrected by weighting adjustments, limited measures available relevant to pain, and the nature of the POM sources assessment. The cross‐sectional data prevent causal inference, and refusal to participate by some approached individuals introduces a degree of bias. Nonetheless, self‐report substance use data are generally valid [44, 45], and protocols followed in the NSDUH further increased data validity (e.g., ACASI interviewing, medication pictures, and use of trade and generic medication names) [46]. The NSDUH does not assess current pain level, chronic pain, and other pain conditions, which limits our understanding of relationships between POM, POM sources, and pain. Furthermore, sources were only assessed at the most recent POM episode, which obscures within‐person changes over time and prevents evaluation of multiple source use. AYAs using multiple POM sources have particularly high rates of other substance use [18, 19, 26], and our inability to capture that is a limitation. Finally, the 2015–2019 logistic regression analyses for adolescent nicotine dependence and young adult suicidal ideation showed evidence of poor model fit; the 2021–2022 analyses had six analyses with poor fit, and these results should be interpreted given poorer model fit.
5. Conclusion
This study provided further evidence of both age‐ and cohort‐based differences in opioid sources for POM. While the changes were generally smaller, opioid theft from family, peers, or medical sources became less common with aging, and the overall magnitude of differences between 14/15 and 24/25‐year‐olds in purchases for POM in the 2021–2022 results was larger than in the 2015–2019 results. Prevalence differences between the 2015–2019 and 2021–2022 POM sources may suggest disruptions to prior networks used to obtain opioid medication at a time when the COVID‐19 pandemic altered mobility, but this is speculative and warrants further investigation. Those who purchase or steal opioids are more likely to engage in other substance use, have a SUD, and have mental health concerns; similarly, those who purchase opioids for POM are more likely to have more POM episodes and a use disorder from POM. These two subgroups are in greatest need of identification, and clinicians are encouraged to ask individuals with signs of POM how they obtained medication for misuse. Coordination of care with psychiatrists and addiction professionals is essential to achieve optimal outcomes not just in those who purchase or steal medication for POM but all AYAs who engage in POM.
Author Contributions
All authors meet the ICMJE criteria for authorship, as they made substantial contributions to the conception or design of the work, drafted and/or reviewed the manuscript critically for important intellectual content, approved of the final version to be published, and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Primary drafting of the article and all statistical analyses were performed by the first author, while all authors contributed to the conception and design of the research.
Ethics Statement
The National Survey on Drug Use and Health (NSDUH) protocol was approved by the Research Triangle International IRB, and the first author's IRB exempted this research from further oversight.
Consent
All participants provided either informed consent (for those 18 years and older) or assent with parental consent (for those under 18 years of age) prior to initiation of any study procedures.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Data S1.
Schepis T. S., West B. T., Ford J. A., Veliz P. T., and McCabe S. E., “US Adolescents’ and Young Adults’ Sources for Prescription Opioid Misuse: Age, Substance Use, and Mental Health Differences,” Pain Practice 25, no. 6 (2025): e70057, 10.1111/papr.70057.
Funding: The NSDUH is funded by the Substance Abuse and Mental Health Services Administration (SAMHSA), and this work was supported by the National Institutes of Health (NIH) via R01DA043691, R01CA276500, and R01DA031160. None of the funders had any role in this study's design, the collection, analysis or interpretation of data, the writing of the report, or the decision to submit the paper for publication.
Data Availability Statement
The National Survey on Drug Use and Health datasets are available from the Substance Abuse and Mental Health Service Administration's data archive, here: https://www.datafiles.samhsa.gov/study‐series/national‐survey‐drug‐use‐and‐health‐nsduh‐nid13517.
References
- 1. D'Amico E. J., Rodriguez A., Dunbar M. S., et al., “Sources of Cannabis Among Young Adults and Associations With Cannabis‐Related Outcomes,” International Journal on Drug Policy 86 (2020): 102971, 10.1016/j.drugpo.2020.102971. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Schepis T. S., Klare D. L., Ford J. A., and McCabe S. E., “Prescription Drug Misuse: Taking a Lifespan Perspective,” Substance Abuse 14 (2020): 1178221820909352, 10.1177/1178221820909352. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Wall M., Cheslack‐Postava K., Hu M. C., Feng T., Griesler P., and Kandel D. B., “Nonmedical Prescription Opioids and Pathways of Drug Involvement in the US: Generational Differences,” Drug and Alcohol Dependence 182 (2018): 103–111, 10.1016/j.drugalcdep.2017.10.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Rigg K. K., McLean K., Monnat S. M., G. E. Sterner, 3rd , and Verdery A. M., “Opioid Misuse Initiation: Implications for Intervention,” Journal of Addictive Diseases 37, no. 3–4 (2018): 111–122, 10.1080/10550887.2019.1609336. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Novak S. P., Bluthenthal R., Wenger L., Chu D., and Kral A. H., “Initiation of Heroin and Prescription Opioid Pain Relievers by Birth Cohort,” American Journal of Public Health 106, no. 2 (2016): 298–300, 10.2105/ajph.2015.302972. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Bonar E. E., Coughlin L., Roche J. S., et al., “Prescription Opioid Misuse Among Adolescents and Emerging Adults in the United States: A Scoping Review,” Preventive Medicine 132 (2020): 105972, 10.1016/j.ypmed.2019.105972. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Bohm M. K. and Clayton H. B., “Nonmedical Use of Prescription Opioids, Heroin Use, Injection Drug Use, and Overdose Mortality in U.S. Adolescents,” Journal of Studies on Alcohol and Drugs 81, no. 4 (2020): 484–488, 10.15288/jsad.2020.81.484. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Compton W. M., Jones C. M., and Baldwin G. T., “Relationship Between Nonmedical Prescription‐Opioid Use and Heroin Use,” New England Journal of Medicine 374, no. 2 (2016): 154–163, 10.1056/NEJMra1508490. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. McCabe S. E., Boyd C. J., Evans‐Polce R. J., McCabe V. V., Schulenberg J. E., and Veliz P. T., “Pills to Powder: A 17‐Year Transition From Prescription Opioids to Heroin Among US Adolescents Followed Into Adulthood,” Journal of Addiction Medicine 15, no. 3 (2021): 241–244, 10.1097/adm.0000000000000741. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Bazazi A. R., Low P., Gomez B. O., et al., “Overdose From Unintentional Fentanyl Use When Intending to Use a Non‐Opioid Substance: An Analysis of Medically Attended Opioid Overdose Events,” Journal of Urban Health 101, no. 2 (2024): 245–251, 10.1007/s11524-024-00852-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Dickson‐Gomez J., Krechel S., Spector A., et al., “The Effects of Opioid Policy Changes on Transitions From Prescription Opioids to Heroin, Fentanyl and Injection Drug Use: A Qualitative Analysis,” Substance Abuse Treatment, Prevention, and Policy 17, no. 1 (2022): 55, 10.1186/s13011-022-00480-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Barrett S. P., Meisner J. R., and Stewart S. H., “What Constitutes Prescription Drug Misuse? Problems and Pitfalls of Current Conceptualizations,” Current Drug Abuse Reviews 1, no. 3 (2008): 255–262, 10.2174/1874473710801030255. [DOI] [PubMed] [Google Scholar]
- 13. McHugh R. K., Votaw V. R., McCarthy M. D., Bichon J. A., Bailey A. J., and Fitzmaurice G. M., “Defining Prescription Drug Misuse: A Naturalistic Evaluation of National Survey on Drug Use and Health Data From 2012‐2014 to 2015‐2017,” Journal of Addiction Medicine 17, no. 3 (2023): 326–332, 10.1097/adm.0000000000001115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Jones C. M., Paulozzi L. J., and Mack K. A., “Sources of Prescription Opioid Pain Relievers by Frequency of Past‐Year Nonmedical Use United States, 2008‐2011,” JAMA Internal Medicine 174, no. 5 (2014): 802–803, 10.1001/jamainternmed.2013.12809. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Rigg K. K., “Prescription Opioid Diversion Sources Among African Americans: Implications for Overdose Prevention,” Substance Use & Misuse 59, no. 8 (2024): 1271–1274, 10.1080/10826084.2024.2330901. [DOI] [PubMed] [Google Scholar]
- 16. Han B., Compton W. M., Blanco C., Crane E., Lee J., and Jones C. M., “Prescription Opioid Use, Misuse, and Use Disorders in U.S. Adults: 2015 National Survey on Drug Use and Health,” Annals of Internal Medicine 167, no. 5 (2017): 293–301, 10.7326/m17-0865. [DOI] [PubMed] [Google Scholar]
- 17. Schepis T. S., McCabe S. E., and Teter C. J., “Sources of Opioid Medication for Misuse in Older Adults: Results From a Nationally Representative Survey,” Pain 159, no. 8 (2018): 1543–1549, 10.1097/j.pain.0000000000001241. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Schepis T. S., Wilens T. E., and McCabe S. E., “Prescription Drug Misuse: Sources of Controlled Medications in Adolescents,” Journal of the American Academy of Child and Adolescent Psychiatry 58, no. 7 (2019): 670–680, 10.1016/j.jaac.2018.09.438. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. McCabe S. E., Veliz P., Wilens T. E., et al., “Sources of Nonmedical Prescription Drug Misuse Among US High School Seniors: Differences in Motives and Substance Use Behaviors,” Journal of the American Academy of Child and Adolescent Psychiatry 58, no. 7 (2019): 681–691, 10.1016/j.jaac.2018.11.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Ford J. A., McCabe S. E., and Schepis T. S., “Sources of Prescription Opioids and Tranquilizers for Misuse Among U.S. Young Adults: Differences Between High School Dropouts and Graduates and Associations With Adverse Outcomes,” Journal of Addictive Diseases 39, no. 1 (2021): 54–65, 10.1080/10550887.2020.1814122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Ford J. A., Pomykacz C., Szalewski A., Esteban McCabe S., and Schepis T. S., “Friends and Relatives as Sources of Prescription Opioids for Misuse Among Young Adults: The Significance of Physician Source and Race/Ethnic Differences,” Substance Abuse 41, no. 1 (2020): 93–100, 10.1080/08897077.2019.1635955. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Purser G. L., “Trends in the Source of Prescription Drugs for Misuse Between 2015 and 2019,” Substance Use & Misuse 58, no. 7 (2023): 871–880, 10.1080/10826084.2023.2191697. [DOI] [PubMed] [Google Scholar]
- 23. McCabe S. E., Pasman E., Wilens T., et al., “Adolescent Use, Diversion Sources, and Perceived Procurement Difficulty of Prescription Medications,” JAMA 332, no. 7 (2024): 587–589, 10.1001/jama.2024.12030. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Frank D., Krawczyk N., Arshonsky J., Bragg M. A., Friedman S. R., and Bunting A. M., “COVID‐19‐Related Changes to Drug‐Selling Networks and Their Effects on People Who Use Illicit Opioids,” Journal of Studies on Alcohol and Drugs 84, no. 2 (2023): 222–229, 10.15288/jsad.21-00438. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Schepis T. S. and Krishnan‐Sarin S., “Sources of Prescriptions for Misuse by Adolescents: Differences in Sex, Ethnicity, and Severity of Misuse in a Population‐Based Study,” Journal of the American Academy of Child and Adolescent Psychiatry 48, no. 8 (2009): 828–836, 10.1097/CHI.0b013e3181a8130d. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. McCabe S. E., Teter C. J., Boyd C. J., Wilens T. E., and Schepis T. S., “Sources of Prescription Medication Misuse Among Young Adults in the United States: The Role of Educational Status,” Journal of Clinical Psychiatry 79, no. 2 (2018): 1–8, 10.4088/JCP.17m11958. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Center for Behavioral Health Statistics and Quality , “2022 National Survey on Drug Use and Health Public Use File Codebook.” Substance Abuse and Mental Health Services Administration. accessed October 8, 2024, https://www.datafiles.samhsa.gov/sites/default/files/field‐uploads‐protected/studies/NSDUH‐2022/NSDUH‐2022‐datasets/NSDUH‐2022‐DS0001/NSDUH‐2022‐DS0001‐info/NSDUH‐2022‐DS0001‐info‐codebook.pdf.
- 28. Center for Behavioral Health Statistics and Quality , “2016 National Survey on Drug Use and Health: Methodological Resource Book (Section 8, Data Collection Final Report).” Substance Abuse and Mental Health Services Administration. accessed October 8, 2024, https://www.samhsa.gov/data/report/nsduh‐2016‐methodological‐resource‐book‐mrb.
- 29. Substance Abuse and Mental Health Services Administration , “NSDUH 2019 Methodological Resource Book (MRB).” Substance Abuse and Mental Health Services Administration. Updated May 11, 2021. accessed February 13, 2024, https://www.samhsa.gov/data/report/nsduh‐2019‐methodological‐resource‐book‐mrb.
- 30. Substance Abuse and Mental Health Services Administration , “NSDUH 2018 Methodological Resource Book (MRB).” Substance Abuse and Mental Health Services Administration. Updated December 21, 2017. accessed February 13, 2024, https://www.samhsa.gov/data/report/nsduh‐2018‐methodological‐resource‐book‐mrb.
- 31. Substance Abuse and Mental Health Services Administration , “NSDUH 2017 Methodological Resource Book (MRB).” Substance Abuse and Mental Health Services Administration. Updated February 23, 2017. accessed February 13, 2024, https://www.samhsa.gov/data/report/nsduh‐2017‐methodological‐resource‐book‐mrb.
- 32. Substance Abuse and Mental Health Services Administration , “NSDUH 2016 Methodological Resource Book (MRB).” Substance Abuse and Mental Health Services Administration. Updated December 29, 2015. accessed February 13, 2024, https://www.samhsa.gov/data/report/nsduh‐2016‐methodological‐resource‐book‐mrb.
- 33. Substance Abuse and Mental Health Services Administration , “NSDUH 2015 Methodological Resource Book (MRB).” Substance Abuse and Mental Health Services Administration. Updated February 18, 2015. accessed February 13, 2024, https://www.samhsa.gov/data/report/nsduh‐2020‐methodological‐resource‐book‐mrb.
- 34. NIH , “National Institute of Alcohol Abuse and Alcoholism,” NIAAA Newsletter, Winter 2004 3, no. 3 (2004): 3. [Google Scholar]
- 35. Center for Behavioral Health Statistics and Quality , “2019 National Survey on Drug Use and Health Public Use File Codebook.” Substance Abuse and Mental Health Services Administration. accessed October 8, 2024, https://www.datafiles.samhsa.gov/sites/default/files/field‐uploads‐protected/studies/NSDUH‐2019/NSDUH‐2019‐datasets/NSDUH‐2019‐DS0001/NSDUH‐2019‐DS0001‐info/NSDUH‐2019‐DS0001‐info‐codebook.pdf.
- 36. Kessler R. C., Barker P. R., Colpe L. J., et al., “Screening for Serious Mental Illness in the General Population,” Archives of General Psychiatry 60, no. 2 (2003): 184–189, 10.1001/archpsyc.60.2.184. [DOI] [PubMed] [Google Scholar]
- 37. Rao J. N. K. and Scott A. J., “On chi‐Squared Tests for Multiway Contingency Tables With Cell Proportions Estimated From Survey Data,” Annals of Statistics 12, no. 1 (1984): 46–60. [Google Scholar]
- 38. Park J. Y. and Wu L. T., “Sources of Misused Prescription Opioids and Their Association With Prescription Opioid Use Disorder in the United States: Sex and Age Differences,” Substance Use & Misuse 55, no. 6 (2020): 928–936, 10.1080/10826084.2020.1713818. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Kutscher E., Barber Grossi M., LaPolla F., and Lee J. D., “Fentanyl Test Strips for Harm Reduction: A Scoping Review,” Journal of Addiction Medicine 18, no. 4 (2024): 373–380, 10.1097/adm.0000000000001321. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Wakeman S. E., “Opioid Use Disorder Diagnosis and Management,” NEJM Evidence 1, no. 4 (2022): EVIDra2200038, 10.1056/EVIDra2200038. [DOI] [PubMed] [Google Scholar]
- 41. Spoth R., Trudeau L., Shin C., et al., “Longitudinal Effects of Universal Preventive Intervention on Prescription Drug Misuse: Three Randomized Controlled Trials With Late Adolescents and Young Adults,” American Journal of Public Health 103, no. 4 (2013): 665–672, 10.2105/ajph.2012.301209. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Spoth R., Trudeau L., Shin C., and Redmond C., “Long‐Term Effects of Universal Preventive Interventions on Prescription Drug Misuse,” Addiction 103, no. 7 (2008): 1160–1168, 10.1111/j.1360-0443.2008.02160.x. [DOI] [PubMed] [Google Scholar]
- 43. Spoth R., Trudeau L., Redmond C., and Shin C., “Replicating and Extending a Model of Effects of Universal Preventive Intervention During Early Adolescence on Young Adult Substance Misuse,” Journal of Consulting and Clinical Psychology 84, no. 10 (2016): 913–921, 10.1037/ccp0000131. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. O'Malley P. M., Bachman J. G., and Johnston L. D., “Reliability and Consistency in Self‐Reports of Drug Use,” International Journal of the Addictions 18 (1983): 805–824, 10.3109/10826088309033049. [DOI] [PubMed] [Google Scholar]
- 45. Johnston L. D. and O'Malley P. M., “Issues of Validity and Population Coverage in Student Surveys of Drug Use,” NIDA Research Monograph 57 (1985): 31–54. [PubMed] [Google Scholar]
- 46. Center for Behavioral Health Statistics and Quality , National Survey on Drug Use and Health (NSDUH): Summary of methodological studies, 1971‐2014 Substance Abuse and Mental Health Services Administration. accessed October 8, 2024, https://www.ncbi.nlm.nih.gov/books/NBK519735/. [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data S1.
Data Availability Statement
The National Survey on Drug Use and Health datasets are available from the Substance Abuse and Mental Health Service Administration's data archive, here: https://www.datafiles.samhsa.gov/study‐series/national‐survey‐drug‐use‐and‐health‐nsduh‐nid13517.