Video Abstract

Video Abstract

Close modal
OBJECTIVES

This study explores the changing prevalence of adolescent handgun carriage, with attention to differences across sociodemographic groups.

METHODS

Data were drawn from repeated cross-sectional, nationally representative surveys conducted annually from 2002 to 2019, the National Survey on Drug Use & Health. The study sample included adolescents aged 12 to 17 (N = 297 055). Logistic regression models estimated the prevalence of past year handgun carriage across cohort and sociodemographic subgroups. Interactions between 4-time cohorts and other variables explored sociodemographic variability in prevalence rates over time.

RESULTS

Handgun carriage increased significantly, particularly among rural, White, and higher-income adolescents. Carriage increased by 41% over cohorts, with predicted prevalence rates increasing from 3.3% in 2002–2006 to 4.6% in 2015–2019. Across cohorts, rural (5.1%), American Indian/Alaskan Native (5.2%), lower-income (<$20 000; 3.9%), male (5.9%), and older (16–17 years old; 4.5%) adolescents were the most likely to report carriage. However, these patterns changed significantly over time, with White and higher-income adolescents (>$75 000) most likely to carry in the most recent cohorts. Predicted carriage rates increased from 3.1% to 5.3% among White adolescents, from 2.6% to 5.1% among higher-income adolescents, and from 4.3% to 6.9% among rural adolescents between the 2002–2006 and 2015–2019 cohorts. Carriage among Black, American Indian/Alaskan Native, and lower-income adolescents decreased.

CONCLUSIONS

Adolescent handgun carriage is increasing, concentrated among particular subgroups of youth, and carriage patterns across sociodemographic groups have changed over time. Programs to address the risk of adolescent gun carriage should be tailored to the specific sociocultural and place-based concerns of diverse adolescents.

What’s Known on the Subject:

Adolescent handgun carriage, a major public health concern, varies across contexts and time and is influenced by both individual and sociodemographic factors.

What This Study Adds:

This study finds a 41% increase in rates of handgun carriage among adolescents from 2002 to 2019. Particularly notable increases emerged among White, rural, and higher-income adolescents, whereas gun carriage rates decreased among Black, American Indian/Alaskan Native, and lower-income adolescents.

Gun carriage by children and adolescents is a pressing public health concern, increasing the risk of firearm-related death or injury for both adolescents and others in their social networks.13  Firearm injuries are the second-leading cause of death among children and adolescents,3  and rates of adolescent firearm-related injuries or deaths have increased in recent years.4,5  Exposure to firearm injuries also has long-term developmental implications for youth, increasing rates of injury6  and the likelihood of engaging in firearm crime.7  Handguns, compared with other firearms, are more likely to be used in homicides,8  and handgun ownership, independent of ownership of other guns, is associated with an elevated risk of suicide.9,10 

Although earlier research on handgun carriage focused primarily on individual-level risk factors,3,4,11  recent research on carriage and exposure to firearm violence has drawn attention to the importance of sociodemographic differences in carriage patterns, particularly those linked to differences in neighborhood or historical contexts and place-based norms around carriage.4,12  Homicide risk has been associated with neighborhood residential segregation13  and income disparities,14  for example, and Southern and Midwestern demographic groups have been found to hold more positive norms around gun carriage.4  Adolescent gun carriage is linked to peer3,12,15  and family16,17  norms around carriage, and research has found substantial heterogeneity in injury risk across different states and regions within the United States.4,5,18  Risk of unintentional or self-inflicted gun injuries among youth, for example, are higher in the South and in rural emergency departments.18 

Although research has drawn attention to sociodemographic differences, little previous work has explored whether national carriage patterns have changed over time or remained stable for different groups. Vaughn and colleagues found that handgun carriage among White youth increased between 2002 and 2013,19  and that carriage among girls increased from 2002 to 2015, with these increases driven by White girls.20  However, recent research on carriage trends among both boys and girls of different racial or ethnic groups, or research focusing on multiyear trends across income groups or rural and urban contexts, is lacking. Understanding such changes is important for several reasons. First, changing patterns in carriage may expand insights into changing patterns in violence, including rising rates of adolescent and rural suicide.5,18  Second, understanding changing patterns of gun carriage may help identify which adolescents are at rising risk of injury, and provide a baseline for examining how risk or protective factors for carriage may vary across ecological levels.21  Finally, changing patterns of gun carriage necessitate the need to reevaluate lessons learned from older research (eg, data drawn from the National Longitudinal Survey of Youth or Add Health), which relied on behaviors of individuals who were adolescents in the early 2000s or 1990s, and may be less applicable to more sociodemographically diverse adolescents today.22 

This study seeks to address gaps in the literature by exploring long-term changes in gun carriage patterns using repeated cross-sectional, nationally representative samples of adolescents. In light of recent national increases in firearm mortality for suicide and homicide,5  differences in firearm mortality across sex and urbanicity,5  previous research documenting increases in carriage among White youth before 2015,19  and research showing that carriage patterns are associated with contextual factors that may change over time,3  we hypothesize that carriage patterns will have changed differentially over time for diverse subgroups of adolescents, with increases found particularly among White, rural, and female adolescents.

We drew data from annual cross-sectional samples of adolescents aged 12 to 17 from the 2002–2019 National Survey of Drug Use and Health (NSDUH) surveys, N = 297 055. The NSDUH completes annual, nationally representative cross-sectional surveys of the US civilian, noninstitutionalized population, aged 12 or older. In contrast to some other national data sets that evaluate adolescent carriage, such as the Youth Risk Behavior Survey (YRBS), the NSDUH is designed to be representative of rural areas, and includes adolescents who are not currently in school.23  Our university institutional review board reviewed this study and considered it exempt.

As part of a sequence of questions involving delinquent activity, adolescents were asked, “During the past 12 months, how many times have you carried a handgun?” Potential responses were 0, 1 to 2 times, 3 to 5 times, 6 to 9 times, and 10 or more times. We recoded this variable in 2 ways: first, as a binary variable to reflect any versus no carriage in the past year; and second, as a count variable to indicate the frequency of carriage (0, 2, 4, 8, or 10).

To assess changing patterns over cohorts and to increase the precision of estimates, surveys were clustered into 4 cohorts: 2002 to 2006, 2007 to 2010, 2011 to 2014, and 2015 to 2019, reflecting trends in the annual data (Supplemental Fig 6).

Urbanicity was reported by NSDUH, which matched respondent addresses to the county’s United States Department of Agriculture Rural Urban Continuum Code (RUCC). Counties were coded as larger metro (county with an urban population of >1 million people), smaller metro (county with an urban population between 20 000 and 1 million), or rural (no or <20 000 urban population). NSDUH used the 2003 RUCC categories up to 2014, and 2013 RUCC categories starting in 2015.

Adolescents reported their race and ethnicity, which were combined into 6 categories: White (not Hispanic), Black (not Hispanic), Hispanic, Asian/Pacific Islander (not Hispanic), American Indian/Alaskan Native (AIAN) (not Hispanic), or multiracial (not Hispanic); as well as sex and age. Annual household income was reported by an adult resident of the household (or by the young person if no adult was available), coded categorically as <$20 000, $20 000 to $49 999, $50 000 to $74 999, or $75 000 and above.24  NSDUH staff impute missing income, race/ethnicity, and age data before the release of the publicly-available data file using procedures developed specifically for the NSDUH.23 

Adolescents missing data on gun carriage (n = 1475; 0.5% of potential sample) were excluded from the analytic sample. Weighted descriptives for the sample are listed in Table 1.

TABLE 1

Weighted Characteristics of 12- to 17-Year-Olds from NSDUH, N = 297 055

nPercent of SamplePercent Reporting Carriage
Gun carriage in past 12 mo 10 968 3.7 3.7 
Cohort    
 2002–2006 83 361 28.1 3.4 
 2007–2010 65 624 22.1 3.2 
 2011–2014 65 954 22.2 3.5 
 2015–2019 82 115 27.6 4.5 
Urbanicity    
 Larger metro 162 123 54.6 3.1 
 Smaller metro 88 539 29.8 4.0 
 Rural 46 394 15.6 5.3 
Race or ethnicity    
 White 169 314 57.0 4.0 
 Black 42 717 14.4 3.4 
 AIAN 1841 0.6 6.4 
 Asian American/Pacific Islander 14 813 5.0 1.7 
 More than 1 race 7284 2.5 4.3 
 Hispanic 61 087 20.6 3.5 
Annual household income, $    
 >75 000 105 941 35.7 3.6 
 50 000–74 999 50 268 16.9 3.8 
 20 000–49 999 91 071 30.7 3.7 
 <20 000 49 775 16.8 3.8 
Sex    
 Female 145 726 49.1 1.5 
 Male 151 329 51.0 5.9 
Age, y    
 12–13 95 246 32.1 2.8 
 14–15 101 246 34.1 3.8 
 16–17 100 563 33.9 4.5 
nPercent of SamplePercent Reporting Carriage
Gun carriage in past 12 mo 10 968 3.7 3.7 
Cohort    
 2002–2006 83 361 28.1 3.4 
 2007–2010 65 624 22.1 3.2 
 2011–2014 65 954 22.2 3.5 
 2015–2019 82 115 27.6 4.5 
Urbanicity    
 Larger metro 162 123 54.6 3.1 
 Smaller metro 88 539 29.8 4.0 
 Rural 46 394 15.6 5.3 
Race or ethnicity    
 White 169 314 57.0 4.0 
 Black 42 717 14.4 3.4 
 AIAN 1841 0.6 6.4 
 Asian American/Pacific Islander 14 813 5.0 1.7 
 More than 1 race 7284 2.5 4.3 
 Hispanic 61 087 20.6 3.5 
Annual household income, $    
 >75 000 105 941 35.7 3.6 
 50 000–74 999 50 268 16.9 3.8 
 20 000–49 999 91 071 30.7 3.7 
 <20 000 49 775 16.8 3.8 
Sex    
 Female 145 726 49.1 1.5 
 Male 151 329 51.0 5.9 
Age, y    
 12–13 95 246 32.1 2.8 
 14–15 101 246 34.1 3.8 
 16–17 100 563 33.9 4.5 

Trends in gun carriage over time and across demographic strata were assessed using multivariable logistic regression models analyzed in Stata 16. Because only 3.7% of the sample reported any previous year carriage, our primary models were estimated using a binary measure of carriage. The first model assessed the main effects of cohort and sociodemographic variables. The second set of models included interactions between cohort and sociodemographic variables added to assess whether variation in gun carriage has shifted over time. Because the social constructs of race and ethnicity are linked to a multitude of contextual factors related to potential gun carriage in the United States,4  we were particularly interested in how the intersection of race and ethnicity with the contexts of family income and urbanicity might shape youth gun carriage. Therefore, a third set of models estimated targeted 3-way interactions to explore the intersections of cohort and urbanicity with income and race/ethnicity. The significance of each set of interactions was assessed with Wald tests. As a secondary analysis, we estimated 0 inflated negative binomial regression models on the count of gun carriage to assess both the likelihood of any carriage and the frequency of carriage among carriers. All analyses were weighted using multiyear probability weights provided by NSDUH to provide nationally representative estimates. Predicted prevalence rates for subgroups were calculated by estimating predicted margins (average predicted probabilities).

For each set of results, we report predicted prevalence rates of carriage with 95% confidence intervals (see Supplemental Tables 4 and 5 for odd ratios and 95% confidence intervals). Previous research has often focused on male-only samples because males are significantly more likely to carry and are at highest risk of being victims or perpetrators of firearm violence.3,4  To provide comparative data, and because overall prevalence estimates may understate the prevalence of carriage among males, we also include predicted prevalence rates for males and females separately across cohort and sociodemographic groups.

Results of main effect models (Table 2) indicate that self-reported handgun carriage by adolescents increased significantly over cohorts, particularly since 2015. Adolescents in the 2015–2019 cohort were 40.8% more likely to report carriage than adolescents in 2002–2006 (predicted prevalence rising from 3.3% to 4.6%). Adolescents in rural and smaller metro areas reported higher levels of carriage than their counterparts in larger urban areas (5.1%, 3.9% vs 3.1%). Race and ethnicity differences emerged, as well: Black (3.5%), Asian American (1.9%), and Hispanic (3.5%) adolescents were less likely to report carriage than White adolescents (3.9%), whereas AIAN adolescents (5.2%) were more likely to carry than White adolescents. Rates of carriage were higher among males (5.9%) than females (1.5%), and higher among older adolescents, with rates of 4.5% and 3.8% among 16- to 17-year-olds and 14- to 15-year-olds, respectively, versus 2.8% among 12- to 13-year-olds. Finally, small but statistically significant differences emerged across income strata. Gun carriage was highest among adolescents in the lowest-income families, with adjusted carriage rates of 3.9% among adolescents in families with annual incomes of <$20 000 in comparison with 3.5% among higher-income adolescents. Across all of these strata, rates of carriage were approximately 4 times higher among males versus females.

TABLE 2

Predicted Prevalence Rates of Carriage Overall and by Sex, Derived from Logistic

Regression ModelsOverall Predicted Prevalence (95% CI)Male Predicted Prevalence (95% CI)Female Predicted Prevalence (95% CI)
Cohort    
 2002–2006a 3.3 (3.2–3.5) 5.3 (5.0–5.5) 1.3 (1.2–1.4) 
 2007–2010 3.2 (3.0–3.4) 5.1 (4.8–5.4) 1.2 (1.2–1.3) 
 2011–2014 3.6 (3.4–3.7) 5.6 (5.4–5.9) 1.4 (1.3–1.5) 
 2015–2019b 4.6 (4.4–4.8b 7.3 (6.9–7.6) 1.8 (1.7–1.9) 
Urbanicity    
 Larger metroa 3.1 (3.0–3.3) 5.0 (4.8–5.2) 1.2 (1.1–1.3) 
 Smaller metro 3.9 (3.8–4.1)b 6.2 (6.0–6.5) 1.5 (1.4–1.6) 
 Rural 5.1 (4.8–5.3)b 8.0 (7.6–8.4) 2.0 (1.9–2.2) 
Race or ethnicity    
 Whitea 3.9 (3.8–4.0) 6.2 (6.0–6.4) 1.5 (1.4–1.6) 
 Black 3.5 (3.3–3.7)b 5.6 (5.2–5.9) 1.4 (1.3–1.5) 
 AIAN 5.2 (4.1–6.3)b 8.2 (6.5–10.0) 2.1 (1.6–2.6) 
 Asian American/Pacific Islander 1.9 (1.5–2.3)b 3.0 (2.4–3.6) 0.7 (0.6–0.9) 
 Multiracial 4.2 (3.6–4.7) 6.6 (5.7–7.4) 1.6 (1.4–1.9) 
 Hispanic 3.5 (3.3–3.8) 5.6 (5.2–6.0) 1.4 (1.3–1.5) 
Household income, $    
 >75 000a 3.5 (3.4–3.7) 5.6 (5.4–5.9) 1.4 (1.3–1.5) 
 50 000–74 999 3.7 (3.5–3.9) 5.9 (5.6–6.2) 1.5 (1.4–1.6) 
 20 000–49 999 3.7 (3.6–3.9) 5.9 (5.6–6.2) 1.5 (1.4–1.6) 
 <20 000b 3.9 (3.7–4.2)b 6.2 (5.8–6.6) 1.5 (1.4–1.7) 
Sex    
 Femalea 1.5 (1.4–1.5) — — 
 Maleb 5.9 (5.7–6.0)b — — 
Age, y    
 12–13a 2.8 (2.6–2.9) 4.4 (4.2–4.6) 1.1 (1.0–1.1) 
 14–15b 3.8 (3.7–4.0)b 6.0 (5.8–6.3) 1.5 (1.4–1.6) 
 16–17b 4.5 (4.3–4.6)b 7.1 (6.8–7.3) 1.8 (1.7–1.9) 
Regression ModelsOverall Predicted Prevalence (95% CI)Male Predicted Prevalence (95% CI)Female Predicted Prevalence (95% CI)
Cohort    
 2002–2006a 3.3 (3.2–3.5) 5.3 (5.0–5.5) 1.3 (1.2–1.4) 
 2007–2010 3.2 (3.0–3.4) 5.1 (4.8–5.4) 1.2 (1.2–1.3) 
 2011–2014 3.6 (3.4–3.7) 5.6 (5.4–5.9) 1.4 (1.3–1.5) 
 2015–2019b 4.6 (4.4–4.8b 7.3 (6.9–7.6) 1.8 (1.7–1.9) 
Urbanicity    
 Larger metroa 3.1 (3.0–3.3) 5.0 (4.8–5.2) 1.2 (1.1–1.3) 
 Smaller metro 3.9 (3.8–4.1)b 6.2 (6.0–6.5) 1.5 (1.4–1.6) 
 Rural 5.1 (4.8–5.3)b 8.0 (7.6–8.4) 2.0 (1.9–2.2) 
Race or ethnicity    
 Whitea 3.9 (3.8–4.0) 6.2 (6.0–6.4) 1.5 (1.4–1.6) 
 Black 3.5 (3.3–3.7)b 5.6 (5.2–5.9) 1.4 (1.3–1.5) 
 AIAN 5.2 (4.1–6.3)b 8.2 (6.5–10.0) 2.1 (1.6–2.6) 
 Asian American/Pacific Islander 1.9 (1.5–2.3)b 3.0 (2.4–3.6) 0.7 (0.6–0.9) 
 Multiracial 4.2 (3.6–4.7) 6.6 (5.7–7.4) 1.6 (1.4–1.9) 
 Hispanic 3.5 (3.3–3.8) 5.6 (5.2–6.0) 1.4 (1.3–1.5) 
Household income, $    
 >75 000a 3.5 (3.4–3.7) 5.6 (5.4–5.9) 1.4 (1.3–1.5) 
 50 000–74 999 3.7 (3.5–3.9) 5.9 (5.6–6.2) 1.5 (1.4–1.6) 
 20 000–49 999 3.7 (3.6–3.9) 5.9 (5.6–6.2) 1.5 (1.4–1.6) 
 <20 000b 3.9 (3.7–4.2)b 6.2 (5.8–6.6) 1.5 (1.4–1.7) 
Sex    
 Femalea 1.5 (1.4–1.5) — — 
 Maleb 5.9 (5.7–6.0)b — — 
Age, y    
 12–13a 2.8 (2.6–2.9) 4.4 (4.2–4.6) 1.1 (1.0–1.1) 
 14–15b 3.8 (3.7–4.0)b 6.0 (5.8–6.3) 1.5 (1.4–1.6) 
 16–17b 4.5 (4.3–4.6)b 7.1 (6.8–7.3) 1.8 (1.7–1.9) 

CI, confidence interval; —, not applicable.

a

Indicates referent group.

b

Indicates significant difference from referent at P < .05.

We next included 2-way interactions between each sociodemographic variable and cohort (Table 3 and Fig 1). There were significant interactions between cohort and all sociodemographic variables (all Wald Tests P < .05), suggesting that the overall sociodemographic differences mask significant changes in patterns among these groups over time. In relation to urbanicity, carriage rates increased more among rural adolescents, increasing from 4.3% in 2002–2006 to 6.9% in 2015–2019, than among their large metro peers (2.9%–3.8%). Gun carriage also increased more among White than among Black, Hispanic, or AIAN adolescents. Predicted carriage by White adolescents increased significantly from 3.1% to 5.3% across cohorts, whereas carriage among Black (4.0%–3.2%) and AIAN adolescents (6.8%–4.4%) decreased and carriage among Hispanic adolescents (3.9%–4.1%) increased more slowly during this period.

FIGURE 1

Adjusted predicted rates of past-year handgun carriage for sociodemographic strata interacted with cohort, including urbanicity (A), race/ethnicity (B), household income (C), sex (D), and age (E).

FIGURE 1

Adjusted predicted rates of past-year handgun carriage for sociodemographic strata interacted with cohort, including urbanicity (A), race/ethnicity (B), household income (C), sex (D), and age (E).

Close modal
TABLE 3

Predicted Prevalence Rates of Carriage for Sociodemographic X Cohort Interactions, Overall and by Sex, Derived from Logistic Regression Models

Overall Predicted Prevalence (95% CI)Male Predicted Prevalence (95% CI)Female Predicted Prevalence (95% CI)
Urbanicity X cohort*    
 Larger metro    
  2002–2006a 2.9 (2.7–3.1) 4.7 (4.3–5.0) 1.1 (1.0–1.2) 
  2007–2010a 2.7 (2.5–3.0) 4.4 (4.0–4.7) 1.1 (1.0–1.2) 
  2011–2014a 3.0 (2.8–3.3) 4.8 (4.4–5.2) 1.2 (1.1–1.3) 
  2015–2019a 3.8 (3.5–4.1) 6.0 (5.6–6.5) 1.5 (1.4–1.6) 
 Smaller Metro    
  2002–2006a 3.5 (3.2–3.8) 5.6 (5.1–6.0) 1.4 (1.2–1.5) 
  2007–2010 3.5 (3.2–3.9) 5.6 (5.1–6.1) 1.4 (1.2–1.5) 
  2011–2014 3.7 (3.4–4.0) 5.9 (5.4–6.4) 1.5 (1.3–1.6) 
  2015–2019 4.8 (4.4–5.1) 7.6 (7.0–8.1) 1.9 (1.7–2.1) 
 Rural    
  2002–2006a 4.3 (3.9–4.6) 6.8 (6.2–7.4) 1.7 (1.5–1.9) 
  2007–2010 4.1 (3.6–4.5) 6.4 (5.7–7.1) 1.6 (1.4–1.8) 
  2011–2014 5.1 (4.6–5.5) 8.0 (7.2–8.7) 2.0 (1.8–2.2) 
  2015–2019b 6.9 (6.3–7.4) 10.8 (9.9–11.7) 2.8 (2.5–3.1) 
Race or ethnicity X cohort***    
 White    
  2002–2006a 3.1 (2.9–3.3) 4.9 (4.6–5.2) 1.2 (1.1–1.3) 
  2007–2010a 3.2 (3.0–3.4) 5.1 (4.8–5.5) 1.3 (1.2–1.4) 
  2011–2014a 3.9 (3.7–4.2) 6.2 (5.8–6.6) 1.5 (1.4–1.7) 
  2015–2019a 5.3 (5.0–5.6) 8.4 (7.9–8.8) 2.1 (2.0–2.3) 
 Black    
  2002–2006a 4.0 (3.6–4.4) 6.4 (5.7–7.0) 1.6 (1.4–1.8) 
  2007–2010 3.5 (3.0–3.9) 5.5 (4.8–6.3) 1.4 (1.2–1.6) 
  2011–2014b 3.0 (2.6–3.5) 4.8 (4.2–5.5) 1.2 (1.0–1.4) 
  2015–2019b 3.2 (2.8–3.7) 5.2 (4.5–5.8) 1.3 (1.1–1.5) 
 AIAN    
  2002–2006a 6.8 (4.2–9.5) 10.8 (6.7–14.8) 2.8 (1.6–3.9) 
  2007–2010b 3.2 (2.0–4.4) 5.1 (3.3–7.0) 1.3 (0.8–1.7) 
  2011–2014 6.0 (3.3–8.6) 9.4 (5.3–13.5) 2.4 (1.3–3.5) 
  2015–2019b 4.4 (2.7–6.2) 7.0 (4.2–9.8) 1.8 (1.0–2.5) 
 Asian American/Pacific Islander    
  2002–2006a 1.9 (1.0–2.8) 3.0 (1.5–4.4) 0.7 (0.4–1.1) 
  2007–2010 1.9 (1.0–2.9) 3.1 (1.6–4.6) 0.7 (0.4–1.1) 
  2011–2014 1.4 (0.8–1.9) 2.2 (1.2–3.1) 0.5 (0.3–0.7) 
  2015–2019 2.3 (1.6–3.0) 3.7 (2.6–4.8) 0.9 (0.6–1.2) 
 Multiracial    
  2002–2006a 3.3 (2.4–4.2) 5.2 (3.8–6.7) 1.3 (0.9–1.7) 
  2007–2010 3.8 (2.5–5.1) 6.1 (4.0–8.1) 1.5 (1.0–2.0) 
  2011–2014 4.4 (3.3–5.6) 7.0 (5.2–8.9) 1.8 (1.3–2.3) 
  2015–2019 5.0 (4.0–5.9) 7.8 (6.4–9.3) 2.0 (1.6–2.4) 
 Hispanic    
  2002–2006a 3.9 (3.4–4.3) 6.1 (5.4–6.8) 1.5 (1.3–1.7) 
  2007–2010b 3.1 (2.7–3.5) 5.0 (4.3–5.6) 1.2 (1.0–1.4) 
  2011–2014b 3.2 (2.7–3.6) 5.0 (4.3–5.7) 1.2 (1.0–1.4) 
  2015–2019b 4.1 (3.6–4.5) 6.4 (5.7–7.2) 1.6 (1.4–1.8) 
Income X cohort***    
 >$75 000    
  2002–2006a 2.6 (2.4–2.9) 4.2 (3.8–4.6) 1.0 (0.9–1.1) 
  2007–2010a 2.7 (2.4–3.0) 4.3 (3.8–4.7) 1.0 (0.9–1.2) 
  2011–2014a 3.4 (3.1–3.7) 5.4 (4.9–5.8) 1.3 (1.2–1.5) 
  2015–2019a 5.1 (4.7–5.4) 8.0 (7.4–8.5) 2.0 (1.8–2.2) 
 $50 000–$74 999    
  2002–2006a 3.0 (2.7–3.3) 4.8 (4.3–5.3) 1.2 (1.0–1.3) 
  2007–2010 3.1 (2.7–3.5) 4.9 (4.3–5.6) 1.2 (1.0–1.4) 
  2011–2014 4.2 (3.7–4.7) 6.7 (5.9–7.4) 1.7 (1.4–1.9) 
  2015–2019b 4.7 (4.2–5.2) 7.4 (6.6–8.2) 1.9 (1.6–2.1) 
 $20 000–$49 999    
  2002–2006a 3.6 (3.3–3.9) 5.7 (5.3–6.2) 1.4 (1.3–1.5) 
  2007–2010 3.5 (3.2–3.8) 5.5 (5.0–6.0) 1.4 (1.2–1.5) 
  2011–2014b 3.4 (3.1–3.7) 5.4 (4.9–5.9) 1.3 (1.2–1.5) 
  2015–2019b 4.2 (3.8–4.6) 6.7 (6.1–7.3) 1.7 (1.5–1.8) 
 <$20 000    
  2002–2006a 4.3 (3.9–4.8) 6.9 (6.2–7.5) 1.7 (1.5–1.9) 
  2007–2010 3.8 (3.4–4.3) 6.1 (5.3–6.8) 1.5 (1.3–1.7) 
  2011–2014b 3.6 (3.1–4.0) 5.6 (4.9–6.4) 1.4 (1.2–1.6) 
  2015–2019b 3.7 (3.2–4.2) 5.9 (5.1–6.7) 1.5 (1.2–1.7) 
Sex X cohort***    
 Female    
  2002–2006a 1.1 (0.9–1.2) — — 
  2007–2010a 1.1 (0.9–1.2) — — 
  2011–2014a 1.4 (1.3–1.6) — — 
  2015–2019a 2.2 (2.0–2.4) — — 
 Male    
  2002–2006a 5.5 (5.2–5.8) — — 
   2007–2010 5.2 (4.9–5.6) — — 
   2011–2014b 5.6 (5.3–5.9) — — 
   2015–2019b 6.9 (6.6–7.3) — — 
Age X cohort**    
  12–13 y    
   2002–2006a 2.4 (2.1–2.6) 3.7 (3.4–4.1) 0.9 (0.8–1.0) 
   2007–2010a 2.1 (1.9–2.3) 3.4 (3.0–3.7) 0.8 (0.7–0.9) 
   2011–2014a 3.0 (2.7–3.3) 4.7 (4.3–5.2) 1.2 (1.0–1.3) 
   2015–2019a 3.6 (3.2–3.9) 5.6 (5.1–6.2) 1.4 (1.2–1.5) 
  14–15 y    
   2002–2006a 3.4 (3.2–3.7) 5.4 (5.0–5.9) 1.3 (1.2–1.5) 
   2007–2010 3.6 (3.2–3.9) 5.7 (5.1–6.2) 1.4 (1.2–1.5) 
   2011–2014 3.7 (3.3–4.0) 5.8 (5.3–6.3) 1.4 (1.3–1.6) 
   2015–2019 4.5 (4.2–4.9) 7.2 (6.6–7.7) 1.8 (1.6–2.0) 
  16–17 y    
   2002–2006a 4.2 (3.9–4.4) 6.6 (6.1–7.0) 1.6 (1.5–1.8) 
   2007–2010 3.9 (3.6–4.2) 6.1 (5.6–6.6) 1.5 (1.4–1.7) 
   2011–2014b 4.0 (3.7–4.4) 6.4 (5.8–6.9) 1.6 (1.4–1.7) 
   2015–2019 5.6 (5.2–6.0) 8.9 (8.3–9.5) 2.3 (2.1–2.5) 
Overall Predicted Prevalence (95% CI)Male Predicted Prevalence (95% CI)Female Predicted Prevalence (95% CI)
Urbanicity X cohort*    
 Larger metro    
  2002–2006a 2.9 (2.7–3.1) 4.7 (4.3–5.0) 1.1 (1.0–1.2) 
  2007–2010a 2.7 (2.5–3.0) 4.4 (4.0–4.7) 1.1 (1.0–1.2) 
  2011–2014a 3.0 (2.8–3.3) 4.8 (4.4–5.2) 1.2 (1.1–1.3) 
  2015–2019a 3.8 (3.5–4.1) 6.0 (5.6–6.5) 1.5 (1.4–1.6) 
 Smaller Metro    
  2002–2006a 3.5 (3.2–3.8) 5.6 (5.1–6.0) 1.4 (1.2–1.5) 
  2007–2010 3.5 (3.2–3.9) 5.6 (5.1–6.1) 1.4 (1.2–1.5) 
  2011–2014 3.7 (3.4–4.0) 5.9 (5.4–6.4) 1.5 (1.3–1.6) 
  2015–2019 4.8 (4.4–5.1) 7.6 (7.0–8.1) 1.9 (1.7–2.1) 
 Rural    
  2002–2006a 4.3 (3.9–4.6) 6.8 (6.2–7.4) 1.7 (1.5–1.9) 
  2007–2010 4.1 (3.6–4.5) 6.4 (5.7–7.1) 1.6 (1.4–1.8) 
  2011–2014 5.1 (4.6–5.5) 8.0 (7.2–8.7) 2.0 (1.8–2.2) 
  2015–2019b 6.9 (6.3–7.4) 10.8 (9.9–11.7) 2.8 (2.5–3.1) 
Race or ethnicity X cohort***    
 White    
  2002–2006a 3.1 (2.9–3.3) 4.9 (4.6–5.2) 1.2 (1.1–1.3) 
  2007–2010a 3.2 (3.0–3.4) 5.1 (4.8–5.5) 1.3 (1.2–1.4) 
  2011–2014a 3.9 (3.7–4.2) 6.2 (5.8–6.6) 1.5 (1.4–1.7) 
  2015–2019a 5.3 (5.0–5.6) 8.4 (7.9–8.8) 2.1 (2.0–2.3) 
 Black    
  2002–2006a 4.0 (3.6–4.4) 6.4 (5.7–7.0) 1.6 (1.4–1.8) 
  2007–2010 3.5 (3.0–3.9) 5.5 (4.8–6.3) 1.4 (1.2–1.6) 
  2011–2014b 3.0 (2.6–3.5) 4.8 (4.2–5.5) 1.2 (1.0–1.4) 
  2015–2019b 3.2 (2.8–3.7) 5.2 (4.5–5.8) 1.3 (1.1–1.5) 
 AIAN    
  2002–2006a 6.8 (4.2–9.5) 10.8 (6.7–14.8) 2.8 (1.6–3.9) 
  2007–2010b 3.2 (2.0–4.4) 5.1 (3.3–7.0) 1.3 (0.8–1.7) 
  2011–2014 6.0 (3.3–8.6) 9.4 (5.3–13.5) 2.4 (1.3–3.5) 
  2015–2019b 4.4 (2.7–6.2) 7.0 (4.2–9.8) 1.8 (1.0–2.5) 
 Asian American/Pacific Islander    
  2002–2006a 1.9 (1.0–2.8) 3.0 (1.5–4.4) 0.7 (0.4–1.1) 
  2007–2010 1.9 (1.0–2.9) 3.1 (1.6–4.6) 0.7 (0.4–1.1) 
  2011–2014 1.4 (0.8–1.9) 2.2 (1.2–3.1) 0.5 (0.3–0.7) 
  2015–2019 2.3 (1.6–3.0) 3.7 (2.6–4.8) 0.9 (0.6–1.2) 
 Multiracial    
  2002–2006a 3.3 (2.4–4.2) 5.2 (3.8–6.7) 1.3 (0.9–1.7) 
  2007–2010 3.8 (2.5–5.1) 6.1 (4.0–8.1) 1.5 (1.0–2.0) 
  2011–2014 4.4 (3.3–5.6) 7.0 (5.2–8.9) 1.8 (1.3–2.3) 
  2015–2019 5.0 (4.0–5.9) 7.8 (6.4–9.3) 2.0 (1.6–2.4) 
 Hispanic    
  2002–2006a 3.9 (3.4–4.3) 6.1 (5.4–6.8) 1.5 (1.3–1.7) 
  2007–2010b 3.1 (2.7–3.5) 5.0 (4.3–5.6) 1.2 (1.0–1.4) 
  2011–2014b 3.2 (2.7–3.6) 5.0 (4.3–5.7) 1.2 (1.0–1.4) 
  2015–2019b 4.1 (3.6–4.5) 6.4 (5.7–7.2) 1.6 (1.4–1.8) 
Income X cohort***    
 >$75 000    
  2002–2006a 2.6 (2.4–2.9) 4.2 (3.8–4.6) 1.0 (0.9–1.1) 
  2007–2010a 2.7 (2.4–3.0) 4.3 (3.8–4.7) 1.0 (0.9–1.2) 
  2011–2014a 3.4 (3.1–3.7) 5.4 (4.9–5.8) 1.3 (1.2–1.5) 
  2015–2019a 5.1 (4.7–5.4) 8.0 (7.4–8.5) 2.0 (1.8–2.2) 
 $50 000–$74 999    
  2002–2006a 3.0 (2.7–3.3) 4.8 (4.3–5.3) 1.2 (1.0–1.3) 
  2007–2010 3.1 (2.7–3.5) 4.9 (4.3–5.6) 1.2 (1.0–1.4) 
  2011–2014 4.2 (3.7–4.7) 6.7 (5.9–7.4) 1.7 (1.4–1.9) 
  2015–2019b 4.7 (4.2–5.2) 7.4 (6.6–8.2) 1.9 (1.6–2.1) 
 $20 000–$49 999    
  2002–2006a 3.6 (3.3–3.9) 5.7 (5.3–6.2) 1.4 (1.3–1.5) 
  2007–2010 3.5 (3.2–3.8) 5.5 (5.0–6.0) 1.4 (1.2–1.5) 
  2011–2014b 3.4 (3.1–3.7) 5.4 (4.9–5.9) 1.3 (1.2–1.5) 
  2015–2019b 4.2 (3.8–4.6) 6.7 (6.1–7.3) 1.7 (1.5–1.8) 
 <$20 000    
  2002–2006a 4.3 (3.9–4.8) 6.9 (6.2–7.5) 1.7 (1.5–1.9) 
  2007–2010 3.8 (3.4–4.3) 6.1 (5.3–6.8) 1.5 (1.3–1.7) 
  2011–2014b 3.6 (3.1–4.0) 5.6 (4.9–6.4) 1.4 (1.2–1.6) 
  2015–2019b 3.7 (3.2–4.2) 5.9 (5.1–6.7) 1.5 (1.2–1.7) 
Sex X cohort***    
 Female    
  2002–2006a 1.1 (0.9–1.2) — — 
  2007–2010a 1.1 (0.9–1.2) — — 
  2011–2014a 1.4 (1.3–1.6) — — 
  2015–2019a 2.2 (2.0–2.4) — — 
 Male    
  2002–2006a 5.5 (5.2–5.8) — — 
   2007–2010 5.2 (4.9–5.6) — — 
   2011–2014b 5.6 (5.3–5.9) — — 
   2015–2019b 6.9 (6.6–7.3) — — 
Age X cohort**    
  12–13 y    
   2002–2006a 2.4 (2.1–2.6) 3.7 (3.4–4.1) 0.9 (0.8–1.0) 
   2007–2010a 2.1 (1.9–2.3) 3.4 (3.0–3.7) 0.8 (0.7–0.9) 
   2011–2014a 3.0 (2.7–3.3) 4.7 (4.3–5.2) 1.2 (1.0–1.3) 
   2015–2019a 3.6 (3.2–3.9) 5.6 (5.1–6.2) 1.4 (1.2–1.5) 
  14–15 y    
   2002–2006a 3.4 (3.2–3.7) 5.4 (5.0–5.9) 1.3 (1.2–1.5) 
   2007–2010 3.6 (3.2–3.9) 5.7 (5.1–6.2) 1.4 (1.2–1.5) 
   2011–2014 3.7 (3.3–4.0) 5.8 (5.3–6.3) 1.4 (1.3–1.6) 
   2015–2019 4.5 (4.2–4.9) 7.2 (6.6–7.7) 1.8 (1.6–2.0) 
  16–17 y    
   2002–2006a 4.2 (3.9–4.4) 6.6 (6.1–7.0) 1.6 (1.5–1.8) 
   2007–2010 3.9 (3.6–4.2) 6.1 (5.6–6.6) 1.5 (1.4–1.7) 
   2011–2014b 4.0 (3.7–4.4) 6.4 (5.8–6.9) 1.6 (1.4–1.7) 
   2015–2019 5.6 (5.2–6.0) 8.9 (8.3–9.5) 2.3 (2.1–2.5) 

Each panel presents results from a separate model. All models include main effects of cohort, urbanicity, race or ethnicity, income, sex, and age. CI, confidence interval; —, not applicable.

Asterisks indicate overall significance for each set of interactions at

*

P < .05,

**

P < .01,

***

P < .001.

a

Indicates referent group.

b

Indicates significant difference from referent at P < .05.

Cohort by income interactions showed diverging patterns, as well. Among adolescents in the highest-income group, predicted carriage nearly doubled between the first and last cohort, increasing from 2.6% to 5.1%, whereas carriage for the lowest-income group declined (4.3%–3.7%) and rates among middle-income groups rose more slowly. Interactions with sex found that, although males continued to report higher gun carriage than females, carriage doubled among females (1.1%–2.2%) and increased more slowly among boys (5.5%–6.9%). A similar pattern emerged in relation to age, with gun carriage increasing more strongly among the youngest adolescents, aged 12 to 13 (2.4%–3.6%) than among their peers aged 16 to 17 (4.2%–5.6%). For males, all predicted prevalence rates are substantially higher; in the most recent cohort, for example, estimated rate of carriage among rural males was 10.8%, with predicted rates among White males at 8.4%, among 16- to 17-year-old males at 8.9%, and among the highest-income males at 8.0%, whereas predicted rates among females were consistently <3%, with the highest rate of 2.8% among rural females in the most recent cohorts.

To explore whether the increases in rural areas or higher-income groups were driven by particular sociodemographic groups, we estimated additional models, including targeted 3-way interactions for both income and race/ethnicity groups by cohort and urbanicity. Wald tests for these sets of 3-way interactions were nonsignificant (results not shown).

Although a small proportion of adolescents (3.7%) reported any gun carriage, there was variability in the frequency of carriage among carriers. To assess this variability, we estimated 0-inflated negative binomial models. Results, presented in Supplemental Table 1, replicate the logistic model results in relation to the inflate portion of the 0-inflated models. However, these results also indicate that, among youth reporting any handgun carriage, the frequency of carriage was lower in 2007–2010 vs 2002–2006. Frequency of carriage was higher among rural and small metro versus large metro areas, and among male versus female youth. Surprisingly, among youth reporting any carriage, 12- to 13-year-olds reported more frequent carriage than their older peers.

In approximately the first two decades of the 21st century from 2002 to 2019, the number of adolescents reporting handgun carriage in the previous year increased by 41%. Increases were most pronounced among White, higher-income, and rural adolescents, with predicted carriage rates nearly doubling among adolescents in the highest-income group. In contrast, predicted rates of carriage decreased among Black, AIAN, and the lowest-income adolescents. Although boys remained significantly more likely to carry than girls, predicted carriage doubled for girls during this time period, a finding consistent with research on earlier time periods.20  The prevalence changes among White and higher-income adolescents were consistent across context, with these adolescents increasingly likely to report carriage across rural, smaller, and larger metropolitan contexts. Notably, rates reflect only past year carriage; other research has found that cumulative rates of carriage are often higher.3,12 

Our findings extend previous research using the NSDUH, which documented increasing carriage among White youth and girls,19,20  and is consistent with research finding higher-reported access to firearms among White and rural youth,15  elevated or increasing firearm-related suicide rates among White1,4,25  and rural4,5,18  adolescents, and higher rates of unintentional pediatric firearm injuries in rural areas.18  These new patterns indicate the importance of extending public health interventions to rural, White, and higher-income populations of youth.

Our findings with regard to overall increases in handgun carriage are at odds with findings by Timsina and colleagues1  using YRBS data, which found consistent rates of adolescent carriage from 1993 to 2017. This discrepancy may be related to differences in the samples used. As noted above, the NSDUH includes youth who are not in school, and is designed to be representative for rural areas. In contrast, the national YRBS only sampled 9th to 12th graders in schools, and because the YRBS does not allow for direct examination of urbanicity or family income differences at the respondent level, Timsina and colleagues explored only state-level urbanicity and median income differences. Notably, they found that, for each 1% increase in a state’s percentage of rural population, there was an accompanying increase in the risk of adolescent report of gun carriage. Finally, the YRBS questions on firearm carriage did not distinguish among kinds of gun carriage, and potentially included hunting rifles or long guns, whereas the NSDUH has consistently asked about handguns only.

To our knowledge, no previous study has used nationally representative data to directly explore the changing prevalence of gun carriage for adolescents of both sexes living in different urbanicity and family income contexts. Our findings suggest that carriage among White, rural, and higher-income adolescents is increasing, but the causes of these increases are far from clear. Previous research on carriage has found heterogeneity in individual risk and protective factors among sexes and within females across rural and urban contexts,20  but has not directly investigated whether the significance of different factors has changed over time, particularly across subgroups. Increased handgun carriage could be driven by changes in underlying stressors leading to carriage, changing norms around carriage among subgroups, differences in the state or local gun policy landscape between urban and rural areas, unequal enforcement of existing gun laws, or other factors. Future research should explore the factors driving these increases and how closely they are linked to risk of injury, while giving due attention to the potential for shifting patterns over time.

Our study has several limitations. First, like much of the research on handgun carriage, our data rely on adolescent self-report of “carriage,” a term which may have different meaning for different adolescents and which is not well defined in the NSDUH survey. Accordingly, a report of carriage could potentially include everything from holding another person’s weapon once, to trips to a gun range with a parent, to carriage for the purposes of committing a crime. Because youth could report at maximum carrying “10 or more” times in the past year, we could not distinguish between those who carried 10 times and those who carried much more frequently, although only 0.53% of the sample reported 10 or more times. Additionally, the racial and ethnic subgroupings included here are panethnic and broad, encompassing an enormous range of experiences, cultures, regional groups, and histories. Because the NSDUH data are cross-sectional, we were able to map shifts across cohorts but not assess shifts in individual behavior over time. Finally, because the NSDUH’s publicly available data do not provide geographic identifiers, our models do not account for differences in state or local gun policies, regional differences, or school or neighborhood effects.

With the goal of decreasing early gun carriage and resulting public health sequelae, our findings underscore the need to develop intervention programs and policy solutions that are tailored to different subgroups of adolescents and which address underlying structural and sociocultural, as well as family and individual, determinants of carriage.4,26  Although rigorous research on effective primary firearm injury prevention is limited,26  programs responsive to particular community needs and cultures have had promising results, including, for example, programs focused on gun storage in rural Alaskan Native communities27  and in an urban county in North Carolina.28  Additional research exploring different individual and family correlates of gun carriage across subgroups and across time is also critical. Research by Vaughn and colleagues, for example, found that parental affirmation was significantly correlated with reduced carriage for White and Hispanic youth, but not for Black youth.19  Similarly, religiosity was associated with a reduction in carriage for urban girls, but not urban boys or rural youth.20  Given the increases in carriage among White, rural, and higher-income youth identified in this study, it is critical to understand which individual or community-level correlates might be related to these increases, and tailor programs and interventions to these needs. Recent research from Colorado, for example, found that adolescents in rural areas reported greater access to firearms,15  and pointed to the potential of public education strategies aimed at parents who are gun owners and implemented by trusted messengers. Pediatric health care providers in both emergency and primary care settings, who treat young people across these contexts, have a critical role to play in providing community-informed screening and education about firearm safety to both adolescents and their families.26 

Ms Carey and Dr Coley conceptualized and designed the study, drafted the initial manuscript, reviewed and revised the manuscript, approved the final manuscript as submitted, and agree to be accountable for all aspects of the work.

FUNDING: No external funding.

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

AIAN

American Indian and Alaskan Native

NSDUH

National Survey on Drug Use & Health

RUCC

Rural Urban Continuum Code

YRBS

Youth Risk Behavior Survey

1
Timsina
LR
,
Qiao
N
,
Mongalo
AC
,
Vetor
AN
,
Carroll
AE
,
Bell
TM
.
National instant criminal background check and youth gun carrying
.
Pediatrics
.
2020
;
145
(
1
):
e20191071
2
Tracy
M
,
Braga
AA
,
Papachristos
AV
.
The transmission of gun and other weapon-involved violence within social networks
.
Epidemiol Rev
.
2016
;
38
(
1
):
70
86
3
Oliphant
SN
,
Mouch
CA
,
Rowhani-Rahbar
A
, et al.
FACTS Consortium
.
A scoping review of patterns, motives, and risk and protective factors for adolescent firearm carriage
.
J Behav Med
.
2019
;
42
(
4
):
763
810
4
Bottiani
JH
,
Camacho
DA
,
Lindstrom Johnson
S
,
Bradshaw
CP
.
Annual research review: youth firearm violence disparities in the United States and implications for prevention
.
J Child Psychol Psychiatry
.
2021
;
62
(
5
):
563
579
5
Goldstick
JE
,
Carter
PM
,
Cunningham
RM
.
Current epidemiological trends in firearm mortality in the United States
.
JAMA Psychiatry
.
2021
;
78
(
3
):
241
242
6
Ranney
M
,
Karb
R
,
Ehrlich
P
,
Bromwich
K
,
Cunningham
R
,
Beidas
RS
.
FACTS Consortium
.
What are the long-term consequences of youth exposure to firearm injury, and how do we prevent them? A scoping review
.
J Behav Med
.
2019
;
42
(
4
):
724
740
7
Schmidt
CJ
,
Rupp
L
,
Pizarro
JM
,
Lee
DB
,
Branas
CC
,
Zimmerman
MA
.
Risk and protective factors related to youth firearm violence: a scoping review and directions for future research
.
J Behav Med
.
2019
;
42
(
4
):
706
723
8
Moe
CA
,
Haviland
MJ
,
Bowen
AG
, %
Rowhani-Rahbar
A
,
Rivara
FP
.
Association of minimum age laws for handgun purchase and possession with homicides perpetrated by young adults aged 18 to 20 Years
.
JAMA Pediatr
.
2020
;
174
(
11
):
1056
1062
9
Bond
AE
,
Anestis
MD
.
Firearm type and number: examining differences among firearm owning suicide decedents [published online ahead of print February 11, 2021]
.
Arch Suicide Research
.
doi:10.1080/13811118.2021. 1885536
10
Studdert
DM
,
Zhang
Y
,
Swanson
SA
, et al
.
Handgun ownership and suicide in California
.
N Engl J Med
.
2020
;
382
(
23
):
2220
2229
11
Vaughn
MG
,
Perron
BE
,
Abdon
A
,
Olate
R
,
Groom
R
,
Wu
LT
.
Correlates of handgun carrying among adolescents in the United States
.
J Interpers Violence
.
2012
;
27
(
10
):
2003
2021
12
Rowhani-Rahbar
A
,
Oesterle
S
,
Skinner
ML
.
Initiation age, cumulative prevalence, and longitudinal patterns of handgun carrying among rural adolescents: a multistate study
.
J Adolesc Health
.
2020
;
66
(
4
):
416
422
13
Wong
B
,
Bernstein
S
,
Jay
J
,
Siegel
M
.
Differences in racial disparities in firearm homicide across cities: the role of racial residential segregation and gaps in structural disadvantage
.
J Natl Med Assoc
.
2020
;
112
(
5
):
518
530
14
Johnson
BT
,
Sisti
A
,
Bernstein
M
, et al
.
Community-level factors and incidence of gun violence in the United States, 2014-2017
.
Soc Sci Med
.
2021
;
280
:
113969
15
Brooks-Russell
A
,
Ma
M
,
Brummett
S
,
Wright-Kelly
E
,
Betz
M
.
Perceived access to handguns among Colorado high school students
.
Pediatrics
.
2021
;
147
(
4
):
e2020015834
16
Johnson
RM
,
Barber
C
,
Azrael
D
,
Clark
DE
,
Hemenway
D
.
Who are the owners of firearms used in adolescent suicides?
Suicide Life Threat Behav
.
2010
;
40
(
6
):
609
611
17
Hemenway
D
.
Comparing gun-owning vs non-owning households in terms of firearm and non-firearm suicide and suicide attempts
.
Prev Med
.
2019
;
119
:
14
16
18
Patel
SJ
,
Badolato
GM
,
Parikh
K
,
Iqbal
SF
,
Goyal
MK
.
Sociodemographic factors and outcomes by intent of firearm injury
.
Pediatrics
.
2021
;
147
(
4
):
e2020011957
19
Vaughn
MG
,
Nelson
EJ
,
Salas-Wright
CP
,
DeLisi
M
,
Qian
Z
.
Handgun carrying among White youth increasing in the United States: new evidence from the National Survey on Drug Use and Health 2002–2013
.
Prev Med
.
2016
;
88
:
127
133
20
Vaughn
MG
,
Oh
S
,
Salas-Wright
CP
,
DeLisi
M
,
Holzer
KJ
,
McGuire
D
.
Sex differences in the prevalence and correlates of handgun carrying among adolescents in the United States
.
Youth Violence Juv Justice
.
2019
;
17
(
1
):
24
41
21
Cunningham
RM
,
Carter
PM
,
Ranney
ML
, et al
.
Prevention of firearm injuries among children and adolescents: consensus-driven research agenda from the Firearm Safety Among Children and Teens (FACTS) Consortium
.
JAMA Pediatr
.
2019
;
173
(
8
):
780
789
22
Neil
R
,
Sampson
RJ
,
Nagin
DS
.
Social change and cohort differences in group-based arrest trajectories over the last quarter-century
.
Proc Natl Acad Sci USA
.
2021
;
118
(
31
):
e2107020118
23
Center for Behavioral Health Statistics and Quality
.
2019 National Survey on Drug Use and Health: public use file codebook
.
24
Center for Behavioral Health Statistics and Quality
.
2019 National Survey on Drug Use and Health (NSDUH): final CAI specifications for programming (English Version)
.
25
Fowler
KA
,
Dahlberg
LL
,
Haileyesus
T
,
Gutierrez
C
,
Bacon
S
.
Childhood firearm injuries in the United States
.
Pediatrics
.
2017
;
140
(
1
):
e20163486
26
Ngo
QM
,
Sigel
E
,
Moon
A
, et al.
FACTS Consortium
.
State of the science: a scoping review of primary prevention of firearm injuries among children and adolescents
.
J Behav Med
.
2019
;
42
(
4
):
811
829
27
Grossman
DC
,
Stafford
HA
,
Koepsell
TD
,
Hill
R
,
Retzer
KD
,
Jones
W
.
Improving firearm storage in Alaska native villages: a randomized trial of household gun cabinets
.
Am J Public Health
.
2012
;
102
(
Suppl 2
):
S291
S297
28
Coyne-Beasley
T
,
Schoenbach
VJ
, %
Johnson
RM
.
“Love our kids, lock your guns”: a community-based firearm safety counseling and gun lock distribution program
.
Arch Pediatr Adolesc Med
.
2001
;
155
(
6
):
659
664

Supplementary data