Adolescent Polysubstance Use and Co-Occurring Weapon Carrying, Bullying Victimization, and Depressive Symptomology: Patterns and Differences in the United States

Abstract

Adolescent substance use commonly co-occurs with poor mental health, bullying victimization and risky behaviors that may lead to violence. The purpose was to describe the United States (US) national prevalence of polysubstance use and co-occurring characteristics and associated demographic characteristics among youth. Middle and high school students in the 2019 CDC YRBS survey reported their demographics and current (≥ 1 days in the last 30 days) substances used (alcohol, cigarette, e-cigarette, cannabis); polysubstance combinations were generated. Cross-sectional weighted logistic regression estimated odds of polysubstance use and frequent use (≥ 6 days in the last 30 days) by weapon carrying, depressive symptoms, bullying victimization, and demographics. Mean age of the sample was 16 years, 51% were boys, 51% were non-Hispanic White. While accounting for 21% of the sample, 22–40% of Multiracial youth reported polysubstance use and frequent use. Odds of frequent polysubstance use (all combinations) were highest for weapon carrying youth.

Introduction

Clarifying the patterns and correlates of adolescent substance use is critical in improving the wellbeing of United States (US) adolescents and informing evidence-based prevention approaches in this population. While much is known about singular substance use among adolescents (i.e., using one drug at a time), the epidemiology and etiology of adolescent polysubstance use (i.e., using two or more substances) remains less known. It is estimated that almost 50% of adolescents reported ever using one or more illicit substances together by age 18 [1–3]. Adolescence, defined as 10 to 19 years of age [4], is a consequential time in human development and engaging in substance use during this time can increase the risk of substance misuse or result in a diagnosis of substance use disorder later in life [5]. Specifically, alcohol and cannabis use have each been associated with misuse of other harmful substances (e.g., cocaine, heroin, or opioids) and with polysubstance use [6–9]. Moreover, heroin and prescription opioid polysubstance use has been associated with fatal outcomes[10, 11] and increased risk of being violently victimized during these formative years [12, 13]. Collectively, previous research shows that polysubstance use during adolescence is a risk factor for more severe consequences relative to singular substance use and is associated with negative mental health, delinquent and/or violent behaviors at school [14–18]. However, much remains unknown as it relates to specific polysubstance use combinations and their association with the complexity of issues faced by US adolescents.

Depressive Symptomology

Depressive symptomology, an indicator of poor mental health [19], has been linked to alcohol use, past month cigarette use [20, 21], higher frequency of e-cigarette use [22, 23], and marijuana use [21] among adolescents. One study showed depressive symptoms associated with first-time use of tobacco and alcohol and increased frequency of these substances [24]; another found robust associations between depressive symptoms and alcohol but not marijuana use [25]. An association between synthetic cannabinoid use and depressive symptomology became statistically non-significant when other substances were controlled for (e.g., tobacco, alcohol, and marijuana) [26] – contrary to another study which found depressive symptoms to be longitudinally associated with synthetic cannabinoid use [27]. Considering the significance of this relationship with some substances and not others, and other conflicting findings, the association between depressive symptomology and polysubstance use warrants further examination.

Bullying Victimization

Bullying victimization may place adolescents at greater risk of engaging in multiple types of substance use due to increased stress [28–32]. One study found those reporting bullying victimization were at increased odds of using three or more substances concurrently relative to singular use [33]. Further, youth who reported bullying victimization were more likely to report cigarette, marijuana, and alcohol use compared to those who had not experienced bullying [18, 34]. In another study, while preliminary models indicated that bullying victimization was positively associated with substance use, once alcohol, tobacco, and cannabis use were controlled for, the association became non-significant [35]. These conflicting findings necessitate further consideration.

Weapon Carrying

Substance use and misuse is also found to be positively associated with weapon carrying [36–42]. Gun carrying, specifically, is found to be associated with frequency of cigarette, alcohol, cocaine, and marijuana use, while carrying a knife or club has been found to be associated with frequency of alcohol and cigarette use [37]. Youth who had recently used alcohol or tobacco were also found to have increased odds of carrying a gun [37, 38] but findings are sometimes contradictory when adjusting for demographics [39]. Therefore, there are gaps regarding these differing associations concerning polysubstance use combinations and weapon carrying.

Study aims

To the authors’ knowledge very few, if any, studies to date have estimated the national prevalence of polysubstance use among a national sample of middle and high school adolescents in the US, or further described correlated conditions and behaviors. While the Centers for Disease Control and Prevention (CDC) have reported singular substance use prevalence among this population [43], there are no (to our knowledge) national reports of concurrent polysubstance use. This analysis aims to: (1) describe the national prevalence of polysubstance use combinations, frequency of use of these combinations, and associated characteristics (including demographics and co-occurring depressive symptomology, weapon carrying, and bullying victimization) among middle and high school adolescents in the US and (2) quantify the association between polysubstance use, frequent polysubstance use, and associated demographic and behavioral correlates.

Methods

Sample

Data from the 2019 Youth Risk Behavior Surveillance System (YRBSS) was used for this analysis. The YRBSS is a biennial cross-sectional survey administered to middle and high school students which aims to capture data on six categories of health-related behaviors that contribute to the leading causes of death and disability among adolescents [44]. The survey employs a sampling design that ensures national representation of demographic groups [45]. All participants were included in this analysis; therefore, appropriate sampling weights were applied and survey estimations were implemented to account for the sampling design [45]. Data used for this secondary analysis are de-identified and publicly available, thus no protocol approval from an institutional review board was necessary.

Measures

The YRBSS captured data on use of substances including alcohol, cigarettes, cigars and cigarillos, chewing tobacco and snuff, prescription opioids, marijuana (cannabis), cocaine, inhalants, heroin, methamphetamine, ecstasy, and LSD. However, the YRBSS only focuses on current usage (past 30 days) for alcohol, cigarettes, cigars and cigarillos, cannabis, electronic nicotine vapor products, chewing tobacco and snuff, and prescription opioids. Alcohol, cigarettes, electronic nicotine vapor products, and cannabis were of interest as these substances are increasingly the most used among this population [3]. Prescription opioids, cigars and cigarillos, and chewing tobacco and snuff were excluded from bivariate analyses due to small sample size after stratification. The primary outcome of interest was polysubstance use combinations, while primary exposures consisted of demographics (age, sex, and race/ethnicity) and correlates (depressive symptoms, bullying victimization, and weapon carrying).

Substance use

The YRBSS asked, “During the past 30 days, on how many days did you smoke cigarettes?” Students in the sample were labeled as current cigarette smokers if they reported smoking cigarettes on at least one day in the past 30 days, while non-users reported zero days. The YRBSS asked, “During the past 30 days, on how many days did you have at least one drink of alcohol?” Students were current alcohol users if they reported drinking alcohol on at least one day in the past 30 days, while non-users reported zero days. The YRBSS asked, “During the past 30 days, how many times did you use marijuana?” Students were current cannabis users if they reported using marijuana at least once in the past 30 days, while non-users reported zero days. The YRBSS asked, “During the past 30 days, how many times did you use an electronic nicotine vapor product?” Students were current electronic nicotine vapor product users if they reported using at least once in the past 30 days, while non-users reported zero days. Polysubstance use contained any combination of alcohol, cigarettes, cannabis, and electronic nicotine vapor products. Frequent use was defined as six to 30 days in the last 30 days based on the definition of frequent use as at least once a week or more established by Shedler & Block [46]; moderate use was defined as one to five days in the last 30 days; no use was zero days in the last 30 days. Moderate and no use was collapsed into one category based on the distribution of the variable.

Co-occurring Characteristics

Weapon Carrying

Weapon carrying was measured on a binary scale. Students were asked, “During the past 30 days, on how many days did you carry a weapon such as a gun, knife, or club?” and could choose from the following responses: 0 days, 1 day, 2 or 3 days, 4 or 5 days, or 6 or more days. Students were considered to carry a weapon if they reported carrying on one or more days and did not carry a weapon if they reported zero days.

Depressive Symptomology

Prolonged sadness or hopelessness was used as an indicator of depressive symptoms which was measured on a binary scale. The YRBSS asked, “During the past 12 months, did you ever feel so sad or hopeless almost every day for two weeks or more in a row that you stopped doing some usual activities?” Students could respond “yes” or “no”. The YRBSS did not include any other measurement of mental health.

Bullying Victimization

Bullying victimization was measured on a binary scale. The YRBSS asked, “During the past 12 months, have you ever been bullied on school property?” Students could respond “yes” or “no”.

Demographics

YRBSS captured data on age (years), gender (boy/girl), and race/ethnicity (Non-Hispanic [NH] White [NHW], NH Black [NHB], Hispanic, Multiracial (defined as those who selected multiple races; both Hispanic and NH), and other [Asian/Pacific Islander/Native Hawaiian/Native Alaskan]). Due to small sample size and in order to increase power, Asians, Pacific Islanders, Native Hawaiians, and Native Alaskans were collapsed into one group. All participants surveyed were middle and high school students; ages ranged from 12 to 18 years. Ages 12 and 13 were excluded from analysis due to small sample size (unweighted n = 87). The total sample surveyed was 13,677, and the final analytical sample was 13,590.

Statistical Approach

Univariate analyses described distribution of demographics and substance and polysubstance use. Bivariate tests (t-tests, chi-square) determined any associations between demographics and polysubstance combinations. Weighted multivariable logistic regression models determined the odds of polysubstance use (Y/N) and frequent use (frequent/no or moderate use) by co-occurring characteristics and key demographics. A sensitivity analysis explored the same multivariate models with cyberbullying victimization as an indicator rather than “conventional” bullying victimization. Results did not change significantly and therefore are not shown. An additional sensitivity analysis explored the same correlates with a multinomial substance use variable as the outcome (0: no substance use, 1: singular substance use of any substance, 2: poly use of any combination). Results are presented in the supplemental materials. Analyses were conducted with survey estimations using Stata statistical software version 15.1 (College Station, TX).

Results

The final analytical sample (unweighted n = 13,590) showed that in 2019, middle and high school students across the US were 49% girls, 51% NHW, 21% Multiracial, and 12% NHB. Approximately 20% reported bullying victimization, 13% reported carrying a weapon on school property at least once, and 37% reported depressive symptomology. About 31% of the sample reported electronic nicotine vapor product use in the past 30 days, 29% of the sample reported drinking alcohol in the past 30 days, 21% of the sample reported using cannabis in the past 30 days, while prevalence of other forms of substance use remained low (< 10%). Approximately 18% of US adolescents reported using both alcohol and cannabis at least once in the past 30 days, 16% reported using cannabis and electronic nicotine vapor products, 13% reported using alcohol and electronic nicotine vapor products, 8% reported using alcohol and cigarettes, 6% reported using cannabis and cigarettes, and 6% reported using cigarettes and electronic nicotine vapor products. Cannabis and electronic nicotine vapor product co-use had the highest prevalence of frequent use at 9%, followed by 6% prevalence of frequent alcohol and electronic nicotine vapor product co-use, while prevalence of other combinations of frequent co-use remained low (< 5%). (Table 1).

Table 1.

Distribution of 2019 YRBSS demographics (Unweighted N = 13,590; weighted N = 14,000)

	unweighted N (weighted %)
Age (mean (SE))	16.0 (0.03)
Age, years
14	1,699 (11.9)
15	3,473 (24.8)
16	3,628 (25.7)
17	3,102 (23.8)
18	1,616 (13.8)
Gender, n (%)
Boys	6,41 (50.6)
Girls	6,885 (49.4)
Race, n (%)
Non-Hispanic [NH] White	6,668 (51.2)
Non-Hispanic Black	2,040 (12.2)
Hispanic	1,009 (9.2)
Multiracial	2,690 (21.4)
Other*	832 (6.1)
Bullying victimization, n (%)
No	10,744 (80.5)
Yes	2,703 (19.5)
Weapon carrying, n (%)
0 days	9,141 (86.8)
At least once	1,403 (13.2)
Depressive symptomology, n (%)
No	8,495 (63.3)
Yes	4,926 (36.7)
Singular substance use, n (%)
Current alcohol use	3,669 (29.2)
Current cigarette use	2,086 (9.9)
Current cannabis use	2,946 (21.2)
Current electronic nicotine vapor product use	4,109 (30.8)
Current opioid misuse	661 (5.7)
Current cigars, cigarillos, or little cigars use	755 (5.3)
Current chewing tobacco, snuff, dip, snus, or dissolvable tobacco products use	544 (3.7)
Polysubstance use, n (%)
Alcohol and cannabis use	1,741 (17.9)
Cannabis and electronic nicotine vapor product use	2,068 (16.0)
Alcohol and electronic nicotine vapor product use	1,316 (13.4)
Alcohol and cigarette use	900 (7.9)
Cannabis and cigarette use	830 (6.3)
Cigarette and electronic nicotine vapor product use	1,013 (6.1)
Frequent polysubstance use^, (n %)
Cannabis and electronic nicotine vapor product use	792 (9.1)
Alcohol and electronic nicotine vapor product use	509 (6.1)
Alcohol and cannabis use	330 (3.6)
Cigarette and electronic nicotine vapor product use	209 (2.6)
Cannabis and cigarette use	171 (1.8)
Alcohol and cigarette use	138 (1.6)

^*Asian, Pacific Islander, Native Hawaiian, and Native Alaskan

^{^}Frequent use: six-30 days in the last 30 days

Table 2 displays current polysubstance use and frequent use prevalence stratified by key demographics. Age was statistically associated (p<0.05) with all combinations of polysubstance use and frequent polysubstance use. Gender was statistically associated (p<0.05) with cannabis and cigarette co-use, frequent cannabis and cigarette co-use, cigarette and electronic nicotine vapor product co-use, frequent alcohol and cannabis co-use, frequent alcohol and electronic nicotine vapor product co-use, and frequent cannabis and electronic nicotine vapor product co-use. Race/ethnicity was statistically associated (p<0.05) with all combinations of polysubstance use and frequent use except alcohol and cigarette use, cannabis and cigarette use, and frequent cigarette and electronic nicotine vapor product use.

Table 2.

Polysubstance use and frequent use^*stratified by key demographics, weapon carrying, and depressive symptomology

Weighted %
	Alcohol and cigarette use	Frequent alcohol and cigarette use	Alcohol and cannabis use	Frequent alcohol and cannabis use	Cannabis and cigarette use	Frequent cannabis and cigarette use	Alcohol and electronic nicotine vapor product use	Frequent alcohol and electronic nicotine vapor product use	Cigarette and electronic nicotine vapor product use	Frequent cigarette and electronic nicotine vapor product use	Cannabis and electronic nicotine vapor product use	Frequent cannabis and electronic nicotine vapor product use
Age, years (%)
14	8.7	6.9	7.1	5.4	8.8	7.8	10.3	5.6	9.8	7.8	8.2	7.8
15	16.4	12.9	17.1	13.8	15.7	10.3	22.9	13.0	17.3	14.2	17.7	14.6
16	22.8	23.4	26.2	24.6	22.1	25.2	29.4	24.6	23.9	26.2	27.4	26.8
17	28.7	31.6	30.2	32.7	30.5	25.1	23.7	30.7	28.4	22.7	28.0	29.8
18	23.3	25.2	19.4	23.6	22.9	31.6	13.7	24.2	22.1	29.0	18.7	21.0
P	< 0.001	0.002	< 0.001	< 0.001	< 0.001	< 0.001	0.003	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001
Gender (%)
Boys	42.3	61.3	52.1	63.4	58.0	64.5	48.3	58.8	58.6	56.8	47.3	61.6
Girls	56.7	38.7	47.9	36.6	42.0	35.5	51.7	41.2	41.4	43.2	52.7	38.4
P	0.12	0.30	0.08	< 0.001	0.006	0.016	0.75	0.015	0.007	0.35	0.10	< 0.001
Race (%)
Non-Hispanic [NH] White	56.5	49.1	58.9	51.8	52.1	45.9	57.0	65.2	55.3	54.2	58.4	54.7
NH Black	7.9	7.7	7.4	9.0	10.5	11.3	9.0	5.5	7.1	7.3	7.7	8.4
Hispanic	7.4	1.5	7.4	5.0	7.6	4.1	8.7	5.0	7.7	4.7	7.7	6.9
Multiracial	24.7	39.8	23.4	30.9	26.1	32.9	21.8	21.9	25.5	28.9	22.9	25.4
Other^	2.5	1.9	3.0	3.3	3.8	5.8	3.6	2.3	4.5	5.0	3.3	4.7
P	0.07	< 0.001	< 0.001	< 0.001	0.22	< 0.001	< 0.001	< 0.001	0.012	0.09	< 0.001	< 0.001
Bullying victimization (%)
Yes	30.4	43.7	25.0	28.5	31.1	44.2	26.6	24.5	31.7	44.8	25.1	25.2
P	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001
Weapon carrying (%)
At least once	37.8	64.4	24.4	48.2	38.2	59.4	13.3	38.1	36.4	56.1	22.4	30.9
P	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	0.009	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001
Depressive symptomology (%)
Yes	55.2	66.1	53.3	52.7	59.5	64.0	44.6	48.5	56.4	66.7	51.6	54.0
P	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001

^*Frequent use: six-30 days in the last 30 days

^{^}Asian, Pacific Islander, Native Hawaiian, and Native Alaskan

Note: bold text indicates statistically significant estimates

Multiracial adolescents had the highest prevalence of polysubstance use across all combinations relative to their proportion in the general sample. Although comprising 21% of the sample, 40% of frequent alcohol and cigarette co-users were Multiracial. Among frequent alcohol and cannabis co-users, 31% were Multiracial; among frequent cannabis and cigarette co-users, 33% were Multiracial.

When comparing to a national prevalence of 20% in the sample (as shown in Table 1), bullying victimization prevalence was greater than 25% among all combinations of polysubstance use and frequent use. Bullying victimization prevalence was highest among frequent cigarette and electronic nicotine vapor product co-users at 45%, followed by frequent alcohol and cigarette co-users (44%) and frequent cannabis and cigarette co-users (44%). Further, bullying victimization was found to be statistically associated with polysubstance use (p < 0.001 for all bullying victimization-related bivariate comparisons).

Compared to a national prevalence of 13% (as shown in Table 1), the prevalence of weapon carrying among all groups of co-users ranged from 13 to 64%. The prevalence of weapon carrying was highest among frequent alcohol and cigarette co-users (64%), second highest among frequent cannabis and cigarette co-users (59%), followed by frequent cigarette and electronic nicotine vapor product co-users (56%). Among alcohol and electronic nicotine vapor product users, the prevalence of weapon carrying was comparable at 13%. Weapon carrying was statistically associated with all combinations of polysubstance use and frequent polysubstance use (p < 0.05 for all weapon carrying-related bivariate comparisons).

Compared to a national prevalence of 37% (as shown in Table 1), the prevalence of depressive symptomology among groups of co-users ranged from 45 to 67%. Prevalence of depressive symptomology was highest among frequent cigarette and electronic nicotine vapor product cousers (67%), then frequent alcohol and cigarette co-users (66%), followed by frequent cannabis and cigarette co-users (64%). The lowest prevalence of depressive symptomology was among alcohol and electronic nicotine vapor product co-users at 45%. (Depressive symptoms was statistically associated with all combinations of polysubstance use and frequent polysubstance use (p < 0.05 for all bivariate comparisons).

Table 3 displays the weighted multivariable odds of polysubstance use combinations co-occurring characteristics and key demographics. The odds of cigarette and electronic nicotine vapor product use was 42% higher among those who experienced bullying victimization (Adjusted odds Ratio [AOR]: 1.42, 95% Confidence Interval [CI]: 1.07, 1.88), compared to those who did not and adjusting for weapon carrying, depressive symptomology, age, gender, and race/ethnicity. Apart from alcohol and electronic nicotine vapor product co-use, the odds of each combination of polysubstance use were significantly greater for those who carried a weapon after adjusting for bullying victimization, depressive symptomology, age, gender, and race/ethnicity. The adjusted odds of each combination of polysubstance use were significantly greater for those who experienced depressive symptomology compared to those who did not. The odds of each combination of polysubstance use were significantly greater with every year increase in age. Middle and high school girls had 28% lower odds of cannabis and cigarette co-use and 27% lower odds of cigarette and electronic nicotine vapor product use than boys, (AOR: 0.72, 95% CI: 0.53, 0.99; and AOR: 0.73, 95% CI: 0.55, 0.98, respectively). Apart from alcohol and cigarette co-use and cannabis and cigarette co-use, adolescents who identified as NH Black had significantly lower adjusted odds of polysubstance use across all combinations, compared to NH White adolescents.

Table 3.

Weighted multivariable odds of polysubstance use (compared to no substance use)

	Adjusted odds ratio [AOR], 95% Confidence Interval [CI]
	Alcohol and cigarette use	Alcohol and cannabis use	Cannabis and cigarette use	Alcohol and electronic nicotine vapor product use	Cigarette and electronic nicotine vapor product use	Cannabis and electronic nicotine vapor product use
Bullying victimization
No	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
Yes	1.34 (0.93, 1.93)	1.11 (0.88, 1.39)	1.29 (0.89, 1.88)	1.50 (1.15, 1.96)	1.42 (1.07, 1.88)	1.09 (0.90, 1.33)
Weapon carrying
0 times	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
At least once	4.40 (3.24, 5.96)	3.04 (2.44, 3.79)	4.08 (3.04, 5.47)	1.26 (0.93, 1.67)	3.50 (2.69, 4.54)	1.98 (1.69, 2.32)
Depressive symptomology
No	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
Yes	2.31 (1.68, 3.18)	2.35 (1.96, 2.83)	2.76 (1.89, 4.02)	1.83 (1.50, 2.23)	2.21 (1.64, 2.98)	2.06 (1.74, 2.45)
Age (years)	1.41 (1.27, 1.57)	1.44 (1.32, 1.56)	1.41 (1.24, 1.61)	1.12 (1.05, 1.20)	1.30 (1.19, 1.43)	1.25 (1.17, 1.34)
Gender
Boys	Ref.	Ref.	Ref.
Girls	0.93 (0.68, 1.27)	1.13 (0.93, 1.38)	0.72 (0.53, 0.99)	0.87 (0.71, 1.05)	0.73 (0.55, 0.98)	0.83 (0.71, 0.96)
Race
Non-Hispanic [NH] White	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
Non-Hispanic Black	0.51 (0.22, 1.20)	0.54 (0.37, 0.80)	1.19 (0.56, 2.53)	0.64 (0.46, 0.91)	0.57 (0.35, 0.94)	0.64 (0.46, 0.90)
Hispanic	0.95 (0.51, 1.79)	0.76 (0.59, 0.98)	1.05 (0.63, 1.74)	0.79 (0.56, 1.12)	0.87 (0.52, 1.46)	0.78 (0.57, 1.09)
Multiracial	0.87 (0.63, 1.23)	0.92 (0.70, 1.19)	1.09 (0.76, 1.57)	0.84 (0.68, 1.07)	1.04 (0.74, 1.45)	0.93 (0.74, 1.17)
Other*	0.46 (0.26, 0.82)	0.39 (0.26, 0.60)	0.65 (0.37, 1.15)	0.46 (0.33, 0.64)	0.73 (0.45, 1.17)	0.51 (0.38, 0.69)

^*Asian, Pacific Islander, Native Hawaiian, and Native Alaskan

Notes: bold text notates statistically significant estimates. Models included bullying victimization, weapon carrying, depressive symptomology, age, gender, and race/ethnicity.

Table 4 displays the weighted multivariate odds of frequent polysubstance use combinations by co-occurring characteristics and key demographics. The adjusted odds of frequent cannabis and cigarette use, and cigarette and electronic nicotine vapor product use were 2.17 and 2.06 times greater (respectively) for those who experienced bullying victimization (AOR: 2.17, 95% CI: 1.34, 3.51; and AOR 2.06, 95% CI: 1.27, 3.36, respectively) compared to those who did not. Across all combinations of frequent polysubstance use, the adjusted odds were significantly greater for those who carried a weapon compared to those who did not. Additionally, across all combinations of frequent polysubstance use, the adjusted odds were significantly greater for those who experienced depressive symptomology compared to those who did not. Multiracial adolescents had higher adjusted odds of frequent alcohol and cannabis co-use and frequent cannabis and cigarette co-use compared to NHW (AOR: 1.56, 95% CI: 1.02, 2.39; and AOR: 2.07, 95% CI: 1.12, 3.82).

Table 4.

Weighted multivariable odds of frequent polysubstance use^*

	Adjusted odds ratio [AOR], 95% Confidence Internal [CI]
	Frequent alcohol and cigarette use	Frequent alcohol and cannabis use	Frequent cannabis and cigarette use	Frequent alcohol and electronic nicotine vapor product use	Frequent cigarette and electronic nicotine vapor product use	Frequent cannabis and electronic nicotine vapor product use
Bullying victimization
No	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
Yes	1.69 (0.82, 3.48)	1.15 (0.76, 1.75)	2.17 (1.34, 3.51)	1.08 (0.76, 1.53)	2.06 (1.27, 3.36)	1.05 (0.76, 1.46)
Weapon carrying
0 times	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
At least once	12.28 (7.35, 20.51)	6.94 (4.46, 10.78)	9.37 (5.93, 14.80)	4.87 (3.39, 6.98)	8.38 (5.45, 12.84)	3.42 (2.59, 4.52)
Depressive symptomology
No	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
Yes	2.51 (1.34, 4.61)	1.99 (1.35, 2.92)	2.69 (1.49, 4.84)	1.72 (1.28, 2.31)	3.33 (1.71, 6.45)	2.84 (2.16, 3.73)
Age (years)	1.50 (1.18, 1.90)	1.52 (1.32, 1.75)	1.69 (1.35, 2.12)	1.58 (1.34, 1.87)	1.60 (1.32, 1.95)	1.46 (1.33, 1.60)
Gender
Boys	Ref.	Ref.	Ref.
Girls	0.93 (0.43, 1.99)	0.76 (0.52, 1.12)	0.60 (0.32, 1.10)	0.82 (0.59, 1.12)	0.75 (0.42, 1.33)	0.64 (0.51, 0.79)
Race
Non-Hispanic [NH] White	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
Non-Hispanic Black	0.97 (0.22, 4.16)	0.87 (0.35, 2.16)	1.97 (0.75, 5.15)	0.41 (0.14, 1.18)	0.67 (0.22, 2.02)	0.80 (0.47, 1.36)
Hispanic	1.35 (0.32, 5.81)	0.78 (0.38, 1.64)	0.71 (0.14, 3.64)	0.47 (0.26, 0.84)	0.63 (0.15, 2.65)	0.73 (0.45, 1.19)
Multiracial	1.57 (0.75, 3.29)	1.56 (1.02, 2.39)	2.07 (1.12, 3.82)	0.77 (0.54, 1.11)	1.06 (0.56, 2.02)	1.10 (0.77, 1.58)
Other^	0.91 (0.24, 3.43)	0.59 (0.21, 1.69)	1.93 (0.66, 5.66)	0.35 (0.16, 0.74)	1.06 (0.36, 3.14)	0.62 (0.39, 1.00)

^*frequent use: six-30 days in the last 30 days; compared to none to moderate substance use (0-five days in the last 30 days)

^{^}Asian, Pacific Islander, Native Hawaiian, and Native Alaskan. Models include bullying victimization, weapon carrying, depressive symptomology, age, gender, race/ethnicity.

Discussion

This study is one of the first to describe the specific polysubstance use combinations and frequency of these specific combinations among middle and high school students in the US, as well as identify demographic differences and associations with co-occurring factors of mental health, bullying victimization, and weapon carrying. Our analysis shows that approximately 20% of adolescents nationwide reported polysubstance use. Multiracial adolescents used multiple substances at disproportionately higher rates to monoracial counterparts, and rates of frequent polysubstance use are high for individuals with depressive symptomology, those who carry weapons, and victims of bullying. These findings quantify, for the first time, an often-overlooked disparity among one of the most rapidly growing racial/ethnic groups in the US (i.e., Multiracial adolescents) and provide new insight into a conglomerate of associated characteristics that may be important in the prevention of adolescent polysubstance use.

Our findings provide evidence that the associations with weapon carrying appear stronger among those who report frequent polysubstance use among US middle and high school adolescents. In Table 3, the odds of weapon carrying were significant and elevated across varying types of polysubstance use apart from alcohol and electronic nicotine vapor product co-use. However, the odds of frequent polysubstance use and weapon carrying are twice as high in comparison to “occasional” polysubstance use and statistical significance between relationships emerged where it was unobserved previously in the context of regular (“occasional”) use (i.e., for alcohol and electronic nicotine vapor product co-use). It could be understood that as polysubstance use may increase from “occasional” use to frequent use, so do the odds of weapon carrying among sampled youth, as other preliminary evidence supports this [47].

Other than alcohol and electronic nicotine vapor product co-use and frequent co-use, the prevalence of depressive symptomology was greater than 50% for almost all groups of polysubstance use and frequent use. Additionally, the odds of depressive symptomology were consistently and significantly elevated for each specific combination of polysubstance use and frequent use. However, for those experiencing depressive symptomology, there were no apparent differences in the magnitude of these associations between general use versus frequent use. Therefore, we conclude that “risk” of depressive symptoms or frequent polysubstance use does not increase with the presence of the other. In other words, levels of depressive symptoms and days co-using substances do not appear to be stronger as observed in our previous study result. Residual confounding may be present that is important in explaining the apparent association between these two adolescent health outcomes. Using latent growth modeling techniques, Felton and colleagues previously found support of a developmental positive feedback effect by which early polysubstance use led to increased depressive symptomology which, in turn, led to more polysubstance use and, in turn, prolonged poor mental health across the life course [48].

The odds of bullying victimization were only significant for the vaping and cigarette co-use model, the frequent vaping and cigarette co-use model, and the frequent cannabis and cigarette co-use models. In these models, odds of co-use were increased for those who were victimized compared to those who were not. The finding of increased odds of vaping and cigarette co-use aligns with previous research [18, 33]. However, previous research has explored this relationship within the context of cyberbullying [33]. Our sensitivity analysis explored cyberbullying in place of “conventional” bullying and results did not change significantly for any of the estimates. This finding provides evidence that odds of co-use are increased for those who are a victim of any type of bullying. Public health interventions and campaigns should consider targeting smoking and bullying behaviors concurrently when developing targeted approaches for adolescent populations. Further, this finding supports previous evidence that victims of bullying may be co-using specific combinations as a coping mechanism; future work should determine if this finding holds while public health interventions should include information regarding healthy ways of coping.

Results from the sensitivity analysis (results shown in supplemental materials) suggest a lack of specificity between risk factors and specific poly use combinations, as indicated by the similar AORs across models. However, one significant distinction is the racial and ethnic differences that appear to be specific to polysubstance use. Further, for depressive symptoms and age, the sensitivity analysis shows that odds increase from singular to polysubstance use. Future research should further elucidate this relationship with longitudinal data to provide further context and temporality.

Racial and Ethnic Differences

Interestingly, Multiracial adolescents had disproportionately higher prevalence of polysubstance use and of frequent use across all combinations of use relative to their proportion in the general sample. This finding signifies a racial/ethnic difference in polysubstance use among this population. Multiracial health is increasingly becoming the focus of health research as the prevalence of individuals identifying as the Multiracial population increases [49]. In the context of Multiracial health, it is important to highlight the ambiguousness of racial categories since individuals with any combination of races can fall into this category and the significance of the influence that cultural background has on substance use [50], especially since many modern scientists with expertise at the intersection race/ethnicity and health have rejected the notion of a genetic basis to racial categories [51]. There are several hypotheses that have been advanced by scholars in the field in attempts to explain the emergence of health disparities among Multiracial adolescents, namely that racial discrimination and microaggressions—subtle comments or actions which may express a prejudiced attitude toward a member of a marginalized group—related to one’s self-actualized and perceived racial and ethnic identity may play a significant influencing role in mental health and mental health-related correlates like substance use [52–54]. The results from the present study do not provide sufficient evidence to support these hypotheses, these results simply provide epidemiological basis that such racial/ethnic differences exist; it is equally plausible that the broader social and environmental factors contribute to these differences. However, these analyses explored this hypothesis in the context of specifically non-NHW Multiracial populations, exact data on combinations of race which were not collected by the YRBSS, and among NHB populations. Future analyses should aim to describe polysubstance use patterns among subcategories of Multiracial identities. In the context of existing literature, substance use among different racial and ethnic backgrounds were found to be specific to the type of substances used.

Limitations

Study results should be interpreted while also considering the limitations. First, the current survey does not capture information on amount of substance used nor simultaneous use. Further, because electronic vapor products typically contain tobacco products, they are referred to as containing nicotine in the present study; however, if YRBSS participants replace the products with illicitly obtained cannabis or other substances, this was not measured by the YRBSS survey. Additionally, collapsing Asian, Pacific Islander, Native Hawaiian, and Native Alaskan into one category obscures patterns of health disparities among these individual populations, as echoed by previous studies [55]. Within these named groups, also including within the Hispanic, NHB, NHW, and Multiracial racial/ethnic groups, there are multitudes of unique cultural, societal, and familial differences that impact individual behavior [56]. Future studies should examine polysubstance use prevalence with more of these mentioned nuances considered, as we were limited to data collected by the YRBSS. Social and environmental factors not accounted for in the present study may be influential, such as family income, community type, discrimination, access to care, or socioeconomic instability. Further, evidence shows that earlier age of initiation and frequency of substance use is associated with sustained use of the substance/lifetime use of a variety of substances [24, 57, 58]. Future studies should account for age of initiation as this was not measured within the YRBSS.

As this was a cross-sectional data set, temporality cannot be established. Past studies have concluded that depressive symptoms predicted increased likelihood of initiation of substance use, but other studies have found a bi-directional association in which early onset of use predicted depressive symptoms and in which substance use predicted bullying victimization [23], while other studies support negative mental health being a predictor of substance use [20, 59]. As demonstrated, substance use, violence, poor mental health, and bullying victimization are closely linked with one another, and causality is difficult to tease out.

Studies have found mediating relationships between these characteristics. While our current study only measures depressive symptoms as a proxy for poor mental health, depressive symptoms have been found to be a risk factor for depression as a disorder [60]. Past research has found depression to act as a mediator between bullying victimization and substance use. Reed et al. used YRBSS data and found that, for girls, the association between victimization and substance use was mediated by depression, however, after controlling for depression, there was no significant direct effect of victimization on substance use [30]. Additionally, Luk et al. found that girls experiencing more depression may be at elevated risk for other harmful experiences such as victimization [61]. In the context of violent and risky behaviors (i.e., carrying a weapon), the association between substance use and gun activity reduced or became nonsignificant after adjusting for mental illness [62]. Furthermore, Luk et al. found that the relationship between bullying victimization and addictive behavior was fully explained by the presence of externalizing and internalizing problems (i.e., anxiety and depressive symptoms) using cross-sectional data [61], while substance use was found to be a mediator between cyber-bullying victimization, violence, and depressive symptomology [30]. Cross-sectional data is a significant limitation in exploring mediation and causality, therefore, future studies utilizing longitudinal data should continue to explore these complex mediating relationships.

Conclusion

Almost 20% of US middle and high school adolescents are engaging in concurrent alcohol and cannabis use, while rates of polysubstance use appear to be high for adolescents who identify as Multiracial. The prevalence of frequent co-use is high among those experiencing depressive symptoms, those carrying a weapon, and those experiencing bullying victimization. Given these epidemiological findings, it is imperative to acknowledge and address these multiple intersecting public health issues in intervention and prevention efforts. In addition, future research should work to understand if and how the Multiracial identity development process and specific environmental stressors experienced by Multiracially-identified adolescents (i.e., identity incongruent discrimination) may contribute to increased risk of using multiple substances [63].

Data Availability

Data is publicly available at https://www.cdc.gov/healthyyouth/data/yrbs/data.htm.