Youth Risk Behavior Surveillance — United States, 2019

Supplement / Vol. 69 / No. 1 August 21, 2020

U.S. Department of Health and Human Services

Centers for Disease Control and Prevention

Morbidity and Mortality Weekly Report

Youth Risk Behavior Surveillance —

United States, 2019

Supplement

The MMWR series of publications is published by the Center for Surveillance, Epidemiology, and Laboratory Services, Centers for Disease Control and Prevention (CDC),

U.S. Department of Health and Human Services, Atlanta, GA 30329-4027.

Suggested citation: [Author names; first three, then et al., if more than six.] [Title]. MMWR Suppl 2020;69(Suppl-#):[inclusive page numbers].

Centers for Disease Control and Prevention

Robert R. Redfield, MD, Director

Anne Schuchat, MD, Principal Deputy Director

Chesley L. Richards, MD, MPH, Deputy Director for Public Health Science and Surveillance

Rebecca Bunnell, PhD, MEd, Director, Office of Science

Arlene Greenspan, PhD, Acting Director, Office of Science Quality, Office of Science

Michael F. Iademarco, MD, MPH, Director, Center for Surveillance, Epidemiology, and Laboratory Services

MMWR Editorial and Production Staff (Serials)

Charlotte K. Kent, PhD, MPH, Editor in Chief

Christine G. Casey, MD, Editor

Mary Dott, MD, MPH, Online Editor

Terisa F. Rutledge, Managing Editor

David C. Johnson, Lead Technical Writer-Editor

Catherine B. Lansdowne, MS, Project Editor

Marella Meadows, Project Editor

Guest Editors: Zewditu Demissie, PhD; J. Michael Underwood, PhD

Martha F. Boyd, Lead Visual Information Specialist

Maureen A. Leahy, Julia C. Martinroe,

Stephen R. Spriggs, Tong Yang,

Visual Information Specialists

Quang M. Doan, MBA, Phyllis H. King,

Terraye M. Starr, Moua Yang,

Information Technology Specialists

MMWR Editorial Board

Timothy F. Jones, MD, Chairman

Michelle E. Bonds, MBA

Matthew L. Boulton, MD, MPH

Carolyn Brooks, ScD, MA

Jay C. Butler, MD

Virginia A. Caine, MD

Katherine Lyon Daniel, PhD

Jonathan E. Fielding, MD, MPH, MBA

David W. Fleming, MD

William E. Halperin, MD, DrPH, MPH

Jewel Mullen, MD, MPH, MPA

Jeff Niederdeppe, PhD

Patricia Quinlisk, MD, MPH

Patrick L. Remington, MD, MPH

Carlos Roig, MS, MA

William Schaffner, MD

Morgan Bobb Swanson, BS

CONTENTS

Overview and Methods for the Youth Risk Behavior

Surveillance System — United States, 2019..............................................1

Condom and Contraceptive Use Among Sexually Active High School

Students — Youth Risk Behavior Survey, United States, 2019

......... 11

Trends in Violence Victimization and Suicide Risk by Sexual Identity

Among High School Students — Youth Risk Behavior Survey,

United States, 2015–2019

............................................................................. 19

Interpersonal Violence Victimization Among High School Students —

Youth Risk Behavior Survey, United States, 2019

...................................... 28

Prescription Opioid Misuse and Use of Alcohol and Other Substances

Among High School Students — Youth Risk Behavior Survey,

United States, 2019

......................................................................................... 38

Suicidal Ideation and Behaviors Among High School Students —

Youth Risk Behavior Survey, United States, 2019

................................. 47

Tobacco Product Use Among High School Students — Youth Risk

Behavior Survey, United States, 2019

....................................................... 56

Dietary and Physical Activity Behaviors Among High School

Students — Youth Risk Behavior Survey, United States, 2019

......... 64

Transportation Risk Behaviors Among High School Students —

Youth Risk Behavior Survey, United States, 2019

................................. 77

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

Overview and Methods for the Youth Risk Behavior

Surveillance System — United States, 2019

J. Michael Underwood, PhD

; Nancy Brener, PhD

; Jemekia Thornton, MPA

; William A. Harris, MM

; Leah N. Bryan, MPH

;

Shari L. Shanklin, MPH

; Nicholas Deputy, PhD

1,2

; Alice M. Roberts, MS

; Barbara Queen, MS

, David Chyen, MS

; Lisa Whittle, MPH

;

Connie Lim, MPA

; Yoshimi Yamakawa, MPH

; Michelle Leon-Nguyen, MPH

; Greta Kilmer, MS

; Jennifer Smith-Grant, MSPH

;

Zewditu Demissie, PhD

; Sherry Everett Jones, PhD, JD

; Heather Clayton, PhD

; Patricia Dittus, PhD

Division of Adolescent and School Health, National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention, CDC;

Epidemic Intelligence Service;

ICF International, Rockville, Maryland;

Westat, Rockville, Maryland

Abstract

Health risk behaviors practiced during adolescence often persist into adulthood and contribute to the leading causes of morbidity

and mortality in the United States. Youth health behavior data at the national, state, territorial, tribal, and local levels help monitor

the effectiveness of public health interventions designed to promote adolescent health. The Youth Risk Behavior Surveillance

System (YRBSS) is the largest public health surveillance system in the United States, monitoring a broad range of health-related

behaviors among high school students. YRBSS includes a nationally representative Youth Risk Behavior Survey (YRBS) and separate

state, local school district, territorial, and tribal school–based YRBSs. This overview report describes the surveillance system and

the 2019 survey methodology, including sampling, data collection procedures, response rates, data processing, weighting, and

analyses presented in this MMWR Supplement. A 2019 YRBS participation map, survey response rates, and student demographic

characteristics are included. In 2019, a total of 78 YRBSs were administered to high school student populations across the United

States (national and 44 states, 28 local school districts, three territories, and two tribal governments), the greatest number of

participating sites with representative data since the surveillance system was established in 1991. The nine reports in this MMWR

Supplement are based on national YRBS data collected during August 2018–June 2019. A full description of 2019 YRBS results

and downloadable data are available (https://www.cdc.gov/healthyyouth/data/yrbs/index.htm).

Efforts to improve YRBSS and related data are ongoing and include updating reliability testing for the national questionnaire,

transitioning to electronic survey administration (e.g., pilot testing for a tablet platform), and exploring innovative analytic methods

to stratify data by school-level socioeconomic status and geographic location. Stakeholders and public health practitioners can

use YRBS data (comparable across national, state, tribal, territorial, and local jurisdictions) to estimate the prevalence of health-

related behaviors among different student groups, identify student risk behaviors, monitor health behavior trends, guide public

health interventions, and track progress toward national health objectives.

Introduction

Adolescence is typically a healthy period of life, and CDC

reports that youths continue to make better decisions for

their health (1). However, some high school–aged youths

experience disparate health risks that increase the possibility

of acquiring a sexually transmitted disease (STD), including

human immunodeficiency virus (HIV) infection, and increase

opportunities for substance use, mental health problems, and

interpersonal violence or self-harm. Risky health behaviors

practiced during adolescence often persist into adulthood

(2). In 2018, CDC reported that the leading causes of death

Corresponding author: J. Michael Underwood, PhD, Division of

Adolescent and School Health, National Center for HIV/AIDS, Viral

Hepatitis, STD, and TB Prevention, CDC. Telephone: 404-718-1471;

E-mail: [email protected].

among U.S. adolescents were attributable to motor-vehicle

crashes, followed by suicide and homicide (3). In contrast,

that same year, a separate study reported the leading causes of

death among persons of all ages were heart disease, followed

by cancer and unintentional injuries (e.g., burns, drowning,

falls, poisoning, and motor-vehicle crashes) (4).

The Youth Risk Behavior Surveillance System (YRBSS)

monitors health behaviors, conditions, and experiences among

high school students throughout the United States. The system

includes a national Youth Risk Behavior Survey (YRBS),

conducted by CDC, and separate state, local school district,

territorial, and tribal school–based YRBSs, which are referred

to as site-level surveys. YRBSS is designed to monitor priority

health risk behaviors that contribute to the leading causes of

mortality, morbidity, and social problems among youths and

adults. The following categories of behaviors are included

in the system: 1) behaviors that contribute to unintentional

Supplement

2 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

injury and violence; 2) tobacco use; 3) alcohol and other

drug use; 4) sexual behaviors that contribute to unintended

pregnancy and STD/HIV infection; 5) dietary behaviors; and

6) physical inactivity.

This report describes the 2019 YRBS methodology,

including sampling, data collection, processing, weighting,

and analyses. Results include a 2019 YRBS participation

map, survey response rates (1991–2019), and student

demographic characteristics from the national survey.

Furthermore, this overview report is one of nine featured

in this MMWR Supplement. Each report uses YRBS data to

assess a priority public health topic among adolescents. In

addition to this overview report, this supplement includes

national YRBS updates regarding condom and contraceptive

use; violence victimization and suicide ideation by sexual

identity; interpersonal violence victimization; opioid, alcohol,

and other substance use behaviors; suicide ideation and

behaviors; tobacco use, including vaping; dietary behaviors

and physical activity; and transportation risk behaviors. Each

report might not include all national YRBS data related to the

topics that were collected in 2019, and this supplement does

not include any data from site-level surveys; however, all the

data are publicly available. (YRBS data and documentation

are available at https://www.cdc.gov/healthyyouth/data/yrbs/

data.htm.) Stakeholders and public health practitioners can

use YRBS data (comparable across national, state, tribal, and

local jurisdictions) to estimate the prevalence of health-related

behaviors among different student groups, identify student

risk behaviors, monitor health behavior trends, guide public

health interventions, and track progress toward national

health objectives.

National YRBS Methodology

Overview

The national YRBS is conducted biennially during the

spring of odd-numbered years and allows CDC to assess how

risk behaviors change temporally among the U.S. high school

population. The national YRBS provides comparable data

across years and allows state and local entities conducting their

own YRBS to demonstrate how the behaviors of their youths

compare with those at the national level. YRBS is conducted

among students in grades 9–12 who attend U.S. public and

private schools. A nationally representative sample of schools

and a random sample of classes within those schools are selected

to participate. The survey is self-administered anonymously by

using a computer-scannable questionnaire booklet and takes

one class period (approximately 45 minutes) to complete.

Questionnaire

In 2019, the YRBS questionnaire consisted of 99 questions.

Eighty-nine of those questions were included in the standard

questionnaire* used by sites. Ten additional questions were

added to the standard questionnaire that reflect areas of interest

for CDC and other stakeholders, forming the 99-question

national YRBS questionnaire. As in all cycles, both the standard

questionnaire and additional national-only questions were

revised to ensure that emerging and prevailing risk behaviors

among high school students were measured. Subject matter

experts from CDC and elsewhere proposed changes, additions,

and deletions to the questionnaire. New and revised questions

were reviewed for format, readability, and clarity and were

subjected to cognitive testing. CDC made further refinements

to the questions on the basis of those testing results.

All questions, except those assessing height, weight, and

race, were multiple choice, with a maximum of eight mutually

exclusive response options and only one possible answer per

question. The survey questions have undergone test-retest

analysis and demonstrated good reliability (5,6). The wording

of each question, including recall periods and response options,

and operational definitions for each variable, are available by

reviewing the 2019 YRBS questionnaire and data user’s guide.

(YRBSS data and documentation are available at https://www.

cdc.gov/healthyyouth/data/yrbs/data.htm.)

Sampling

The 2019 YRBS sampling frame consisted of all regular

public (including charter schools), parochial, and other

nonpublic schools with students in at least one of grades 9–12

in the 50 U.S. states and the District of Columbia. Alternative

schools, special education schools, schools operated by the U.S.

Department of Defense, the Bureau of Indian Education, and

vocational schools serving only students who also attended

another school were excluded. Schools with an enrollment

of ≤40 students across grades 9–12 also were excluded. The

sampling frame was based on data sets obtained from Market

Data Retrieval, Inc., and the National Center for Education

Statistics (NCES). NCES data sets were based on the Common

Core of Data (https://nces.ed.gov/ccd) for public schools

and the Private School Universe Survey (https://nces.ed.gov/

surveys/pss) for nonpublic schools.

A three-stage cluster sampling design was used to produce

a nationally representative sample of students in grades 9–12

* The standard YRBSS questionnaire includes 89 questions. YRBS coordinators

(located in CDC-funded states, local school districts, territories, and tribes)

voted for or against each proposed change, addition, and deletion. Final content

of the standard YRBS questionnaire was decided on the basis of the results of

this voting process.

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 3

US Department of Health and Human Services/Centers for Disease Control and Prevention

who attend public and private schools. The first-stage sampling

frame comprised 1,257 primary sampling units (PSUs),

consisting of entire counties, groups of smaller adjacent

counties, or parts of larger counties. The 1,257 PSUs were

categorized into 16 strata according to their metropolitan

statistical area status (e.g., urban or rural) and the percentages

of non-Hispanic black (black) and Hispanic students in each

PSU. From the 1,257 PSUs, 54 were sampled with probability

proportional to overall school enrollment size for that PSU. For

the second-stage sampling, secondary sampling units (SSUs)

were defined as a physical school with grades 9–12 or a school

created by combining nearby schools to provide all four grades.

From the 54 PSUs, 162 SSUs were sampled with probability

proportional to school enrollment size. To provide adequate

coverage of students in small schools, an additional 15 small

SSUs were selected from a subsample of 15 PSUs from the 54

PSU sample. These 177 SSUs corresponded to 184 physical

schools. The third stage of sampling comprised random

sampling of one or two classrooms in each of grades 9–12

from either a required subject (e.g., English or social studies)

or a required period (e.g., homeroom or second period).

All students in sampled classes were eligible to participate.

Schools, classes, and students who refused to participate were

not replaced in the sampling design.

Data Collection Procedures

CDC’s Institutional Review Board approved the protocol

for the YRBS. Survey procedures were designed to protect

students’ privacy by allowing for anonymous and voluntary

participation. Before survey administration, local parental

permission procedures were followed. During survey

administration, students completed the self-administered

questionnaire during one class period and recorded their

responses directly on a computer-scannable booklet.

Response Rates and Data Processing

For the 2019 YRBS, 13,872 questionnaires were completed

in 136 schools. The national data set was cleaned and edited

for inconsistencies. Missing data were not statistically imputed.

A questionnaire failed quality control when <20 responses

remained after editing or when it contained the same answer

to ≥15 consecutive questions. Among the 13,872 completed

questionnaires, 195 failed quality control and were excluded

from analysis, resulting in 13,677 usable questionnaires. The

school response rate was 75.1%; the student response rate was

80.3%; and the overall response rate (i.e., [student response

rate] × [school response rate]) was 60.3%.

Race/ethnicity was ascertained from two questions:

1) “Are you Hispanic or Latino?” (with response options of

“yes” or “no”) and 2) “What is your race?” (with response

options of “American Indian or Alaska Native,” “Asian,” “black

or African American,” “Native Hawaiian or other Pacific

Islander,” or “white”). For the second question, students

could select more than one response option. For this report,

students were classified as Hispanic/Latino and are referred

to as Hispanic if they answered “yes” to the first question,

regardless of how they answered the second question. Students

who answered “no” to the first question and selected only black

or African American to the second question were classified as

black or African American and are referred to as black. Students

who answered “no” to the first question and selected only white

to the second question were classified and are referred to as

white. Race/ethnicity was classified as missing for students

who did not answer the first question and for students who

answered “no” to the first question but did not answer the

second question.

To obtain a sufficient sample size for analyses of health-

related behaviors by sexual identity and sex of sexual contacts,

students were divided into groups (Table 1). Students who had

no sexual contact were excluded from analyses related to sexual

behaviors, female students who had sexual contact with only

females were excluded from analyses on condom use and birth

control use, and male students who had sexual contact with

only males were excluded from analyses on birth control use.

Weighting

A weight based on student sex, race/ethnicity, and grade

was applied to each record to adjust for school and student

nonresponse and oversampling of black and Hispanic students.

The overall weights were scaled so that the weighted count

of students equals the total sample size, and the weighted

proportions of students in each grade match the national

population proportions. Therefore, weighted estimates

are nationally representative of all students in grades 9–12

attending U.S. public and private schools.

Analytic Methods

Findings presented in this MMWR Supplement and Youth

Online (https://nccd.cdc.gov/Youthonline/App/Default.aspx),

an interactive data analysis tool that allows access to all YRBSS

data, follow analytic methods similar to what is described in

this overview report. For more information regarding the

analyses presented in this supplement (e.g., variables analyzed,

custom measures, and data years), see the Methods section in

each individual report.

All statistical analyses were conducted on weighted data by

using SAS (version 9.4; SAS Institute) and SUDAAN (version

11.0.1; RTI International) software to account for the complex

Supplement

4 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 1. Processing of sexual identity and sex of sexual contacts questions — Youth Risk Behavior Survey, United States, 2019

Question Student response Analytic description

Sexual identity

Which of the following best describes you?

1) Heterosexual (straight), 2) gay or lesbian, 3) bisexual, or

4) not sure

Heterosexual (straight)

Gay or lesbian or bisexual

Not sure

Heterosexual students

Lesbian, gay, or bisexual students

Not-sure students

Sex of sexual contacts

During your life, with whom have you had sexual contact?

1) I have never had sexual contact, 2) females, 3) males, or

4) females and males

What is your sex?

1) Male or 2) female

I have never had sexual contact* Students who had no sexual contact

Contact:

Female

Male

Student:

Male

Female

Students who had sexual contact with only the

opposite sex

Contact:

Male

Females and males

Female

Females and males

Student:

Male

†

Male

Female

†

Female

Students who had sexual contact with only the

same sex or with both sexes

* Excluded from analyses on sexual behaviors.

†

Excluded from analyses on birth control use and condom use.

sampling designs. In all reports, prevalence estimates and

confidence intervals were computed for variables in the YRBS

data set. Pairwise differences between populations (e.g., sex, race/

ethnicity, grade, sexual identity, and sex of sexual contacts) were

determined using t-tests. Prevalence estimates were considered

statistically significant if the t-test p value was <0.05.

In reports that analyzed data related to temporal trends,

prevalence estimates for variables assessed with identically

worded questions were examined. Logistic regression analyses

were used to account for all available estimates; control for

sex, grade, and racial/ethnic changes over time; and assess

long-term linear and quadratic trends. A p value associated

with the regression coefficient that was <0.05 was considered

statistically significant. Linear and quadratic time variables

were treated as continuous and were coded by using orthogonal

coefficients calculated with PROC IML in SAS. A minimum

of 3 survey years was required for calculating linear trends,

and a minimum of 6 survey years was required to calculate

quadratic trends. Separate regression models were used to

assess linear and quadratic trends for every variable. When a

significant quadratic trend was identified, Joinpoint software

was used to automate identification of the year when the

nonlinear (i.e., quadratic) trend changed. Regression models

were used to identify linear trends occurring in each segment.

Cubic and higher-order trends were not assessed. A quadratic

trend indicates a statistically significant but nonlinear trend

in prevalence over time. A long-term temporal change that

includes a significant linear and quadratic trend demonstrates

nonlinear variation (e.g., leveling off or change in direction) in

addition to an overall increase or decrease over time.

In reports that analyzed 2-year changes in health-related

behaviors, prevalence estimates from 2017 and 2019 were

compared by using t-tests for variables assessed with identically

worded questions in both survey years. Prevalence estimates were

considered statistically different if the t-test p value was <0.05.

Data Availability and Dissemination

YRBS data (1991–2019) are available from the YRBSS

data and documentation website (https://www.cdc.gov/

healthyyouth/data/yrbs/data.htm), as are additional resources,

including data documentation and analysis guides. Data are

available in both Access and ASCII formats. SAS and SPSS

programs are provided for converting the ASCII data into

SAS and SPSS data sets. Variables are standardized to facilitate

trend analyses and for combining data. YRBSS data are also

available online by using Youth Online (https://nccd.cdc.gov/

Youthonline/App/Default.aspx), a tool that allows point-and-

click data analysis and creation of customized tables, graphs,

maps, and fact sheets. Youth Online also performs statistical

tests by health topic and filters and sorts data by race/ethnicity,

sex, grade, and sexual orientation (sexual identity and sex of

sexual contacts). Finally, YRBS Explorer is a new application

featuring user-friendly options to view and compare national,

state, and local data via tables and graphs (https://yrbs-explorer.

services.cdc.gov). Data requests and other YRBSS-related

questions can be sent to CDC by using the data request form

(https://www.cdc.gov/healthyyouth/data/yrbs/contact.htm).

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 5

US Department of Health and Human Services/Centers for Disease Control and Prevention

State, Local School District,

Territorial, and Tribal YRBS

Methodology

Overview

Biennial administration of site-level YRBSs allows state

and local education and health agencies to assess how risk

behaviors change temporally among the high school population

in their respective jurisdiction. Site-level YRBS data provide

comparable data across years and allow comparisons of student

behaviors across jurisdictions (e.g., national or state). Site-

level surveys are conducted among students in grades 9–12

attending public schools by using samples representative of

the state, local, territorial, or tribal jurisdiction where they are

administered. The survey is self-administered anonymously

and takes one class period (approximately 45 minutes) to

complete. State and local institutional review boards approved

the protocol for their respective YRBSs. Survey methodology

for data collection, processing, and analytic methods were the

same as those described for the national YRBS.

Questionnaires

The 2019 YRBS standard questionnaire contained

89 questions and was used as the starting point for site-level

YRBS questionnaires. Sites could add or delete questions but

were required to use at least 60 of the questions on the standard

questionnaire. This flexibility allows YRBS coordinators and

other local stakeholders the opportunity to pursue topics of

interest by customizing their survey.

Sampling

Sites used a two-stage cluster sampling design to produce

a representative sample of students in grades 9–12 in their

jurisdiction. In 41 states, three local school districts, and one

territory, in the first sampling stage, public schools with any

of grades 9–12 were sampled with probability proportional

to school enrollment size. In two states, 25 local school

districts, and two territories, all schools in the jurisdiction

were selected to participate (i.e., a census of schools). In the

second sampling stage, intact classes from either a required

subject (e.g., English or social studies) or a required period

(e.g., homeroom or second period) were sampled randomly.

In three sites (Vermont, the District of Columbia, and Palau),

a census of students was eligible to participate.

Response Rates and Nonresponse

Bias Analyses

Before the 2019 YRBS cycle, CDC required a minimum

60% overall response rate for data from a jurisdiction to be

weighted. As response rates in federal surveys continue to

decline (7), a better understanding of the complex association

between nonresponse and nonresponse bias is needed. In 2019,

CDC chose three YRBS sites with overall response rates of

50%–60% (Nebraska; Texas; and Spartanburg County, South

Carolina) to pilot nonresponse bias analyses to evaluate data

representativeness. Because of data limitations, comparisons

were limited to responding and nonresponding schools by

school size and responding and nonresponding students by

grade. Weighted sample and population percentages by grade,

sex, and race/ethnicity were also compared. Overall, few

statistically significant differences between comparison groups

were found, which suggested that the data were generally

representative of their respective populations. For the 2019

cycle, CDC used nonresponse bias analysis results to help

determine whether data were weighted for sites with overall

response rates <60%.

Weighting

YRBS data were weighted if sites collected data from a

representative sample of students (determined either by an

overall response rate of ≥60% or nonresponse bias analysis

indicating no significant bias). A weight based on student sex,

race/ethnicity, and grade was applied to each record to adjust

for school and student nonresponse in each jurisdiction. The

weighted count of students equals the student population in

each jurisdiction. Data from 44 states and 28 local school

districts were weighted. In 26 states and 13 local school

districts, weighted estimates are representative of all students

in grades 9–12 attending regular public schools, and in

13 states and eight local school districts, weighted estimates are

representative of regular public school students plus students in

grades 9–12 in other types of public schools (e.g., alternative

or vocational schools).

Data Availability and Dissemination

A combined data set including national, state, local school

district, territorial, and tribal YRBS data (1991–2019) is available

from the YRBSS data and documentation website (https://nccd.

cdc.gov/Youthonline/App/Default.aspx). Availability of site

data depends on survey participation, data quality, and data-

sharing policies. Information about YRBSS data is available at

the participation maps and history website (https://www.cdc.

gov/healthyyouth/data/yrbs/participation.htm). Data requests

Supplement

6 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

and other YRBS-related questions can be sent to CDC by

using the data request form. (The YRBSS question, comment,

and data request form is available at https://www.cdc.gov/

healthyyouth/data/yrbs/contact.htm.) Site-level YRBS data

(from high school and middle school surveys) collected during

1991–2019 are available through Youth Online (https://nccd.

cdc.gov/Youthonline/App/Default.aspx) and YRBS Explorer

(https://yrbs-explorer.services.cdc.gov).

YRBS Response Rates and 2019

Demographic Characteristics

During 1991–2019, national YRBS overall response rates

remained at >60% (Figure 1). They reached a high of 71%

during the 2009 and 2011 YRBS cycles, followed by steady

decreases; response rates have remained in the low 60% range

during the 2015–2019 cycles. Since 1991, school response

rates have varied from 70% to the low 80% range, whereas

student participation rates have been consistent at 80%–90%.

Data were weighted to match national population

proportions. Thus, 50.6% of students were male, and 26.6%

were in 9th grade; 25.5% were in 10th grade; 24.2% were in

11th grade; and 23.5% were in 12th grade (Table 2). In regard

to race/ethnicity, the majority of students were no-Hispanic

FIGURE 1. Overall, school, and student response rates for the national Youth Risk Behavior Surveys — United States, 1991–2019

100

1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 2017 2019

Percentage

Year

School response rate

Student response rate

Overall response rate

white (white) (51.2%), followed by Hispanic (26.1%), black

(12.2%), and other (10.6%), which is defined as American

Indian or Alaska Native, Asian, Native Hawaiian or other

Pacific Islander, or multiracial but non-Hispanic.

Nationwide, 84.4% of students self-identified as heterosexual,

2.5% as gay or lesbian, and 8.7% as bisexual; 4.5% were not

sure of their sexual identity (Table 3). In 2019, 45.4% of

students had sexual contact with only the opposite sex, 2.2%

with only the same sex, and 4.8% with both sexes; 47.6% had

had no sexual contact.

2019 Site-Level YRBS Participation and

Student Response Rates

In 2019, a total of 44 states, 28 local school districts, three

territories, and two tribal governments had representative

data (Figure 2). In 2019, the median response rate for state

YRBSs with representative data was 65.0% (Figure 3), which

has typically remained at 60%–70% since 1991. The median

response rate for local school district YRBSs with representative

data was 76.5% (Figure 3) and has typically remained at

70%–80% since 1991. Since the inception of YRBSS in 1991,

the number of sites with representative data has increased,

reaching a high of 77 in 2019 (Figure 4).

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 7

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 2. Youth Risk Behavior Survey student demographic

characteristics — United States, 2019

Characteristic No. (%)

Participating schools 136 (100)

Student sample size 13,677* (100)

Response rates

Schools (75.1)

Students (80.3)

Total (60.3)

Sex

†

Male 6,641 (50.6)

Female 6,885 (49.4)

Race/Ethnicity

†,§

White, non-Hispanic 6,668 (51.2)

Black, non-Hispanic 2,040 (12.2)

Hispanic 3,038 (26.1)

Other 1,493 (10.6)

Grade

†,§

9 3,637 (26.6)

10 3,717 (25.5)

11 3,322 (24.2)

12 2,850 (23.5)

* Among the 13,872 completed questionnaires, 195 failed quality control and

were excluded from analysis, resulting in 13,677 usable questionnaires.

†

Does not included students who responded “ungraded” or “other grade.”

Percentages might not total 100% because of rounding.

Discussion

YRBSS is the largest public health surveillance system in the

United States, monitoring multiple health-related behaviors

among high school students. Since 1991, YRBSS has collected

data from approximately 4.9 million high school students in

approximately 2,100 separate surveys. Survey response rates

have remained slightly above 60%, since YRBSS inception.

Consistent and relatively high response rates allow for

long-term trend analyses of student health behaviors and

experiences. During the 2019 cycle, 78 separate jurisdictions

successfully collected YRBS data from a broad diversity of

high school students. Nationally representative data from

adolescents of various demographic profiles (e.g., sex, race

and ethnicity, sexual identity) provide information regarding

TABLE 3. Number and percentage of students, by sexual identity and sex of sexual contacts — Youth Risk Behavior Survey, United States, 2019

Characteristic

Total Male Female

No. (%) 95% CI No. (%) 95% CI No. (%) 95% CI

Sexual identity

Heterosexual 10,853 (84.4) 83.4–85.3 5,728 (91.2) 90.1–92.3 5,048 (77.6) 75.9–79.3

Gay or lesbian 380 (2.5) 2.1–3.0 157 (2.1) 1.6–2.7 211 (2.9) 2.3–3.6

Bisexual 1,151 (8.7) 8.0–9.4 201 (3.4) 2.8–4.1 929 (13.9) 12.7–15.2

Not sure 591 (4.5) 3.9–5.0 223 (3.2) 2.7–3.9 350 (5.6) 4.7–6.6

Sex of sexual contacts

Opposite sex only 4,856 (45.4) 42.8–48.1 2,642 (49.5) 46.2–52.8 2,214 (41.3) 38.7–44.0

Same sex only 292 (2.2) 1.8–2.7 99 (1.6) 1.2–2.0 193 (2.8) 2.2–3.6

Both sexes 526 (4.8) 4.2–5.5 90 (1.8) 1.4–2.3 436 (7.8) 6.7–9.1

No sexual contact 4,953 (47.6) 44.8–50.4 2,346 (47.1) 43.9–50.4 2,607 (48.0) 45.1–50.9

Abbreviation: CI=confidence interval.

disparities in health-related behaviors and highlight long-term

trends in the prevalence of these behaviors.

In 2019, CDC launched the Public Health Data

Modernization Initiative to enhance the potential of using

data for disease detection and elimination. The initiative

envisions a future in which data drives action efficiently,

flexibly, rapidly, and with impact. CDC leverages technology,

knowledge, leadership, access, and collaboration to harness

the life-saving power of data. YRBSS has both longstanding

and newly implemented features that align with the

modernization initiative. CDC scientists provide technical

support to help state and local education and health agencies

administer their YRBS. Flexibility in the questionnaire

design process allows stakeholders to collect data of interest

across student populations. Detailed YRBS site reports are

rapidly returned to state and local departments of health and

education, often within 16 weeks of survey administration.

In 2019, YRBSS reach (measured by the number of sites with

representative data) has increased to 78 sites including the

national survey, the most in YRBSS history. These data will

help identify student risk behaviors, affect decision-making,

and guide public health interventions.

The public release of YRBS data coincides with the

publication of this nine-part MMWR Supplement and is an

agencywide collaboration. Subject matter experts from selected

CDC programs contributed to this supplement to highlight

public health concerns among U.S. high school students.

YRBS data dissemination is managed through online requests,

Youth Online, and YRBS Explorer. This year, CDC updated

Youth Online to strengthen data presentation, improve user

experience, and ultimately expand reach for YRBS data. These

improvements to data dissemination will improve YRBS access,

expand usage, and maximize impact.

CDC continually works to strengthen YRBSS, and new

developments are under way. In 2019, CDC launched a

project to update reliability testing for the national YRBS

questionnaire. As other school-based surveys move toward

Supplement

8 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

FIGURE 2. State, local school district, territorial, and tribal government Youth Risk Behavior Surveys — selected U.S. sites, 2019

State: representative data

State: no representative data

Local school district, territory,

or tribal government: representative data

Did not participate

New York City

Newark

Philadelphia

District of Columbia

Boston

Cleveland

Genesee

Eaton

Chicago

Duval County

Orange County

Palm Beach County

Broward County

Gaston County

Spartanburg County

Nashville

Shelby County

Northern Mariana Islands

Guam

Puerto Rico

San Francisco

Oakland

San Diego

Los Angeles

Pasco County

Hillsborough County

Houston

Fort Worth

Albuquerque

Cherokee Nation

Winnebago Tribe of Nebraska

electronic platforms (e.g., computer, smart phone, or tablet),

some site-level YRBSs have also transitioned to electronic

survey administration. CDC recently completed pilot

testing for a tablet-based survey administration of the YRBS

questionnaire and is considering using tablets for future

YRBSs. Finally, CDC is exploring innovative analytic methods

to stratify YRBS data by school-level socioeconomic status

and geographic location. A recent study using this approach

reported students attending schools in low socioeconomic

areas were more likely to experience violence, poor emotional

well-being, and suicidality (8).

Limitations

Reports in this supplement include a limitations section

describing the analyses pertaining to that particular report. In

general, YRBSS findings are subject to at least six limitations.

First, these data apply only to youths who attend school and

therefore are not representative of all persons in this age group.

In 2019, approximately 5% of high school–aged youths (ages

14–17 years) were not enrolled in school (9). Those youths

might engage in riskier health behaviors than their peers, and

those behaviors are not captured in the school-administered

YRBS. Second, the extent of underreporting or overreporting

of health-related behaviors cannot be determined, although

the survey questions demonstrate good test–retest reliability

(5,6). Third, not all states and local school districts administer

YRBS, and those that did administer it might not include all

the standard questions on their YRBS questionnaire; therefore,

data for certain variables are not available for some sites. Fourth,

YRBS data analyses are based on cross-sectional surveys and

can only provide an indication of association, not causality.

Moreover, the survey is descriptive and not designed to explain

the reasons behind any observed trends. Fifth, limitations exist

related to assessment of sexual and gender identity. Students

might not be fully aware of their sexual identity at the time of

assessment or might not have understood the sexual identity

question. The category of students who are not sure of their

sexual identity might encompass students who are unsure of

their sexuality, students who were uncomfortable answering

the question, or students who did not understand the question.

In addition, although some sites asked questions about

transgender students, the national YRBS does not include a

question about gender identity; therefore, national prevalence

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 9

US Department of Health and Human Services/Centers for Disease Control and Prevention

FIGURE 3. National, state, and local school district Youth Risk Behavior Survey response rates — United States and selected U.S. sites,

1991–2019*

100

1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 2017 2019

National response rate

State median response rate

Local school district median response rate

Year

Percentage

* Does not include Youth Risk Behavior Survey data from U.S. territories and tribal governments.

estimates for this population of students cannot be assessed.

Finally, a limitation exists regarding the aggregation of race

and ethnicity data. The national YRBS aggregates these data

into broad categories of white, black, and Hispanic. All other

students are classified as “other.” More detailed racial/ethnic

information, as published elsewhere, provides valuable data

regarding health disparities among high school students (10).

Conclusion

YRBSS is the best source for quality data at the national,

state, territorial, tribal, and local school district levels for

monitoring health-related behaviors that contribute to the

leading causes of mortality and morbidity among U.S. high

school students and that can lead to health problems as adults.

A recent report from the National Academies of Sciences,

Engineering, and Medicine used YRBS as its data source

on the basis of the strengths of the system (11). In 2019, in

addition to the national data, 44 states, 28 local school districts,

three territories, and two tribal governments received data

representative of their high school student populations.

This overview report describes YRBSS methods for guiding

the analyses presented in this MMWR Supplement. A full

description of 2019 YRBS results and downloadable data

are available (https://www.cdc.gov/healthyyouth/data/yrbs/

index.htm).

Conicts of Interest

All authors have completed and submitted the International

Committee of Medical Journal Editors form for disclosure of

potential conflicts of interest. No potential conflicts of interest

were disclosed.

References

1. CDC. Youth Risk Behavior Survey: data summary and trends report,

2007–2017. Atlanta, GA: US Department of Health and Human

Services, CDC; 2018. https://www.cdc.gov/healthyyouth/data/yrbs/pdf/

trendsreport.pdf

2. Wiium N, Breivik K, Wold B. Growth trajectories of health behaviors

from adolescence through young adulthood. Int J Environ Res Public

Health 2015;12:13711–29. https://doi.org/10.3390/ijerph121113711

3. Murphy SL, Xu J, Kochanek KD, Arias E. Mortality in the United States,

2017. NCHS Data Brief 2018; (328):1–8.

4. Curtin SC, Heron M, Miniño AM, Warner M. Recent increases in injury

mortality among children and adolescents aged 10–19 years in the United

States: 1999–2016. Natl Vital Stat Rep 2018;67:1–16.

5. Brener ND, Kann L, McManus T, Kinchen SA, Sundberg EC, Ross JG.

Reliability of the 1999 youth risk behavior survey questionnaire.

J Adolesc Health 2002;31:336–42. https://doi.org/10.1016/

S1054-139X(02)00339-7

Supplement

10 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

FIGURE 4. Number of states, local school districts, territories, and tribal governments with representative Youth Risk Behavior Survey data, by

year of survey — selected U.S. sites, 1991–2019

0 10 20 30 40 50 60 70 80 90 100

1991

1993

1995

1997

1999

2001

2003

2005

2007

2009

2011

2013

2015

2017

2019

Number of representative sites

Year

States Districts Territories Tribal governments

6. Brener ND, Mcmanus T, Galuska DA, Lowry R, Wechsler H. Reliability

and validity of self-reported height and weight among high school

students. J Adolesc Health 2003;32:281–7. https://doi.org/10.1016/

S1054-139X(02)00708-5

7. Czajka JL, Beyler A. Declining response rates in federal surveys: trends

and implications. Final report. Vol. I. Washington, DC: Mathematica

Policy Research; 2016. https://aspe.hhs.gov/system/files/pdf/255531/

Decliningresponserates.pdf

8. Everett Jones S, Underwood JM, Pampati S, et al. School-level poverty

and persistent feelings of sadness or hopelessness, suicidality, and

experiences with violence victimization among public high school

students. J Health Care Poor Underserved 2020;30 [In press].

9. US Department of Education, National Center for Education Statistics.

Fast facts: enrollment trends. Washington, DC: US Department of

Education; 2019. https://nces.ed.gov/fastfacts/display.asp?id=65

10. Lowry R, Eaton DK, Brener ND, Kann L. Prevalence of health-risk

behaviors among Asian American and Pacific Islander high school

students in the U.S., 2001–2007. Public Health Rep 2011;126:39–49.

https://doi.org/10.1177/003335491112600108

11. National Academies of Sciences, Engineering, and Medicine. Promoting

positive adolescent health behaviors and outcomes: thriving in the 21st

century. Washington, DC: The National Academies Press; 2019. http://

www.nap.edu/catalog/25552/promoting-positive-adolescent-health-

behaviors-and-outcomes-thriving-in-the

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 11

US Department of Health and Human Services/Centers for Disease Control and Prevention

Condom and Contraceptive Use Among Sexually Active High School

Students — Youth Risk Behavior Survey, United States, 2019

Leigh E. Szucs, PhD

; Richard Lowry, MD

; Amy M. Fasula, PhD

; Sanjana Pampati, MPH

; Casey E. Copen, PhD

; Khaleel S. Hussaini, PhD

;

Rachel E. Kachur, MPH

; Emilia H. Koumans, MD

; Riley J. Steiner, PhD

Division of Adolescent and School Health, National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention,

CDC;

Office of the Director, National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention, CDC;

Division of Reproductive Health, National Center for Chronic Disease Prevention and Health Promotion, CDC;

Oak Ridge Institute for Science and Education;

Division of STD Prevention, National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention CDC

Abstract

Preventing unintended pregnancy and sexually transmitted diseases (STDs), including human immunodeficiency virus (HIV)

infection, among adolescents is a public health priority. This report presents prevalence estimates for condom and contraceptive

use among sexually active U.S. high school students from the 2019 Youth Risk Behavior Survey. Behaviors examined included

any condom use, primary contraceptive method use, and condom use with a more effective contraceptive method, all reported at

last sexual intercourse. Analyses were limited to sexually active students (i.e., those who had sexual intercourse with one or more

persons during the 3 months before the survey). Except for any condom use, students reporting only same-sex sexual contact were

excluded from analyses. Weighted prevalence estimates were calculated, and bivariate differences in prevalence were examined

by demographic characteristics (sex, race/ethnicity, and grade) and other sexual risk behaviors (age of sexual initiation, previous

3-month and lifetime number of sex partners, and substance use before last sexual intercourse). Nationwide, 27.4% of high school

students reported being sexually active (n = 3,226). Among sexually active students who reported having had sexual contact with

someone of the opposite sex (n = 2,698), most students (89.7%) had used a condom or a primary contraceptive method at last

sexual intercourse. Prevalence of any condom use at last sexual intercourse was 54.3%, and condoms were the most prevalent

primary contraceptive method (43.9% versus 23.3% for birth control pills; 4.8% for intrauterine device [IUD] or implant; and

3.3% for shot, patch, or ring). Approximately 9% had used condoms with an IUD, implant, shot, patch, ring, or birth control pills.

Using no pregnancy prevention method was more common among non-Hispanic black (23.2%) and Hispanic (12.8%) students

compared with non-Hispanic white students (6.8%); compared with Hispanic students, using no pregnancy prevention method

was more common among non-Hispanic black students. Prevalence of condom use was consistently lower among students with

other sexual risk behaviors. Results underscore the need for public health professionals to provide quality sexual and reproductive

health education and clinical services for preventing unintended pregnancy and STDs/HIV and decreasing disparities among

sexually active youths.

Introduction

Preventing unintended pregnancy and sexually transmitted

diseases (STDs), including human immunodeficiency virus

(HIV) infection, is a U.S. public health priority, particularly

among adolescents (1). U.S. birth rates among youths aged

15–19 years have decreased to record lows; evidence suggests

that increasing use of a range of contraceptive options,

including intrauterine devices (IUDs) and implants, also

known as long-acting reversible contraception, is a contributing

factor (2). However, U.S. birth rates among adolescents

remain higher than rates in comparable Western industrialized

nations (3). In 2018, U.S. birth rates for persons aged 15–17

Corresponding author: Leigh E. Szucs, PhD, Division of Adolescent

and School Health, National Center for HIV/AIDS, Viral Hepatitis,

STD, and TB Prevention, CDC. Telephone: 404-718-6785; E-mail:

[email protected].

and 18–19 years were 7.2 and 32.3 births per 1,000 females,

respectively (4). Moreover, racial/ethnic, geographic, and

socioeconomic disparities persist (4). For example, in 2018,

birth rates among non-Hispanic black (black) (26.3) and

Hispanic (26.7) persons aged 15–19 years were almost two

times the rate for non-Hispanic white (white) (12.1) persons (4).

Contraceptive methods vary in effectiveness and highly and

moderately effective methods do not prevent STDs, which

disproportionately affect adolescents (5). Highly effective

reversible contraceptive methods (IUDs and implants) are

associated with a <1% failure rate during the first year of typical

use; moderately effective contraceptive methods (injectables,

patches, rings, and birth control pills) are associated with a

4%–7% failure rate during the first year of typical use; and

less effective methods (condoms, diaphragm, and spermicides)

are associated with a >10% failure rate during the first year

of typical use (6). Condoms, although categorized as a less

Supplement

12 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

effective method of pregnancy prevention (6), remain vital

for STD/HIV prevention and promoting condom use is

particularly important given increasing STD rates in the

United States (5). Professional medical organizations (7,8) and

federal agencies, including CDC, recommend using condoms

for STD/HIV prevention with a more effective method of

contraception for optimal protection against unintended

pregnancy (9). However, recent decreases in condom use

have been documented, and the proportion of adolescents

using condoms with more effective methods of contraception

has been consistently low, with recent national estimates of

approximately 9% of sexually active high school students (10).

Because of these challenges to pregnancy- and STD/HIV-

prevention goals, monitoring condom and contraceptive use

behaviors among sexually active youths is essential. This study

reports prevalence estimates from the 2019 Youth Risk Behavior

Survey (YRBS) for any condom use at last sexual intercourse

among sexually active U.S. high school students. In addition,

prevalence estimates of primary contraceptive method use and

condom use with more effective methods of contraception

at last sexual intercourse among sexually active students who

had sexual contact with the opposite sex during their lifetime

are reported. Variations in these behaviors by demographic

characteristics and sexual risk behaviors were examined to

support public health professionals in implementing quality

sexual and reproductive health education and clinical services

that prevent STDs/HIV and unintended pregnancy.

Methods

Data Source

This report includes data from the 2019 YRBS, a cross-

sectional, school-based survey conducted biennially since

1991. Each survey year, CDC collects data from a nationally

representative sample of public and private school students in

grades 9–12 in the 50 U.S. states and the District of Columbia.

Additional information about YRBS sampling, data collection,

response rates, and processing is available in the overview report

of this supplement (11). The prevalence estimates for all sexual

behavior questions for the overall study population and by

sex, race/ethnicity, grade, and sexual orientation are available

at https://nccd.cdc.gov/youthonline/App/Default.aspx. The

full YRBS questionnaire is available at https://www.cdc.gov/

healthyyouth/data/yrbs/pdf/2019/2019_YRBS-National-HS-

Questionnaire.pdf.

Measures

Behaviors analyzed included any condom use, primary

contraceptive method, and condom use with more effective

methods of contraception, all reported at last sexual intercourse.

Any condom use was assessed by the question, “The last time

you had sexual intercourse, did you or your partner use a

condom?” Response options included the following: I have

never had sexual intercourse, yes, or no. Primary contraceptive

method was assessed through a separate question, “The last

time you had sexual intercourse, what one method did you or

your partner use to prevent pregnancy?” Respondents could

select only one response from the following list of options:

I have never had sexual intercourse; no method was used to

prevent pregnancy; birth control pills; condoms; an IUD

(such as Mirena or ParaGard) or implant (such as Implanon

or Nexplanon); a shot (such as Depo-Provera), patch (such

as Ortho Evra), or birth control ring (such as NuvaRing);

withdrawal or some other method; or not sure. Dichotomous

(yes versus no) variables for each response option were created,

except for “not sure”; although participants selecting this

response (n = 93; 3.9%) were included in the analytic sample,

prevalence estimates for this category are not reported.

A dichotomous (yes versus no) variable for any condom

use with an IUD, implant, shot, patch, ring, or birth control

pills was constructed by using the separate items for any

condom use and primary contraceptive method at last sexual

intercourse. These two items were also used to create the

following dichotomous (yes versus no) indicators: condom

use only (yes to any condom use and condoms or no method

for pregnancy prevention); highly or moderately effective

contraceptive use only (no to any condom use and an IUD,

implant, shot, patch, ring, or birth control pills for pregnancy

prevention); withdrawal or some other contraceptive method

use only (no to any condom use and withdrawal or some other

method for pregnancy prevention); and use of no condom and

no primary contraceptive method (no to any condom use and

no method for pregnancy prevention).

Condom and contraceptive use were examined by

demographic characteristics and sexual risk behaviors.

Demographic characteristics included sex (female or male),

race/ethnicity (non-Hispanic white [white], non-Hispanic

black [black], or Hispanic; other/multiple responses are not

reported), and grade (9, 10, 11, or 12). Four dichotomous

sexual risk behaviors were created: age of sexual initiation

(<13 years versus ≥13 years); lifetime number of sex partners

(1–3 versus ≥4); number of sex partners during the previous

3 months (1 versus ≥2); and alcohol or drug use before last

sexual intercourse (yes versus no).

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 13

US Department of Health and Human Services/Centers for Disease Control and Prevention

Analysis

The analytic sample was restricted to sexually active students

(i.e., those who reported having had sexual intercourse with

one or more persons during the 3 months before the survey).

Analyses involving pregnancy prevention methods excluded

students who only had same-sex sexual contacts during their

lifetime, on the basis of an item about respondents’ sex (“What

is your sex?” with response options including female or male)

and another item assessing the sex of sexual contacts (“During

your life, with whom have you had sexual contact?” with

response options including I have never had sexual contact,

females, males, and females and males).

All analyses were conducted using SUDAAN (version 11.0.0;

RTI International) to account for the complex sampling design.

Weighted prevalence estimates and 95% confidence intervals

were calculated for each outcome. Chi-square statistics

were used to examine bivariate differences by demographic

characteristics and sexual risk behaviors. For significant overall

differences by race/ethnicity and grade, t-tests were used to

identify pairwise differences. Differences were considered

significant if p<0.05.

Results

Among the 27.4% of sexually active students (n = 3,226),

approximately half were female (52.2%) and white (52.3%);

approximately one third were in grade 12 (36.9%) (Table 1).

Regarding sexual risk behaviors among those sexually active

students, 7.0% had sexual intercourse for the first time before

age 13 years (3.0% of all YRBS respondents reported having

had sexual intercourse for the first time before age 13 years);

26.9% had sexual intercourse with ≥4 persons during their

lifetime (8.6% of all YRBS respondents reported having had

sexual intercourse with ≥4 persons during their lifetime);

20.5% had sexual intercourse with ≥2 persons during the

previous 3 months; and 21.2% had drunk alcohol or used

drugs before last sexual intercourse.

Among sexually active students, prevalence of any condom

use at last sexual intercourse was 54.3% (Table 2). Among

sexually active students who reported having had sexual

contact with someone of the opposite sex (i.e., excluding

those who reported only same-sex sexual contact) (n = 2,698),

condoms (43.9%) were the most prevalent primary pregnancy

prevention method, based on responses to the distinct item

assessing pregnancy prevention method, followed by birth

control pills (23.3%); withdrawal or other method (10.1%);

IUD or implant (4.8%); and shot, patch, or ring (3.3%). (Of

participants excluded from the analytic sample for primary

method of pregnancy prevention, 95 students reported having

had only same-sex sexual contact and 433 students did not

answer the questions, “What is your sex?” or “During your life,

with whom have you had sexual contact?”) Approximately one

tenth (10.7%) had not used a pregnancy prevention method

at last sexual intercourse; 9.1% had used a condom with an

IUD, implant, shot, patch, ring, or birth control pills at last

sexual intercourse. Prevalence of condom and IUD or implant

use (<1.0%) and condom and shot, patch, or ring use (<1.0%)

was lower than condom and birth control pills use (7.5%).

Prevalence estimates for mutually exclusive categories that

reflect both pregnancy and STD/HIV prevention effectiveness

and account for any condom use in addition to a primary

pregnancy prevention method indicate that condom use

only was most common (44.3%), followed by highly or

moderately effective contraceptive method use only (22.2%)

(Figure). Prevalence of condom use with an IUD, implant,

shot, patch, ring, or birth control pills (9.1%) was similar to

prevalence of using withdrawal or other method only (9.5%)

and using no condom and no primary pregnancy prevention

method (10.3%).

Analyses revealed significant differences in any condom

use and primary pregnancy prevention method at last sexual

intercourse by demographic characteristics (Table 2). By

sex, no differences occurred in not using any method of

pregnancy prevention (i.e., no method); however, differences

were identified in type of method used. Compared with male

students’ report of contraceptive use by their female partner,

prevalence as reported by female students was higher for shot,

patch, or ring (4.5% versus 2.1%) and birth control pills

(26.1% versus 20.2%). In contrast, prevalence of condom use

as the primary method for pregnancy prevention reported by

male students (49.4%) was higher than female students’ report

of condom use by their male partner (38.8%), as was any

condom use at last sexual intercourse (60.0% versus 49.6%).

A similar pattern emerged when examining prevalence of

any condom and primary contraceptive method use by grade.

The prevalence of using no method was the same across grades;

however, differences occurred in method type. Any condom

use and condom use as the primary pregnancy prevention

method was more prevalent in lower versus higher grades. In

contrast, use of an IUD or implant; birth control pills; and

condom with an IUD or implant, shot, patch, ring, or birth

control pills was typically more prevalent in higher versus lower

grades. For example, condom use as the primary pregnancy

prevention method was more common among students in

grades 9 (55.3%), 10 (47.7%), and 11 (45.3%) versus students

in grade 12 (37.4%) (and grade 9 versus grade 11), whereas

IUD or implant use was less common among 9th-grade

Supplement

14 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 1. Prevalence of demographic characteristics and sexual risk behaviors among sexually active* high school students — Youth Risk

Behavior Survey, United States, 2019

Characteristic No.

†

) 95% CI

Sex

Female 1,679 (52.2) 49.4–55.0

Male 1,510 (47.8) 45.0–50.6

Race/Ethnicity

Black, non-Hispanic 474 (11.2) 8.9–14.0

Hispanic 771 (28.4) 22.3–35.5

White, non-Hispanic 1,602 (52.3) 46.4–58.1

Grade

9 389 (11.3) 9.8–13.0

10 741 (21.4) 19.3–23.6

11 967 (30.4) 27.8–33.2

12 1,089 (36.9) 33.4–40.4

Sexual risk behavior

Had sexual intercourse before age 13 years 242 (7.0) 5.7–8.5

Had sexual intercourse with ≥4 persons during their lifetime 854 (26.9) 24.3–29.7

Had sexual intercourse with ≥2 persons during the previous 3 months 658 (20.5) 18.5–22.7

Had drunk alcohol or used drugs before last sexual intercourse 652 (21.2) 18.8–23.9

Abbreviation: CI = confidence interval.

* Defined as having had sexual intercourse with at least one person during the 3 months before the survey (n = 3,226).

†

Unweighted.

Weighted estimates.

Race/ethnicity values do not total 100% because “other/multiple” responses are not reported (i.e., American Indian/Alaska Native, Asian, Native Hawaiian/Other

Pacific Islander, and multiple race).

students (<1.0%), compared with students in grades 10 (3.3%),

11 (3.2%), and 12 (8.2%). Prevalence of IUD or implant use

among 10th- and 11th-grade students was also lower than

among 12th-grade students.

In contrast with sex and grade, not using a pregnancy

prevention method differed by race/ethnicity, with higher

prevalence of no method among black (23.2%) and Hispanic

(12.8%) students, compared with white students (6.8%);

compared with Hispanic students, using no pregnancy

prevention method was more common among black students.

Additional racial/ethnic differences in type of method were

identified, with the general pattern that prevalence of using

a more effective method of contraception was lower among

black and Hispanic students compared with white students.

Specifically, prevalence among black and Hispanic students

was lower than among white students for use of an IUD or

implant (2.0% and 1.6% versus 6.7%, respectively); birth

control pills (12.1% and 15.5% versus 29.7%, respectively);

and condom use with an IUD, implant, shot, patch, ring, or

birth control pills (7.5% and 4.8% versus 12.4%, respectively).

In contrast, prevalence of withdrawal or other method use was

higher among black (13.9%) and Hispanic (13.1%) students

than among white students (7.7%). Condom use as the

primary method for pregnancy prevention was higher among

Hispanic students (49.6%), compared with black (37.2%) and

white (42.3%) students, and any condom use at last sexual

intercourse was higher among Hispanic (56.2%) and white

(55.8%) students compared with black students (48.2%).

Differences by sexual risk behaviors in the prevalence of using no

contraceptive method and in the type of method used, including

any condom use, also were observed (Table 3). Comparing

students who had initiated sex before age 13 years with students

who had not, differences in no method use were not significant;

however, prevalence was lower for any condom use at last sexual

intercourse (40.9% versus 55.4%), condom use as the primary

method of pregnancy prevention (30.4% versus 44.8%), and

withdrawal or other method use (5.5% versus 10.4%). Students

who had ≥4 lifetime partners had higher prevalence of no method

use (14.7% versus 9.2%) and lower prevalence of any condom

use (46.6% versus 57.1%); condom use as the primary pregnancy

prevention method (36.2% versus 46.6%); and condom use with

an IUD, implant, shot, patch, ring, or birth control pills (6.5%

versus 10.1%) compared with students who had <4 lifetime sex

partners. A similar pattern was observed for students who reported

having had ≥2 recent partners, although no significant differences

in no method use were observed. Comparing students who had

drunk alcohol or used drugs before last sexual intercourse with

students who had not, use of no method was higher (14.7%

versus 9.6%), whereas any condom use (47.4% versus 56.0%)

and condom use as the primary pregnancy prevention method

(39.3% versus 45.1%) were lower.

Discussion

This report provides the most recent nationally representative

estimates of condom and contraceptive use among sexually

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 15

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 2. Prevalence of condom and primary contraceptive use at last sexual intercourse among sexually active* high school students, by

demographic characteristics — Youth Risk Behavior Survey, United States, 2019

Demographic

characteristic

Any

condom use

†

Primary contraceptive method Condoms and IUD,

implant, shot,

patch, ring, or

birth control pills

IUD or

implant

Shot, patch,

or ring

Birth control

pills

Condom

Withdrawal or

other method

method

(95%

CI)

value

(95%

CI)

value

(95%

CI)

value

(95%

CI)

value

(95%

CI)

value

(95%

CI)

value

(95%

CI)

value

(95%

CI)

value

Total 54.3

(52.0–56.6)

NA 4.8

(3.3–7.0)

NA 3.3

(2.3–4.7)

NA 23.3

(19.8–27.2)

NA 43.9

(40.6–47.3)

NA 10.1

(8.5–12.0)

NA 10.7

(8.8–12.8)

NA 9.1

(7.4–11.2)

Sex NA <0.01 NA 0.20 NA <0.05 NA <0.01 NA <0.01 NA 0.34 NA 0.21 NA 0.10

Female 49.6

(45.6–53.6)

NA 5.6

(4.0–7.6)

NA 4.5

(2.9–6.8)

NA 26.1

(22.1–30.5)

NA 38.8

(34.0–44.0)

NA 10.8

(8.9–13.0)

NA 11.9

(9.1–15.3)

NA 10.3

(8.3–12.7)

Male 60.0

(57.0–63.0)

NA 4.0

(2.1–7.4)

NA 2.1

(1.2–3.6)

NA 20.2

(16.4–24.7)

NA 49.4

(45.8–53.1)

NA 9.3

(7.1–12.2)

NA 9.3

(7.1–12.1)

NA 7.9

(5.8–10.7)

Grade NA <0.05 NA <0.01 NA 0.29 NA <0.01 NA <0.01 NA 0.20 NA 0.63 NA <0.01

9 61.3**

(54.6–67.5)

NA 0.1**

,††,§§

(0.0–0.7)

NA 2.4

(0.9–6.2)

NA 10.9**

,††

(6.0–19.1)

NA 55.3**

,††

(47.4–62.9)

NA 10.6

(6.9–15.9)

NA 14.1

(9.1–21.2)

NA 4.7**

(2.7–8.2)

10 55.4

(50.2–60.4)

NA 3.3**

(2.0–5.6)

NA 2.1

(1.0–4.4)

NA 18.2**

,††

(13.4–24.3)

NA 47.7**

(41.2–54.3)

NA 12.5

(9.5–16.4)

NA 10.5

(7.2–15.1)

NA 7.0**

(4.7–10.3)

11 56.3

(51.9–60.6)

NA 3.2

(1.8–5.8)

NA 4.2

(2.6–6.7)

NA 25.8

(21.0–31.3)

NA 45.3**

(39.4–51.4)

NA 7.9

(5.9–10.5)

NA 10.1

(7.9–12.8)

NA 8.9

(6.1–12.9)

12 50.3

(46.9–53.8)

NA 8.2

(5.5–12.2)

NA 3.6

(2.0–6.3)

NA 27.7

(23.3–32.5)

NA 37.4

(33.4–41.6)

NA 10.3

(7.9–13.3)

NA 10.2

(6.9–14.9)

NA 11.6

(8.9–15.0)

Race/Ethnicity NA <0.05 NA <0.01 NA 0.07 NA <0.01 NA <0.01 NA <0.01 NA <0.01 NA <0.01

Black, non-Hispanic 48.2

¶¶,

***

(43.2–53.3)

NA 2.0

¶¶

(1.0–4.0)

NA 5.4

(2.9–9.9)

NA 12.1

¶¶

(8.7–16.5)

NA 37.2***

(31.2–43.6)

NA 13.9

¶¶

(8.4–22.2)

NA 23.2

¶¶,

***

(19.2–27.7)

NA 7.5

¶¶

(5.1–10.8)

Hispanic 56.2

(52.0–60.3)

NA 1.6

¶¶

(0.7–3.4)

NA 1.4

(0.6–3.2)

NA 15.5

¶¶

(11.5–20.5)

NA 49.6

¶¶

(44.7–54.4)

NA 13.1

¶¶

(10.0–17.0)

NA 12.8

¶¶

(9.1–17.8)

NA 4.8

¶¶

(3.1–7.4)

White, non-Hispanic 55.8

(52.9–58.6)

NA 6.7

(5.0–9.0)

NA 4.0

(2.5–6.4)

NA 29.7

(25.7–34.0)

NA 42.3

(38.2–46.5)

NA 7.7

(6.1–9.8)

NA 6.8

(5.3–8.6)

NA 12.4

(10.1–15.2)

Abbreviations: CI = confidence interval; IUD = intrauterine device; NA = not applicable.

* Defined as having had sexual intercourse with at least one person during the 3 months before the survey (n = 3,226). Except for any condom use at last sexual intercourse, students

reporting only same-sex sexual contact use were excluded; therefore, the analytic sample was restricted to sexually active students who reported having had sexual contact with someone

of the opposite sex (n = 2,698). Among sexually active students, excluding those who only had same-sex sexual contact, a total of 93 (3.9%) students answered the pregnancy prevention

question “not sure”; findings are not presented for this group.

†

Any condom use at last sexual intercourse was measured by a separate item from condoms as the primary method used for preventing pregnancy.

Weighted estimates.

Significance is defined as p<0.05, by chi-square test.

** Significantly different than grade 12, by linear contrast t-test

††

Significantly different than grade 11, by linear contrast t-test.

§§

Significantly different than grade 10, by linear contrast t-test.

¶¶

Significantly different than white, non-Hispanic race/ethnicity, by linear contrast t-test.

*** Significantly different than Hispanic race/ethnicity, by linear contrast t-test.

active U.S. high school students. In addition, notable differences

in these behaviors by demographic characteristics and sexual

risk behaviors are identified that can support implementation

of interventions to improve condom and contraceptive use

among adolescents most in need. Doing so will help to achieve

unintended pregnancy and STD/HIV prevention goals,

including reducing disparities by race/ethnicity.

Overall, most (89.7%) sexually active students (excluding

those who only reported same-sex sexual contact) used a

condom or a primary contraceptive method at last sexual

intercourse, yet approximately one fifth (19.8%) reported

using withdrawal or some other method only or no condom

and no primary contraceptive method. Moreover, prevalence

estimates by method type, as well as differences by demographic

characteristics and sexual risk behaviors, underscore the

importance of meeting the unintended pregnancy and STD/

HIV prevention needs of all sexually active high school students.

Only 9.1% of sexually active students (excluding those who

only reported same-sex sexual contact) reported having used a

condom with a more effective contraceptive method, which is

the recommended approach for preventing both unintended

pregnancy and STDs/HIV because the most effective forms of

contraception confer no STD/HIV protection (7–9). Although

use of condoms alone can prevent both adverse outcomes and

was the most prevalent method used, only approximately half

of sexually active students reported any condom use at last

sexual intercourse, which is concerning given the high risk

for STDs among this population (5). Moreover, condoms are

categorized as a less effective pregnancy prevention method,

given that they are associated with a 13.0% pregnancy risk

during the first year of typical use (6), and prevalence of

any highly or moderately effective method use at last sexual

intercourse was only 31.4%.

Supplement

16 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

FIGURE. Prevalence of condom and primary contraceptive use* at last sexual intercourse among sexually active

†

high school students —

Youth Risk Behavior Survey, United States, 2019

100

Condom with highly or

moderately eective

contraceptive

method

Condom

only

Highly or

moderately eective

contraceptive

method only

Method used

Withdrawal or other

contraceptive

method only

No condom and

no primary

contraceptive

method

Prevalence (%)

95% Condence interval

* Condom with highly or moderately effective contraceptive method = students who responded “yes” to any condom use at last sexual intercourse and intrauterine

device or implant; shot, patch, or ring; or birth control pills (i.e., highly or moderately effective methods) as primary pregnancy prevention method. Condom only =

students who responded “yes” to any condom use at last sexual intercourse and condom or no method as primary pregnancy prevention method. Highly or

moderately effective contraceptive method only = students who responded “no” to any condom use at last sexual intercourse and intrauterine device or implant;

shot, patch, or ring; or birth control pills (i.e., highly or moderately effective methods) as primary pregnancy prevention method. Withdrawal or other contraceptive

method only = students who responded “no” to any condom use at last sexual intercourse and withdrawal or some other method as primary pregnancy prevention

method. No condom and no primary contraceptive method = students who responded “no” to any condom use at last sexual intercourse and no method of

pregnancy prevention.

†

Defined as having had sexual intercourse with at least one person during the 3 months before the survey (n = 2,698). Students reporting only same-sex sexual

contact were excluded from the analytic sample.

Notable demographic differences in condom and contraceptive

use warrant particular attention. Compared with white students,

black and Hispanic students had higher prevalence of no

pregnancy prevention method use and lower prevalence of

highly and moderately effective contraceptive method use. Black

students also had lower prevalence of any condom use at last sexual

intercourse than white and Hispanic students. On the basis of

these findings and the documented racial/ethnic disparities in birth

and STD rates among adolescents (4,5), meeting the unintended

pregnancy and STD/HIV prevention needs of black and Hispanic

youths is vital. Understanding and addressing structural barriers

that might contribute to the observed differences are important

next steps. As for grade, differences indicate that younger students

are more likely to use condoms, whereas older students are more

likely to use an IUD or implant, birth control pills, and condoms

with a more effective contraceptive method. Therefore, improving

younger adolescents’ knowledge of, comfort with, and access to the

most effective methods of pregnancy and STD/HIV prevention is

needed. Whereas findings related to race/ethnicity and grade have

clear practice implications, patterns by sex might largely reflect

reporting differences on the basis of who uses a given method. As

compared with female students, the proportion of male students

reporting condom use was higher, and the proportion reporting

their partners’ use of shot, patch, or ring, and birth control pill use

was lower. For the latter female-controlled methods, self-report

by females is considered more accurate (12).

Finally, differences in condom and contraceptive use by

sexual risk behaviors reveal that use of preventive strategies

is suboptimal among high school students who engage in

those behaviors. The general pattern was that students with a

given risk indicator, compared with those without, had lower

prevalence of condom use and higher prevalence of using

no method of contraception, although not all differences

were significant. Such findings might reflect potential

disempowerment in sexual interactions (13) and the challenge

of using condoms correctly and consistently while under the

influence of alcohol or drugs (14). Because number of partners

is an indicator of STD/HIV risk, findings that students

with ≥2 recent or ≥4 lifetime partners had lower prevalence

of condom use, alone or with a highly or moderately effective

contraceptive method, are particularly concerning.

Collectively, these findings from the 2019 YRBS highlight the

importance of programmatic efforts that can improve condom

and contraceptive use among adolescents. The effectiveness of

sexual risk reduction education is well documented (15); because

of given decreasing attention to condom-related topics in school-

based instruction (16), efforts to strengthen implementation

are warranted. Such education should ensure that highly and

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 17

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 3. Prevalence of condom and primary contraceptive use at last sexual intercourse among sexually active* high school students, by

sexual risk behaviors — Youth Risk Behavior Survey, United States, 2019

Sexual risk behavior

Any

condom use

†

Primary contraceptive method Condom and IUD,

implant, shot,

patch, ring, or

birth control pills

IUD or

implant

Shot, patch,

or ring

Birth control

pills

Condom

Withdrawal or

other method

method

(95%

CI)

value

(95%

CI)

value

(95%

CI)

value

(95%

CI)

value

(95%

CI)

value

(95%

CI)

value

(95%

CI)

value

(95%

CI)

value

Had sexual

intercourse before

age 13 years

NA <0.05 NA 0.05 NA 0.83 NA 0.81 NA <0.05 NA <0.05 NA 0.05 NA 0.12

Yes 40.9

(30.4–52.4)

NA 2.6

(1.1–6.0)

NA 3.0

(1.0–8.2)

NA 22.1

(13.4–34.1)

NA 30.4

(19.3–44.2)

NA 5.5

(2.9–10.3)

NA 22.8

(12.6–37.7)

NA 4.8

(1.7–13.3)

No 55.4

(52.9–57.9)

NA 5.0

(3.4–7.2)

NA 3.4

(2.3–4.9)

NA 23.4

(19.7–27.5)

NA 44.8

(41.4–48.3)

NA 10.4

(8.6–12.5)

NA 9.8

(8.0–12.0)

NA 9.4

(7.6–11.6)

Had sexual

intercourse with

≥4 persons during

their lifetime

NA <0.01 NA <0.01 NA 0.60 NA 0.43 NA <0.01 NA 0.32 NA <0.05 NA <0.05

Yes 46.6

(42.9–50.2)

NA 7.4

(5.0–10.8)

NA 2.9

(1.6–5.2)

NA 21.5

(17.0–26.9)

NA 36.2

(31.0–41.7)

NA 11.5

(8.7–15.0)

NA 14.7

(10.9–19.6)

NA 6.5

(4.3–9.8)

No 57.1

(54.3–59.8)

NA 3.9

(2.6–5.8)

NA 3.5

(2.3–5.4)

NA 23.9

(19.7–28.6)

NA 46.6

(43.1–50.1)

NA 9.6

(7.7–11.9)

NA 9.2

(7.3–11.7)

NA 10.1

(8.1–12.4)

Had sexual

intercourse with

≥2 persons during

the previous

3 months

NA <0.01 NA 0.69 NA <0.05 NA 0.13 NA <0.05 NA 0.14 NA 0.20 NA <0.01

Yes 47.1

(43.1–51.1)

NA 5.0

(3.1–8.0)

NA 1.7

(0.6–4.5)

NA 19.9

(15.1–25.8)

NA 39.3

(35.1–43.7)

NA 12.7

(9.2–17.3)

NA 14.0

(9.5–20.0)

NA 5.2

(3.2–8.4)

No 56.2

(53.4–58.9)

NA 4.7

(3.3–6.8)

NA 3.7

(2.6–5.3)

NA 24.1

(20.4–28.3)

NA 45.1

(41.3–48.9)

NA 9.4

(7.7–11.4)

NA 9.8

(8.2–11.7)

NA 10.1

(8.3–12.4)

Had drunk alcohol or

used drugs before

last sexual

intercourse

NA <0.05 NA 0.35 NA 0.35 NA 0.10 NA <0.05 NA 0.68 NA <0.05 NA 0.10

Yes 47.4

(42.0–52.9)

NA 5.7

(3.6–8.9)

NA 2.3

(1.1–5.1)

NA 20.5

(16.2–25.7)

NA 39.3

(33.4–45.5)

NA 10.6

(7.4–15.0)

NA 14.7

(11.0–19.3)

NA 6.4

(3.9–10.5)

No 56.0

(53.1–58.8)

NA 4.7

(3.1–6.9)

NA 3.5

(2.3–5.4)

NA 24.0

(20.4–28.1)

NA 45.1

(41.7–48.6)

NA 9.7

(7.9–11.9)

NA 9.6

(7.6–12.2)

NA 9.6

(7.6–12.1)

Abbreviations: CI = confidence interval; IUD = intrauterine device; NA = not applicable.

* Defined as having had sexual intercourse with at least one person during the 3 months before the survey (n = 3,226). Except for any condom use at last sexual intercourse, students reporting

only same-sex sexual contact use were excluded; therefore, the analytic sample was restricted to sexually active students who reported having had sexual contact with someone of the

opposite sex (n = 2,698). Among sexually active students, excluding those who only had same-sex sexual contact, a total of 93 (3.9%) students answered the pregnancy prevention question

“not sure”; findings are not presented for this group.

†

Any condom use at last sexual intercourse was measured by a separate item from condoms as the primary method used for preventing pregnancy.

Weighted estimates.

Statistical significance is defined as p<0.05, by chi-square test.

moderately effective contraceptive methods are clearly addressed,

including in earlier grades (e.g., middle school). Doing so in the

context of broader education about health services might be a

developmentally appropriate approach.

Engaging directly with communities most affected by

unintended pregnancy and STD/HIV can be one strategy to help

identify and address social determinants of health that contribute

to disparities in condom and contraceptive use. Furthermore,

education and clinical services can be delivered through

community- and school-based programs tailored to serve young

persons most in need. Fostering community–clinic partnerships

through youth-serving organizations is one strategy for reaching

the most vulnerable adolescents. Such partnerships can help

address barriers and improve access to sexual and reproductive

health care, either through referral or service integration (17).

In addition to access, delivery of comprehensive, client-

centered, and adolescent-friendly care by well-trained

providers is essential. For example, same-day initiation of

long-acting reversible contraception methods (i.e., providing

the method during the initial appointment) is a best practice

that can facilitate adolescents’ access to these methods (17).

Another example is provider counseling about condom use

with more effective contraceptive methods, which has been

associated with adolescents’ use of this prevention strategy (18).

Integrating unintended pregnancy and STD/HIV prevention

in school-, clinic-, and community-based health promotion

likely requires explicit attention to individual prevention goals as

well as preferences related to the various prevention strategies (19).

Supplement

18 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

Limitations

General limitations for the YRBS are available in the overview

report of this supplement (11). The findings in this report are

subject to at least five additional limitations. First, male students’

report of their female partners’ contraceptive use might not be

accurate (11). Second, distinguishing the intended purpose of

condom use in relation to pregnancy and STD/HIV prevention

is not feasible. Although YRBS assesses condom use as a primary

method for pregnancy prevention, condom use for STD/HIV

prevention is not explicitly measured. Third, condom use with

a more effective contraceptive method might be underestimated

because respondents could only select one method of pregnancy

prevention at last sexual intercourse. Fourth, the estimates

for highly and moderately effective contraception could be

underestimated if respondents viewed a less effective option (i.e.,

condoms or withdrawal or some other method) as their primary

contraceptive method used at last sexual intercourse. Finally,

because the sex of last sex partner is not measured, the analytic

sample might include students with same-sex partners at last

sexual intercourse for whom pregnancy prevention is not needed.

Conclusion

Ongoing national surveillance will remain important to

understanding the population-level effects of public health and

clinical approaches to preventing unintended pregnancy and

STDs/HIV among young persons. To complement these efforts,

implementation science and observational research should

address unresolved questions (e.g., young men’s role in condom

and contraceptive use, barriers and facilitators to integration of

pregnancy and STD/HIV prevention, and effective strategies

for addressing disparities, including racial/ethnic differences).

Taken together, these data can be used to improve condom and

contraceptive use for all sexually active adolescents.

Conflicts of Interest

All authors have completed and submitted the International

Committee of Medical Journal Editors form for disclosure of

potential conflicts of interest. No potential conflicts of interest

were disclosed.

References

1. US Department of Health and Human Services. Healthy people 2020:

sexually transmitted diseases. Washington, DC: US Department of

Health and Human Services; 2020. https://www.healthypeople.

gov/2020/topics-objectives/topic/sexually-transmitted-diseases

2. Lindberg LD, Santelli JS, Desai S. Changing patterns of contraceptive

use and the decline in rates of pregnancy and birth among US

adolescents, 2007–2014. J Adolesc Health 2018;63:253–6. https://doi.

org/10.1016/j.jadohealth.2018.05.017

3. Sedgh G, Finer LB, Bankole A, Eilers MA, Singh S. Adolescent

pregnancy, birth, and abortion rates across countries: levels and recent

trends. J Adolesc Health 2015;56:223–30. https://doi.org/10.1016/j.

jadohealth.2014.09.007

4. Martin J, Hamilton B, Osterman M, Driscoll A. Births: final data for

2018. Hyattsville, MD: National Center for Health Statistics; 2019.

https://www.cdc.gov/nchs/data/nvsr/nvsr68/nvsr68_13-508.pdf

5. CDC. Sexually transmitted disease surveillance 2018. Atlanta, GA: US

Department of Health and Human Services, CDC; 2019. https://www.

cdc.gov/std/stats18/default.htm

6. Trussell J, Aiken ARA, Micks E, Guthrie KA. Efficacy, safety, and personal

considerations. In: Contraceptive technology. 21st ed. Atlanta, GA:

Managing Contraception; 2018.

7. Committee on Adolescent Health Care Long-Acting Reversible

Contraception Working Group, American College of Obstetricians and

Gynecologists. Committee opinion no. 539: adolescents and long-acting

reversible contraception: implants and intrauterine devices. Obstet

Gynecol 2012;120:983–8. https://doi.org/10.1097/

AOG.0b013e3182723b7d

8. Committee on Adolescence. Contraception for adolescents. Pediatrics

2014;134:e1244-56. https://doi.org/10.1542/peds.2014-2299

9. Gavin L, Moskosky S, Carter M, et al. Providing quality family planning

services: recommendations of CDC and the U.S. Office of Population

Affairs. MMWR Recomm Rep 2014;63(No. RR-04).

10. Kann L, McManus T, Harris WA, et al. Youth risk behavior surveillance—

United States, 2017. MMWR Surveill Summ 2018;67(No. SS-8).

https://doi.org/10.15585/mmwr.ss6708a1

11. Underwood JM, Brener N, Thornton J, et al. Overview and methods

for the Youth Risk Behavior Surveillance System—United States, 2019.

In: Youth Risk Behavior Surveillance—United States, 2019. MMWR

Suppl 2020;69(No. Suppl 1).

12. Brauner-Otto S, Yarger J, Abma J. Does it matter how you ask? Question

wording and males’ reporting of contraceptive use at last sex. Soc Sci

Res 2012;41:1028–36. https://doi.org/10.1016/j.ssresearch.2012.04.004

13. Seth P, DiClemente RJ, Loworn AE. State of the evidence: intimate

partner violence and HIV/STI risk among adolescents. Curr HIV Res

2013;11:528–35.

14. Parkes A, Wight D, Henderson M, Hart G. Explaining associations

between adolescent substance use and condom use. J Adolesc Health

2007;40:180.e1-18.

15. Chin HB, Sipe TA, Elder R, et al; Community Preventive Services Task

Force. The effectiveness of group-based comprehensive risk-reduction

and abstinence education interventions to prevent or reduce the risk of

adolescent pregnancy, human immunodeficiency virus, and sexually

transmitted infections: two systematic reviews for the Guide to

Community Preventive Services. Am J Prev Med 2012;42:272–94.

https://doi.org/10.1016/j.amepre.2011.11.006

16. Lindberg LD, Maddow-Zimet I, Boonstra H. Changes in adolescents’

receipt of sex education, 2006–2013. J Adolesc Health 2016;58:621–7.

https://doi.org/10.1016/j.jadohealth.2016.02.004

17. Brittain AW, Tevendale HD, Mueller T, et al. The teen access and quality

initiative: improving adolescent reproductive health best practices in

publicly funded health centers. J Community Health 2020;45:615–25.

https://doi.org/10.1007/s10900-019-00781-z

18. Morroni C, Heartwell S, Edwards S, Zieman M, Westhoff C. The impact

of oral contraceptive initiation on young women’s condom use in 3

American cities: missed opportunities for intervention. PLoS One

2014;9:e101804. https://doi.org/10.1371/journal.pone.0101804

19. Steiner RJ, Liddon N, Swartzendruber AL, Pazol K, Sales JM. Moving

the message beyond the methods: toward integration of unintended

pregnancy and sexually transmitted infection/HIV prevention. Am J

Prev Med 2018;54:440–3. https://doi.org/10.1016/j.amepre.2017.10.022

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 19

US Department of Health and Human Services/Centers for Disease Control and Prevention

Trends in Violence Victimization and Suicide Risk by Sexual Identity

Among High School Students — Youth Risk Behavior Survey,

United States, 2015–2019

Michelle M. Johns, PhD

; Richard Lowry, MD

; Laura T. Haderxhanaj, PhD

; Catherine N. Rasberry, PhD

; Leah Robin, PhD

; Lamont Scales, MA

;

Deborah Stone, ScD

; Nicolas A. Suarez, MPH

Division of Adolescent and School Health, National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention, CDC;

Office of the Director, National

Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention, CDC;

Division of STD Prevention, National Center for HIV/AIDS, Viral Hepatitis, STD,

and TB Prevention, CDC;

Division of HIV/AIDS Prevention, National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention, CDC;

Division of Injury Prevention, National Center for Injury Prevention and Control, CDC

Abstract

Lesbian, gay, and bisexual (LGB) youths continue to experience more violence victimization and suicide risk than heterosexual

youths; however, few studies have examined whether the proportion of LGB youths affected by these outcomes has varied over

time, and no studies have assessed such trends in a nationally representative sample. This report analyzes national trends in violence

victimization and suicide risk among high school students by self-reported sexual identity (LGB or heterosexual) and evaluates

differences in these trends among LGB students by sex (male or female) and race/ethnicity (non-Hispanic black, non-Hispanic

white, or Hispanic). Data for this analysis were derived from the 2015, 2017, and 2019 cycles of CDC’s Youth Risk Behavior Survey

(YRBS), a cross-sectional, school-based survey conducted biennially since 1991. Logistic regression models assessed linear trends

in prevalence of violence victimization and indicators of suicide risk among LGB and heterosexual students during 2015–2019;

in subsequent models, sex-stratified (controlling for race/ethnicity and grade) and race/ethnicity-stratified (controlling for sex

and grade) linear trends were examined for students self-identifying as LGB during 2015–2019. Results demonstrated that LGB

students experienced more violence victimization and reported more suicide risk behaviors than heterosexual youths. Among LGB

youths, differences in the proportion reporting violence victimization and suicide risk by sex and race/ethnicity were found. Across

analyses, very few linear trends in these outcomes were observed among LGB students. Results highlight the continued need for

comprehensive intervention strategies within schools and communities with the express goal of reducing violence victimization

and preventing suicide risk behaviors among LGB students.

Introduction

Lesbian, gay, and bisexual (LGB) youths experience more

violence victimization and suicide risk than heterosexual youths

(1–3). In 2015, CDC’s Youth Risk Behavior Surveillance

System (YRBSS) added two new questions to the national

questionnaire regarding sexual identity and sexual behavior.

These questions facilitated the first nationally representative

estimates of the health behaviors and experiences of sexual

minority youths (students who identify as LGB or those who

have sexual contact with persons of the same or both sexes)

and affirmed the presence of substantial health disparities (i.e.,

differences in health outcomes between social groups driven

by unequal social or environmental circumstances) in violence

victimization and suicide risk between LGB and heterosexual

youths . Findings from the 2017 Youth Risk Behavior Survey

Corresponding author: Michelle M. Johns, PhD, National Center

for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention, CDC.

Telephone: 404-718-8858. E-mail: [email protected].

(YRBS) revealed that LGB high school students experienced

more bullying at school (33% among LGB students and 17%

among heterosexual students), more sexual dating violence

by dating partners (LGB, 16%; heterosexual, 6%), and more

suicide attempts (LGB, 23%; heterosexual, 5%) (3) than their

heterosexual peers.

Notably, the 2019 YRBS data collection cycle presented

the first opportunity for examining linear trends in violence

victimization and suicide risk trends for LGB students across

time in a nationally representative sample. Few studies have

examined whether prevalence of violence victimization and

suicide risk varies among LGB youths over time (4). School

environments in the United States might be improving in their

ability to meet the needs of LGB youths (5); for example, recent

surveillance data from CDC’s 2018 School Health Profiles,

which include representative data from 43 states, provide

evidence that many schools are implementing supportive

practices. In the 2018 School Health Profiles, an average of

78.5% of schools across U.S. states included in the sample

reported identifying safe spaces for LGB and transgender

Supplement

20 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

and questioning youths, and an average of 96.1% of schools

across these states prohibited harassment based on a student’s

perceived or actual sexual orientation or gender identity

(6). A recent study (4) that pooled local YRBS data during

2009–2017 to examine trends in suicide risk found that reports

among LGB youths of suicide risk might be decreasing but

that LGB students still are as much as three times more likely

to have attempted suicide than heterosexual students. National

U.S. trends have not been estimated.

LGB youths are a heterogenous population with intersecting

social identities (e.g., sex, race/ethnicity, or gender identity), and

important differences might exist among LGB youths regarding

risk for violence and suicide. For example, LGB females appear

to be at higher risk for dating and sexual violence than LGB

males (7,8). In examinations of racial/ethnic differences among

LGB youths, clear patterns of differences in experiences of

violence victimization and suicide risk are less consistent (9,10).

For example, one study of interpersonal violence among sexual

minorities reported that physical victimization from an intimate

partner was 1–4 times higher among non-white youths than

among white youths (9). Another study reported that non-

Hispanic white and Hispanic LGB youths were more likely to

be bullied than non-Hispanic white heterosexual youths but

that non-Hispanic black LGB youths were not more likely to

be bullied than non-Hispanic white heterosexual youths (10).

This same study reported that all LGB youths, regardless of

race/ethnicity, were at increased risk for suicidal ideation (10).

More systematic evaluations of the within-group differences in

violence victimization and suicide risk behaviors among sexual

minority youths are warranted.

This analysis contributes to the evidence base regarding LGB

students, violence victimization, and suicide risk. YRBS data

were used to examine national trends in violence victimization

and suicide risk among high school students by self-reported

sexual identity and evaluated differences among LGB students

by sex and race/ethnicity. The analysis was guided by the

following four questions:

1. How did the prevalence of violence victimization

and suicide risk among LGB students vary during

2015–2019?

2. To what extent did violence victimization and suicide

risk trends differ from these trends among heterosexual

students during the same period?

3. Among LGB students, to what extent did violence

victimization and suicide risk trends vary by sex (male

or female)?

4. Among LGB students, to what extent did violence

victimization and suicide risk trends vary by race/ethnicity

(non-Hispanic black, non-Hispanic white, or Hispanic)?

Methods

Data Source

This report includes data from the 2015 (n = 15,624), 2017

(n = 14,765), and 2019 (n = 13,677) cycles of the national

YRBS (pooled n = 44,066), a cross-sectional, school-based

survey conducted biennially since 1991. Each survey year,

CDC collects data from a nationally representative sample

of public and private school students in grades 9–12 in the

50 U.S. states and the District of Columbia. Additional

information about YRBS sampling, data collection, response

rates, and processing is available in the overview report of this

supplement (11). The prevalence estimates for all questions

on violence victimization and suicide risk for the overall

study population and by sex, race/ethnicity, grade, and sexual

orientation are available at https://nccd.cdc.gov/youthonline/

App/Default.aspx. The full YRBS questionnaire is available at

https://www.cdc.gov/healthyyouth/data/yrbs/pdf/2019/2019_

YRBS-National-HS-Questionnaire.pdf.

Measures

All measures analyzed for this report are provided

(Table 1). Students responded to seven questions about

violence victimization, including ever experiencing forced

sexual intercourse; experiencing sexual dating violence, physical

dating violence, bullying at school, electronic bullying, and

being threatened or injured with a weapon at school during

the previous 12 months; and missing school because of feeling

unsafe at or on the way to or from school during the previous

30 days. Students responded to five questions about suicide

risk during the previous 12 months, including having felt

persistently sad or hopeless; having seriously considered suicide;

and having made a suicide plan, having attempted suicide,

or having made a suicide attempt that had to be treated by

a doctor or nurse. Students responded to five demographic

questions relating to sex, sexual identity, grade, race, and

ethnicity, which were used as covariates and to create relevant

strata in all trend analyses.

Analysis

Data from the 2015, 2017, and 2019 national YRBS were

examined for trends in the prevalence among LGB students in

experiences of violence victimization and indicators of suicide

risk. Data were analyzed by using SAS (version 9.4; SAS

Institute) and SUDAAN (version 11.0.0; RTI International) to

account for the complex sampling designs. Data were assessed

using complete case analysis; missing data were not imputed.

All outcomes were dichotomized as either yes or no, ≥1 time

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 21

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 1. Measures for demographic characteristics, violence victimization, and suicide risk behaviors among high school students — Youth

Risk Behavior Survey, United States, 2019

Construct Measure

Demographic characteristics

Sexual identity Which of the following best describes you?

A. Heterosexual (straight)

B. Gay or lesbian

C. Bisexual

D. Not sure

Sex at birth What is your sex?

A. Female

B. Male

Race What is your race? (Select one or more responses.)

A. American Indian or Alaska Native

B. Asian

C. Black or African American

D. Native Hawaiian or Other Pacific Islander

E. White

Ethnicity Are you Hispanic or Latino?

A. Yes

B. No

Grade In what grade are you?

A. 9th grade

B. 10th grade

C. 11th grade

D. 12th grade

E. Ungraded or other grade

Violence victimization*

Forced sex Have you ever been physically forced to have sexual intercourse when you did not want to?

Sexual dating violence During the past 12 months, how many times did someone you were dating or going out with force you to do sexual things

that you did not want to do? (Count such things as kissing, touching, or being physically forced to have sexual intercourse.)

Physical dating violence During the past 12 months, how many times did someone you were dating or going out with physically hurt you on

purpose? (Count such things as being hit, slammed into something, or injured with an object or weapon.)

Bullying at school During the past 12 months, have you ever been bullied on school property?

Electronic bullying During the past 12 months, have you ever been electronically bullied? (Count being bullied through texting,

Instagram, Facebook, or other social media.)

Felt unsafe at, to, or from school During the past 30 days, on how many days did you not go to school because you felt you would be unsafe at school

or on your way to or from school?

Threatened or injured with a weapon

at school

During the past 12 months, how many times has someone threatened or injured you with a weapon, such as a gun,

knife, or club, on school property?

Suicide risk behaviors*

Persistent feelings of

sadness/hopelessness

During the past 12 months, did you ever feel so sad or hopeless almost every day for 2 weeks or more in a row that you

stopped doing some usual activities?

Seriously considered suicide During the past 12 months, did you ever seriously consider attempting suicide?

Made a suicide plan During the past 12 months, did you make a plan about how you would attempt suicide?

Attempted suicide During the past 12 months, how many times did you actually attempt suicide?

Suicide attempt requiring

medical treatment

If you attempted suicide during the past 12 months, did any attempt result in an injury, poisoning, or overdose that

had to be treated by a doctor or nurse?

* All violence victimization and suicide risk measures were dichotomized as either “yes” (i.e., ≥1 time, ≥1 day) or “no” (i.e., 0 days, 0 times).

or 0 times, or ≥1 day or 0 days. Weighted prevalence estimates

with 95% confidence intervals (CIs) were calculated by using

Taylor series linearization to produce nationally representative

prevalence estimates for each survey year.

Logistic regression models were used to assess linear trends

in the prevalence of violence victimization and indicators

of suicide risk among LGB and heterosexual students for

2015–2019, controlling for sex, race/ethnicity, and grade.

Main effects odds ratios (ORs) comparing LGB students with

heterosexual students also were calculated for the 2015–2019

period. In subsequent models, sex-stratified (controlling

for race/ethnicity and grade) and race/ethnicity-stratified

(controlling for sex and grade) linear trends, were examined

for students self-identifying as LGB on the survey. Main

effects ORs comparing sex and race/ethnicity groups also were

calculated for these subsequent regression models. Linear trends

were considered statistically significant if p<0.05. Main effects

ORs were considered statistically significant if 95% CIs did

not include 1.0.

Supplement

22 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

Results

Violence Victimization

Among all students (Table 2), LGB students had greater

odds of violence victimization than heterosexual students

across all seven indicators, as evidenced by statistically

significant main effects of sexual identity on each indicator

(Table 2). Among LGB students, the percentage who reported

experiencing physical dating violence during 2015–2019

significantly decreased from 17.5% to 13.1%. No other

violence victimization outcomes varied significantly among

LGB students in this period.

Among LGB students stratified by sex (Table 3), male

students reported greater odds of feeling unsafe at or on the way

to or from school (aOR: 1.61) and being threatened or injured

with a weapon (aOR: 1.54) than female students. Conversely,

male LGB students reported reduced odds of electronic

bullying (aOR: 0.71), sexual dating violence (aOR: 0.66), and

forced sex (aOR: 0.51) than female LGB students. Among male

LGB students, the percentage reporting being threatened or

injured with a weapon at school significantly increased from

2015 (11.6%) to 2019 (15.9%), as did the percentage reporting

forced sex (2015: 8.0%; 2019: 15.6%). Among female LGB

students, the percentage reporting physical dating violence

significantly decreased from 2015 (16.9%) to 2019 (12.1%).

Among LGB students stratified by race (Table 4), non-

Hispanic black (black) and Hispanic students reported higher

odds of feeling unsafe at or on the way to or from school

than non-Hispanic white (white) students (aOR: 1.63 and

aOR: 1.46, respectively), and black students also reported

greater odds of being threatened or injured with a weapon than

white students (aOR: 1.60). With regard to bullying, black

and Hispanic LGB students reported reduced odds of both

bullying at school (black, aOR: 0.31; Hispanic, aOR: 0.56) and

electronic bullying (black, aOR: 0.41; Hispanic, aOR: 0.55),

compared with white LGB students. Black LGB students also

reported reduced odds of sexual dating violence, compared

with white LGB students (aOR: 0.44). The only significant

trend among violence models stratified by race/ethnicity was

among Hispanic LGB students, who had reduced percentage

of reporting experiencing physical dating violence in 2019

(9.8%), compared with 2015 (22.6%).

Suicide Risk

Among all students (Table 2), LGB students had greater

odds of suicide risk than heterosexual students across all five

indicators, as evidenced by significant main effects for each

variable. The percentage of LGB students reporting these

outcomes did not vary significantly during 2015–2019.

Among LGB students stratified by sex (Table 3), male

students had lower odds of all five suicide risk indicators than

female students. Among female LGB students, the percentage

reporting suicide attempts decreased significantly from 2015

(32.8%) to 2019 (23.6%). All other trends in suicide risk in

these sex-stratified models remained stable.

Among LGB students stratified by race (Table 4), black

and Hispanic students had lower odds than white students

of reporting persistent feelings of sadness or hopelessness

(black, aOR: 0.42; Hispanic, aOR: 0.69) and seriously

considering attempting suicide (black, aOR: 0.43; Hispanic,

aOR: 0.65). Black LGB students also had lower odds than

white LGB students of making a suicide plan (aOR: 0.61).

The percentage of LGB students reporting these outcomes in

the race/ethnicity-stratified models did not vary significantly

during 2015–2019.

Discussion

Overall, these results underscore that LGB students continue

to have a greater prevalence of violence victimization and

suicidal behavior than their heterosexual peers. The higher

prevalence of violence and suicide among LGB students is

consistent with results from other studies regarding sexual

minorities and minority stress (12,13). Minority stress is the

preeminent framework for understanding disparities among

sexual minorities and refers to the process by which social

stigma directed toward LGB and other nonheterosexual

persons is enacted through external stressors (e.g., violence,

discrimination, or harassment) and internal stressors (e.g.,

identity concealment or expectations of rejection) (12). Both

types of stress shape mental and physical health (12,14), and

the impact of violence victimization on LGB youths (15) and

its connection to elevated suicide risk is well-documented

(16). LGB students’ disproportionate experience of violence

victimization and suicide risk, compared with their heterosexual

peers in this study, underscores the continued relevance of

minority stress among LGB youths and the continued public

health need for action that addresses these sizeable disparities.

Notably, the proportion of LGB students experiencing

violence victimization or suicide risk remained fairly stable

during 2015–2019. One exception is reports of physical dating

violence; fewer LGB students reported experiencing physical

dating violence in 2019 than in 2015. This downward trajectory

of physical dating violence appears to be a continuation of

an already documented population trend of a decrease in

experiences of dating violence among adolescents (17), and its

detection among LGB youths is promising. Regarding suicide

risk, a recent study examined local trends since 2009 and

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 23

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 2. Trends in the prevalence of violence victimization and suicide risk behaviors among high school students, by self-identied sexual

identity — Youth Risk Behavior Survey, United States, 2015–2019*

Health risk behavior

Main effect 2015 2017 2019 Linear trend

aOR (95% CI) % (95% CI) % (95% CI) % (95% CI) Beta p value

†

Violence victimization

Feeling unsafe at school (past 30 days)

Lesbian, gay, or bisexual 1.98 (1.70–2.30) 12.5 (10.2–15.3) 10.0 (8.1–12.3) 13.5 (11.0–16.5) 0.0619 0.65

Heterosexual 1.0 (Ref.) 4.6 (3.9–5.4) 6.1 (5.1–7.3) 7.5 (6.3–8.9) 0.3749 0.00

Threatened or injured with a weapon at school (past 12 months)

Lesbian, gay, or bisexual 2.09 (1.80–2.43) 10.0 (7.9–12.7) 9.4 (7.4–11.8) 11.9 (9.3–15.2) 0.2463 0.12

Heterosexual 1.0 (Ref.) 5.1 (4.5–5.9) 5.4 (4.8–6.0) 6.3 (5.5–7.3) 0.1629 0.02

Bullied at school (past 12 months)

Lesbian, gay, or bisexual 2.10 (1.87–2.37) 34.2 (29.6–39.0) 33.0 (27.4–39.0) 32.0 (29.5–34.6) –0.0847 0.28

Heterosexual 1.0 (Ref.) 18.8 (17.3–20.3) 17.1 (16.1–18.2) 17.1 (15.7–18.7) –0.0800 0.10

Electronically bullied (past 12 months)

Lesbian, gay, or bisexual 1.94 (1.72–2.20) 28.0 (24.0–32.3) 27.1 (23.1–31.4) 26.6 (23.3–30.2) –0.0775 0.43

Heterosexual 1.0 (Ref.) 14.2 (13.1–15.3) 13.3 (12.4–14.4) 14.1 (12.9–15.4) –0.0047 0.92

Physical dating violence (past 12 months)

Lesbian, gay, or bisexual 2.06 (1.77–2.40) 17.5 (14.4–21.2) 17.2 (14.3–20.5) 13.1 (10.5–16.1) –0.2264 0.04

Heterosexual 1.0 (Ref.) 8.3 (7.5–9.3) 6.4 (5.8–7.1) 7.2 (6.2–8.3) –0.1448 0.047

Sexual dating violence (past 12 months)

Lesbian, gay, or bisexual 2.08 (1.69–2.57) 22.7 (18.0–28.2) 15.8 (12.3–20.1) 16.4 (12.7–20.9) –0.2420 0.15

Heterosexual 1.0 (Ref.) 9.1 (8.2–10.0) 5.5 (4.8–6.3) 6.7 (5.9–7.5) –0.2785 <0.001

Forced sexual intercourse (lifetime)

Lesbian, gay, or bisexual 3.31 (2.90–3.77) 17.8 (14.4–21.8) 21.9 (19.0–25.0) 19.4 (16.2–23.1) 0.0650 0.59

Heterosexual 1.0 (Ref.) 5.4 (4.6–6.4) 5.4 (4.7–6.2) 5.5 (4.9–6.2) 0.0200 0.82

Suicide risk behaviors

Persistent feelings of sadness or hopelessness (past 12 months)

Lesbian, gay, or bisexual 3.60 (3.22–4.03) 60.4 (55.1–65.4) 63.0 (59.5–66.5) 66.3 (62.2–70.2) 0.1566 0.13

Heterosexual 1.0 (Ref.) 26.4 (24.6–28.4) 27.5 (25.9–29.2) 32.2 (30.8–33.7) 0.1949 0.00

Seriously considered attempting suicide (past 12 months)

Lesbian, gay, or bisexual 4.51 (4.07–4.99) 42.8 (38.4–47.3) 47.7 (43.7–51.8) 46.8 (43.1–50.6) 0.0936 0.26

Heterosexual 1.0 (Ref.) 14.8 (13.7–15.9) 13.3 (12.5–14.3) 14.5 (13.4–15.7) –0.0242 0.62

Made a suicide plan (past 12 months)

Lesbian, gay, or bisexual 4.28 (3.84–4.77) 38.2 (34.0–42.6) 38.0 (34.5–41.7) 40.2 (36.6–44.0) 0.0646 0.45

Heterosexual 1.0 (Ref.) 11.9 (10.8–13.1) 10.4 (9.3–11.7) 12.1 (11.1–13.1) 0.0031 0.96

Attempted suicide (past 12 months)

Lesbian, gay, or bisexual 4.54 (3.89–5.28) 29.4 (25.7–33.3) 23.0 (18.6–28.0) 23.4 (20.0–27.1) –0.1901 0.06

Heterosexual 1.0 (Ref.) 6.4 (5.6–7.3) 5.4 (4.6–6.4) 6.4 (5.6–7.4) –0.0148 0.85

Suicide attempt requiring medical treatment (past 12 months)

Lesbian, gay, or bisexual 3.78 (3.02–4.73) 9.4 (7.3–12.1) 7.5 (5.7–9.8) 6.3 (4.8–8.3) –0.2852 0.07

Heterosexual 1.0 (Ref.) 2.0 (1.5–2.7) 1.7 (1.4–2.1) 1.7 (1.4–2.2) –0.1197 0.40

Abbreviations: aOR = adjusted odds ratio; CI = confidence interval; Ref. = referent group.

* Logistic regression models were used to assess linear trends in the prevalence of violence victimization, and indicators of suicide risk among lesbian, gay, or bisexual

students and heterosexual students for 2015–2019, controlling for sex, race/ethnicity, and grade.

†

Statistical significance is defined as p<0.05 or a 95% CI that does not include 1.0.

reported a decrease in reported suicide risk behaviors among

LGB students (4). The national trends reported in this analysis

warrant continued monitoring over time to assess whether the

downward trajectory in local contexts (4) reflects the general

trajectory of suicide risk and LGB youths nationally.

Results from sex-stratified models highlight important

differences between male and female experiences of violence

victimization among LGB students. In this sample, male LGB

students were more likely to report feeling unsafe at school

and being threatened with a weapon; conversely, female LGB

students were more likely to report bullying both at school

and electronically. This finding is consistent with observational

studies of bullying during adolescence; males tend to report

more physical forms of bullying and harassment, whereas

females tend to report experiencing more verbal and social

bullying (18). In addition, female LGB students had a greater

prevalence of sexual dating violence and forced sex than male

LGB students. This sex difference is also consistent with what

is known about dating and sexual violence among LGB youths

(7) and mirrors national trends in dating and sexual violence,

in which females are consistently disproportionately affected

by these types of victimization (19).

Of concern, the percentage of male LGB students who

reported being threatened or injured with a weapon at school

and who reported forced sexual intercourse significantly

increased over time. Although both male and female LGB

students are negatively affected by violence, these percentages

highlight an increasing trend in violence among male LGB

Supplement

24 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 3. Trends in violence victimization and suicide risk behaviors among lesbian, gay, and bisexual high school students, by sex and sexual

identity — Youth Risk Behavior Survey, United States, 2015–2019*

Health risk behavior

Main effect 2015 2017 2019 Linear trend

aOR (95% CI) % (95% CI) % (95% CI) % (95% CI) Beta p value

†

Violence victimization

Feeling unsafe at school (past 30 days)

Gay or bisexual male 1.61 (1.14–2.28) 15.5 (9.5–24.4) 12.3 (7.4–19.6) 18.3 (12.4–26.1) 0.1623 0.55

Lesbian or bisexual female 1.0 (Ref.) 10.8 (8.6–13.5) 9.1 (6.9–11.9) 11.5 (9.5–14.0) 0.0435 0.72

Threatened or injured with a weapon at school (past 12 months)

Gay or bisexual male 1.54 (1.14–2.08) 11.6 (7.5–17.5) 14.6 (9.8–21.2) 15.9 (11.4–21.8) 0.3973 0.04

Lesbian or bisexual female 1.0 (Ref.) 9.1 (6.6–12.4) 7.4 (5.6–9.7) 10.6 (8.1–13.9) 0.1944 0.29

Bullied at school (past 12 months)

Gay or bisexual male 0.86 (0.69–1.08) 26.3 (19.4–34.7) 35.0 (25.4–45.9) 31.7 (25.7–38.4) 0.0942 0.58

Lesbian or bisexual female 1.0 (Ref.) 37.2 (32.7–42.0) 32.2 (26.9–38.1) 32.0 (28.6–35.7) –0.1550 0.09

Electronically bullied (past 12 months)

Gay or bisexual male 0.71 (0.57–0.89) 22.4 (16.3–30.1) 22.3 (16.5–29.4) 25.5 (18.7–33.8) 0.0786 0.71

Lesbian or bisexual female 1.0 (Ref.) 30.5 (26.0–35.4) 28.5 (24.4–33.1) 27.1 (23.7–30.7) –0.1076 0.28

Physical dating violence (past 12 months)

Gay or bisexual male 1.06 (0.72–1.58) 19.9 (12.9–29.4) 16.8 (10.0–27.0) 15.9 (9.4–25.6) –0.0798 0.78

Lesbian or bisexual female 1.0 (Ref.) 16.9 (13.9–20.4) 16.9 (13.5–21.0) 12.1 (9.3–15.6) –0.2638 0.04

Sexual dating violence (past 12 months)

Gay or bisexual male 0.66 (0.44–0.98) 20.9 (12.7–32.6) 13.5 (7.5–23.0) 10.3 (5.6–18.3) –0.4638 0.20

Lesbian or bisexual female 1.0 (Ref.) 22.6 (18.0–27.9) 16.3 (12.8–20.6) 18.2 (13.6–23.8) –0.2166 0.18

Forced sexual intercourse (lifetime)

Gay or bisexual male 0.51 (0.38–0.68) 8.0 (4.8–13.1) 15.6 (10.3–22.9) 15.6 (10.7–22.0) 0.4388 0.047

Lesbian or bisexual female 1.0 (Ref.) 21.1 (17.0–25.9) 23.7 (20.6–27.2) 21.0 (17.3–25.4) –0.0203 0.87

Suicide risk behaviors

Persistent feelings of sadness or hopelessness (past 12 months)

Gay or bisexual male 0.39 (0.33–0.47) 43.9 (35.9–52.3) 45.5 (38.9–52.2) 53.5 (46.3–60.4) 0.2667 0.12

Lesbian or bisexual female 1.0 (Ref.) 66.5 (61.4–71.2) 68.8 (65.1–72.2) 70.5 (66.6–74.2) 0.1167 0.25

Seriously considered attempting suicide (past 12 months)

Gay or bisexual male 0.59 (0.47–0.73) 32.7 (23.6–43.3) 37.0 (31.5–42.8) 40.4 (33.9–47.2) 0.1960 0.30

Lesbian or bisexual female 1.0 (Ref.) 46.6 (42.1–51.1) 51.0 (46.1–55.9) 49.0 (44.8–53.3) 0.0553 0.55

Made a suicide plan (past 12 months)

Gay or bisexual male 0.57 (0.46–0.71) 27.0 (20.3–34.9) 28.7 (22.8–35.5) 33.0 (26.4–40.3) 0.2350 0.21

Lesbian or bisexual female 1.0 (Ref.) 42.0 (37.1–47.2) 40.8 (36.8–45.0) 42.4 (38.4–46.4) 0.0130 0.89

Attempted suicide (past 12 months)

Gay or bisexual male 0.73 (0.55–0.96) 19.4 (13.6–27.0) 18.3 (11.5–27.9) 23.8 (17.8–31.1) 0.1626 0.45

Lesbian or bisexual female 1.0 (Ref.) 32.8 (28.1–37.9) 23.7 (19.4–28.5) 23.6 (20.0–27.6) –0.2929 0.01

Suicide attempt requiring medical treatment (past 12 months)

Gay or bisexual male 0.63 (0.42–0.97) 7.0 (3.6–13.1) 3.8 (1.9–7.3) 5.9 (3.2–10.6) –0.2535 0.53

Lesbian or bisexual female 1.0 (Ref.) 10.3 (7.8–13.4) 8.2 (6.2–10.7) 6.6 (5.0–8.7) –0.2977 0.05

Abbreviations: aOR = adjusted odds ratio; CI = confidence interval; Ref. = referent group.

Logistic regression models were used to assess linear trends in the prevalence of violence victimization and indicators of suicide risk among lesbian, gay, and bisexual

high school students, by sex and self-identified sexual identity for 2015–2019, controlling for race/ethnicity and grade.

†

Statistical significance is defined as p<0.05 or a 95% CI that does not include 1.0.

students. Among adults, gay men are at greater risk for physical

violence than lesbians (20), and the increasing prevalence

in these types of violence among male LGB students might

suggest an increasing disparity between sexual minority men

and women in violence victimization. Continued monitoring of

this trend is needed, in both adolescent and adult populations.

Female LGB students reported fewer experiences of physical

dating violence over time, whereas male LGB students’ reports

of experiencing physical dating violence remained stable. This

pattern might indicate that the overall reduction in physical

dating violence in the population is not occurring among

sexual minority males, which might be supported by the data

regarding being threatened or injured with a weapon and

experiencing forced sex. An assessment of the ways that violence

in schools and in dating relationships affects sexual minority

males is warranted, both through research to understand

underlying mechanisms and in practice to ensure violence

prevention programming is directly addressing the needs of

sexual minority males.

Despite a trend of decreasing suicide attempts among

LGB females during 2015–2019, LGB females consistently

reported more suicide risk behaviors than LGB males. This

pattern echoes larger population trends in which both adult

and youth females report more suicidal ideation than adult and

youth males (21). Notably, this same literature finds that males

experience more deaths by suicide (i.e., completed suicide

attempts) than females (21); thus, an important remaining

question for LGB youths is whether these sex-specific patterns

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 25

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 4. Trends in violence victimization and suicide risk behaviors among lesbian, gay, and bisexual high school students, by

race/ethnicity — Youth Risk Behavior Survey, United States, 2015–2019*

Health risk behavior

Main effect 2015 2017 2019 Linear trend

aOR (95% CI) % (95% CI) % (95% CI) % (95% CI) Beta p value

†

Violence victimization

Feeling unsafe at school (past 30 days)

Black, non-Hispanic 1.63 (1.13–2.35) 17.8 (11.4–26.6) 12.5 (7.4–20.4) 15.2 (8.3–26.2) –0.0269 0.93

Hispanic 1.46 (1.07–1.99) 15.6 (11.5–21.0) 12.4 (8.4–18.0) 13.7 (9.5–19.5) –0.0688 0.75

White, non-Hispanic 1.0 (Ref.) 9.0 (6.7–12.0) 8.6 (6.4–11.6) 11.1 (8.0–15.1) 0.1309 0.41

Threatened or injured with a weapon at school (past 12 months)

Black, non-Hispanic 1.60 (1.07–2.41) 15.6 (8.0–28.1) 15.7 (11.7–20.7) 12.9 (7.2–22.0) –0.0494 0.89

Hispanic 0.89 (0.63–1.27) 9.0 (5.5–14.3) 9.4 (6.7–13.0) 7.7 (4.9–11.8) –0.1270 0.61

White, non-Hispanic 1.0 (Ref.) 8.2 (5.5–12.0) 7.1 (4.9–10.1) 12.9 (8.6–18.9) 0.4292 0.09

Bullied at school (past 12 months)

Black, non-Hispanic 0.31 (0.22–0.44) 21.4 (12.2–34.7) 17.2 (10.6–26.9) 18.2 (12.2–26.2) –0.0651 0.82

Hispanic 0.56 (0.45–0.70) 31.2 (25.1–38.0) 26.6 (21.0–33.1) 27.6 (23.2–32.5) –0.1077 0.44

White, non-Hispanic 1.0 (Ref.) 42.2 (34.8–50.0) 40.8 (32.8–49.3) 37.6 (33.6–41.7) –0.1265 0.31

Electronically bullied (past 12 months)

Black, non-Hispanic 0.41 (0.31–0.54) 17.0 (12.6–22.6) 16.3 (12.1–21.5) 20.5 (12.5–31.6) 0.1336 0.61

Hispanic 0.55 (0.42–0.71) 24.8 (17.8–33.4) 17.7 (13.3–23.2) 25.4 (20.1–31.5) 0.0619 0.75

White, non-Hispanic 1.0 (Ref.) 36.0 (29.1–43.5) 36.0 (30.1–42.4) 28.2 (22.8–34.1) –0.2484 0.09

Physical dating violence (past 12 months)

Black, non-Hispanic 1.19 (0.81–1.75) 14.2 (8.8–22.2) 23.8 (15.5–34.7) 11.6 (7.0–18.6) –0.1425 0.53

Hispanic 1.13 (0.83–1.54) 22.6 (16.3–30.4) 19.1 (14.2–25.2) 9.8 (5.8–16.2) –0.7080 0.003

White, non-Hispanic 1.0 (Ref.) 15.3 (11.9–19.5) 14.1 (10.4–18.7) 13.8 (10.6–17.6) –0.0883 0.55

Sexual dating violence (past 12 months)

Black, non-Hispanic 0.44 (0.27–0.72) 20.4 (11.5–33.7) 6.4 (3.3–12.0) 10.3 (5.5–18.7) –0.7987 0.11

Hispanic 0.97 (0.68–1.39) 23.0 (14.8–34.1) 18.6 (11.4–29.0) 18.3 (11.3–28.2) –0.2334 0.41

White, non-Hispanic 1.0 (Ref.) 22.3 (16.8–29.1) 18.2 (13.7–23.8) 16.7 (11.8–23.1) –0.2426 0.21

Forced sexual intercourse (lifetime)

Black, non-Hispanic 0.92 (0.66–1.29) 16.0 (8.3–28.7) 23.9 (17.7–31.6) 15.4 (9.8–23.4) 0.0759 0.77

Hispanic 1.10 (0.85–1.43) 24.0 (18.7–30.3) 21.8 (17.6–26.8) 19.1 (13.3–26.7) –0.2740 0.17

White, non-Hispanic 1.0 (Ref.) 15.5 (11.5–20.6) 21.0 (16.8–26.1) 21.3 (16.6–26.9) 0.2531 0.12

Suicide risk behaviors

Persistent feelings of sadness or hopelessness (past 12 months)

Black, non-Hispanic 0.42 (0.33–0.55) 44.8 (35.2–54.7) 52.1 (42.6–61.4) 51.1 (44.6–57.5) 0.1107 0.56

Hispanic 0.69 (0.54–0.89) 58.2 (50.6–65.5) 61.2 (52.9–69.0) 64.1 (56.0–71.4) 0.1766 0.30

White, non-Hispanic 1.0 (Ref.) 67.4 (60.3–73.8) 66.3 (60.9–71.3) 71.6 (65.7–76.8) 0.1679 0.24

Seriously considered attempting suicide (past 12 months)

Black, non-Hispanic 0.43 (0.34–0.55) 34.4 (25.7–44.3) 28.4 (21.3–36.7) 35.1 (29.2–41.4) –0.0389 0.85

Hispanic 0.65 (0.54–0.79) 40.7 (34.7–46.9) 45.3 (38.7–52.0) 39.2 (33.2–45.6) –0.0625 0.63

White, non-Hispanic 1.0 (Ref.) 48.9 (42.2–55.7) 54.1 (50.4–57.6) 52.4 (47.1–57.7) 0.0959 0.44

Made a suicide plan (past 12 months)

Black, non-Hispanic 0.61 (0.46–0.82) 32.9 (23.6–43.7) 24.1 (16.8–33.2) 36.0 (28.8–43.9) 0.0564 0.81

Hispanic 0.86 (0.69–1.06) 37.9 (31.3–45.0) 35.5 (29.5–42.0) 40.0 (32.8–47.7) 0.0765 0.62

White, non-Hispanic 1.0 (Ref.) 40.1 (34.5–46.0) 42.8 (37.7–48.0) 40.3 (35.8–45.0) 0.0044 0.97

Attempted suicide (past 12 months)

Black, non-Hispanic 0.97 (0.70–1.34) 29.2 (23.1–36.1) 20.7 (12.5–32.3) 27.2 (18.0–38.8) 0.0293 0.91

Hispanic 1.06 (0.83–1.37) 31.4 (26.4–36.9) 24.6 (18.3–32.2) 23.2 (17.3–30.4) –0.2610 0.12

White, non-Hispanic 1.0 (Ref.) 28.6 (23.1–34.7) 21.8 (16.4–28.4) 22.3 (18.1–27.3) –0.2078 0.13

Suicide attempt requiring medical treatment (past 12 months)

Black, non-Hispanic 0.83 (0.50–1.40) 5.9 (3.2–10.7) 6.4 (3.1–12.9) 7.3 (3.0–16.6) 0.2889 0.50

Hispanic 1.20 (0.83–1.75) 13.3 (9.2–19.0) 8.7 (5.2–14.3) 6.8 (4.3–10.5) –0.4683 0.05

White, non-Hispanic 1.0 (Ref.) 9.3 (6.4–13.2) 7.5 (5.2–10.8) 5.6 (3.5–8.7) –0.3545 0.11

Abbreviations: aOR = adjusted odds ratio; CI = confidence interval; Ref. = referent group.

* Logistic regression models were used to assess linear trends in the prevalence of violence victimization and indicators of suicide risk among lesbian, gay, and bisexual

high school students, by race/ethnicity for 2015–2019, controlling for sex and grade.

†

Statistical significance is defined as p<0.05 or a 95% CI that does not include 1. 0.

in deaths by suicide hold in this group; however, reliable data

regarding sexual orientation and rates of death by suicide are

unavailable. Such data could aid in further illuminating how

LGB youths are affected by suicide risk behaviors and guide

interventions for addressing this public health concern.

In models stratified by race/ethnicity, black and Hispanic

LGB students were more likely to feel unsafe and were more

likely to be threatened or injured with a weapon than white

LGB students. This finding might highlight black and Hispanic

LGB students being at greater risk for the forms of victimization

Supplement

26 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

that directly compromise physical safety (18). Conversely,

white LGB students were more likely to report school and

electronic bullying, indicating they might be at greater risk

for verbal and social victimization. Although the types of

racial/ethnic disparities in violence victimization presented in

this report do not mirror those reported in previous studies

(9,10), these findings underscore that differences by race/

ethnicity among sexual minority youths exist. Schools seeking

to address victimization through policies and practices designed

to address safety concerns for LGB students can benefit from

acknowledging differences in the experiences of LGB youths

across races/ethnicities and ensuring all youths are served

through these intervention strategies.

Regarding suicide risk, although a significantly lower

percentage of black and Hispanic LGB youths reported feeling

sad and hopeless or considering suicide than white LGB

youths, no differences existed among races/ethnicities in suicide

attempts or medically serious suicide attempts. These findings

are similar to those from other studies highlighting that all

LGB youths are at increased risk for suicide, regardless of race/

ethnicity (10) and might again highlight the mental health

impact of minority stress among all racial/ethnic groups (12).

Limitations

General limitations for the YRBS are available in the

overview report of this supplement (11). The findings in this

report are subject to at least five additional limitations. First,

although three cycles of national data to examine trends among

LGB youths are available, the brief 2015–2019 period might

be inadequate to assess trends. Continued monitoring of these

indicators over time to detect progress regarding disparities

experienced by LGB high school students is needed. Second,

the overall proportion of students identifying as LGB was

small: 2015, 8.3% (n = 1,246); 2017, 10.9% (n = 1,494);

and 2019, 11.7% (n = 1,531). Therefore, these analyses

might be underpowered for detecting statistical differences

in trends in models stratified by sex and race/ethnicity. As

more data are collected from LGB youths in future cycles

of the national YRBS, pooling data across cycles to improve

statistical power will be essential for increasing the likelihood

of detecting trends in stratified models. Third, this report does

not include differences in violence victimization and suicide

risk for students who identified their sexual identity as “not

sure” or across sexual behavior categories; future studies might

benefit from assessing these youths to further understand the

experiences of sexual minority students, violence victimization,

and suicide risk. Fourth, by pooling 2015–2019 data, the aOR

for the difference between groups on all outcomes might mask

heterogeneity over time within each subpopulation (e.g., the

size of the difference between LGB and heterosexual students

might vary between years); however, a disparity between LGB

and heterosexual students on these outcomes has been observed

since sexual identity data began to be collected on the national

YRBS in 2015. Finally, three survey measures had relatively large

amounts of missing data in 2019: forced sex (approximately

2,400 observations), sexual dating violence (approximately 3,400

observations), and attempted suicide with injury (approximately

4,900 observations). Most of these missing data can be attributed

to some selected schools administering YRBS questionnaire

versions that did not include these questions. Consequently, not

all students in the national sample were given the opportunity

to answer these questions and were counted as missing.

Conclusion

These findings highlight the continued need for policies and

practices within school environments that reduce victimization

and bolster the mental health of LGB students. Substantial

evidence exists for the role of antiharassment policies, gay-

straight alliances (or other student-led clubs designed to

support sexual minority students), and programs aimed at

improving staff support of LGB students in improving school

environments for these students (22). In addition to in-school

programs and policies, schools might consider engagement

with community organizations and stakeholders to collaborate

on implementation of comprehensive violence and suicide

prevention strategies that address a range of risk and protective

factors at the individual, relationship, community, and societal

levels. Comprehensive packages designed to inform these

prevention efforts are available from CDC (https://www.

cdc.gov/violenceprevention/pub/technical-packages.html).

For example, comprehensive approaches to suicide reduction

help to prevent suicide risk, support persons at increased risk,

prevent reattempts, and help survivors of suicide loss. When

refining such practices to meet the needs of LGB students,

special consideration of the impact of physical violence on LGB

males, suicide risk among LGB females, and the interactions

between race/ethnicity and these outcomes is warranted.

Furthermore, continued monitoring of these disparities

between LGB and heterosexual students over time is needed

until these disparities can be eradicated.

Conflicts of Interest

All authors have completed and submitted the International

Committee of Medical Journal Editors form for disclosure of

potential conflicts of interest. No potential conflicts of interest

were disclosed.

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 27

US Department of Health and Human Services/Centers for Disease Control and Prevention

References

1. Johns MM, Lowry R, Rasberry CN, et al. Violence victimization,

substance use, and suicide risk among sexual minority high school

students—United States, 2015–2017. MMWR Morb Mortal Wkly Rep

2018;67:1211–5. https://doi.org/10.15585/mmwr.mm6743a4

2. Kann L, Olsen EOM, McManus T, et al. Sexual identity, sex of sexual

contacts, and health-related behaviors among students in grades 9–12—

United States and selected sites, 2015. MMWR Surveill Summ 2016;65.

https://doi.org/10.15585/mmwr.ss6509a1

3. Kann L, McManus T, Harris WA, et al. Youth risk behavior surveillance—

United States, 2017. MMWR Surveill Summ 2018;67(No. SS-8).

https://doi.org/10.15585/mmwr.ss6708a1

4. Raifman J, Charlton BM, Arrington-Sanders R, et al. Sexual

orientation and suicide attempt disparities among US adolescents:

2009–2017. Pediatrics 2020;145:e20191658. https://doi.org/10.1542/

peds.2019-1658

5. Demissie Z, Rasberry CN, Steiner RJ, Brener N, McManus T. Trends

in secondary schools’ practices to support lesbian, gay, bisexual,

transgender, and questioning students, 2008–2014. Am J Public Health

2018;108:557–64. https://doi.org/10.2105/AJPH.2017.304296

6. CDC. School Health Profiles 2018: characteristics of health programs

among secondary schools. Atlanta, GA: US Department of Health and

Human Services, CDC; 2019. https://www.cdc.gov/healthyyouth/data/

profiles/pdf/2018/CDC-Profiles-2018.pdf

7. Dank M, Lachman P, Zweig JM, Yahner J. Dating violence experiences

of lesbian, gay, bisexual, and transgender youth. J Youth Adolesc

2014;43:846–57. https://doi.org/10.1007/s10964-013-9975-8

8. Walters ML, Chen J, Breiding MJ. The National Intimate Partner

and Sexual Violence Survey (NISVS): 2010 findings on victimization

by sexual orientation. Atlanta, GA: US Department of Health and

Human Services, CDC, National Center for Injury Prevention and

Control; 2013. https://www.cdc.gov/violenceprevention/pdf/NISVS_

Report2010-a.pdf

9. Whitton SW, Newcomb ME, Messinger AM, Byck G, Mustanski B. A

longitudinal study of IPV victimization among sexual minority youth. J Interpers

Violence 2016;34:912–45. https://doi.org/10.1177/0886260516646093

10. Mueller AS, James W, Abrutyn S, Levin ML. Suicide ideation and

bullying among US adolescents: examining the intersections of

sexual orientation, gender, and race/ethnicity. Am J Public Health

2015;105:980–5. https://doi.org/10.2105/AJPH.2014.302391

11. Underwood JM, Brener N, Thornton J, et al. Overview and methods

for the Youth Risk Behavior Surveillance System—United States, 2019.

In: Youth Risk Behavior Surveillance—United States, 2019. MMWR

Suppl 2020;69(No. Suppl 1).

12. Meyer IH, Frost DM. Minority stress and the health of sexual minorities.

In: Patterson CJ, D’Augelli AR, eds. Handbook of psychology and sexual

orientation. New York, NY: Oxford University Press; 2013: 252–66.

13. Bouris A, Everett BG, Heath RD, Elsaesser CE, Neilands TB. Effects

of victimization and violence on suicidal ideation and behaviors

among sexual minority and heterosexual adolescents. LGBT Health

2016;3:153–61. https://doi.org/10.1089/lgbt.2015.0037

14. Frost DM, Lehavot K, Meyer IH. Minority stress and physical health

among sexual minority individuals. J Behav Med 2015;38:1–8. https://

doi.org/10.1007/s10865-013-9523-8

15. Zaza S, Kann L, Barrios LC. Lesbian, gay, and bisexual adolescents:

population estimate and prevalence of health behaviors. JAMA

2016;316:2355–6. https://doi.org/10.1001/jama.2016.11683

16. Burton CM, Marshal MP, Chisolm DJ, Sucato GS, Friedman MS.

Sexual minority-related victimization as a mediator of mental health

disparities in sexual minority youth: a longitudinal analysis. J Youth

Adolesc 2013;42:394–402. https://doi.org/10.1007/s10964-012-9901-5

17. CDC. Youth Risk Behavior Survey: data summary & trends report

2007–2017. Atlanta, GA: US Department of Health and Human

Services, CDC, Division of Adolescent and School Health; 2018. https://

www.cdc.gov/healthyyouth/data/yrbs/pdf/trendsreport.pdf

18. Carbone-Lopez K, Esbensen F-A, Brick BT. Correlates and consequences

of peer victimization: gender differences in direct and indirect forms

of bullying. Youth Violence Juv Justice 2010;8:332–50. https://doi.

org/10.1177/1541204010362954

19. Smith SG, Zhang X, Basile KC, et al. The National Intimate Partner and

Sexual Violence Survey (NISVS): 2015 data brief—updated release. Atlanta,

GA: National Center for Injury Prevention and Control, CDC; 2018.

https://www.cdc.gov/violenceprevention/pdf/2015data-brief508.pdf

20. Stotzer RL. Comparison of hate crime rates across protected and

unprotected groups—an update. Los Angeles, CA: University of

California at Los Angeles, The Williams Institute; 2012. https://

escholarship.org/uc/item/43z1q49r

21. Beautrais AL. Gender issues in youth suicidal behaviour. Emerg Med (Fremantle)

2002;14:35–42. https://doi.org/10.1046/j.1442-2026.2002.00283.x

22. Johns MM, Poteat VP, Horn SS, Kosciw J. Strengthening our schools to

promote resilience and health among LGBTQ youth: emerging evidence

and research priorities from The State of LGBTQ Youth Health and

Wellbeing Symposium. LGBT Health 2019;6:146–55. https://doi.

org/10.1089/lgbt.2018.0109

Supplement

28 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

Interpersonal Violence Victimization Among High School Students —

Youth Risk Behavior Survey, United States, 2019

Kathleen C. Basile, PhD

; Heather B. Clayton, PhD

; Sarah DeGue, PhD

; John W. Gilford, PhD

; Kevin J. Vagi, PhD

;

Nicolas A. Suarez, MPH

; Marissa L. Zwald, PhD

; Richard Lowry, MD

Division of Violence Prevention, National Center for Injury Prevention and Control, CDC;

Division of Adolescent and School Health, National Center for HIV/AIDS,

Viral Hepatitis, STD, and TB Prevention, CDC;

Office of the Director, National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention, CDC

Abstract

Adolescent interpersonal violence victimization is an adverse childhood experience and a serious public health problem for

youths, their families, and communities. Violence victimization includes dating violence, sexual violence, and bullying. Youth

Risk Behavior Survey data for 2019 were used to examine physical and sexual dating violence; sexual violence by anyone; and

bullying victimization, whether on school property or electronic, of U.S. high school students by sex, race/ethnicity, and sexual

identity. In addition, this report explores frequency of dating violence and frequency of sexual violence among students who

reported these forms of victimization and presents composites of dating violence and bullying. Findings reveal that 8.2% of

students reported physical dating violence; 8.2% reported sexual dating violence; 10.8% reported sexual violence by anyone, of

which 50% of cases were by a perpetrator other than a dating partner; 19.5% reported bullying on school property; and 15.7%

reported electronic bullying victimization during the previous 12 months. Approximately one in eight students reported any

dating violence, and one in four reported any bullying victimization. Female students; lesbian, gay, and bisexual students; and

students not sure of their sexual identity reported the highest prevalence estimates across all five violence victimization types, any

and both forms of dating violence, and any bullying victimization. Non-Hispanic white students reported the highest prevalence

of bullying victimization. Among students experiencing physical or sexual dating violence or sexual violence by anyone, the most

common frequency reported was one time during the previous year; higher frequency was more prevalent among male students

compared with female students. These findings provide a contextual understanding of the prevalence of interpersonal violence

of U.S. high school students, highlighting those with highest prevalence. Findings can be used by public health professionals to

guide prevention efforts with youths in schools and communities.

Introduction

Interpersonal violence, or aggression perpetrated by another

person, including dating violence, sexual violence, and bullying,

is a serious problem for students, schools, and communities.

Violence can reoccur across the lifespan and is associated with

multiple health effects and negative health behaviors (e.g.,

risky sexual behaviors, substance misuse, and physical health

symptoms) (1). Victimization often begins during adolescence

and can be viewed as an adverse childhood experience (ACE). For

example, nationally representative data from adults during 2015

indicate that 43.2% of females and 51.3% of males who had

been raped were first raped before age 18 years (2). Prevalence

studies of adolescents confirm this finding. For example, a survey

Corresponding author: Kathleen C. Basile, PhD, Division of Violence

Prevention, National Center for Injury Prevention and Control, CDC.

Telephone: 404-398-8317; E-mail: [email protected].

of students in grades 7–12 found that 56% of females and 48%

of males reported some form of sexual violence victimization by

a peer (e.g., unwelcome comments, touching, or being forced

to do something sexual) during the 2010–11 school year (3).

Approximately 20% of adolescents reported physical dating

violence and 9% reported sexual dating violence (4). These

studies indicate that sexual violence during adolescence occurs

inside and outside of the dating context. In addition, 20% of

students in grades 6–12 reported bullying victimization during

the 2017 school year (5).

Scientific literature indicates that certain groups (e.g., females,

racial/ethnic minorities, and sexual minority youths)

disproportionately experience interpersonal violence during

adolescence (1). For instance, in a sample of northeastern

10th-grade students, sexual minority youths reported more

bullying, sexual violence, and dating violence victimization

than heterosexual youths, with sexual minority females

reporting particularly high levels (91% of sexual minority

females and 79% of sexual minority males reported at least

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 29

US Department of Health and Human Services/Centers for Disease Control and Prevention

one form of victimization) (6). Furthermore, in a study of

sexual violence victimization of college students, females had

higher odds of victimization than did males, and non-Hispanic

black (black) students and students of other races/ethnicities

had higher odds of victimization than did non-Hispanic

white (white) students; moreover, these racial differences were

greater for males. For females, Hispanics had lower odds of

sexual violence victimization than whites, and for males, no

substantial differences existed between Hispanics and whites

(7). Understanding these disparities in the experience of

violence victimization is crucial for identifying those at highest

risk and for guiding prevention efforts. Contextual factors also

are valuable in describing victimization (e.g., frequency of

victimization or co-occurrences of violence subtypes). These

factors increase understanding of these violence types and

further contextualize prevalence estimates. For example, in

a report using 2013 data, approximately 21% of female and

10% of male high school students who reported dating in the

previous year experienced sexual or physical dating violence,

and 6% of females and 3% of males experienced both physical

and sexual dating violence (8).

This report presents 2019 prevalence estimates for dating

violence, sexual violence, and bullying victimization of U.S.

high school students by sex, race/ethnicity, and sexual identity,

and includes frequency of dating violence and sexual violence

victimization by demographic characteristics. Combined

prevalence of different forms of dating violence and bullying

also is presented to provide the most current estimates of each

violence type. These findings can guide prevention efforts

in addressing adolescent interpersonal violence at different

levels of the social ecology (i.e., individual, relationship, and

community or societal levels).

TABLE 1. Violence victimization measures — Youth Risk Behavior Survey, United States, 2019

Violence victimization Questionnaire item Coding for analysis

Physical dating violence

victimization

“During the past 12 months, how many times did someone you were dating or going out

with physically hurt you on purpose? (Count such things as being hit, slammed into

something, or injured with an object or weapon.)” [Question excludes students who did

not date or go out with anyone during the previous 12 months.]

≥1 time versus 0 times;

1 time, 2 or 3 times, ≥4 times

Sexual dating violence

victimization*

“During the past 12 months, how many times did someone you were dating or going out

with force you to do sexual things that you did not want to do? (Count such things as

kissing, touching, or being physically forced to have sexual intercourse.)” [Question

excludes students who did not date or go out with anyone during the previous

12 months.]

≥1 time versus 0 times;

1 time, 2 or 3 times, ≥4 times

Sexual violence victimization

by anyone

†

“During the past 12 months, how many times did anyone force you to do sexual things

that you did not want to do? (Count such things as kissing, touching, or being physically

forced to have sexual intercourse.)”

≥1 time versus 0 times;

1 time, 2 or 3 times, ≥4 times

Bullied on school property “During the past 12 months, have you ever been bullied on school property?” Yes versus no

Electronically bullied “During the past 12 months, have you ever been electronically bullied?” Yes versus no

Abbreviation: YRBS = Youth Risk Behavior Survey.

* A total of 3,324 students had missing data for this variable, mostly attributed to the use of different versions of the YRBS questionnaire that did not include the

sexual violence questions in certain selected schools.

†

A total of 3,439 students had data missing for this variable, mostly attributed to the use of different versions of the YRBS questionnaire that did not include the sexual

violence questions in certain selected schools.

Methods

Data Source

This report includes data from CDC’s 2019 Youth Risk

Behavior Survey (YRBS), a cross-sectional, school-based

survey conducted biennially since 1991. Each survey year,

CDC collects data from a nationally representative sample

of public and private school students in grades 9–12 in the

50 U.S. states and the District of Columbia (N = 13,677).

Additional information about YRBS sampling, data collection,

response rates, and processing is available in the overview

report of this supplement (9). The prevalence estimates for

all violence questions for the overall study population and by

sex, race/ethnicity, grade, and sexual orientation are available

at https://nccd.cdc.gov/youthonline/App/Default.aspx. The

full YRBS questionnaire is available at https://www.cdc.gov/

healthyyouth/data/yrbs/pdf/2019/2019_YRBS-National-HS-

Questionnaire.pdf.

Measures

This analysis included five standard measures of violence

victimization and three composite variables created from those

standard measures. The standard measures included 1) having

experienced physical dating violence, 2) having experienced

sexual dating violence, 3) having experienced sexual violence

by anyone, 4) having been bullied on school property, and

5) having been bullied electronically (Table 1). For each of

these five standard measures, dichotomous categories were

created: ≥1 time versus 0 times for all sexual violence and dating

violence measures and “yes” versus “no” for both bullying

Supplement

30 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

victimization measures. The manner in which the data were

collected (see Limitations) means that approximately 25% of

respondents were missing data for sexual violence victimization

by anyone (3,439) out of a sample of 13,677 students. The

denominators for dating violence victimization measures are

students who reported dating during the 12 months before

the survey (66.1% [n = 8,703 students] for physical dating

violence victimization and 66.2% [n = 6,847 students] for

sexual dating violence victimization), whereas the denominator

for the sexual violence by anyone and bullying victimization

measures are the full sample of students for which data were

available. Three of these standard measures included levels

of victimization frequency. For each of three measures (i.e.,

physical dating violence, sexual dating violence, and sexual

violence by anyone), frequencies were collapsed into three

levels: 1 time, 2 or 3 times, or ≥4 times.

The two dating violence victimization measures were

combined into composite measures: experienced any dating

violence victimization and experienced both physical and

sexual dating violence victimization. Because of the manner

in which the data were collected, approximately 25% of

respondents were missing data for sexual dating violence

victimization (3,324 observations out of a sample of 13,677

students). When calculating the “any dating violence

victimization” measure, responses missing data for either the

sexual or the physical dating violence measure were removed

from the analysis. Any “yes” responses to either the physical

dating violence measure or the sexual dating violence measure

were combined for the numerator, with all responses without

missing data as the denominator. Similarly, to create the

“both physical and sexual dating violence” measure, “yes”

responses to both physical dating violence and sexual dating

violence were required for the numerator, with all nonmissing

responses in the denominator. A similar strategy was also used

for creating a bullying victimization “any” measure. “Any

bullying victimization” included any “yes” response to either

experiencing bullying at school or experiencing electronic

bullying, with all nonmissing responses in the denominator.

The option of exploring “both bullying at school and electronic

bullying” was not pursued. Use of personal electronic devices

in the school setting is increasing; therefore, the amount of

overlap between electronic bullying and bullying at school

might be considerable and combining these items could result

in an overestimate of their prevalence. Additional analysis

examined overlap between the sexual dating violence measure

and the sexual violence by anyone measure.

Three demographic characteristics were included in the

analyses: student sex (male or female), race/ethnicity (white,

black, Hispanic, or other), and sexual identity (heterosexual;

lesbian, gay, or bisexual [LGB]; or not sure). Although students

of multiple or other race/ethnicity are included in these

analyses, data are not presented for this group because small

sample sizes and unknown heterogeneity within this group

resulted in limited interpretability.

Analysis

Weighted prevalence estimates and corresponding 95%

confidence intervals were determined for all violence victimization

measures. Comparisons by demographic characteristics were

conducted with the chi-square test (p<0.05). When differences

among groups were demonstrated, additional t-tests were

performed to determine pairwise differences between groups.

Differences between prevalence estimates were considered

statistically significant if the t-test p value was <0.05 for main

effects (sex, race/ethnicity, or sexual identity).

Results

Among the approximately two thirds of U.S. high school

students who reported dating during the 12 months before the

survey, 8.2% reported experiencing physical dating violence,

and 8.2% experienced sexual dating violence (Table 2).

Sexual violence victimization perpetrated by anyone during

the 12 months before the survey was reported by 10.8% of

students. When comparing the sexual dating violence measure

with the sexual violence by anyone measure, half (50%) of the

10.8% of students who reported sexual violence by anyone

were victimized only by someone other than a dating partner.

Experiences of bullying victimization during the 12 months

before the survey varied, with 15.7% of students reporting

experiencing electronic bullying and 19.5% reporting bullying

on school property. For all violence victimization measures,

the prevalence varied by both sex and sexual identity, and

variation by race/ethnicity was only observed for bullying

victimization. Specifically, female students, LGB students,

and students not sure of their sexual identity consistently

had the highest prevalence across all five of the violence

victimization indicators. In addition, compared with Hispanic

or black students, white students had the highest prevalence

of experiencing bullying victimization at school and electronic

bullying. The prevalence of electronic bullying among Hispanic

students was also significantly greater than the prevalence

among black students.

Among students who experienced physical dating violence,

sexual dating violence, or sexual violence by anyone during the

previous year, the most common frequency reported was 1 time for

each (Figure). The pattern of frequency for violence victimization

differed by type of victimization. The distribution of frequency

for physical dating violence victimization was U-shaped, with

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 31

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 2. Percentage of high school students who experienced violence victimization,* by demographic characteristics and type of violence —

Youth Risk Behavior Survey, United States, 2019

Characteristic

Experienced physical dating violence

†

Experienced sexual dating violence

Experienced sexual violence by anyone

% (95% CI) p value** % (95% CI) p value** % (95% CI) p value**

Total 8.2 (7.2–9.4) NA 8.2 (7.4–9.1) NA 10.8 (9.9–11.7) NA

Sex

Female 9.3 (8.0–10.8) 0.01 12.6 (11.2–14.2) <0.01 16.6 (15.1–18.2) <0.01

Male 7.0

††

(5.8–8.4) NA 3.8

††

(3.1–4.7) NA 5.2

††

(4.4–6.1) NA

Race/Ethnicity

White, non-Hispanic 7.5 (6.4–8.7) 0.43 8.1 (6.9–9.6) 0.11 10.2 (9.1–11.4) 0.23

Black, non-Hispanic 8.2 (6.1–10.8) NA 6.2 (4.5–8.6) NA 10.3 (8.0–13.1) NA

Hispanic 8.9 (7.4–10.8) NA 8.7 (6.9–10.8) NA 12.2 (10.6–14.0) NA

Sexual identity

Heterosexual 7.2 (6.2–8.3) 0.01 6.7 (5.9–7.5) <0.01 9.0 (8.2–9.9) <0.01

Lesbian, gay, or bisexual 13.1

§§

(10.5–16.1) NA 16.4

§§

(12.7–20.9) NA 21.5

§§

(18.2–25.2) NA

Not sure 16.9

§§

(11.1–24.9) NA 15.0

§§

(9.5–23.0) NA 16.2

§§

(11.7–22.0) NA

Characteristic Bullied on school property Electronically bullied — —

Total 19.5 (18.2–20.9) NA 15.7 (14.6–16.9) NA — —

Sex

Female 23.6 (21.8–25.5) <0.01 20.4 (18.9–22.0) <0.01 — —

Male 15.4

§§

(14.0–16.9) NA 10.9

††

(9.6–12.4) NA — —

Race/Ethnicity

White, non-Hispanic 23.1 (21.4–24.8) <0.01 18.6 (17.1–20.2) <0.01 — —

Black, non-Hispanic 15.1

¶¶

(13.1–17.4) NA 8.6

¶¶

(7.4–10.0) NA — —

Hispanic 14.8

¶¶

(12.8–17.1) NA 12.7

¶¶,

*** (11.1–14.5) NA — —

Sexual identity

Heterosexual 17.1 (15.7–18.7) <0.01 14.1 (12.9–15.4) <0.01 — —

Lesbian, gay, or bisexual 32.0

§§

(29.5–34.6) NA 26.6

§§

(23.3–30.2) NA — —

Not sure 26.9

§§

(22.2–32.2) NA 19.4

§§,†††

(15.5–24.0) NA — —

Abbreviations: CI=confidence interval; NA = not applicable; YRBS = Youth Risk Behavior Survey.

* During the 12 months before the survey.

†

Being physically hurt on purpose (counting such things as being hit, slammed into something, or injured with an object or weapon) by someone they were dating

or going out with, ≥1 time, among the 66.1% (n = 8,703) of students nationwide who dated or went out with someone during the 12 months before the survey.

Being forced to do “sexual things” (counting such things as kissing, touching, or being physically forced to have sexual intercourse) they did not want to do by

someone they were dating or going out with, ≥1 time, among the 66.2% (n = 6,847) of students nationwide who dated or went out with someone during the

12 months before the survey. Of 13,677 students, this variable was missing for 3,324, mostly attributed to the use of different versions of the YRBS questionnaire

that did not include the sexual violence questions in certain selected schools. This resulted in complete data for 10,353 students, of which 66.2% (6,847) reported

dating in the 12 months before the survey.

Being forced to do “sexual things” (counting such things as kissing, touching, or being physically forced to have sexual intercourse) they did not want to do by

anyone, ≥1 time, during the 12 months before the survey. Data were missing for 3,439 students for this variable, mostly attributed to the use of different versions

of the YRBS questionnaire that did not include the sexual violence questions in certain selected schools.

** Chi-square test (p<0.05).

††

Significantly different from female students, based on t-test (p<0.05).

§§

Significantly different from heterosexual students, based on t-test (p<0.05).

¶¶

Significantly different from white students, based on t-test (p<0.05).

*** Significantly different from black students, based on t-test (p<0.05).

†††

Significantly different from lesbian, gay, or bisexual students, based on t-test (p<0.05).

the highest levels of frequency at 1 time and ≥4 times, whereas

for both sexual dating violence victimization and sexual violence

victimization by anyone, the most common frequency was 1 time,

with a decreasing prevalence as the frequency increased.

The frequency of physical and sexual dating violence varied

significantly by sex (Table 3). Specifically, the prevalence of

physical dating violence was significantly greater at higher

frequency levels (≥4 times) among male students compared

with female students (41.6% versus 21.6%, respectively).

This frequency distribution pattern was similar for sexual

dating violence. The prevalence at the higher end of frequency

for sexual dating violence was significantly greater for male

students compared with female students (41.0% versus 20.8%,

respectively). Higher frequency (≥4 times) was also reported

for sexual violence by anyone for male students compared

with female students (33.9% versus 18.6%, respectively). No

significant differences existed by race/ethnicity in frequency

of physical and sexual dating violence or sexual violence by

anyone. These analyses could not include sexual identity

because of limited data (i.e., group counts <30).

Overall, 12.2% of students experienced any type of dating

violence victimization, and 3.0% experienced both types

(Table 4). Both dating violence composite measures varied

substantially by sex and sexual identity but not by race/ethnicity.

Supplement

32 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

The prevalence of the dating violence composite variables was

significantly greater for female students compared with male

students (16.4% versus 8.2% for any dating violence type;

3.8% versus 2.1% for both dating violence types). Students

who did not identify as heterosexual had substantially greater

prevalence of both dating violence composites. For any type of

FIGURE. Percentage of high school students who experienced violence,

by type of victimization (physical dating violence, sexual dating

violence, or sexual violence by anyone) and by number of times during

the previous year — Youth Risk Behavior Survey, United States, 2019

1 time

2–3 times

≥4 times

Physical dating

violence

(n = 799)

Sexual dating

violence

(n = 558)

Sexual violence

by anyone

(n = 1,159)

Percentage

Type of victimization

100

dating violence, the prevalence was 22.3% for LGB students

and 18.7% for students who were not sure of their sexual

identity versus 10.5% for heterosexual students. For both types

of dating violence, the prevalence was 5.8% for LGB students

and 9.4% for students not sure of their sexual identity versus

2.4% for heterosexual students.

The prevalence of experiencing any type of bullying

victimization was 24.8% (Table 4), and prevalence varied

significantly by sex, race/ethnicity, and sexual identity. The

prevalence of experiencing any bullying victimization was

significantly greater for female students compared with male

students (30.2% versus 19.2%, respectively) and significantly

greater for white (28.8%) compared with black (18.0%) or

Hispanic (19.2%) students. Both LGB students (39.5%)

and students not sure of their sexual identity (32.7%) had

significantly higher prevalence of any bullying compared with

heterosexual students (22.2%), with LGB students reporting

greater prevalence than students not sure of their sexual identity.

Discussion

This report describes the 2019 prevalence and frequency

of different forms of interpersonal violence victimization

experienced by U.S. high school students. Similar to

TABLE 3. Frequency of types of violence victimization,* by demographic characteristics among high school students reporting experiencing

specific types of violence — Youth Risk Behavior Survey, United States, 2019

Type of violence victimization

Sex Race/Ethnicity

Male

% (95% CI)

Female

% (95% CI) p value

†

White, non-

Hispanic

% (95% CI)

Black, non-Hispanic

% (95% CI)

Hispanic

% (95% CI) p value

†

Experienced physical dating violence

NA NA <0.01 NA NA NA 0.21

1 time 38.0 (32.2–44.2) 51.7 (44.2–59.2) NA 47.9 (39.7–56.2) 47.5 (37.6–57.7) 40.9 (31.3–51.3) NA

2 or 3 times 20.4 (14.2–28.4) 26.7 (21.6–32.5) NA 25.5 (18.7–33.8) 16.7 (10.3–25.9) 27.0 (19.5–36.0) NA

≥4 times 41.6 (34.6–48.9) 21.6 (16.9–27.1) NA 26.6 (20.1–34.3) 35.8 (25.0–48.2) 32.1 (24.9–40.3) NA

Experienced sexual dating violence

NA NA 0.05 NA NA NA 0.39

1 time 33.3 (23.8–44.4) 44.0 (36.5–51.8) NA 42.2 (33.7–51.2) 29.0 (15.5–47.6) 45.0 (33.3–57.3) NA

2 or 3 times 25.7 (16.8–37.2) 35.2 (28.4–42.6) NA 32.3 (25.8–39.4) 38.6 (23.8–56.0) 33.3 (22.3–46.5) NA

≥4 times 41.0 (28.0–55.3) 20.8 (15.3–27.6) NA 25.5 (18.8–33.7) 32.4 (15.4–55.7) 21.6 (14.4–31.2) NA

Experienced sexual violence by anyone** NA NA 0.006 NA NA NA 0.36

1 time 36.6 (28.7–45.4) 47.3 (42.8–52.0) NA 47.6 (41.2–54.1) 39.7 (30.0–50.2) 44.0 (36.6–51.6) NA

2 or 3 times 29.5 (21.8–38.6) 34.1 (29.9–38.5) NA 31.2 (26.3–36.6) 34.9 (27.1–43.7) 34.5 (28.1–41.6) NA

≥4 times 33.9 (25.3–43.8) 18.6 (15.2–22.5) NA 21.2 (16.1–27.4) 25.4 (17.1–36.0) 21.5 (15.7–28.7) NA

Abbreviations: CI=confidence interval; NA =not applicable; YRBS = Youth Risk Behavior Survey.

* During the 12 months before the survey.

†

Chi-square test (p<0.05).

Being physically hurt on purpose (counting such things as being hit, slammed into something, or injured with an object or weapon) by someone they were dating

or going out with, ≥1 time, among the 66.1% (n = 8,703) of students nationwide who dated or went out with someone during the 12 months before the survey.

Being forced to do “sexual things” (counting such things as kissing, touching, or being physically forced to have sexual intercourse) they did not want to do by

someone they were dating or going out with, ≥1time, among the 66.2% (n = 6,847) of students nationwide who dated or went out with someone during the

12 months before the survey. Of 13,677 students, this variable was missing for 3,324, mostly attributed to the use of different versions of the YRBS questionnaire

that did not include the sexual violence questions in certain selected schools. This resulted in complete data for 10,353 students, of which 66.2% (6,847) reported

dating in the 12 months before the survey.

** Being forced to do “sexual things” (counting such things as kissing, touching, or being physically forced to have sexual intercourse) they did not want to do by

anyone during the 12 months before the survey. These data were missing for 3,439 students for this variable, mostly attributed to the use of different versions of

the YRBS questionnaire that did not include the sexual violence questions in certain selected schools.

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 33

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 4. Percentage of high school students who experienced any dating violence or both physical and sexual dating violence* and any form

of bullying victimization,

†

by demographic characteristics — Youth Risk Behavior Survey, United States, 2019

Characteristic

Dating violence composite variables Bullying victimization composite

Experienced any dating violence

Experienced both physical and sexual

dating violence

Experienced any bullying**

% (95% CI) p value

††

% (95% CI) p value

††

% (95% CI) p value

††

Total 12.2 (11.3–13.3) NA 3.0 (2.5–3.7) NA 24.8 (23.4–26.3) NA

Sex

Female 16.4 (14.7–18.2) <0.01 3.8 (3.0–5.0) 0.006 30.2 (28.4–32.1) <0.01

Male 8.2

§§

(7.1–9.4) NA 2.1

§§

(1.6–2.9) NA 19.2

§§

(17.6–20.9) NA

Race/Ethnicity

White, non-Hispanic 12.1 (10.8–13.5) 0.42 2.8 (2.2–3.5) 0.51 28.8 (26.9–30.7) <0.01

Black, non-Hispanic 10.6 (7.9–14.1) NA 3.0 (1.7–5.2) NA 18.0

¶¶

(15.7–20.6) NA

Hispanic 12.7 (11.1–14.6) NA 3.3 (2.1–5.1) NA 19.2

¶¶

(17.4–21.1) NA

Sexual identity

Heterosexual 10.5 (9.5–11.6) <0.01 2.4 (2.0–2.9) 0.007 22.2 (20.6–23.8) <0.01

Lesbian, gay, or bisexual 22.3***

(17.9–27.5) NA 5.8***

(3.9–8.4) NA 39.5*** (36.6–42.5) NA

Not sure 18.7*** (13.2–26.0) NA 9.4*** (5.0–16.9) NA 32.7***

,†††

(27.6–38.3) NA

Abbreviations: CI=confidence interval; NA = not applicable.

* During the 12 months before the survey, among students who dated or went out with someone during the 12 months before the survey.

†

During the 12 months before the survey.

Combined any “yes” responses to physical dating violence and sexual dating violence. Because of the manner in which this variable was calculated, missing values

in both the physical dating violence and sexual dating violence measures resulted in 3,355 missing values in the “experienced any dating violence” composite measure.

Combined where responses to both physical dating violence and sexual dating violence were “yes.” Because of the manner in which this variable was calculated,

the missing values in both the physical dating violence and sexual dating violence measures resulted in 3,355 missing observations in the “experienced both

physical and sexual dating violence” composite measure.

** Combined any “yes” responses to bullied at school and electronic bullying.

††

Chi-square test (p<0.05).

§§

Significantly different from female students, based on t-test (p<0.05).

¶¶

Significantly different from white, non-Hispanic students, based on t-test (p<0.05).

*** Significantly different from heterosexual students, based on t-test (p<0.05).

†††

Significantly different from lesbian, gay, or bisexual students, based on t-test (p<0.05).

findings from previous YRBSs (https://www.cdc.gov/

violenceprevention/pdf/2012FindingsonSVinYouth-508.pdf),

physical dating violence, sexual dating violence, sexual violence

by anyone, bullying on school property, and electronic bullying

victimization are adverse childhood experiences (ACEs) that are

occurring at high rates. Examining their prevalence individually

and in combination by key demographic characteristics

provides an overall observation and contextual understanding

of interpersonal violence experienced by U.S. high school

students and helps identify disparities in health and safety

among U.S. youths, which can guide prevention efforts.

All five types of victimization, including any or both forms of

dating violence and any form of bullying, were more common

among female and sexual minority students, highlighting their

more frequent victimization. These findings are consistent

with previous studies that reported disparities in interpersonal

violence victimization, particularly dating violence and sexual

violence, by sex and sexual identity (6,7). Although findings

did not reveal substantially greater prevalence for racial/ethnic

minority youths for the forms of violence examined, research

has consistently shown that racial/ethnic minority youths are

at greater risk for homicides and other community violence

victimization (https://www.cdc.gov/violenceprevention/

pub/technical-packages.html). Disparities in health and risk

for violence have been linked to sexism, homophobia, and

structural disadvantage (10).

Half of students who reported sexual violence victimization

by anyone did not report sexual violence by a dating partner,

indicating that students who experience sexual violence are often

victimized by someone other than a dating partner. This finding

is consistent with previous research (3) documenting that sexual

violence happening in school during adolescence is frequently

perpetrated by peers and not necessarily by dating partners.

Indeed, perpetrators of sexual violence during youth can be

acquaintances, family members, persons in a position of authority,

and strangers, in addition to dating partners (https://www.cdc.

gov/violenceprevention/pdf/2012FindingsonSVinYouth-508.

pdf). This indicates that efforts might need to be focused on

preventing sexual violence both inside and outside the context

of dating relationships to be most helpful.

Males who experienced dating violence or sexual violence

reported high frequencies of victimization (≥4 times during

Supplement

34 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

the previous year) substantially more often than did females.

That is, although male students do not report higher prevalence

of victimization than do female students, when they do

report it, they report experiencing it at a higher frequency.

Previous research has documented that, among youths at

high risk (i.e., previously exposed to violence in the home or

community), adolescent males reported higher frequency of

victimization than did females for sexual dating violence (11).

However, male adolescents might also be more likely to disclose

dating violence and sexual violence when the victimization has

happened more than once.

In this study, bullying victimization was the only type

of violence victimization examined for which racial/ethnic

differences existed, with substantially higher prevalence

occurring among white students compared with black or

Hispanic students. This result for bullying is supported in

part by previous research (12). In addition, Hispanic students

reported substantially higher prevalence of electronic bullying

victimization compared with black students. Other research

has indicated that black students might underreport bullying

victimization when presented with a definition-based measure

of bullying that includes a form of the word “bully,” as is used

in YRBS, as opposed to behaviorally specific measures that

describe the victimization behaviors but do not use the word

“bully” (13). The measurement of bullying in this study might

have differentially affected reporting across racial/ethnic groups.

Overall, these findings highlight the importance of early

engagement in effective, evidence-based efforts for preventing

violence victimization and perpetration before they begin or

stopping them from continuing. Findings from this study

also demonstrate substantial differences in exposure to these

types of violence by sex, race/ethnicity, and sexual identity,

highlighting the need for prevention efforts that address the

unique needs of these groups. To help communities focus

their prevention efforts on what works and to address risk

and protective factors for violence and other ACEs across the

social ecology, CDC developed a series of technical packages

that identify key violence prevention strategies and approaches

on the basis of the best available research evidence. (CDC’s

technical packages for violence prevention are available at

https://www.cdc.gov/violenceprevention/pub/technical-

packages.html.) This series includes packages focused on sexual

violence, intimate partner violence (including dating violence),

and youth violence (including bullying). Preventing Adverse

Childhood Experiences (ACEs): Leveraging the Best Available

Evidence compiles evidence focused on ACEs from across the

technical packages (https://www.cdc.gov/violenceprevention/

pdf/preventingACES.pdf ).

Multiple evidence-based interpersonal violence prevention

approaches are directly related to the findings in this study.

For example, social-emotional learning programs that

support development of skills for communication, emotion

regulation, empathy, and respect and that target risk factors for

interpersonal violence (e.g., impulsivity or drug use) have been

reported to decrease adolescent sexual violence perpetration

and homophobic name-calling, with indirect effects on peer

bullying, cyberbullying, and sexual harassment perpetration

when mediated by delinquency (14,15). By addressing shared

risk and protective factors across types of violence, social-

emotional learning programs can build the skills youths

need for engaging in healthy relationships with family, peers,

dating partners, and others, thus preventing multiple forms of

adolescent interpersonal violence and long-term consequences

into adulthood. In addition, bystander programs teach youths

how to safely act when they see behaviors that increase risk

for violence and change social norms within their peer groups.

Although originally conceptualized as a means of challenging

heterosexist attitudes to prevent sexual and dating violence

(16), such programs might also prevent other forms of

adolescent violence, including bullying and violence targeting

sexual, gender, and racial minorities by focusing the training

on recognizing and challenging these specific harmful attitudes

and behaviors (17,18).

Modifying the social and physical environment in schools

and neighborhoods might improve safety and reduce risk

for violence for more of the population than individual- or

relationship-level approaches alone. For example, one school-

based prevention approach that includes a building-level

intervention (e.g., addressing physical areas in the school

identified by students as less safe) has been reported to reduce

sexual violence victimization and perpetration by peers

and dating partners (19). In addition, the development of

safe and supportive environments in schools that promote

protective factors (e.g., school connectedness and professional

development regarding lesbian, gay, bisexual, and transgender

[LGBT] youths) can help create accepting school environments

for LGBT youths and reduce the risk for bullying and other

violence (20). Results from this report indicate that LGB

youths, specifically, are at a disproportionately higher risk

for interpersonal violence victimization compared with

heterosexual youths. As of 2019, gender identity has not been

assessed by the YRBS nationwide. However, during 2017,

gender identity was assessed in YRBSs conducted in 10 states

and nine large urban school districts; these data show that

transgender students consistently report greater prevalence

of violence victimization than their cisgender peers (21).

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 35

US Department of Health and Human Services/Centers for Disease Control and Prevention

Promotion of gay-straight alliances and support of LGBT

students can help provide these youths with an accepting school

environment, which might also reduce the risk for school-based

violence against these youths (22). (Information about CDC’s

current school health programs is available at https://www.

cdc.gov/healthyyouth/fundedprograms/1807/resources.htm.)

CDC is engaged in ongoing research and programmatic

activities for expanding the research evidence and adding to

the knowledge base of effective primary prevention programs,

policies, and practices available to communities for preventing

interpersonal violence among youths. For example, CDC’s

Dating Matters: Strategies to Promote Healthy Teen Relationships

is a comprehensive adolescent dating violence prevention

model. Dating Matters includes multiple integrated prevention

strategies that address risk factors for youths and their families,

schools, and neighborhoods with demonstrated effects on

adolescent dating violence, bullying, and peer violence in

middle school. (Additional information about Dating Matters

is available at https://www.cdc.gov/violenceprevention/

intimatepartnerviolence/datingmatters/index.html.)

In addition, since 2001, CDC has provided funding for

primary prevention of sexual violence through the Rape

Prevention and Education Program to state health departments

in all 50 states, the District of Columbia, and four U.S.

territories. Funded organizations implement initiatives that

address youths in their communities, including community-

and societal-level approaches (e.g., improving education and

leadership opportunities for girls). (Additional information

about the Rape Prevention and Education Program is available

at https://www.cdc.gov/violenceprevention/sexualviolence/rpe/

index.html.) CDC also sponsors youth violence prevention

research through its National Centers of Excellence in Youth

Violence Prevention. Their goal is to build the scientific

infrastructure and community partnerships necessary for

stimulating new youth violence prevention research and

practice across the country, including a focus on the impact

of structural factors (e.g., housing, education, or systemic

discrimination) that limit access to positive social determinants

of health.

Prevention of interpersonal violence among adolescents

might be most successful when a comprehensive strategy is

used that addresses these ACEs at multiple levels of the social

ecology simultaneously and recognizes that these different

forms of victimization can be co-occurring (1). The findings

reported here also highlight the importance of acknowledging

the disproportionate prevalence of these forms of victimization

on certain youths (i.e., females and sexual minorities) and

addressing these disparities in prevention efforts.

Limitations

General limitations for the YRBS are available in the overview

report of this supplement (9).The findings in this report are

subject to at least five additional limitations. First, substantial

overlap likely existed in the measures that examined experiences

of sexual violence victimization (i.e., sexual dating violence

victimization and sexual violence victimization by anyone),

and among the bullying victimization measures (i.e., electronic

bullying and bullied at school). For these reasons, composites

for the sexual violence measures and a “both” composite for

bullying (i.e., experienced both electronic bullying and bullying

at school) were not created. Second, because of the breadth of

topics included in the YRBS, violence victimization subtype

measures included in the YRBS tend to be broad in nature and,

in this study, were assessed by single items. More specific and

detailed measures of violence victimization would allow for a

comprehensive analysis of the prevalence and overlap between

different forms of interpersonal violence victimization. Third,

the YRBS bullying items include the word “bullied,” which

might have decreased disclosure (13). Fourth, the interpersonal

violence victimization types that could be included in this

study (i.e., dating violence, sexual violence, and bullying), as

a whole, do not reflect the breadth of interpersonal violence

victimization experienced by youths (i.e., other forms of youth

violence experienced in the community) and might partially

explain why few racial/ethnic differences were found. Finally,

the sexual violence measures (and composite measures that

were created with the sexual violence measures) in this report

had a relatively large amount of missing data (approximately

3,400 observations) in 2019. Most of this missing data can

be attributed to the use of different versions of the YRBS

questionnaire that did not include the sexual violence questions

in certain selected schools. Consequently, not all students in

the national sample were given the opportunity to answer the

sexual violence questions and were counted as missing. When

constructing the composite measures for any dating violence,

and both physical and sexual dating violence victimization, the

analytic sample was restricted to students who had complete

data for both physical and sexual dating violence victimization,

which reduced the potential for biased estimates.

Future Directions

To increase understanding of the differential experiences

of adolescent interpersonal violence victimization, future

research that focuses in more detail on the demographic

groups highlighted in this study can be beneficial. For

example, on the basis of these findings, additional research

Supplement

36 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

to better understand the characteristics and consequences

of these forms of interpersonal violence on sexual minority

youths is warranted. Research exploring sex differences in the

frequency of victimization across additional types of violence

can add to the findings reported here. Future studies that

include more detailed measures of dating violence, sexual

violence, and bullying for capturing and isolating understudied

subtypes of these forms of violence (e.g., psychological dating

violence, nonconsensual sexting, or relational bullying) would

increase knowledge of the full prevalence of these forms of

violence among youths. Finally, studies that examine the

co-occurrence and cumulative impact of different forms of

violence victimization during adolescence and into adulthood

can guide more comprehensive prevention efforts.

Conclusion

Interpersonal violence victimization experiences of high

school students are a form of ACEs and represent a substantial

public health problem in the United States. Multiple forms

of interpersonal violence, including dating violence, sexual

violence, and bullying, negatively affect youths and can

continue to have damaging effects throughout a person’s

life. The findings in this report are consistent with those in

previous studies about disparities in interpersonal violence

victimization by demographic characteristics; the report also

provides additional insight about the specific groups of students

who are at highest risk for particular types of interpersonal

violence and who might benefit most from prevention

efforts. In addition, the findings increase understanding of

the contextual factors associated with interpersonal violence

victimization (e.g., frequency, location, and co-occurrence of

subtypes) and can guide how violence prevention professionals

select and implement prevention approaches for addressing

dating violence, sexual violence, and bullying. Prevention

approaches at the individual, relationship, and school or

community levels (e.g., those that seek to increase youths’ skills

in preventing violence, change social norms related to violence,

and modify the physical and social environment in schools and

communities to increase protection against violence) are crucial

for building a comprehensive strategy to reduce interpersonal

violence victimization among youths.

Conicts of Interest

All authors have completed and submitted the International

Committee of Medical Journal Editors form for disclosure of

potential conflicts of interest. No potential conflicts of interest

were disclosed.

References

1. CDC. Preventing multiple forms of violence: a strategic vision for

connecting the dots. Atlanta, GA: US Department of Health and Human

Services, CDC, National Center for Injury Preventon and Control,

Division of Violence Prevention; 2016.

2. Smith SG, Zhang X, Basile KC, et al. The National Intimate Partner

and Sexual Violence Survey (NISVS): 2015 data brief. Atlanta, GA: US

Department of Health and Human Services, CDC, National Center for

Injury Prevention and Control; 2018. https://www.cdc.gov/

violenceprevention/datasources/nisvs/2015NISVSdatabrief.html

3. Hill C, Kearl H. Crossing the line: sexual harassment at school. Washington,

DC: American Association of University Women; 2011. https://www.aauw.

org/resources/research/crossing-the-line-sexual-harassment-at-school

4. Wincentak K, Connolly J, Card N. Teen dating violence: a meta-analytic

review of prevalence rates. Psychol Violence 2017;7:224–41. https://

doi.org/10.1037/a0040194

5. Seldin M, Yaez C. Student reports of bullying: results from the 2017

School Crime Supplement to the National Crime Victimization Survey.

Report no.: NCES 2019–054. Washington, DC: US Department of

Education, National Center for Education Statistics; 2019. https://nces.

ed.gov/pubsearch/pubsinfo.asp?pubid=2019054

6. Norris AL, Orchowski LM. Peer victimization of sexual minority and

transgender youth: a cross-sectional study of high school students.

Psychol Violence 2020;10:201–11. https://doi.org/10.1037/vio0000260

7. Coulter RWS, Mair C, Miller E, Blosnich JR, Matthews DD,

McCauley HL. Prevalence of past-year sexual assault victimization among

undergraduate students: exploring differences by and intersections of

gender identity, sexual identity, and race/ethnicity. Prev Sci

2017;18:726–36. https://doi.org/10.1007/s11121-017-0762-8

8. Vagi KJ, O’Malley Olsen E, Basile KC, Vivolo-Kantor AM. Teen dating

violence (physical and sexual) among US high school students: findings

from the 2013 National Youth Risk Behavior Survey. JAMA Pediatr

2015;169:474–82. https://doi.org/10.1001/jamapediatrics.2014.3577

9. Underwood JM, Brener N, Thornton J, et al. Overview and methods

for the Youth Risk Behavior Surveillance System—United States, 2019.

In: Youth Risk Behavior Surveillance—United States, 2019. MMWR

Suppl 2020;69(No. Suppl 1).

10. Popkin SJ, Bogle M, Zweig JM, Saxena PD, Breslav L, Michie M. Let

girls be girls: how coercive sexual environments affect girls who live in

disadvantaged communities and what we can do about it. Washington,

DC: Urban Institute; 2015. https://www.urban.org/sites/default/files/

publication/72466/2000490-Let-Girls-Be-Girls.pdf

. Reidy DE, Kearns MC, Houry D, Valle LA, Holland KM, Marshall KJ.

Dating violence and injury among youth exposed to violence. Pediatrics

2016;137:e20152627. https://doi.org/10.1542/peds.2015-2627

. Spriggs AL, Iannotti RJ, Nansel TR, Haynie DL. Adolescent bullying

involvement and perceived family, peer and school relations:

commonalities and differences across race/ethnicity. J Adolesc Health

2007;41:283–93. https://doi.org/10.1016/j.jadohealth.2007.04.009

. Sawyer AL, Bradshaw CP, O’Brennan LM. Examining ethnic, gender,

and developmental differences in the way children report being a victim

of “bullying” on self-report measures. J Adolesc Health 2008;43:106–14.

https://doi.org/10.1016/j.jadohealth.2007.12.011

. Espelage DL, Low S, Ryzin MJV, Polanin JR. Clinical trial of Second

Step Middle School Program: impact on bullying, cyberbullying,

homophobic teasing, and sexual harassment perpetration. School Psych

Rev 2015;44:464–79. https://doi.org/10.17105/spr-15-0052.1

. Espelage DL, Low S, Polanin JR, Brown EC. The impact of a middle

school program to reduce aggression, victimization, and sexual violence.

J Adolesc Health 2013;53:180–6. https://doi.org/10.1016/j.

jadohealth.2013.02.021

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 37

US Department of Health and Human Services/Centers for Disease Control and Prevention

16. Storer HL, Casey E, Herrenkohl T. Efficacy of bystander programs to

prevent dating abuse among youth and young adults: a review of the

literature. Trauma Violence Abuse 2016;17:256–69. https://doi.

org/10.1177/1524838015584361

. Nelson JK, Dunn KM, Paradies Y. Bystander anti‐racism: a review of

the literature. Anal Soc Issues Public Policy 2011;11:263–84. https://

doi.org/10.1111/j.1530-2415.2011.01274.x

. National Academies of Sciences. Preventing bullying through science,

policy, and practice. Rivara F, Le Menestrel S, eds. Washington, DC:

The National Academies Press; 2016. https://www.nap.edu/read/23482/

chapter/1

. Taylor BG, Stein ND, Mumford EA, Woods D. Shifting boundaries:

an experimental evaluation of a dating violence prevention program in

middle schools. Prev Sci 2013;14:64–76. https://doi.org/10.1007/

s11121-012-0293-2

. Bonell C, Hinds K, Dickson K, et al. What is positive youth development

and how might it reduce substance use and violence? A systematic review

and synthesis of theoretical literature. BMC Public Health 2016;16:135.

https://doi.org/10.1186/s12889-016-2817-3

. Johns MM, Lowry R, Andrzejewski J, et al. Transgender identity and

experiences of violence victimization, substance use, suicide risk, and

sexual risk behaviors among high school students—19 states and large

urban school districts, 2017. MMWR Morb Mortal Wkly Rep

2019;68:67–71.

22. Marx RA, Kettrey HH. Gay-straight alliances are associated with lower

levels of school-based victimization of LGBTQ+ youth: a systematic

review and meta-analysis. J Youth Adolesc 2016;45:1269–82. https://

doi.org/10.1007/s10964-016-0501-7

Supplement

38 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

Prescription Opioid Misuse and Use of Alcohol and Other Substances

Among High School Students — Youth Risk Behavior Survey,

United States, 2019

Christopher M. Jones, PharmD, DrPH

; Heather B. Clayton, PhD

; Nicholas P. Deputy, PhD

2,3

; Douglas R. Roehler, PhD

; Jean Y. Ko, PhD

;

Marissa B. Esser, PhD

; Kathryn A. Brookmeyer, PhD

, Marci Feldman Hertz, MS

Office of the Director, National Center for Injury Prevention and Control, CDC;

Division of Adolescent and School Health, National Center for HIV/AIDS,

Viral Hepatitis, STD, and TB Prevention, CDC;

Epidemic Intelligence Service, CDC;

Division of Overdose Prevention, National Center for Injury Prevention

and Control, CDC;

Division of Reproductive Health, National Center for Chronic Disease Prevention and Health Promotion, CDC;

Division of Population

Health, National Center for Chronic Disease Prevention and Health Promotion, CDC;

Division of STD Prevention, National Center for HIV/AIDS, Viral

Hepatitis, STD, and TB Prevention, CDC

Abstract

Adolescence is an important period of risk for substance use initiation and substance use–related adverse outcomes. To examine

youth substance use trends and patterns, CDC analyzed data from the 2009–2019 Youth Risk Behavior Survey. This report

presents estimated prevalence of current (i.e., previous 30-days) marijuana use, prescription opioid misuse, alcohol use, and binge

drinking and lifetime prevalence of marijuana, synthetic marijuana, cocaine, methamphetamine, heroin, injection drug use, and

prescription opioid misuse among U.S. high school students. Logistic regression and Joinpoint analyses were used to assess 2009–

2019 trends. Prevalence of current and lifetime substance use by demographics, frequency of use, and prevalence of co-occurrence

of selected substances among students reporting current prescription opioid misuse are estimated using 2019 data. Multivariable

logistic regression analysis was used to determine demographic and substance use correlates of current prescription opioid misuse.

Current alcohol, lifetime cocaine, methamphetamine, heroin, and injection drug use decreased during 2009–2019. Lifetime use

of synthetic marijuana (also called synthetic cannabinoids) decreased during 2015–2019. Lifetime marijuana use increased during

2009–2013 and then decreased during 2013–2019. In 2019, 29.2% reported current alcohol use, 21.7% current marijuana use,

13.7% current binge drinking, and 7.2% current prescription opioid misuse. Substance use varied by sex, race/ethnicity, grade,

and sexual minority status (lesbian, gay, or bisexual). Use of other substances, particularly current use of alcohol (59.4%) and

marijuana (43.5%), was common among students currently misusing prescription opioids. Findings highlight opportunities for

expanding evidence-based prevention policies, programs, and practices that aim to reduce risk factors and strengthen protective

factors related to youth substance use, in conjunction with ongoing initiatives for combating the opioid crisis.

Introduction

Substance use and associated adverse outcomes contribute

to substantial morbidity, mortality, and economic costs to

society each year in the United States (1). Data from national

surveys indicate the majority of adolescents will engage in

some form of substance use before they graduate from high

school (https://www.samhsa.gov/data/report/2018-nsduh-

detailed-tables). During adolescence, areas of the brain

associated with emotional responses and reward systems

develop before those associated with executive functioning,

judgement, and decision making (2). This uneven maturation

results in increased susceptibility for engaging in risky and

impulsive behaviors, including substance use, and increases

Corresponding author: Christopher M. Jones, PharmD, DrPH, Office

of the Director, National Center for Injury Prevention and Control,

CDC. Telephone: 404-498-0756; E-mail: [email protected].

vulnerability to reinforcing and rewarding effects of substances

(2,3). Preventing or delaying substance use initiation among

youths can reduce later risk for substance use and substance

use disorders (1,3,4). Beyond the individual negative effects

of substance use during youth and into adulthood, substance

use among youths also increases the likelihood for negative

consequences that affect peers, families, and communities

(5). Youth substance use is associated with increased risk for

delinquency, academic underachievement, teenage pregnancy,

sexually transmitted diseases, perpetrating or experiencing

violence, injuries, and mental health problems (1,3–6).

As the United States confronts its decades-long opioid

overdose epidemic (1,2), preventing opioid misuse among

youth is a public health imperative. Previous research has

documented that misuse of prescription opioids among

youths is associated with multiple adverse health outcomes

and risk behaviors, including use of alcohol and other illicit

drugs, injection drug use, suicidal ideation, youth violence,

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 39

US Department of Health and Human Services/Centers for Disease Control and Prevention

delinquency, having four or more lifetime sexual partners, not

using a condom at last sexual intercourse, increased risk for

acquisition of human immunodeficiency virus infection and

sexually transmitted diseases (6), and increasing overdoses (7).

Studies also have demonstrated that prescription opioid misuse

among youths is strongly linked with subsequent initiation

and use of heroin and increased risk for injecting prescription

opioids and developing an opioid use disorder (8–10).

Preventing substance use among youths is necessary because

of the health and social effects of youth substance use. To

inform substance use prevention initiatives and to improve

understanding of youth substance use patterns, including

misuse of prescription opioids and other substances, this

analysis 1) examines trends and patterns in substance use

among high school students overall and by demographic

characteristics, 2) characterizes the frequency of use of

specific substances among high school students, 3) explores

co-occurring substance use among high school students who

misuse prescription opioids, and 4) examines the demographic

and substance use correlates of prescription opioid misuse

among high school students. Findings from this analysis can

help inform efforts by public health practitioners, clinicians,

and the substance use prevention community to expand

the implementation of evidence-based prevention policies,

programs, and practices that aim to reduce risk factors and

strengthen protective factors related to youth substance use.

Methods

Data Source

This report includes data from CDC’s 2009–2019 Youth

Risk Behavior Survey (YRBS), a cross-sectional, school-based

survey conducted biennially since 1991. Each survey year,

CDC collects data from a nationally representative sample

of public- and private-school students in grades 9–12 in

the 50 U.S. states and the District of Columbia. Additional

information about YRBS sampling, data collection, response

rates, and processing is available in the overview report of this

supplement (11). The prevalence estimates for all substance

use questions for the overall study population and by sex,

race/ethnicity, grade, and sexual orientation are available at

https://nccd.cdc.gov/youthonline/App/Default.aspx. The

full YRBS questionnaire is available at https://www.cdc.gov/

healthyyouth/data/yrbs/pdf/2019/2019_YRBS-National-HS-

Questionnaire.pdf.

Measures

This report addresses four current (i.e., previous 30 days

before the survey) and seven lifetime substance use behaviors.

The four current substance use behaviors include 1) marijuana

use (ascertained by the question, “During the past 30 days,

how many times did you use marijuana?”), 2) alcohol use

(“During the past 30 days, on how many days did you have

at least one drink of alcohol?”), 3) binge drinking (“During

the past 30 days, on how many days did you have 4 or more

drinks of alcohol in a row, that is, within a couple of hours

[if you are a female] or 5 or more drinks of alcohol in a row,

that is, within a couple of hours [if you are a male]?”), and

4) prescription opioid misuse (“During the past 30 days, how

many times have you taken prescription pain medicine without

a doctor’s prescription or differently than how a doctor told you

to use it?”). The current prescription opioid misuse question

is new for the 2019 YRBS, providing opportunities to explore

substance use patterns and individual characteristics associated

with this variable for the first time.

The seven lifetime substance use behaviors include

1) marijuana use (“During your life, how many times have you

used marijuana?”), 2) synthetic marijuana (also called synthetic

cannabinoids) use (“During your life, how many times have you

used synthetic marijuana?”), 3) cocaine use (“During your life,

how many times have you used any form of cocaine, including

powder, crack, or freebase?”), 4) methamphetamine use (“During

your life, how many times have you used methamphetamines

[also called speed, crystal meth, crank, ice, or meth]?”), 5) heroin

use (“During your life, how many times have you used heroin

[also called smack, junk, or China White]?”), 6) prescription

opioid misuse (“During your life, how many times have you

taken prescription pain medicine without a doctor’s prescription

or differently than how a doctor told you to use it?”), and

7) injection drug use (“During your life, how many times have

you used a needle to inject any illegal drug into your body?”).

Substance use behaviors were dichotomized to indicate

current or lifetime use versus no use. With three exceptions,

frequency of use for each substance was categorized as

1–2 times, 3–9 times, 10–39 times, or ≥40 times. Frequency of

current alcohol use and current binge drinking were categorized

as 1–2 days, 3–9 days, 10–19 days, or ≥20 days. For injection

drug use, frequency of use was categorized as 1 time or ≥2 times.

Four demographic characteristics were included in the analysis:

sex (male or female), race/ethnicity (non-Hispanic white [white],

non-Hispanic black [black], Hispanic, or other), grade (9/10

or 11/12), and sexual identity (heterosexual; lesbian, gay, or

bisexual; or not sure). Students reporting “other” race/ethnicity

are included in all analyses; however, data are not presented for

that group because of limited interpretability.

Supplement

40 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

Analysis

First, annual prevalence of each substance use behavior was

estimated for all years with available data. Second, to identify

temporal trends, logistic regression analyses were used to

model linear and quadratic time effects while controlling for

sex, grade, and racial/ethnic group changes over time; for

significant quadratic time effects, Joinpoint software was used

to identify the year the trend changed direction (11). Trends

were assessed during 2009–2019 for current alcohol, current

marijuana, lifetime marijuana, lifetime cocaine, lifetime

methamphetamine, lifetime heroin, and lifetime injection

drug use. Synthetic marijuana use was first assessed by YRBS in

2015; therefore, trend analysis for this variable was conducted

for 2015–2019. Third, to identify 2-year changes in substance

use behaviors, prevalence estimates from 2017 and 2019 were

compared by using t-tests; changes were considered statistically

different if the p value was <0.05.

Four additional analyses were conducted by using 2019

YRBS data only. First, prevalence estimates and associated

95% confidence intervals (CIs) for each substance use behavior

were calculated by demographic characteristics. Statistically

significant pairwise differences between demographic groups

for each of the substance use behaviors were determined by

t-tests; differences were considered statistically significant if the

p value was <0.05. Second, to examine frequency of use among

students who reported engaging in each substance use behavior,

prevalence estimates and 95% CIs of students reporting each

frequency of use category were calculated. Third, prevalence

estimates of co-occurring use of selected substances among

students reporting current prescription opioid misuse were

estimated. Finally, multivariable logistic regression analysis

was used to determine demographic and substance use

correlates of current prescription opioid misuse. Because the

use of one substance is generally strongly associated with use

of one or more other substances, it is important to account

for multiple substance use behaviors during the modeling

process. Therefore, all demographic and substance use variables

were included in a single model to examine the independent

effect of each variable on current prescription opioid misuse.

This modeling strategy is consistent with previous research

examining substance use behaviors among youths (6).

To improve model stability during multivariable analyses,

three composite substance use variables were created. A

composite variable regarding alcohol consumption was created

with three levels: 1) no previous 30-day use, 2) previous

30-day use (current drinking but no binge drinking), and

3) previous 30-day binge alcohol use. A marijuana composite

variable also was created with three levels: 1) no lifetime use,

2) lifetime use but no previous 30-day use, and 3) previous

30-day use. A composite lifetime use of cocaine, heroin,

or methamphetamine variable was created by combining

answers of “1 or more times” for each of the three constituent

variables. Because substance use variables are known to be

highly correlated with each other, the Variance Inflation Factor

was used to assess multicollinearity. None was observed (i.e.,

no values >10).

Adjusted prevalence ratios (aPRs) and corresponding

95% CIs were calculated; estimates were considered

statistically significant if the 95% CI did not include 1.0. All

analyses were conducted by using SAS-callable SUDAAN

(version 11.0.1; RTI International) to account for survey

weights and the complex sample design of the YRBS. No

imputation methods were used for data that were missing.

Results

Substance use was common among U.S. high school

students during 2019 and varied by substance, year, and

demographic groups (Table 1). Among current substance

use measures, the highest prevalence estimates were for

alcohol (29.2%) and marijuana use (21.7%). Current binge

drinking was reported by 13.7% of high school students,

and 7.2% reported current prescription opioid misuse.

Among lifetime use measures, marijuana use was reported

by 36.8% of high school students, followed by misuse of

prescription opioids (14.3%) and use of synthetic marijuana

(7.3%), cocaine (3.9%), methamphetamine (2.1%), or heroin

(1.8%). Lifetime injection drug use was reported by 1.6% of

high school students.

Trend data were available for eight of the 11 substance

use measures included in the analyses. Among these

measures, current alcohol use, lifetime cocaine, lifetime

methamphetamine, lifetime heroin, and lifetime injection drug

use decreased during 2009–2019. Lifetime use of synthetic

marijuana decreased during 2015–2019. The prevalence

of lifetime marijuana use increased during 2009–2013

(36.8%–40.7%) and then decreased during 2013–2019

(40.7%–36.8%). No statistically significant changes from 2017

to 2019 were observed for any of the substance use behaviors.

Compared with females, males had a significantly higher

prevalence of lifetime use of cocaine (4.9% versus 2.7%),

methamphetamine (2.7% versus 1.5%), heroin (2.3% versus

1.0%), and injection drug use (2.1% versus 1.1%) (Table 2).

Compared with males, females had a significantly higher

prevalence of current alcohol use (31.9% versus 26.4%), binge

drinking (14.6% versus 12.7%), current prescription opioid

misuse (8.3% versus 6.1%), and lifetime prescription opioid

misuse (16.1% versus 12.4%). Among racial/ethnic groups,

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 41

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 1. Prevalence of and trends in prevalence of lifetime and current use of specific substances and use behaviors among high school

students — Youth Risk Behavior Survey, United States, 2009–2019

Behavior

Prevalence

Linear change* Quadratic change*

Change

from

2017 to 2019

†

2009 2011 2013 2015 2017 2019

Current use

Marijuana 20.8 23.1 23.4 21.7 19.8 21.7 No change No change No change

Alcohol 41.8 38.7 34.9 32.8 29.8 29.2 Decreased 2009–2019 No change No change

Binge drinking — — — — 13.5 13.7 NA

No change

Prescription opioid misuse — — — — — 7.2 NA

Lifetime use

Marijuana 36.8 39.9 40.7 38.6 35.6 36.8 No change Increased 2009–2013

Decreased 2013–2019

No change

Cocaine 6.4 6.8 5.5 5.2 4.8 3.9 Decreased 2009–2019 No change No change

Methamphetamine 4.1 3.8 3.2 3.0 2.5 2.1 Decreased 2009–2019 No change No change

Heroin 2.5 2.9 2.2 2.1 1.7 1.8 Decreased 2009–2019 No change No change

Injection drug use 2.1 2.3 1.7 1.8 1.5 1.6 Decreased 2009–2019 No change No change

Synthetic marijuana — — — 9.2 6.9 7.3 Decreased 2015-2019 NA

No change

Prescription opioid misuse — — — — 14.0 14.3 NA

No change

Abbreviation: NA = not available.

* Based on trend analyses by using a logistic regression model controlling for sex, race/ethnicity, and grade (p<0.05).

†

Based on t-test analysis (p<0.05).

Previous 30 days before the survey.

Insufficient years of data to assess trends.

TABLE 2. Prevalence of lifetime and current use of specific substances and use behaviors among high school students, by demographic

characteristics — Youth Risk Behavior Survey, United States, 2019

Behavior

Sex Race/Ethnicity Grade Sexual identity

Male

(n = 6,641)

(95% CI)

Female

(n = 6,885)

(95% CI)

White,

non-Hispanic

(n = 6,668)

(95% CI)

Black,

non-Hispanic

(n = 2,040)

(95% CI)

Hispanic

(n = 3,038)

(95% CI)

9/10

(n = 7,354)

(95% CI)

11/12

(n = 6,172)

(95% CI)

Heterosexual

(n = 10,853)

(95% CI)

LGB

(n = 1,531)

(95% CI)

Not sure

(n = 591)

(95% CI)

Current use*

Marijuana 22.5

(20.6–24.5)

20.8

(18.7–23.1)

22.1

(19.9–24.6)

21.7

(19.1–24.5)

22.4

(20.4–24.6)

17.1

(15.5–18.8)

26.6

†

(23.6–29.7)

20.9

(19.0–23.0)

31.1

(27.4–35.1)

19.5

(14.8–25.3)

Alcohol 26.4

(24.4–28.6)

31.9

(29.6–34.3)

34.2

(31.7–36.8)

16.8

††

(13.5–20.7)

28.4

††,§§

(26.1–30.8)

22.8

(20.6–25.2)

36.0

†

(33.8–38.3)

28.8

(26.8–30.8)

33.9

(29.8–38.2)

25.3

(20.0–31.4)

Binge drinking 12.7

(11.0–14.6)

14.6

(13.2–16.2)

17.3

(15.1–19.7)

6.2

††

(4.2–9.2)

12.4

††,§§

(11.0–14.0)

8.9

(7.4–10.7)

18.8

†

(17.0–20.8)

13.4

(12.0–15.0)

15.6

(12.8–18.8)

13.1

(9.0–18.8)

Prescription

opioid misuse

6.1

(5.3–7.1)

8.3

(7.0–9.9)

5.5

(4.4–6.9)

8.7

††

(6.5–11.6)

9.8

††

(8.2–11.6)

7.0

(5.8–8.4)

7.3

(6.1–8.8)

6.4

(5.4–7.5)

12.0

(9.6–14.9)

11.5

(8.2–15.9)

Lifetime use

Marijuana 37.0

(34.2–40.0)

36.5

(34.1–38.9)

36.8

(33.9–39.8)

37.5

(34.0–41.1)

39.2

(36.5–41.9)

29.2

(26.7–31.8)

44.8

†

(41.5–48.2)

36.0

(33.3–38.7)

49.6

(45.1–54.1)

27.5

§,¶

(22.4–33.3)

Cocaine 4.9

(4.2–5.8)

2.7

(2.0–3.7)

2.9

(2.2–3.7)

4.0

(2.7–5.9)

5.6

††

(4.5–6.9)

2.8

(2.0–3.7)

5.0

†

(4.1–6.1)

3.3

(2.7–4.0)

7.0

(4.8–10.1)

7.6

(4.3–12.9)

Methamphetamine 2.7

(2.1–3.4)

1.5

(1.0–2.2)

1.2

(0.9–1.6)

3.8

††

(2.4–6.0)

2.7

††

(1.8–4.0)

1.5

(1.0–2.3)

2.6

†

(1.9–3.3)

1.5

(1.2–1.9)

5.0

(3.1–7.9)

6.1

(3.4–10.8)

Heroin 2.3

(1.8–3.1)

1.0

(0.6–1.8)

0.9

(0.6–1.2)

3.4

††

(2.2–5.3)

2.4

††

(1.5–3.9)

1.6

(1.0–2.5)

1.8

(1.3–2.5)

1.2

(0.9–1.6)

3.8

(2.1–7.0)

6.2

(3.4–11.0)

Injection drug use 2.1

(1.5–2.9)

1.1

(0.6–1.9)

0.8

(0.6–1.2)

2.9

††

(1.5–5.5)

2.5

††

(1.8–3.5)

1.6

(1.1–2.3)

1.5

(1.0–2.4)

1.1

(0.8–1.6)

3.5

(2.1–5.7)

5.1

(2.5–10.2)

Synthetic marijuana 7.2

(6.2–8.4)

7.4

(6.2–8.7)

6.7

(5.6–8.0)

5.7

(4.4–7.4)

9.8

††,§§

(8.6–11.3)

6.2

(5.3–7.3)

8.3

†

(7.2–9.7)

6.7

(5.8–7.7)

11.6

(9.0–14.7)

10.4

(6.9–15.5)

Prescription

opioid misuse

12.4

(11.0–14.1)

16.1

(14.1–18.4)

12.7

(10.9–14.7)

15.3

(12.9–18.1)

16.0

(13.5–18.8)

13.6

(11.9–15.5)

14.9

(13.2–16.7)

12.7

(11.2–14.4)

23.9

(19.9–28.3)

19.1

(14.6–24.5)

Abbreviations: CI = confidence interval; LGB = lesbian, gay, or bisexual.

* Previous 30 days before the survey.

†

Significantly different from 9/10 grade students, based on t-test analysis (p<0.05).

Significantly different from heterosexual students, based on t-test analysis (p<0.05).

Significantly different from lesbian, gay, or bisexual students, based on t-test analysis (p<0.05).

Significantly different from male students, based on t-test analysis (p<0.05).

††

Significantly different from white students, based on t-test analysis (p<0.05).

§§

Significantly different from black students, based on t-test analysis (p<0.05).

Supplement

42 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

notable differences in prevalence estimates were identified for

current use of alcohol, binge drinking, current prescription

opioid misuse, and lifetime use of cocaine, methamphetamine,

heroin, injection drug use, and synthetic marijuana. However,

no clear pattern emerged. For example, the prevalence of

current prescription opioid misuse was significantly lower

among white students (5.5%) compared with black (8.7%)

or Hispanic students (9.8%). Conversely, the prevalence of

current alcohol use was lower among black students (16.8%)

compared with white (34.2%) or Hispanic students (28.4%).

Approximately half of the substance use behaviors varied

substantially by grade, with consistently higher prevalence

among 11th-

and 12th-grade students compared with 9th- and

10th-grade students for current marijuana use, current alcohol

use and binge drinking, lifetime marijuana use, lifetime cocaine

use, lifetime methamphetamine use, and lifetime synthetic

marijuana. Prevalence of all but one of the substance use

behaviors (i.e., binge drinking) varied considerably by sexual

identity. Students who identified as lesbian, gay, or bisexual

had a higher prevalence of all substance use behaviors, except

binge drinking, compared with students who identified as

heterosexual. Similarly, students who identified as not sure of

their sexual identity also had higher prevalence of approximately

TABLE 3. Frequency of lifetime and current use among high school students reporting use of specific substances — Youth Risk Behavior Survey,

United States, 2019

Behavior

Frequency

1–2 times

% (95% CI)

3–9 times

% (95% CI)

10–39 times

% (95% CI)

≥40 times

% (95% CI)

Current use*

Marijuana (n = 2,946) 36.7 (33.7–39.8) 21.8 (19.7–24.1) 23.5 (21.4–25.8) 18.0 (15.1–21.3)

Prescription opioid misuse (n = 661) 53.2 (47.9–58.5) 23.3 (18.9–28.3) 13.7 (11.0–16.9) 9.8 (6.3–14.8)

Lifetime use

Marijuana (n = 4,219) 24.6 (22.4–27.1) 20.9 (19.3–22.6) 20.8 (19.2–22.5) 33.6 (30.5–36.9)

Prescription opioid misuse (n = 2,000) 48.8 (45.7–51.9) 24.7 (22.4–27.2) 15.9 (14.0–18.1) 10.6 (8.8–12.7)

Synthetic marijuana (n = 955) 48.8 (44.9–52.6) 20.8 (17.8–24.3) 18.5 (14.9–22.8) 11.9 (9.2–15.3)

Cocaine (n = 557) 45.0 (38.7–51.5) 20.3 (15.7–25.9) 18.5 (14.4–23.6) 16.1 (11.8–21.7)

Methamphetamine (n = 351) 42.9 (34.7–51.5) 15.5 (10.8–21.7) 13.7 (9.0–20.3) 27.9 (19.1–39.0)

Heroin (n = 316) 31.7 (24.4–40.1) 18.6 (12.8–26.2) 16.9 (11.8–23.6) 32.9 (21.7–46.3)

Behavior

Frequency

1–2 days

% (95% CI)

3–9 days

% (95% CI)

10–19 days

% (95% CI)

≥20 days

% (95% CI)

Current use

Alcohol use (n = 3,669) 54.8 (52.6–57.0) 36.6 (34.4–38.8) 5.1 (4.1–6.4) 3.5 (2.6–4.7)

Binge drinking (n = 1,657) 61.2 (56.5–65.7) 31.1 (27.7–34.7) 4.1 (2.7–6.2) 3.6 (2.4–5.3)

Behavior

Frequency

1 time

% (95% CI)

≥2 times

% (95% CI)

Lifetime use

Injection drug use (n = 200) 47.8 (35.4–60.4) 52.2 (39.6–64.6)

Abbreviation: CI = confidence interval.

* Previous 30 days before the survey.

half of the substance use behaviors compared with heterosexual

students, including current prescription opioid misuse, lifetime

cocaine use, lifetime methamphetamine use, lifetime heroin

use, lifetime injection drug use, and lifetime prescription opioid

misuse. However, students who identified as not sure of their

sexual identity had lower prevalence of certain substance use

behaviors compared with students identifying as lesbian, gay,

or bisexual, including current marijuana use, current alcohol

use, and lifetime marijuana use.

Frequency of use (i.e., number of times used or number

of days used) varied across specific substance use behaviors

(Table 3). Among students reporting marijuana use during the

30 days before the survey (i.e., current use), 18.0% reported

using it ≥40 times; 23.5%, 10–39 times; 21.8%, 3–9 times;

and 36.7%, 1–2 times. For current prescription opioid misuse,

9.8% reported misuse ≥40 times; 13.7%, 10–39 times; 23.3%,

3–9 times, and 53.2%, 1–2 times. Among students reporting

lifetime use of specific substances, marijuana had the highest

percentage of students reporting use ≥40 times (33.6%),

followed by heroin (32.9%), methamphetamine (27.9%),

and cocaine (16.1%). Lifetime prescription opioid misuse and

lifetime synthetic cannabinoid use were the two substances

with the highest percentages reporting use 1–2 times (48.8%

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 43

US Department of Health and Human Services/Centers for Disease Control and Prevention

each), followed by cocaine (45.0%), and methamphetamine

(42.9%). Among students reporting current alcohol use or

current binge drinking, more than half of students (54.8%

and 61.2%, respectively) reported those behaviors on 1–2 days.

Among students who had ever injected drugs (1.2%), 47.8%

reported injecting drugs 1 time, and 52.2% reported injecting

drugs ≥2 times.

Students reporting current prescription opioid misuse

commonly indicated use of other substances (Figure). Overall,

43.5% of students reporting current prescription opioid misuse

also reported current marijuana use, 59.4% reported current

alcohol use, and 30.3% reported current binge drinking.

Lifetime use of other substances among students reporting

current prescription opioid misuse was 62.9% for marijuana,

30.3% for synthetic marijuana, 20.5% for cocaine, 15.0%

for methamphetamine, and 14.0% for heroin. Approximately

12.4% of students reporting current prescription opioid misuse

also reported lifetime injection drug use.

In adjusted analyses, current prescription opioid misuse

varied by sex, race/ethnicity, and sexual identity (Table 4).

Specifically, males were significantly less likely to report

engaging in current prescription opioid misuse (aPR: 0.69;

95% CI: 0.57–0.84) compared with females (referent group);

black and Hispanic students were significantly more likely to

have engaged in prescription opioid misuse (black students,

aPR: 1.49; 95% CI: 1.05–2.10; Hispanic students, aPR: 1.52;

95% CI: 1.12–2.05) compared with white students (referent

group); and students identifying as lesbian, gay, or bisexual

were more likely to report current prescription opioid misuse

(aPR: 1.35; 95% CI: 1.02–1.79) compared with students

identifying as heterosexual (referent group). All substance

use behaviors included in the model, except for marijuana

use, were significantly associated with current prescription

opioid misuse, ranging from aPR= 2.13 (95% CI: 1.59–

2.86) for lifetime synthetic marijuana use and aPR = 2.13

(95% CI: 1.58–2.86) for previous 30-day binge drinking, to

aPR=5.08 (95% CI: 2.72–9.49) for lifetime injection drug use.

Discussion

This report provides key insights into substance use

behaviors of U.S. high school students during 2009–2019.

Encouraging findings include decreasing prevalence of current

alcohol use and decreases in the prevalence of lifetime use of

marijuana, cocaine, methamphetamine, heroin, synthetic

marijuana, and injection drug use. However, the findings in

this report underscore that substance use among high school

students remains common, with approximately one in three

students reporting current alcohol use, one in five reporting

current marijuana use, and one in seven reporting current

binge drinking. Because of the ongoing U.S. opioid crisis, of

particular concern are the high rates of lifetime (one in seven

students) and current prescription opioid misuse (one in 14

students) and high rates of co-occurring substance use among

students currently misusing prescription opioids.

Notable demographic differences and patterns in substance

use among high school students are identified in this report.

Specifically, males had substantially higher rates of cocaine,

methamphetamine, heroin, and injection drug use compared

with females, and females had substantially higher rates of

current alcohol use and current binge drinking. In addition,

females had higher rates of current prescription opioid misuse

compared with males, and this pattern persisted in multivariable

models where males had lower adjusted prevalence ratios for

current prescription opioid misuse compared with females.

Differences also occurred in substance use patterns across

racial/ethnic groups. For example, black and Hispanic students

reported higher rates of current prescription opioid misuse

compared with white students; in contrast, white students

reported the highest rates of current alcohol use and binge

drinking, followed by Hispanic and black students. These

substance use patterns by racial/ethnic groups are similar to

those identified in other U.S. youth substance use surveys

(https://www.samhsa.gov/data/report/2018-nsduh-detailed-

tables). This heterogeneity in substance use patterns among

demographic groups can be used to guide development of

tailored and targeted prevention messages and interventions.

Particularly noteworthy were the universally elevated rates

of substance use among self-identified sexual minority youths

compared with heterosexual youths, which is consistent with

previous research (12). In addition to findings regarding

broader substance use patterns, this report provides actionable

information on prescription opioid misuse among high school

students that can be applied to ongoing efforts for preventing

opioid misuse, use disorders, and overdoses. Specifically, the

high rates of co-occurring substance use, especially alcohol and

marijuana use, among students currently misusing prescription

opioids highlights the importance of prevention efforts that focus

on general substance use risk and protective factors. Notably,

these associations are not limited to high school students because

binge drinking and marijuana use are associated with increased

prescription opioid misuse among both adults and adolescents

(13). Finally, sexual minority youths also had significantly

higher prevalence of current prescription opioid misuse even

after controlling for other demographic and substance use

characteristics, which is consistent with their overall pattern

of higher rates of substance use in this study. It also further

emphasizes the importance of identifying tailored prevention

strategies to address disparities among this vulnerable population.

Supplement

44 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

FIGURE. Percentage of co-occurring substance use behaviors among high school students who reported previous 30-day prescription opioid

misuse* — Youth Risk Behavior Survey, United States, 2019

Percentage

100

30-day

marijuana

use

30-day

alcohol

use

30-day

binge

alcohol

use

Lifetime

marijuana

use

Lifetime

synthetic

marijuana

use

Lifetime

cocaine

use

Lifetime

meth-

amphetamine

use

Lifetime

heroin

use

Lifetime

injection

drug

use

95% Condence interval

Substance use behaviors

* Unweighted N = 661.

Scientific evidence for the effective prevention of substance

use indicates the importance of interventions that target risk

and protective factors at the individual, family, and community

levels to maximize their public health impact (1–3). Risk

factors include adverse childhood experiences at the individual

level, limited parental monitoring and involvement and

active substance use in the home at the family level, and easy

availability and accessibility to alcohol and other substances

and community norms favorable toward use of alcohol and

other substances at the community level (1,2,5). In addition,

studies have demonstrated that youth alcohol use is associated

with adult alcohol use, and that both community-level and

individual-level alcohol use are affected by population-level

alcohol policies (e.g., those that reduce the availability and

accessibility of alcohol and increase its price) (14).

The ability to reach young persons during early elementary

ages, before they begin using substances, and throughout

adolescence makes the school environment well-suited

for prevention programming. School-based substance use

prevention programs that focus on broad-based skill building

(e.g., psychosocial development, life-skills development, and

social-emotional learning and connectedness) have greater

promise than substance-specific programs (15,16). In addition,

multifaceted programs that incorporate aspects of individual,

school, and family interventions (e.g., the Promoting School-

community-university Partnerships to Enhanced Resilience

[PROSPER] program and Communities That Care [CTC])

have demonstrated effectiveness at reducing or preventing

youth substance use (17,18).

Broader prevention policies for changing the environment

in which youths live (e.g., those that reduce the availability of

substances) can also be used as part of a comprehensive approach

for reducing youth substance use. The U.S. Community

Preventive Services Task Force recommends certain population-

level strategies (e.g., increasing alcohol taxes and regulating

the number and concentration of places that sell alcohol as

interventions for reducing excessive alcohol use, including

alcohol use among youths) (https://www.thecommunityguide.

org/resources/what-works-preventing-excessive-alcohol-

consumption). Enhanced enforcement of existing substance

use policies (e.g., prescription drug monitoring programs

that are used universally with near–real-time data and laws

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 45

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 4. Multivariable logistic regression model examining

individual-level characteristics associated with previous 30-day

prescription opioid misuse among high school students — Youth

Risk Behavior Survey, United States, 2019

Characteristic

Adjusted*

prevalence ratios

(95% CI)

Demographics

Sex

Female Referent

Male 0.69 (0.57–0.84)

Race/Ethnicity

White, non-Hispanic Referent

Black, non-Hispanic 1.49 (1.05–2.10)

Hispanic 1.52 (1.12–2.05)

Grade

9 or 10 Referent

11 or 12 0.85 (0.66–1.10)

Sexual identity

Heterosexual Referent

Lesbian, gay, or bisexual 1.35 (1.02–1.79)

Not sure 1.37 (0.86–2.17)

Substance use and use behaviors

Alcohol use

No previous 30-day use Referent

Previous 30-day nonbinge drinking 2.28 (1.63–3.19)

Previous 30-day binge drinking 2.13 (1.58–2.86)

Marijuana use

No lifetime use Referent

Lifetime use, but no previous 30-day use 1.21 (0.89–1.65)

Previous 30-day use 1.31 (0.95–1.80)

Lifetime synthetic marijuana use

No Referent

Yes 2.13 (1.59–2.86)

Lifetime use of cocaine, heroin, or methamphetamine

No Referent

Yes 2.49 (1.89–3.27)

Lifetime injection drug use

No Referent

Yes 5.08 (2.72–9.49)

Abbreviation: CI = confidence interval.

* Adjusted prevalence ratios were calculated from a single logistic regression

model that included all covariates listed in this table.

prohibiting sales of alcohol to persons aged <21 years) also

can help reduce substance use among youths (19,20). In

addition, strategies for expanding access to evidence-based pain

treatment and improving prescribing of prescription opioids

through safer prescribing practices can help reduce opioid

misuse and overdoses. Improving opioid prescribing can have

dual benefits by reducing the environmental availability of

prescription opioids for diversion and misuse and reducing

the risk for misuse associated with the prescription of opioids

to youths (2).

Limitations

General limitations for the YRBS are available in the overview

report of this supplement (11). The findings in this report are

subject to at least three additional limitations. First, the questions

assessing lifetime and current prescription opioid misuse refer

to prescription pain medicine; however, the questions provide

examples of opioid-containing prescription medications only.

Therefore, if students considered nonopioid prescription pain

medications when answering, an overestimation of prescription

opioid misuse prevalence might have occurred. Second, many

of the substance use questions included common street names

for drugs; however, newly introduced street names or street

names specific to certain geographic areas were not included,

which might have resulted in underreporting of substance

use behaviors. Finally, there was variation in the amount of

missing data for some substance use variables (e.g., the largest

amount missing was for current prescription opioid misuse

[5,000 missing observations]). Missing data might result from

a variety of factors, such as students choosing not to answer

questions or inconsistent responses to similar questions that

are set to missing during the data cleaning process (11). In

addition, schools selected to participate in the national YRBS

and in a state or local YRBS only complete the local version

of the survey; as a result, questions included on the national

survey but not the local survey are set to missing.

Conclusion

The findings in this report indicate that youth substance use

has declined in recent years; however, substance use, including

misuse of prescription opioids, remains common among

U.S. high school students. Opportunities exist for bringing

to scale evidence-based policies, programs, and practices

that aim to reduce risk factors and strengthen protective

factors among youths in conjunction with initiatives already

underway for combating the U.S. opioid overdose epidemic.

Disproportionately affected populations (e.g., sexual minority

youths) might benefit from tailored substance use interventions

combined with more widespread implementation of broader

population-level policy strategies.

Conicts of Interest

All authors have completed and submitted the International

Committee of Medical Journal Editors form for disclosure of

potential conflicts of interest. No potential conflicts of interest

were disclosed.

Supplement

46 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

References

1. Office of the Surgeon General. Facing addiction in America: the Surgeon

General’s report on alcohol, drugs, and health. Rockville, MD: US

Department of Health and Human Services; 2016. https://addiction.

surgeongeneral.gov/sites/default/files/surgeon-generals-report.pdf

2. Compton WM, Jones CM, Baldwin GT, Harding FM, Blanco C,

Wargo EM. Targeting youth to prevent later substance use disorder:

an underutilized response to the U.S. opioid crisis. Am J Public

Health 2019;109(Suppl3):S185–9. https://doi.org/10.2105/

AJPH.2019.305020

3. Steinberg L. Risk taking in adolescence: new perspectives from brain

and behavioral science. Curr Dir Psychol Sci 2007;16:55–9. https://doi.

org/10.1111/j.1467-8721.2007.00475.x

4. DuPont RL, Han B, Shea CL, Madras BK. Drug use among youth:

National survey data support a common liability of all drug use. Prev

Med 2018;113:68–73. https://doi.org/10.1016/j.ypmed.2018.05.015

5. Hawkins JD, Jenson JM, Catalano R, et al. Unleashing the power

of prevention. Am J Med Res (N Y) 2016;3:39–74. https://doi.

org/10.22381/AJMR3120162

6. Clayton HB, Bohm MK, Lowry R, Ashley C, Ethier KA. Prescription

opioid misuse associated with risk behaviors among adolescents. Am J Prev

Med 2019;57:533–9. https://doi.org/10.1016/j.amepre.2019.05.017

7. Gaither JR, Leventhal JM, Ryan SA, Camenga DR. National trends in

hospitalizations for opioid poisonings among children and adolescents,

1997 to 2012. JAMA Pediatr 2016;170:1195–201. https://doi.

org/10.1001/jamapediatrics.2016.2154

8. Cerda M, Santaella J, Marshall BD, Kim JH, Martins SS. Nonmedical

prescription opioid use in childhood and early adolescence predicts

transitions to heroin use in young adulthood: a national study. J Pediatr

2015;167:605–12e1–2.

9. Jones CM. Trends and key correlates of prescription opioid injection

misuse in the United States. Addict Behav 2018;78:145–52. https://doi.

org/10.1016/j.addbeh.2017.10.018

10. Jones CM, Christensen A, Gladden RM. Increases in prescription

opioid injection abuse among treatment admissions in the United States,

2004-2013. Drug Alcohol Depend 2017;176:89–95. https://doi.

org/10.1016/j.drugalcdep.2017.03.011

11. Underwood JM, Brener N, Thornton J, et al. Overview and methods

for the Youth Risk Behavior Surveillance System—United States, 2019.

In: Youth Risk Behavior Surveillance—United States, 2019. MMWR

Suppl 2020;69(No. Suppl 1).

12. Rostad WL, Clayton HB, Estefan LF, Johns MM. Substance use and

disparities in teen dating violence victimization by sexual identity

among high school students. Prev Sci 2020;21:398–407. https://doi.

org/10.1007/s11121-019-01049-7

13. Esser MB, Guy GP Jr, Zhang K, Brewer RD. Binge drinking and

prescription opioid misuse in the U.S., 2012–2014. Am J Prev Med

2019;57:197–208. https://doi.org/10.1016/j.amepre.2019.02.025

14. Xuan Z, Blanchette JG, Nelson TF, et al. Youth drinking in the

United States: relationships with alcohol policies and adult drinking.

Pediatrics 2015;136:18–27. https://doi.org/10.1542/peds.2015-0537

15. Griffin KW, Botvin GJ. Evidence-based interventions for preventing

substance use disorders in adolescents. Child Adolesc Psychiatr Clin N

Am 2010;19:505–26. https://doi.org/10.1016/j.chc.2010.03.005

16. Faggiano F, Minozzi S, Versino E, Buscemi D. Universal school-

based prevention for illicit drug use. Cochrane Database Syst Rev

2014;12:CD003020. https://doi.org/10.1002/14651858.CD003020.

pub3

17. Spoth R, Redmond C, Shin C, Greenberg MT, Feinberg ME, Trudeau L.

PROSPER delivery of universal preventive interventions with young

adolescents: long-term effects on emerging adult substance misuse and

associated risk behaviors. Psychol Med 2017;47:2246–59. https://doi.

org/10.1017/S0033291717000691

18. Oesterle S, Kuklinski MR, Hawkins JD, Skinner ML, Guttmannova K,

Rhew IC. Long-term effects of the Communities That Care trial on

substance use, antisocial behavior, and violence through age 21 years.

Am J Public Health 2018;108:659–65. https://doi.org/10.2105/

AJPH.2018.304320

19. Elder R, Lawrence B, Janes G, et al. Enhanced enforcement of laws

prohibiting sale of alcohol to minors: systematic review of effectiveness

for reducing sales and underage drinking. Washington, DC:

Transportation Research Board; 2007:181–8. Transportation Research

Circular no. E-C123.

20. Haegerich TM, Jones CM, Cote PO, Robinson A, Ross L. Evidence

for state, community and systems-level prevention strategies to address

the opioid crisis. Drug Alcohol Depend 2019;204:107563. https://doi.

org/10.1016/j.drugalcdep.2019.107563

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 47

US Department of Health and Human Services/Centers for Disease Control and Prevention

Suicidal Ideation and Behaviors Among High School Students —

Youth Risk Behavior Survey, United States, 2019

Asha Z. Ivey-Stephenson, PhD

; Zewditu Demissie, PhD

; Alexander E. Crosby, MD

; Deborah M. Stone, ScD

; Elizabeth Gaylor, MPH

;

Natalie Wilkins, PhD

; Richard Lowry, MD

; Margaret Brown, DrPH

Division of Injury Prevention, National Center for Injury Prevention and Control, CDC;

Division of Adolescent and School Health, National Center for HIV/AIDS,

Viral Hepatitis, STD, and TB Prevention, CDC;

Office of the Director, National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention, CDC

Abstract

Suicide is the second leading cause of death among high school-aged youths 14–18 years after unintentional injuries. This

report summarizes data regarding suicidal ideation (i.e., seriously considered suicide) and behaviors (i.e., made a suicide plan,

attempted suicide, and made a suicide attempt requiring medical treatment) from CDC’s 2019 Youth Risk Behavior Survey.

Results are reported overall and by sex, grade, race/ethnicity, sexual identity, and sex of sexual contacts, overall and within sex

groups. Trends in suicide attempts during 2009–2019 are also reported by sex, race/ethnicity, and grade. During 2009–2019,

prevalence of suicide attempts increased overall and among female, non-Hispanic white, non-Hispanic black, and 12th-grade

students. Data from 2019 reflect substantial differences by demographics regarding suicidal ideation and behaviors. For example,

during 2019, a total of 18.8% of students reported having seriously considered suicide, with prevalence estimates highest among

females (24.1%); white non-Hispanic students (19.1%); students who reported having sex with persons of the same sex or with

both sexes (54.2%); and students who identified as lesbian, gay, or bisexual (46.8%). Among all students, 8.9% reported having

attempted suicide, with prevalence estimates highest among females (11.0%); black non-Hispanic students (11.8%); students

who reported having sex with persons of the same sex or with both sexes (30.3%); and students who identified as lesbian, gay, or

bisexual (23.4%). Comprehensive suicide prevention can address these differences and reduce prevalence of suicidal ideation and

behaviors by implementing programs, practices, and policies that prevent suicide (e.g., parenting programs), supporting persons

currently at risk (e.g., psychotherapy), preventing reattempts (e.g., emergency department follow-up), and attending to persons

who have lost a friend or loved one to suicide.

Introduction

Suicidal behavior presents a major challenge to public

health in the United States and globally (1). Although fatal

(i.e., suicide) and nonfatal (e.g., suicide attempts) suicidal

behaviors are a public health concern across the life span,

they are of particular concern for youths and young adults

aged 10–24 years. During 2018, a total of 48,344 persons

(all ages) died from suicide, and suicide was the 10th leading

cause of death overall in the United States, accounting for

approximately 1.7% of all deaths (2). Among high school–

aged youths (14–18 years), 2,039 suicides occurred that year,

making it the second leading cause of death for this age group

after unintentional injuries (n = 2,590). Suicide accounted

for approximately 33.9% or approximately one of every

three injury-related deaths among this age group (2). During

2009–2018, suicide rates among youths aged 14–18 years

Corresponding author: Asha Z. Ivey-Stephenson, PhD, Division of

Injury Prevention, National Center for Injury Prevention and Control,

CDC. Telephone: 770-488-0940; E-mail: [email protected].

increased by 61.7% from 6.0 to 9.7 per 100,000 population

(2). Although suicide is a major public health problem, many

more youths make suicide attempts and struggle with suicidal

ideation. For example, during 2018, according to data from

a nationally representative sample of emergency departments

(EDs), approximately 95,000 youths aged 14–18 years visited

EDs for self-harm injuries (2).

One objective of the Healthy People 2020 Mental Health

and Mental Disorders is to reduce suicide attempts by

adolescents that resulted in an injury, poisoning, or overdose

that had to be treated by a doctor or nurse (3). The Youth Risk

Behavior Survey (YRBS) monitors six categories of priority

health behaviors and experiences among adolescents, with

four questions specifically related to suicide (4). This report

summarizes 2019 YRBS data regarding suicidal ideation and

behaviors among high school students and presents trends in

suicide attempts among this population during 2009–2019.

The report is intended for decision-makers, prevention

program practitioners, and those who work in youth-serving

organizations so that they can identify vulnerable youths and

take appropriate action to direct prevention resources to those

young persons.

Supplement

48 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

Methods

Data Source

This report includes data from the 2009–2019 cycles of

the YRBS, a cross-sectional, school-based survey conducted

biennially since 1991. Each survey year, CDC collects data

from a nationally representative sample of public and private

school students in grades 9–12 in the 50 U.S. states and the

District of Columbia. Additional information about YRBS

sampling, data collection, response rates, and processing

is available in the overview report of this supplement (4).

The overview report also includes information about the

classification of sexual identity and sex of sexual contacts and

standard data analysis methods. The prevalence estimates for

all suicidal ideation and behavior questions for the overall

study population and by sex, race/ethnicity, grade, and sexual

orientation are available at https://nccd.cdc.gov/youthonline/

App/Default.aspx. The full YRBS questionnaire is available at

https://www.cdc.gov/healthyyouth/data/yrbs/pdf/2019/2019_

YRBS-National-HS-Questionnaire.pdf.

Measures

Four suicidal ideation and behavior variables are included in

this report. Suicidal ideation was measured with the question,

“During the past 12 months, did you ever seriously consider

attempting suicide?” Making a suicide plan was measured

with the question, “During the past 12 months, did you

make a plan about how you would attempt suicide?” (These

two questions had “yes” or “no” response options.) Suicide

attempts were measured with the question, “During the

past 12 months, how many times did you actually attempt

suicide?” Suicide attempts were assessed by frequency of

attempts, but the variable was dichotomized into yes or no

responses for analytic purposes. Lastly, students were asked,

“If you attempted suicide during the past 12 months, did any

attempt result in an injury, poisoning, or overdose that had to

be treated by a doctor or nurse?” This question is referred to

in this report as, “made a suicide attempt requiring medical

treatment.” The response options for the last question were,

“I did not attempt suicide during the past 12 months,” “yes,”

or “no”; however, this variable was also dichotomized into yes

or no responses for analysis.

Analysis

Analyses of these suicidal ideation and behavior variables

included examining associations between each item and

demographic characteristics, including sex (male/female), race/

ethnicity (non-Hispanic white [white], non-Hispanic black

[black], or Hispanic), grade (9, 10, 11, or 12), sexual identity

(heterosexual; lesbian, gay, or bisexual [LGB]; or not sure), or

sex of sexual contacts (sexual contact with only the opposite

sex, sexual contact with only the same sex or both sexes, and

no sexual contact). Associations by race/ethnicity, grade, sexual

identity, and sex of sexual contacts were calculated for the

overall study population but also separately for male and female

students. Statistical differences were determined by using

chi-square analyses at the p <0.05 level of significance. Linear

trends for 2009–2019 were examined for attempted suicide by

sex, race/ethnicity, and grade. All analyses of suicidal ideation

and behaviors were conducted among the full sample, and

analysis of behavior variables was not limited to students who

reported suicidal ideation (i.e., analysis conducted among the

full sample). Additional information about the methods used

to conduct YRBS trend analyses are provided in the overview

report of this supplement (4).

Results

Suicidal Ideation and Behaviors,

Overall and by Sex

During the 12 months before the survey, 18.8% of students

nationwide reported seriously considered attempting suicide

(prevalence significantly higher among female [24.1%] than

male [13.3%] students), and among students nationwide,

15.7% of students had made a plan about how they would

attempt suicide (prevalence significantly higher among female

[19.9%] than male [11.3%] students), and 8.9% of students

had attempted suicide ≥1 time (prevalence significantly higher

among female [11.0%] than male [6.6%] students) (Table 1).

Nationwide, 2.5% of students had made a suicide attempt

requiring medical treatment, with a prevalence significantly

higher among female (3.3%) than male (1.7%) students.

Suicidal Ideation and Behaviors by Race/

Ethnicity and Grade, Overall and by Sex

Overall, a significant difference occurred in having seriously

considered attempting suicide by race/ethnicity (white: 19.1%;

black: 16.9%; Hispanic: 17.2%) (Table 2), with a significant

difference by race/ethnicity among male students (white:

13.8%; black: 10.7%; Hispanic: 11.4%) but not among female

students. No significant differences (overall or by sex) occurred

in having seriously considered attempting suicide by grade.

Among students reporting having made a suicide plan, a

significant difference occurred by race and ethnicity overall

(white: 15.7%; black: 15.0%; Hispanic: 14.7%) but not among

male or female students. No significant difference occurred in

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 49

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 1. Percentage of high school students who had seriously considered attempting suicide, had made a suicide plan, had attempted

suicide, or had made a suicide attempt requiring medical treatment during the 12 months before the survey, by sex — Youth Risk Behavior

Survey, United States, 2019

Behavior

Female

% (95% CI)

Male

% (95% CI)

Total

% (95% CI)

Chi-square

(p value)

Seriously considered attempting suicide 97.922 (<0.001)

Yes 24.1 (22.3–26.0) 13.3 (12.2–14.5) 18.8 (17.6–20.0) NA

No 75.9 (74.0–77.7) 86.7 (85.5–87.8) 81.2 (80.0–82.4) NA

Made a suicide plan 109.568 (<0.001)

Yes 19.9 (18.4–21.6) 11.3 (10.3–12.4) 15.7 (14.6–16.9) NA

No 80.1 (78.4–81.6) 88.7 (87.6–89.7) 84.3 (83.1–85.4) NA

Attempted suicide 27.037 (<0.001)

Yes 11.0 (9.7–12.5) 6.6 (5.5–8.1) 8.9 (7.9–10.0) NA

No 89.0 (87.5–90.3) 93.4 (91.9–94.5) 91.1 (90.0–92.1) NA

Made a suicide attempt requiring medical treatment* 10.313 (0.003)

Yes 3.3 (2.6–4.2) 1.7 (1.3–2.3) 2.5 (2.1–3.0) NA

No 96.7 (95.8–97.4) 98.3 (97.7–98.7) 97.5 (97.0–97.9) NA

Abbreviations: CI = confidence interval; NA = not applicable.

* Made a suicide attempt that resulted in an injury, poisoning, or overdose that had to be treated by a doctor or nurse.

having made a suicide plan by grade overall or among female

students, but a significant difference was identified among male

students (9th grade: 9.5%; 10th grade: 10.4%; 11th grade:

12.1%; 12th grade: 13.6%). Students who had attempted

suicide were significantly different by race/ethnicity overall

(white: 7.9%; black: 11.8%; Hispanic: 8.9%) and among

female students (white: 9.4%; black: 15.2%; Hispanic: 11.9%)

but not among male students. No significant differences

existed in having attempted suicide by grade (overall or by

sex). In addition, no significant difference in having made a

suicide attempt requiring medical treatment was noted by race/

ethnicity or grade, overall or by sex.

Suicidal Ideation and Behaviors by Sexual

Identity and Sex of Sexual Contacts,

Overall and by Sex

A significant difference occurred in having seriously

considered attempting suicide by sexual identity overall

(heterosexual: 14.5%; LGB: 46.8%; not sure: 30.4%) (Table 3)

and among both female (heterosexual: 18.0%; LGB: 49.0%;

not sure: 35.9%) and male (heterosexual: 11.4%; LGB:

40.4%; not sure: 21.7%) students. Similarly, having seriously

considered attempting suicide varied by sex of sexual contacts,

overall (had sexual contact with only the opposite sex: 19.4%;

had sexual contact with only the same sex or both sexes:

54.2%; had no sexual contact: 13.0%), among female (had

sexual contact with only the opposite sex: 25.3%; had sexual

contact with only the same sex or both sexes: 59.2%; had no

sexual contact: 16.2%), and among male (had sexual contact

with only the opposite sex: 14.6%; had sexual contact with

only the same sex or both sexes: 39.1%; had no sexual contact:

9.7%) students.

Overall, a significant difference occurred in having made

a suicide plan by sexual identity (heterosexual: 12.1%; LGB:

40.2%; not sure: 23.9%), with a significant difference among

both female (heterosexual: 14.6%; LGB: 42.4%; not sure:

28.1%) and male (heterosexual: 9.9%; LGB: 33.0%; not sure:

17.4%) students. Similarly, a significant difference was noted

among students having made a suicide plan by sex of sexual

contacts, overall (had sexual contact with only the opposite

sex: 16.5%; had sexual contact with only the same sex or both

sexes: 44.0%; had no sexual contact: 10.9%), with a significant

difference among both female (had sexual contact with only

the opposite sex: 20.7%; had sexual contact with only the same

sex or both sexes: 48.2%; had no sexual contact: 13.8%) and

male (had sexual contact with only the opposite sex: 12.9%;

had sexual contact with only the same sex or both sexes: 31.2%;

had no sexual contact: 7.9%) students.

A significant difference existed in having attempted suicide

by sexual identity, overall (heterosexual: 6.4%; LGB: 23.4%;

not sure: 16.1%) and among both female (heterosexual: 7.9%;

LGB: 23.6%; not sure: 15.2%) and male (heterosexual: 5.1%;

LGB: 23.8%; not sure: 16.4%) students. Similarly, a significant

difference was identified in having attempted suicide by sex

of sexual contacts, overall (had sexual contact with only the

opposite sex: 9.3%; had sexual contact with only the same

sex or both sexes: 30.3%; no sexual contact: 4.8%), with a

significant difference among both female (had sexual contact

with only the opposite sex: 11.4%; had sexual contact with

only the same sex or both sexes: 31.4%; no sexual contact:

6.1%) and male (had sexual contact with only the opposite

sex: 7.5%; had sexual contact with only the same sex or both

sexes: 26.5%; no sexual contact: 3.5%) students.

Finally, a significant difference occurred in having made a

suicide attempt requiring medical treatment by sexual identity,

Supplement

50 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 2. Percentage of high school students who had seriously considered attempting suicide, had made a suicide plan, had attempted

suicide, or had made a suicide attempt requiring medical treatment during the 12 months before the survey, by sex, race/ethnicity, and

grade — Youth Risk Behavior Survey, United States, 2019

Behavior

Female

% (95% CI)

Chi-square

(p value)

Male

% (95% CI)

Chi-square

(p value)

Total

% (95% CI)

Chi-square

(p value)

Seriously considered attempting suicide

Race/Ethnicity —* 1.504 (0.230) — 4.989 (0.005) — 5.870 (0.002)

White, non-Hispanic 24.3 (21.9–26.9) — 13.8 (12.3–15.3) — 19.1(17.6–20.8) —

Black, non-Hispanic 23.7 (20.7–27.1) — 10.7 (8.2–13.7) — 16.9 (15.3–18.7) —

Hispanic 22.7 (19.3–26.5) — 11.4 (9.8–13.3) — 17.2 (15.2–19.4) —

Grade — 0.209 (0.889) — 0.790 (0.507) — 0.820 (0.491)

9 23.7 (20.7–27.0) — 11.9 (9.9–14.2) — 17.7 (15.7–19.8) —

10 23.6 (20.3–27.3) — 13.2 (11.1–15.8) — 18.5 (16.1–21.1) —

11 24.9 (22.5–27.6) — 13.6 (11.5–16.0) — 19.3 (17.7–21.1) —

12 24.0 (20.7–27.6) — 14.9 (12.4–17.7) — 19.6 (17.5–21.9) —

Made a suicide plan

Race/Ethnicity — 1.652 (0.194) — 2.358 (0.087) — 3.043 (0.041)

White, non-Hispanic 19.2 (16.9–21.8) — 12.0 (10.6–13.5) — 15.7 (14.1–17.4) —

Black, non-Hispanic 20.4 (17.6–23.5) — 10.1 (7.3–13.9) — 15.0 (12.9–17.5) —

Hispanic 19.6 (16.9–22.6) — 9.6 (8.0–11.4) — 14.7 (13.0–16.7) —

Grade — 0.461 (0.711) — 3.195 (0.035) — 0.652 (0.587)

9 20.4 (17.9–23.2) — 9.5 (7.9–11.4) — 14.8 (13.1–16.6) —

10 20.3 (17.2–23.7) — 10.4 (8.6–12.4) — 15.4 (13.4–17.7) —

20.4 (17.6–23.5) — 12.1 (10.3–14.2) — 16.4 (14.5–18.5) —

12 18.5 (15.7–21.6) — 13.6 (11.4–16.1) — 16.2 (14.3–18.3) —

Attempted suicide

Race/Ethnicity — 2.973 (0.044) — 1.505 (0.229) — 2.866 (0.050)

White, non-Hispanic 9.4 (7.8–11.3) — 6.4 (5.1–7.8) — 7.9 (6.9–9.1) —

Black, non-Hispanic 15.2 (10.8–20.9) — 8.5 (5.6–12.9) — 11.8 (8.7–15.9) —

Hispanic 11.9 (9.0–15.6) — 5.5 (3.9–7.6) — 8.9 (7.1–11.1) —

Grade — 1.878 (0.150) — 0.384 (0.765) — 0.311 (0.817)

9 12.8 (10.7–15.3) — 6.0 (4.5–7.9) — 9.4 (7.9–11.1) —

10 11.0 (9.1–13.3) — 6.5 (4.7–9.0) — 8.8 (7.4–10.5) —

11 10.4 (8.1–13.3) — 6.7 (5.2–8.8) — 8.6 (7.1–10.4) —

12 9.4 (6.9–12.6) — 7.3 (5.2–10.0) — 8.5 (6.8–10.6) —

Made a suicide attempt requiring medical treatment

†

Race/Ethnicity — 0.446 (0.721) — 1.583 (0.210) — 1.387 (0.262)

White, non-Hispanic 2.9 (1.9–4.4) — 1.2 (0.8–1.9) — 2.1 (1.5–2.8) —

Black, non-Hispanic 3.8 (2.3–6.2) — 2.9 (1.5–5.5) — 3.3 (2.2–4.9) —

Hispanic 3.6 (2.6–4.9) — 2.3 (1.4–3.9) — 3.0 (2.3–3.8) —

Grade — 0.406 (0.750) — 0.571 (0.638) — 0.274 (0.844)

9 3.3 (2.3–4.8) — 1.3 (0.7–2.3) — 2.3 (1.7–3.1) —

10 3.6 (2.3–5.5) — 1.7 (0.9–3.3) — 2.7 (1.8–3.9) —

11 2.7 (1.7–4.3) — 2.0 (1.2–3.2) — 2.3 (1.7–3.3) —

12 3.4 (2.2–5.3) — 1.9 (1.0–3.9) — 2.7 (2.0–3.7) —

Abbreviation: CI = Confidence interval.

* Not applicable.

†

Made a suicide attempt that resulted in an injury, poisoning, or overdose that had to be treated by a doctor or nurse.

overall (heterosexual: 1.7%; LGB: 6.3%; not sure: 5.2%) and

among both female (heterosexual: 2.3%; LGB: 6.6%; not

sure: 3.8%) and male (heterosexual: 1.3%; LGB: 5.9%; not

sure: 7.6%) students. A significant difference also was noted

in having made a suicide attempt requiring medical treatment

by sex of sexual contacts, overall (had sexual contact with only

the opposite sex: 2.6%; had sexual contact with only the same

sex or both sexes: 10.2%; had no sexual contact: 1.0%) and

among both female (had sexual contact with only the opposite

sex: 3.4%; had sexual contact with only the same sex or both

sexes: 10.4%; had no sexual contact: 1.4%) and male (had

sexual contact with only the opposite sex: 1.9%; had sexual

contact with only the same sex or both sexes: 9.4%; had no

sexual contact: 0.5%) students.

Trends in Suicide Attempts, Overall and by

Sex, Race/Ethnicity, and Grade

Among the total student population, the percentage of

students who had attempted suicide ≥1 time during the

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 51

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 3. Percentage of high school students who had seriously considered attempting suicide, had made a suicide plan, had attempted sui-

cide, or had made a suicide attempt requiring medical treatment during the 12 months before the survey, by sex, sexual identity, and sex of

sexual contacts — Youth Risk Behavior Survey, United States, 2019

Behavior

Female

% (95% CI)

Chi-square

(p value)

Male

% (95% CI)

Chi-square

(p value)

Total

% (95% CI)

Chi-square

(p value)

Seriously considered attempting suicide

Sexual identity — 75.728 (<0.001) — 22.231 (<0.001) — 88.194 (<0.001)

Heterosexual 18.0 (16.3–20.0) —* 11.4 (10.4–12.6) — 14.5 (13.4–15.7) —

LGB 49.0 (44.8–53.3) — 40.4 (33.9–47.1) — 46.8 (43.1–50.6) —

Not sure 35.9 (29.5–42.9) — 21.7 (14.8–30.5) — 30.4 (25.4–35.9) —

Sex of sexual contacts — 64.007 (<0.001) — 13.972 (<0.001) — 66.938 (<0.001)

Opposite sex only 25.3 (22.8–28.0) — 14.6 (12.9–16.5) — 19.4 (17.6–21.4) —

Same sex only or both sexes 59.2 (52.5–65.6) — 39.1 (29.3–49.9) — 54.2 (49.0–59.3) —

No sexual contact 16.2 (14.2–18.3) — 9.7 (8.1–11.7) — 13.0 (11.8–14.3) —

Made a suicide plan

Sexual identity — 66.568 (<0.001) — 19.732 (<0.001) — 90.368 (<0.001)

Heterosexual 14.6 (13.2–16.0) — 9.9 (8.9–11.0) — 12.1 (11.1–13.1) —

LGB 42.4 (38.4–46.4) — 33.0 (26.4–40.3) — 40.2 (36.6–44.0) —

Not sure 28.1 (22.1–35.0) — 17.4 (11.8–24.8) — 23.9 (19.4–29.0) —

Sex of sexual contacts — 56.442 (<0.001) — 18.435 (<0.001) — 62.470 (< 0.001)

Opposite sex only 20.7 (18.4–23.3) — 12.9 (11.5–14.6) — 16.5 (14.9–18.1) —

Same sex only or both sexes 48.2 (42.8–53.6) — 31.2 (23.8–39.7) — 44.0 (39.7–48.4) —

No sexual contact 13.8 (12.3–15.6) — 7.9 (6.7–9.4) — 10.9 (9.8–12.1) —

Attempted suicide

Sexual identity — 26.919 (<0.001) — 15.972 (<0.001) — 40.352 (<0.001)

Heterosexual 7.9 (6.6–9.4) — 5.1 (4.2–6.3) — 6.4 (5.6–7.4) —

LGB 23.6 (20.0–27.6) — 23.8 (17.8–31.1) — 23.4 (20.0–27.1) —

Not sure 15.2 (9.6–23.3) — 16.4 (9.9–26.0) — 16.1 (11.1–22.8) —

Sex of sexual contacts — 58.123 (<0.001) — 12.379 (<0.001) — 66.202 (<0.001)

Opposite sex only 11.4 (9.5–13.5) — 7.5 (5.8–9.6) — 9.3 (7.9–10.8) —

Same sex only or both sexes 31.4 (27.0–36.1) — 26.5 (17.5–38.0) — 30.3 (25.9–35.0) —

No sexual contact 6.1 (4.8–7.8) — 3.5 (2.6–4.8) — 4.8 (4.0–5.8) —

Made a suicide attempt requiring medical treatment

†

Sexual identity — 7.893 (0.001) — 5.592 (0.008) — 13.034 (<0.001)

Heterosexual 2.3 (1.6–3.2) — 1.3 (0.9–1.9) — 1.7 (1.4–2.2) —

LGB 6.6 (5.0–8.7) — 5.9 (3.2–10.6) — 6.3 (4.8–8.3) —

Not sure 3.8 (1.6–8.4) — 7.6 (3.6–15.2) — 5.2 (3.0–9.0) —

Sex of sexual contacts — 14.728 (<0.001) — 10.517 (<0.001) — 23.046 (<0.001)

Opposite sex only 3.4 (2.4–4.8) — 1.9 (1.3–2.9) — 2.6 (2.0–3.3) —

Same sex only or both sexes 10.4 (7.5–14.2) — 9.4 (4.9–17.6) — 10.2 (7.6–13.4) —

No sexual contact 1.4 (0.8–2.4) — 0.5 (0.3–1.1) — 1.0 (0.6–1.5) —

Abbreviations: CI = confidence interval; LGB = lesbian, gay, or bisexual.

* Not applicable.

†

Made a suicide attempt that resulted in an injury, poisoning, or overdose that had to be treated by a doctor or nurse.

12 months before the survey experienced a significant linear

increase from 6.3% during 2009 to 8.9% during 2019

(Figure 1–3). Among female students, a significant linear

increase (from 8.1% to 11.0%) occurred in the prevalence of

having attempted suicide. No significant linear change was

observed for the prevalence of having attempted suicide among

male students. By race/ethnicity, significant linear increases in

having attempted suicide were observed for white (from 5.0%

to 7.9%) and black (from 7.9% to 11.8%) but not Hispanic

students. By grade, a significant linear increase in having

attempted suicide was observed only for 12th-grade students

(from 4.2% to 8.5%).

Discussion

During 2019, approximately one in five (18.8%) youths had

seriously considered attempting suicide, one in six (15.7%)

had made a suicide plan, one in 11 (8.9%) had made an

attempt, and one in 40 (2.5%) had made a suicide attempt

requiring medical treatment. Linear trends in suicide attempts

have increased during 2009–2019 overall and among certain

demographic groups.

The 2019 YRBS data highlight considerable differences

in suicidal ideation, plans, attempts, and attempts requiring

medical treatment. Consistent with previous research, during

2019, females had more suicidal ideation, suicide plans,

and suicide attempts, including attempts requiring medical

Supplement

52 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

FIGURE 1. Percentage of high school students who attempted suicide during the 12 months before the survey, overall and by sex — Youth Risk

Behavior Survey, United States, 2009–2019

100

2009 2011 2013 2015 2017 2019

Female

Overall

Male

Percentage

Year

FIGURE 2. Percentage of high school students who attempted suicide during the 12 months before the survey, by race/ethnicity — Youth Risk

Behavior Survey, United States, 2009–2019

100

2009 2011 2013 2015 2017 2019

Percentage

Year

White, non-Hispanic

Black, non-Hispanic

Hispanic

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 53

US Department of Health and Human Services/Centers for Disease Control and Prevention

FIGURE 3. Percentage of high school students who attempted suicide during the 12 months before the survey, by grade — Youth Risk Behavior

Survey, United States, 2009–2019

100

2009 2011 2013 2015 2017 2019

Percentage

Year

9th

10th

11th

12th

treatment, than males (5). Certain racial/ethnic differences

also were identified. For example, black male students had

the lowest prevalence estimates of suicidal ideation. Regarding

suicide attempts, black students (male and female) had the

highest prevalence estimates. This finding is consistent with

previous research (6). Also consistent with previous research are

the study findings regarding sexual orientation and sex of sexual

contacts (5). Namely, prevalence estimates of suicidal ideation,

suicide plans, attempts, and attempts requiring medical

treatment were highest among sexual minority youths, those

who identified as LGB, and youths who reported having had

sexual contact with the same or with both sexes during 2019.

Adolescence is a developmental stage often characterized

by rapid and extensive physical and psychosocial changes (7).

It also represents a time for expanded identity development,

with sexual identity development representing a complex,

multidimensional, and often stressful process for youths (8).

The potential dissonance between sexual identity and behavior

and the social rejection sexual minority youths often experience

can contribute to increased suicidal ideation and behaviors

along with an increased risk for suicide (9,10). Because of the

high prevalence of suicidal ideation and behaviors among sexual

minority youths, additional research is needed to determine

how best to support this vulnerable group. Such research

might evaluate strategies designed to reduce sexual minority

stress (e.g., discrimination and victimization resulting from

sharing one’s sexual orientation) (11) and unhealthy behaviors

(e.g., substance use) and the resultant impact on suicidal

ideation and behaviors.

Suicide attempts are a known risk factor for and the greatest

predictor of death by suicide (12), which is the rationale for

investigating trends only on this outcome. The number of

children and adolescents who sought medical treatment at EDs

for suicide attempts increased sharply from 2007 (540,000)

to 2015 (960,000) (13). Either a linear increase or no change

in suicide attempts by variables reported here (i.e., sex, race/

ethnicity, and grade) was identified for 2009–2019. Although

for a different period (1991–2017), other researchers also

have reported that suicide attempts among black students

increased significantly (6). More specifically, previous findings

indicated that suicide attempts increased at an accelerating rate

among black females, and black male youths had a substantial

increase in attempts requiring medical treatment during the

period (6). Future studies are needed to continue monitoring

trends in suicidal ideation and behavior for black students and

other race/ethnicity groups. For example, such studies might

include more detailed analyses among the American Indian/

Alaska Native youth population who have been reported to

be at increased risk for suicidal behaviors (6).

In this analysis, one notable finding emerged by sex and

grade; a substantial increase in making a suicide plan occurred

among males as grade increased. To address this trend, schools

can consider a sex-by-grade–specific approach to implementing

suicide prevention or intervention activities.

Supplement

54 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

Limitations

General limitations for the YRBS are available in the

overview report of this supplement (4). The findings in this

report are subject to at least one additional limitation. This

analysis is conducted among all students (i.e., does not separate

ideation from behaviors); suicide patterns might differ between

those who experienced suicidal ideation and those who did not.

Future Directions

To address the health differences in suicidal ideation and

behaviors observed by student demographics and to decrease

these outcomes overall, a comprehensive approach to suicide

prevention, including programs, practices, and policies based

on the best available evidence, is needed. Such an approach

addresses the range of risk and protective factors occurring

across the individual, relationship, community, and societal

levels. A comprehensive approach also seeks to prevent suicide

risk, identify and support youths at increased risk, prevent

attempts and reattempts, and help survivors of suicide loss (i.e.,

those grieving the death of a friend or loved one). States and

communities, including school communities, can use strategies

with such best available evidence as that documented in the

CDC Preventing Suicide Technical Package (14).

Preventing adverse childhood experiences (e.g., child

maltreatment) can help reduce suicide risk among adolescents

through strategies that promote safe, stable, nurturing

relationships and environments in childhood (15). Other

strategies in a comprehensive approach to suicide prevention

include supporting families by strengthening economic

supports and teaching coping and problem-solving skills

among children, adolescents, and their parents; promoting

connectedness between youths and their schools, teachers,

peers, and family; creating protective environments in schools

and at home (e.g., limiting access to such lethal means among

students at risk as medications and firearms); promoting help-

seeking behaviors; reducing stigma; and training teachers and

adults in recognizing signs of suicide and responding effectively

through referrals to evidence-based treatment (e.g., cognitive-

behavioral therapy) (14). Finally, schools and the media should

respond to and report on suicides in ways that are supportive

and responsible (e.g., not sensationalizing deaths), thereby

avoiding additional suicides (i.e., suicide contagion) (14).

Conclusion

Suicide is a leading cause of death among youths; however,

many more youths are at risk for suicide as a result of

experiencing suicidal ideation, making suicide plans, and

attempting suicide, making a focus on nonfatal suicidal

behavior a crucial public health priority. During 2009–2019,

trends in suicide attempts among adolescents increased overall

and among many demographic groups. Prevalence estimates

of suicidal ideation, suicide plans, attempts, and attempts

requiring medical treatment were highest among sexual

minority youths and youths who reported having had sexual

contact with the same or with both sexes. Regarding differences

by race/ethnicity, black students had the highest prevalence

estimates for attempted suicide. Factors at the individual,

relationship, community, and societal levels likely contribute

to the differences in suicide attempts among different racial/

ethnic groups and sexual minority youths and the differences

observed by sex and grade. More research is needed to better

understand the risk and protective factors to determine which

suicide prevention strategies might best serve each group.

The findings in this report underscore the importance of a

comprehensive approach to suicide prevention, which would

provide necessary support to those at risk, decrease suicidal

ideation and behaviors, and ultimately prevent suicide among

youths and save lives.

Conflicts of Interest

All authors have completed and submitted the International

Committee of Medical Journal Editors form for disclosure of

potential conflicts of interest. No potential conflicts of interest

were disclosed.

References

1. World Health Organization. Preventing suicide: a global imperative.

Geneva, Switzerland: World Health Organization; 2014. https://apps.

who.int/iris/bitstream/handle/10665/131056/9789241564779_eng.pdf

2. CDC. Web-based Injury Statistics Query and Reporting System

(WISQARS). Atlanta, GA: US Department of Health and Human

Services, CDC, National Center for Injury Prevention and Control;

2020. https://www.cdc.gov/injury/wisqars/index.html

3 US Department of Health and Human Services. Healthy people 2020.

Washington, DC: US Department of Health and Human Services,

Office of Disease Prevention and Health Promotion; 2020. https://www.

healthypeople.gov

4. Underwood JM, Brener N, Thornton J, et al. Overview and methods

for the Youth Risk Behavior Surveillance System—United States, 2019.

In: Youth Risk Behavior Surveillance—United States, 2019. MMWR

Suppl 2020;69(No. Suppl 1).

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 55

US Department of Health and Human Services/Centers for Disease Control and Prevention

5. Kann L, McManus T, Harris WA, et al. Youth risk behavior surveillance—

United States, 2017. MMWR Surveill Summ 2018;67(No. SS-8):1–114.

https://doi.org/10.15585/mmwr.ss6708a1

6. Lindsey MA, Sheftall AH, Xiao Y, Joe S. Trends of suicidal behaviors

among high school students in the United States: 1991–2017. Pediatrics

2019;144:e20191187. https://doi.org/10.1542/peds.2019-1187

7. Tsang SK, Hui EK, Law BC. Positive identity as a positive youth

development construct: a conceptual review. ScientificWorldJournal

2012;2012:529691. https://doi.org/10.1100/2012/529691

8. Rosario M, Schrimshaw EW, Hunter J, Braun L. Sexual identity

development among lesbian, gay, and bisexual youths: consistency and

change over time. J Sex Res 2006;43:46–58. https://doi.

org/10.1080/00224490609552298

9. Annor FB, Clayton HB, Gilbert LK, et al. Sexual orientation discordance

and nonfatal suicidal behaviors in U.S. high school students. Am J Prev

Med 2018;54:530–8. https://doi.org/10.1016/j.amepre.2018.01.013

10. Hong JS, Espelage DL, Kral MJ. Understanding suicide among sexual

minority youth in America: an ecological systems analysis. J Adolesc

2011;34:885–94. https://doi.org/10.1016/j.adolescence.2011.01.002

11. Meyer IH. Minority stress and mental health in gay men. J Health Soc

Behav 1995;36:38–56. https://doi.org/10.2307/2137286

12. Franklin JC, Ribeiro JD, Fox KR, et al. Risk factors for suicidal thoughts

and behaviors: A meta-analysis of 50 years of research. Psychol Bull

2017;143:187–232. https://doi.org/10.1037/bul0000084

13. Burstein B, Agostino H, Greenfield B. Suicidal attempts and ideation

among children and adolescents in US emergency departments, 2007–

2015. JAMA Pediatr 2019;173:598–600. https://doi.org/10.1001/

jamapediatrics.2019.0464

14. Stone DM, Holland KM, Bartholow B, Crosby AE, Davis S, Wilkins

N. Preventing suicide: a technical package of policies, programs, and

practices. Atlanta, GA: US Department of Health and Human Services,

CDC, National Center for Injury Prevention and Control; 2017. https://

www.cdc.gov/violenceprevention/pdf/suicideTechnicalPackage.pdf

15. CDC. Preventing adverse childhood experiences (ACEs): leveraging the

best available evidence. Atlanta, GA: US Department of Health and

Human Services, CDC, National Center for Injury Prevention and

Control; 2019. https://www.cdc.gov/violenceprevention/pdf/

preventingACES.pdf

Supplement

56 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

Tobacco Product Use Among High School Students — Youth Risk

Behavior Survey, United States, 2019

MeLisa R. Creamer, PhD

; Sherry Everett Jones, PhD, JD

; Andrea S. Gentzke, PhD

; Ahmed Jamal, MBBS

; Brian A. King, PhD

Office on Smoking and Health, National Center for Chronic Disease Prevention and Health Promotion, CDC;

Division of Adolescent and School Health,

National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention, CDC

Abstract

Tobacco product use is the leading cause of preventable disease, disability, and death in the United States. This report used data

from the 2019 Youth Risk Behavior Survey to assess the following among U.S. high school students: ever use of cigarettes and

electronic vapor products, current use (≥1 day during the 30 days before the survey) of tobacco products, frequent use (≥20 days

during the 30 days before the survey) among current users of tobacco products, trends in use over time, and usual source of

electronic vapor products among current electronic vapor product users. In 2019, a total of 50.1% of U.S. high school students

had ever used electronic vapor products, and 24.1% had ever tried cigarette smoking. Current electronic vapor product use was

32.7%, current cigarette smoking was 6.0%, current cigar smoking was 5.7%, and current smokeless tobacco use was 3.8%.

Approximately 36.5% of students were current users of any tobacco product, and 8.2% were current users of two or more tobacco

products. Frequent use among users of individual products was 32.6% for electronic vapor products, 28.5% for smokeless tobacco,

22.2% for cigarettes, and 18.4% for cigars. Among current electronic vapor product users who were aged ≤17 years, the most

commonly reported source was borrowing them from someone else (42.8%). Significant decreases occurred in current cigarette

smoking (1991: 27.5%; 2019: 6.0%), cigar smoking (1997: 22.0%; 2019: 5.7%), and smokeless tobacco use (2017: 5.5%;

2019: 3.8%). However, significant increases occurred in current electronic vapor product use (2015: 24.1%; 2019: 32.7%)

and any tobacco product use (2017: 19.5%; 2019: 36.5%). Although current cigarette smoking, cigar smoking, and smokeless

tobacco use has decreased among high school students, the increased prevalence of electronic vapor product use among youths

is concerning. Continued surveillance for all tobacco product use is warranted for guiding and evaluating public health policy at

the local, state, tribal, and national levels.

Introduction

Smoking is the leading cause of preventable premature

disease and death in the United States (1). An estimated 88%

of adult daily cigarette smokers report first trying cigarette

smoking before age 18 years (2). Previous reports indicate

decreases in current cigarette smoking (i.e., use during the

30 days before the survey) among U.S. high school students

from a high of 36.4% in 1997 to 8.8% in 2017 (3). However,

there are a variety of tobacco products, including smokeless

tobacco products, cigars, and most recently, electronic vapor

products (e.g., e-cigarettes).

Electronic vapor products have evolved since entering the

U.S. marketplace in 2007. Initial products were disposable,

resembled the size and shape of conventional cigarettes,

and used free-base nicotine; however, newer products are

rechargeable, resemble common objects (e.g., USB flash

drives), and typically deliver nicotine salts (4,5), which

Corresponding author: Andrea S. Gentzke, PhD, Office on Smoking and

Health, National Center for Chronic Disease Prevention and Health

Promotion, CDC. Telephone: 404-498-1795; E-mail: [email protected].

allow higher levels of nicotine to be inhaled more easily by

the user (6). Sales of these newer generation, or “pod-mod,”

products have increased in the United States during recent

years. For example, sales of JUUL, the most commonly sold

e-cigarette in the United States since December 2017, increased

approximately 600% during 2016–2017 from 2.2 million

unit sales to 16.2 million unit sales (7). By December 2018,

JUUL accounted for an estimated 76% of the $322.1 million

total e-cigarettes sales that occurred that month in the United

States (8). The popularity of these electronic vapor products

among youths is likely the result of multiple factors, including

advertising exposure, availability of youth-appealing flavors,

curiosity, and social exposure through friends and others (4–6).

In 2014, prevalence of electronic vapor product use among

high school students surpassed prevalence of cigarette smoking

(9), and according to data from the 2017 Youth Risk Behavior

Survey (YRBS), 13.2% of high school students had used

electronic vapor products during the previous 30 days (3).

These findings align with increases in use observed in other

national surveys of youth in the United States. For example,

according to the National Youth Tobacco Survey (NYTS),

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 57

US Department of Health and Human Services/Centers for Disease Control and Prevention

current electronic vapor product use among high school students

increased 78% (11.7% to 20.8%) during 2017–2018 (9).

Youth use of tobacco products in any form is unsafe (1,4).

Cigarette smoking harms nearly every organ in the body, and

smokeless tobacco product use is associated with multiple health

risks, including cancers of the mouth (1). Moreover, the aerosol

in electronic vapor products can contain harmful ingredients,

including heavy metals, ultrafine particles, and nicotine (4).

Nicotine is highly addictive, can harm the developing adolescent

brain, and can prime the brain for addiction to other drugs

(4,10). In addition, a growing body of scientific literature

suggests that youths who use e-cigarettes are more likely to smoke

conventional cigarettes in the future (4,10).

Surveillance for tobacco product use among youths is crucial

for guiding and evaluating tobacco control strategies at local,

state, tribal, and national levels. This report presents the latest

data from the 2019 YRBS to assess the following among U.S.

high school students: ever use of cigarettes and electronic

vapor products; current use (≥1 day during the 30 days before

the survey) of tobacco products (electronic vapor products,

cigarettes, cigars [cigars/cigarillos/little cigars], smokeless

tobacco [chewing tobacco, snuff, dip, snus, or dissolvable

tobacco products], any tobacco product, and two or more

products); frequent use (≥20 days during the 30 days before

the survey) of tobacco products among current users of those

products; trends in tobacco product use over time; and usual

source of obtaining electronic vapor products among current

electronic vapor product users.

Methods

Data Source

This report includes data from the 1991–2019 cycles of

CDC’s national YRBS, a cross-sectional, school-based survey

conducted biennially since 1991. Each survey year, CDC

collects data from a nationally representative sample of public

and private school students in grades 9–12 in the 50 U.S. states

and the District of Columbia. Additional information about

YRBS sampling, data collection, response rates, and processing

is available in the overview report for this supplement (11). The

prevalence estimates for all tobacco product use questions for

the overall study population and by sex, race/ethnicity, grade,

and sexual orientation are available at https://nccd.cdc.gov/

youthonline/App/Default.aspx. The full YRBS questionnaire

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

pdf/2019/2019_YRBS-National-HS-Questionnaire.pdf.

Measures

Ever use, which was defined as having used the product

at least one time during their lifetime, was assessed for two

distinct tobacco products: cigarettes and electronic vapor

products. Ever cigarette smoking was assessed by the question,

“Have you ever tried cigarette smoking, even one or two

puffs?” Ever electronic vapor product use was assessed by the

question, “Have you ever used an electronic vapor product?”

with a preamble that read, “The next 3 questions ask about

electronic vapor products, such as JUUL, Vuse, MarkTen, and

blu. Electronic vapor products include e-cigarettes, vapes, vape

pens, e-cigars, e-hookahs, hookah pens, and mods.”

Current use (≥1 day during the 30 days before the survey)

was assessed for four tobacco products: 1) current electronic

vapor product use was assessed by the question, “During the

past 30 days, on how many days did you use an electronic

vapor product?” 2) current cigarette smoking was assessed by

the question, “During the past 30 days, on how many days did

you smoke cigarettes?” 3) current cigar smoking was assessed

by the question, “During the past 30 days, on how many days

did you smoke cigars, cigarillos, or little cigars?” and 4) current

smokeless tobacco use was assessed by the question, “During

the past 30 days, on how many days did you use chewing

tobacco, snuff, dip, snus, or dissolvable tobacco products, such

as Copenhagen, Grizzly, Skoal, or Camel Snus? (Do not count

any electronic vapor products.)” Response options for each of

the four questions were 0 days, 1–2 days, 3–5 days, 6–9 days,

10–19 days, 20–29 days, and all 30 days. Among current users

of each individual product, frequent use was also calculated.

Frequent use was defined as having used the respective product

on ≥20 days during the 30 days before the survey.

Two composite measures were also investigated in this

analysis. Any current tobacco product use was defined as any

use of electronic vapor products, cigarettes, cigars, or smokeless

tobacco during the 30 days before the survey. Use of two or

more products was defined as current use of two or more of

the four assessed tobacco products.

Respondents also were asked how they usually obtained

electronic vapor products by the question (referred to as source

hereinafter), “During the past 30 days, how did you usually

get your own electronic vapor products? (Select only one

response.)” Response options were as follows: I did not use any

electronic vapor products during the past 30 days; I bought

them in a store such as a convenience store, supermarket,

discount store, gas station, or vape store; I got them on the

Internet; I gave someone else money to buy them for me; I

borrowed them from someone else; a person who can legally

buy these products gave them to me; I took them from a store

Supplement

58 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

or another person; or I got them some other way. Analysis of this

variable was limited to current electronic vapor product users.

The demographic characteristics of students analyzed for

this report included sex (female or male), grade (9, 10, 11, or

12), age (≤15 years, 16 or 17 years, or ≥18 years), and sexual

identity (heterosexual; lesbian, gay, or bisexual; or not sure).

In addition, students were classified into four racial/ethnic

categories: non-Hispanic white (white); non-Hispanic black

(black); Hispanic or Latino of any race (Hispanic); and other

or multiple races (non-Hispanic). The numbers of students

in the other or multiple racial/ethnic groups were too small

to produce statistically stable estimates; therefore, those data

are not presented as a separate group in this report but were

retained in the overall analytic sample.

Analysis

Prevalence of use for each respective tobacco product was

estimated for all years for which data were available. For 2019,

statistically significant pairwise differences by sex, grade, race/

ethnicity, age, and sexual identity were determined for each of

the assessed tobacco product use behaviors by using t-tests. For

each tobacco product, changes in prevalence were compared for

2017 and 2019 by using t-tests. In addition, t-tests were used

to compare how students who were ≤17 years and ≥18 years

usually obtained their electronic vapor products; these age

groups were used because age 18 years was the federal legal

age of sale for tobacco products at the time of the survey.

Prevalence estimates were considered statistically different if

the p value was <0.05.

To identify temporal trends, logistic regression analyses

were used to model linear and quadratic time effects while

controlling for sex, grade, and race/ethnicity. Linear time

effects were analyzed for current electronic vapor products

use (2015–2019), and both linear and quadratic time effects

were analyzed for current cigarette smoking (1991–2019) and

current cigar smoking (1997–2019). Because of substantial

changes in the question wording for smokeless tobacco

products in 2017, trends were not assessed for smokeless

tobacco. Additional information about the methods used to

conduct YRBS trend analyses are provided in the overview

report of this supplement (11).

Results

Among U.S. high school students in 2019, a total of 50.1%

(95% confidence interval [CI]: 48.1–52.2) had ever used

electronic vapor products, and 24.1% (CI: 21.3–27.0) had

ever tried cigarette smoking (data not shown). Prevalence of

current use was 32.7% for electronic vapor products, 6.0% for

cigarettes, 5.7% for cigars, and 3.8% for smokeless tobacco. In

addition, 36.5% of students had currently used any tobacco

products, and 8.2% had currently used two or more tobacco

products (Table 1).

Prevalence of tobacco product use varied by demographic

groups, with current use of cigarettes, cigars, smokeless tobacco,

and two or more tobacco products being higher among male

students than female students. Although differences in tobacco

product use varied by grade, prevalence of current use of each

individual product, any tobacco product, and two or more

tobacco products was higher among 12th-grade students than

9th-grade students. Prevalence of current use of electronic

vapor products, cigarettes, any tobacco product, and two or

more tobacco products was higher among white and Hispanic

students than black students, and the prevalence of electronic

vapor products and any tobacco product use was higher among

white than Hispanic students. Prevalence of current cigar use

was higher among students aged ≥18 years than those aged 16

or 17 years and those aged ≤15 years. For all other individual

products, any tobacco product, and two or more tobacco

products, prevalence increased in each age category. Among

sexual identity groups, prevalence of electronic vapor product

use was higher among heterosexual students and lesbian, gay, or

bisexual students than not-sure students. Prevalence of current

use of cigarettes, cigars, any tobacco product, and two or more

tobacco products was higher among lesbian, gay, or bisexual

students than heterosexual students. Finally, the prevalence

of any tobacco product use was higher among lesbian, gay, or

bisexual students than not-sure students.

In 2019, among the 32.7% of current electronic vapor

product users, 32.6% were frequent users; among the 5.7%

current cigarette smokers, 22.2% were frequent users; among

the 3.8% current cigar smokers, 18.4% were frequent users;

and among the 6.0% current smokeless tobacco product users,

28.5% were frequent users. From 2017 to 2019, among current

electronic vapor product users, a significant increase occurred

in frequent use (from 25.1% to 32.6%), and among current

cigarette smokers, a significant decrease occurred in frequent

use (from 30.0% in 2017 to 22.2% in 2019) (Figure 1). No

significant changes in frequent use of smokeless tobacco or

cigars were observed among users of these products from

2017 to 2019.

The usual source of electronic vapor products among current

users varied by age (Table 2). Among current electronic vapor

product users who were aged ≤17 years, the most commonly

reported usual source of electronic vapor products was

borrowing them from someone else (42.8%). Among those

aged ≥18 years, the most commonly reported source was

buying them in a store (56.4%). Compared with students

aged ≤17 years, a higher prevalence of students aged ≥18 years

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 59

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 1. Percentage of high school students who were current tobacco users, by selected characteristics and type of tobacco product — Youth

Risk Behavior Survey, United States, 2019

Characteristic

Electronic vapor products* Cigarettes

†

Cigars

Smokeless tobacco

Any tobacco product** ≥2 products

††

% (95% CI) % (95% CI) % (95% CI) % (95% CI) % (95% CI) % (95% CI)

Total 32.7 (30.7–34.8) 6.0 (5.0–7.2) 5.7 (4.8–6.7) 3.8 (3.2–4.6) 36.5 (33.6–39.5) 8.2 (7.0–9.5)

Sex

§§

Male 32.0 (29.7–34.3) 6.9 (5.7–8.4) 7.4 (6.4–8.6) 5.8 (4.7–7.1) 36.3 (33.3–39.3) 10.4 (9.0–11.9)

Female 33.5 (30.9–36.1) 4.9 (3.8–6.4) 3.8 (2.8–5.1) 1.6 (1.2–2.1) 36.6 (33.1–40.2) 5.8 (4.5–7.5)

Grade

¶¶

9 25.0 (22.8–27.4) 3.8 (2.8–5.1) 3.8 (2.7–5.2) 2.0 (1.4–3.0) 27.7 (24.8–30.9) 5.3 (4.2–6.6)

10 30.5 (27.3–33.8) 5.2 (3.9–6.9) 4.7 (3.5–6.2) 3.6 (2.6–5.0) 34.3 (30.3–38.6) 7.3 (5.6–9.6)

11 35.9 (32.3–39.8) 5.9 (4.5–7.7) 6.0 (4.6–7.8) 3.9 (3.0–5.1) 39.8 (35.7–44.1) 8.4 (6.7–10.4)

12 40.4 (37.5–43.4) 9.0 (7.6–10.7) 8.5 (6.9–10.4) 5.5 (4.3–7.1) 45.0 (41.3–48.7) 11.9 (10.3–13.7)

Race/Ethnicity***

Black, non-Hispanic 19.7 (16.9–22.8) 3.3 (2.3–4.6) 5.3 (4.1–6.8) 2.8 (1.8–4.4) 24.7 (21.3–28.4) 4.8 (3.7–6.2)

Hispanic 31.2 (28.6–33.8) 6.0 (4.3–8.4) 6.1 (4.7–8.0) 3.1 (2.3–4.3) 33.8 (31.1–36.7) 7.9 (6.2–10.0)

White, non-Hispanic 38.3 (36.0–40.7) 6.7 (5.3–8.4) 5.9 (4.7–7.4) 4.4 (3.3–5.7) 42.0 (38.3–45.9) 9.5 (7.8–11.5)

Age group (yrs)

†††

≤15 25.9 (24.1–27.9) 4.2 (3.2–5.4) 4.2 (3.1–5.6) 2.7 (2.0–3.8) 29.1 (26.2–32.1) 5.8 (4.6–7.2)

16 or 17 35.2 (32.3–38.3) 6.0 (4.8–7.4) 5.7 (4.5–7.0) 3.7 (3.0–4.6) 38.8 (35.2–42.4) 8.4 (6.8–10.1)

≥18 42.8 (39.0–46.7) 10.9 (8.6–13.6) 10.2 (8.1–12.7) 7.2 (5.5–9.2) 49.1 (44.9–53.4) 14.2 (12.0–16.7)

Sexual identity

§§§

Heterosexual 32.8 (30.5–35.2) 5.2 (4.3–6.3) 5.2 (4.4–6.1) 3.7 (3.1–4.4) 36.1 (33.1–39.2) 7.8 (6.7–9.0)

Lesbian, gay, or bisexual 34.1 (30.8–37.6) 10.4 (7.8–13.7) 8.1 (5.9–11.1) 3.2 (2.0–5.2) 40.3 (36.2–44.4) 10.4 (8.0–13.5)

Not sure 24.9 (19.8–30.7) 7.4 (4.8–11.3) 7.2 (4.3–12.0) 5.5 (3.1–9.5) 30.0 (23.3–37.6) 8.1 (5.4–11.9)

Abbreviation: CI = confidence interval.

* Percentage of students who used an electronic vapor product, including e-cigarettes, e-cigars, e-pipes, vape pipes, vaping pens, e-hookahs, and hookah pens

(e.g., blu, NJOY, Vuse, MarkTen, Logic, Vapin Plus, eGo, and Halo), on ≥1 day during the 30 days before the survey.

†

Percentage of students who smoked cigarettes on ≥1 day during the 30 days before the survey.

Percentage of students who smoked cigars, cigarillos, or little cigars on ≥1 day during the 30 days before the survey.

Percentage of students who used smokeless tobacco, including chewing tobacco, snuff, dip, snus, or dissolvable tobacco products (e.g., Red Man, Levi Garrett,

Beechnut, Skoal, Skoal Bandits, Copenhagen, Camel Snus, Marlboro Snus, General Snus, Ariva, Stonewall, or Camel Orbs), but not including any electronic vapor

products, on ≥1 day during the 30 days before the survey.

** Percentage of students who smoked cigarettes or cigars or used smokeless tobacco or an electronic vapor product, on ≥1 day during the 30 days before the survey.

††

Percentage of students who used ≥2 of the following tobacco products: cigarettes, cigars (cigars, cigarillos, or little cigars), an electronic vapor product, or smokeless

tobacco, on ≥1 day during the 30 days before the survey.

§§

Sex pairwise comparisons assessed by t-test (p<0.05): for cigarettes, cigars, smokeless tobacco, and ≥2 products, male students were significantly different (p<0.05)

from female students.

¶¶

Grade pairwise comparisons assessed by t-test (p<0.05): for electronic vapor products and any tobacco product: all pairwise comparisons were significantly

different (p<0.05); for cigarettes, cigars, and ≥2 products: 12th grade was significantly different (p<0.05) than 9th, 10th, and 11th grades; 11th grade was significantly

different (p<0.05) than 9th grade; for smokeless tobacco: 12th grade was significantly different (p<0.05) than 9th, 10th, and 11th grades; 10th and 11th grades

were significantly different (p<0.05) than 9th grade.

*** Race/ethnicity pairwise comparisons assessed by t-test (p<0.05): for electronic vapor products and any tobacco product: all pairwise comparisons were significantly

different (p<0.05); for cigarettes and ≥2 products: white and Hispanic were significantly different (p<0.05) than black.

†††

Age pairwise comparisons assessed by t-test (p<0.05): for electronic vapor products, cigarettes, smokeless tobacco, any tobacco product, and ≥2 products: all

pairwise comparisons were significantly different (p<0.05); for cigars: ≥18 years was significantly different (p<0.05) than 16–17 years and ≤15 years.

§§§

Sexual identity pairwise comparisons assessed by t-test (p<0.05): for electronic vapor products: heterosexual and lesbian, gay, or bisexual were significantly

different (p<0.05) than not-sure students; for cigarettes, cigars, and ≥2 products: lesbian, gay, or bisexual was significantly different (p<0.05) than heterosexual;

for any tobacco product: lesbian, gay, or bisexual was significantly different (p<0.05) than heterosexual and not-sure students.

usually bought electronic vapor products in a store. In contrast,

compared with older students, a higher prevalence of students

aged ≤17 years got them on the Internet, gave someone else

money to buy them, borrowed them from someone else, got

them from a person who could legally buy them, or got them

some other way.

Trend analyses indicated that during 2015–2019, a significant

linear increase occurred in prevalence of current electronic

vapor products use (from 24.1% to 32.7%) (Figure 2). Trend

analyses also indicated that during 1991–2019, a significant

linear decrease in current cigarette smoking was observed (from

27.5% to 6.0%). A significant quadratic trend in cigarette

smoking also was identified: a 6-year increase in prevalence

(from 27.5% in 1991 to 36.4% in 1997) was followed by

a 22-year decrease (from 36.4% in 1997 to 6.0% in 2019).

Additionally, during 1997–2019, a significant linear decrease

(from 22.0% to 5.7%) occurred in the overall prevalence

of current cigar smoking. A significant quadratic trend also

was identified: a 16-year decrease in prevalence (from 22.0%

in 1997 to 12.6% in 2013) was followed by another 6-year

decrease, but at a different rate of decrease (from 12.6% in

2013 to 5.7% in 2019).

During 2017–2019, a significant increase occurred in current

electronic vapor products use (from 13.2% to 32.7%) and

Supplement

60 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

FIGURE 1. Prevalence of frequent tobacco use* among current users, by type of tobacco product

†

— Youth Risk Behavior Survey, United States,

2017 and 2019

100

Electronic vapor products Cigarettes Cigars Smokeless tobacco

Prevalence (%)

Tobacco product

2019

2017

* Frequent use was defined as use on ≥20 days during the 30 days before the survey.

†

Frequent use was assessed among respondents who reported current use (on ≥1 day during the 30 days before the survey) of each tobacco product. In 2017, among

the 13.2% of students nationwide who used electronic vapor products on ≥1 day during the 30 days before the survey; among the 8.8% of students nationwide

who smoked cigarettes on ≥1 day during the 30 days before the survey; among the 8.0% of students nationwide who smoked cigars on ≥1 day during the 30 days

before the survey; among the 5.5% of students nationwide who used smokeless tobacco on ≥1 day during the 30 days before the survey. In 2019, among the 32.7%

of students nationwide who used electronic vapor products on ≥1 day during the 30 days before the survey; among the 6.0% of students nationwide who smoked

cigarettes on ≥1 day during the 30 days before the survey; among the 5.7% of students nationwide who smoked cigars on ≥1 day during the 30 days before the

survey; among the 3.8% of students nationwide who used smokeless tobacco on ≥1 day during the 30 days before the survey.

Differences from 2017 to 2019 were assessed by t-test (p<0.05): A significant increase occurred in frequent use of electronic vapor products; a significant decrease

occurred in frequent use of cigarettes; and no change occurred in frequent use of cigars/cigarillos/little cigars and smokeless tobacco.

any tobacco product use (from 19.5% to 36.5%). During

2017–2019, significant decreases were observed in current

cigarette smoking (from 8.8% to 6.0%), current cigar smoking

(from 8.0% to 5.7%), and current smokeless tobacco use (from

5.5% to 3.8%). No change occurred in use of two or more

tobacco products during 2017–2019.

Discussion

In 2019, a total of 36.5% of high school students currently

used any tobacco product, with electronic vapor products being

the most commonly used product. This reflects an increase in

use of electronic vapor products from 2017 to 2019, findings

that are consistent with those from other national surveillance

systems, including NYTS (9,12) and Monitoring the Future

(13). For example, NYTS results demonstrated that, among

high school students, e-cigarette use increased from 11.7%

in 2017 to 27.5% in 2019 (9,12). These increases align

with the increasing popularity of newer electronic vapor

product devices, including JUUL (7). The dramatic increase

in electronic vapor product use among high school students

has led to increases in overall tobacco product use among

U.S. youths, erasing gains made in previous years and leading

the U.S. Surgeon General to declare youth e-cigarette use an

epidemic in the United States (10).

Use of any tobacco product among youth is unsafe, regardless

of frequency of use or number of products used. Although the

2019 national YRBS results indicate that most current youth

tobacco product users are infrequent users, variations exist by

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 61

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 2. Usual source* of obtaining electronic vapor products among

current electronic vapor product users,

†

by age — Youth Risk

Behavior Survey, United States, 2019

Usual source

Age group

≥18 yrs ≤17 yrs

% (95% CI) % (95% CI)

Bought them in a store (e.g., a

convenience store, supermarket,

discount store, gas station, or

vape store)

56.4 (51.0–1.6) 8.1 (6.8–9.6)

Got them on the Internet 1.8 (0.9–3.4) 3.6 (2.8–4.6)

Gave someone else money to buy

them for me

3.1 (1.5–6.1) 21.3 (19.5–23.2)

Borrowed them from someone else 27.5 (23.4–32.0) 42.8 (40.2–45.4)

A person who can legally buy these

products gave them to me

3.9 (2.4–6.3) 11.1 (9.9–12.3)

Took them from a store or

another person

2.0 (0.8–5.0) 1.6 (1.1–2.4)

Got them some other way 5.4 (3.3–8.8) 11.6 (10.1–13.4)

Abbreviation: CI = confidence interval.

* Students were limited to selecting only one response.

†

Including e-cigarettes, e-cigars, e-pipes, vape pipes, vaping pens, e-hookahs,

or hookah pens (e.g., blu, NJOY, Vuse, MarkTen, Logic, Vapin Plus, eGo, and

Halo) among students who used electronic vapor products during the 30 days

before the survey.

Comparisons between age groups were assessed by t-test (p<0.05). All

comparisons were statistically different with the exception of “took them from

a store or another person.”

product; for example, frequent use ranged from 18.4% for

cigars to 32.6% for electronic vapor products. In addition, these

results indicate that frequent use of electronic vapor products

increased during 2017–2019; whereas frequent use of other

products decreased or did not change. Even infrequent tobacco

product use, particularly cigarette smoking, is predictive of

progression to daily smoking (14). Nearly all tobacco products

include nicotine, and even infrequent use of tobacco products

has been linked to symptoms of nicotine dependence (15).

Further, 8.2% of high school students currently used two or

more tobacco products in 2019. Multiple tobacco product use is

associated with substance use disorders (16) and might increase

nicotine exposure and risk for nicotine dependence (15).

In 2019, electronic vapor product users aged ≤17 years

usually obtained their products from social sources (e.g.,

by borrowing them from someone). This is consistent with

results from both the Population Assessment of Tobacco and

Health Study and NYTS, which also determined that social

sources were the most common way for adolescents to obtain

electronic vapor products (17,18). These social sources might

include older students who are of legal age for purchasing

the products in their state or community. In 2016, electronic

vapor products were deemed to be tobacco products under

the Family Smoking Prevention Tobacco Control Act (https://

www.federalregister.gov/documents/2016/05/10/2016-10685/

deeming-tobacco-products-to-be-subject-to-the-federal-food-

drug-and-cosmetic-act-as-amended-by-the), thus setting the

federal minimum purchase age for electronic vapor products

at 18 years. However, on December 20, 2019, federal

legislation increased the minimum age of sales of tobacco

products from 18 to 21 years nationwide; the law does not

preempt more stringent state or local age of sale laws (https://

www.fda.gov/tobacco-products/retail-sales-tobacco-products/

selling-tobacco-products-retail-stores). Before this federal

law, 19 states, the District of Columbia, Guam, and Palau

had enacted laws that increased the age of sale for tobacco

products to 21 years, including 13 laws enacted during 2019

(19). Such laws might limit the ability for high school students

to obtain tobacco products from their peers, including those

older students who were of legal age to purchase these products

in their state or community before the law’s implementation.

Multiple factors continue to promote and influence tobacco

product use among youths, including exposure to tobacco

product advertising and imagery through media, as well as the

availability of flavored tobacco products. The sustained and

comprehensive implementation of population-based strategies,

in coordination with the regulation of tobacco products by

the U.S. Food and Drug Administration (FDA), can reduce

all forms of tobacco product use and initiation among U.S.

youths. Such strategies include increasing the price of tobacco

products, implementing comprehensive smoke-free policies,

implementing advertising and promotion restrictions and

national antitobacco public education media campaigns,

restricting youth access to flavored tobacco products, and

implementing policies that increase the minimum age of

purchase for tobacco products to 21 years (1,2,4,10). In

addition to population-level policies for preventing and

reducing initiation of tobacco product use among youths,

tools from the National Cancer Institute (e.g., https://

teen.smokefree.gov) and the Truth Initiative (e.g., https://

truthinitiative.org/thisisquitting) provide resources to help

youth quit tobacco product use.

Limitations

Limitations for YRBS overall are available in the overview

report of this supplement (11). This report is subject to at least

three additional limitations. First, changes in question wording

for smokeless tobacco use in 2017 prohibit comparability with

previous years’ data and long-term trend analyses for prevalence

of smokeless tobacco use, any tobacco product use, and use of

two or more tobacco products. Second, the question addressing

how students usually obtained electronic vapor products requires

that respondents select only one response, although they might

have obtained these products through multiple sources; therefore,

the full scope of the sources students use to access these products

Supplement

62 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

FIGURE 2. Prevalence of current tobacco product use, by year — Youth Risk Behavior Survey, United States, 1991–2019*

* Logistic regression analyses were used to model linear and quadratic time effects while controlling for sex, grade, and race/ethnicity. Electronic vapor products:

significant linear increase (2015–2019); cigarettes: significant linear decrease (1991–2019); significant quadratic trend: increase during 1991–1997, decrease during

1997–2019; cigars/cigarillos/little cigars: significant linear decrease (1997–2019); significant quadratic trend: decrease 1997–2013; decrease 2013–2019 (different

rate of decrease). Differences from 2017 to 2019 were assessed by t-test (p<0.05): A significant increase occurred in use of electronic vapor products and any tobacco

product; a significant decrease occurred in use of cigarettes, cigars/cigarillos/little cigars and smokeless tobacco; and no change occurred in use of ≥2 products.

100

1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 2017 2019

Prevalence (%)

Year

Any tobacco product

Cigarettes

Cigars/cigarillos/little cigars

Electronic vapor products

Smokeless tobacco

≥2 products

might not have been addressed. Finally, the questions related to

electronic vapor products and cigars do not specifically exclude

the possibility of marijuana use in either product (e.g., blunt use).

Conclusion

Although current use of cigarettes, cigars, and smokeless

tobacco among U.S. high school students has decreased,

tobacco product usage has evolved, and the increasing

prevalence of electronic vapor product use among youths

during recent years is concerning. Implementing evidence-

based tobacco control strategies, combined with FDA’s

regulatory efforts, is important for preventing and reducing

all forms of tobacco product use among youths. In addition,

continued surveillance of all tobacco products is warranted for

guiding and evaluating public health policy at the local, state,

tribal, and national levels.

Conflicts of Interest

All authors have completed and submitted the International

Committee of Medical Journal Editors form for disclosure of

potential conflicts of interest. No potential conflicts of interest

were disclosed.

References

1. US Department of Health and Human Services. The health consequences of

smoking—50 years of progress: a report of the Surgeon General. Atlanta, GA:

US Department of Health and Human Services, CDC; 2014. https://www.

ncbi.nlm.nih.gov/books/NBK179276/pdf/Bookshelf_NBK179276.pdf

2. US Department of Health and Human Services. Preventing tobacco use

among youth and young adults. Atlanta, GA: US Department of Health

and Human Services, CDC; 2012. https://www.cdc.gov/tobacco/data_

statistics/sgr/2012/index.htm

3. Kann L, McManus T, Harris WA, et al. Youth risk behavior surveillance—

United States, 2017. MMWR Surveill Summ 2018;67(No. SS-8). https://

doi.org/10.15585/mmwr.ss6708a1

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 63

US Department of Health and Human Services/Centers for Disease Control and Prevention

4. US Department of Health and Human Services. E-cigarette use among

youth and young adults: a report of the Surgeon General. Atlanta, GA:

US Department of Health and Human Services, CDC; 2016. https://

www.cdc.gov/tobacco/data_statistics/sgr/e-cigarettes/pdfs/2016_sgr_

entire_report_508.pdf

5. Spindle TR, Eissenberg T. Pod mod electronic cigarettes—an emerging

threat to public health. JAMA Netw Open 2018;1:e183518. https://

doi.org/10.1001/jamanetworkopen.2018.3518

6. Jackler RK, Ramamurthi D. Nicotine arms race: JUUL and the high-

nicotine product market. Tob Control 2019;28:623–8. https://doi.

org/10.1136/tobaccocontrol-2018-054796

7. King BA, Gammon DG, Marynak KL, Rogers T. Electronic cigarette

sales in the United States, 2013–2017. JAMA 2018;320:1379–80.

https://doi.org/10.1001/jama.2018.10488

8. King BA, Gammon DG, Coats EM, Marynak K, Loomis BR, Rogers

T. The rise of the pod mod: trends in e-cigarette sales in the United

States, 2013–2017. Society for Research on Nicotine and Tobacco annual

meeting. San Francisco, CA; February 20–23, 2019.

9. Gentzke AS, Creamer M, Cullen KA, et al. Vital signs: tobacco product

use among middle and high school students—United States, 2011–2018.

MMWR Morb Mortal Wkly Rep 2019;68:157–64. https://doi.

org/10.15585/mmwr.mm6806e1

10. US Department of Health and Human Services. Surgeon General’s

advisory on e-cigarette use among youth. Washington, DC: US

Department of Health and Human Services, Office of the Surgeon

General; 2018. https://e-cigarettes.surgeongeneral.gov/documents/

surgeon-generals-advisory-on-e-cigarette-use-among-youth-2018.pdf

11. Underwood JM, Brener N, Thornton J, et al. Overview and methods

for the Youth Risk Behavior Surveillance System—United States, 2019.

In: Youth Risk Behavior Surveillance—United States, 2019. MMWR

Suppl 2020;69(No. Suppl 1).

12. Wang TW, Gentzke AS, Creamer MR, et al. Tobacco product use and

associated factors among middle and high school students—United

States, 2019. MMWR Surveill Summ 2019;68(No. SS-12). https://doi.

org/10.15585/mmwr.ss6812a1

13. Miech R, Johnston L, O’Malley PM, Bachman JG, Patrick ME. Trends

in adolescent vaping, 2017–2019. N Engl J Med 2019;381:1490–1.

https://doi.org/10.1056/NEJMc1910739

14. Saddleson ML, Kozlowski LT, Giovino GA, Homish GG, Mahoney

MC, Goniewicz ML. Assessing 30-day quantity-frequency of U.S.

adolescent cigarette smoking as a predictor of adult smoking 14 years

later. Drug Alcohol Depend 2016;162:92–8. https://doi.org/10.1016/j.

drugalcdep.2016.02.043

15. Apelberg BJ, Corey CG, Hoffman AC, et al. Symptoms of tobacco

dependence among middle and high school tobacco users: results from

the 2012 National Youth Tobacco Survey. Am J Prev Med

2014;47(Suppl 1):S4–14. https://doi.org/10.1016/j.amepre.2014.04.013

16. Cavazos-Rehg PA, Krauss MJ, Spitznagel EL, Grucza RA, Bierut LJ.

Youth tobacco use type and associations with substance use disorders.

Addiction 2014;109:1371–80. https://doi.org/10.1111/add.12567

17. Tanski S, Emond J, Stanton C, et al. Youth access to tobacco products

in the United States: findings from wave 1 (2013–2014) of the

population assessment of tobacco and health study. Nicotine Tob Res

2019;21:1695–9. https://doi.org/10.1093/ntr/nty238

18. Liu ST, Snyder K, Tynan MA, Wang TW. Youth access to tobacco

products in the United States, 2016–2018. Tob Regul Sci 2019;5:491–501.

https://doi.org/10.18001/TRS.5.6.2

19. Marynak K, Mahoney M, Williams KS, Tynan MA, Reimels E, King

BA. State and territorial laws prohibiting sales of tobacco products to

persons aged <21 years, United States, December 31, 2019. MMWR

Morb Mortal Wkly Rep 2020;69:189–92. https://doi.org/10.15585/

mmwr.mm6907a3

Supplement

64 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

Dietary and Physical Activity Behaviors Among High School

Students — Youth Risk Behavior Survey, United States, 2019

Caitlin L. Merlo, MPH

; Sherry Everett Jones, PhD, JD

; Shannon L. Michael, PhD

; Tiffany J. Chen, MSPH

;

Sarah A. Sliwa, PhD

; Seung Hee Lee, PhD

; Nancy D. Brener, PhD

; Sarah M. Lee, PhD

; Sohyun Park, PhD

Division of Population Health, National Center for Chronic Disease Prevention and Health Promotion, CDC;

Division of Adolescent and School Health,

National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention, CDC;

Division of Nutrition, Physical Activity, and Obesity, National Center for

Chronic Disease Prevention and Health Promotion, CDC

Abstract

Establishing healthy dietary and physical activity patterns among youths is an important public health strategy for improving

health and preventing chronic diseases; however, few adolescents meet U.S. government recommendations for dietary or physical

activity behaviors, and disparities by sex and race/ethnicity exist. CDC analyzed data from the 2019 Youth Risk Behavior Survey to

update estimates of dietary and physical activity behaviors among U.S. high school students overall and by sex and race/ethnicity.

In addition, 2-year comparisons (2017 and 2019) and trends in prevalence of these behaviors during 2009–2019 were examined.

In 2019, overall, during the 7 days before the survey, 41.8% of students had eaten fruit or drunk 100% fruit juices <1 time/day;

40.7% had eaten vegetables <1 time/day; and 16.7% had not eaten breakfast on all 7 days. Moreover, although 57.4% of students

had played on ≥1 sports team during the 12 months before the survey, less than half of students had been physically active for

≥60 minutes/day on all 7 days (23.2%), had exercised to strengthen or tone their muscles on ≥3 days/week (49.5%), had met both

aerobic and muscle-strengthening physical activity guidelines (16.5%), or had attended physical education classes on all 5 days in

an average school week (25.9%). Trend data indicate limited progress in shifting dietary and physical activity behaviors. That is,

with the exception of decreases in the percentage of students who had consumed soda ≥1 time/day (2009: 29.2%; 2019: 15.1%),

sports drinks ≥1 time/day (2015: 13.8%; 2019: 10.6%), and <3 glasses/day of plain water (2015: 50.5%; 2019: 44.6%), high

school students’ dietary and physical activity behaviors have not improved and, in certain cases, have worsened. These findings

support the need for multicomponent approaches, including policy and environmental changes, and opportunities for adolescents

to learn about and practice making healthy choices.

Introduction

Nutrition and physical activity are important for optimal

growth and development and chronic disease prevention.

Approximately half of U.S. adults have a chronic disease that

is related to inadequate diet quality and physical activity,

including type 2 diabetes, cardiovascular diseases, or obesity

(1). Such diseases can affect productivity and quality of life and

contribute to high health-care costs. Dietary and physical activity

behaviors develop during childhood and can create a trajectory

that continues into adulthood (1,2). Establishing healthy dietary

and physical activity behaviors early in life is a vital public health

strategy for promoting lifelong physical health.

The U.S. government establishes recommendations for

healthy dietary and physical activity patterns for persons of

different ages, including adolescents. The Dietary Guidelines

Corresponding author: Caitlin L. Merlo, MPH, Division of Population

Health, National Center for Chronic Disease Prevention and Health

Promotion, CDC. Telephone: 770-488-6171; E-mail: [email protected].

for Americans 2015–2020 outlines key recommendations for

following a healthy eating pattern within calorie levels that are

appropriate for a person’s age, sex, height, weight, and physical

activity level (1). Recommendations include eating a variety

of vegetables, fruits, and whole grains, and limiting sodium,

added sugars, and saturated and trans fats.

The Physical Activity Guidelines for Americans, 2nd Edition,

recommends that children and adolescents ages 6–17 years

engage in ≥60 minutes of moderate-to-vigorous physical

activity daily (2). Most of this daily physical activity should

be aerobic activity, whereas muscle- and bone-strengthening

physical activity should each be done ≥3 days each week (2).

Previous data indicate that most adolescents are not meeting

recommendations for healthy eating (1) or physical activity

(3), which increases the risk for chronic diseases later in life.

In addition, disparities by sex and race/ethnicity exist (4).

To update national estimates of dietary and physical activity

behaviors among U.S. high school students overall and by

sex and race/ethnicity and to determine how these behaviors

have changed over time, CDC analyzed data from the 2019

Youth Risk Behavior Survey (YRBS) and examined trends in

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 65

US Department of Health and Human Services/Centers for Disease Control and Prevention

prevalence of these behaviors during the previous 10 years.

Public health and school health researchers and practitioners

can use these findings to inform policies and practices that

support healthy eating and physical activity among adolescents.

Methods

Data Source

This report includes data from the 2009–2019 cycles of

the YRBS, a cross-sectional, school-based survey conducted

biennially since 1991. Each survey year, CDC collects data

from a nationally representative sample of public and private

school students in grades 9–12 in the 50 U.S. states and the

District of Columbia. Additional information about YRBS

sampling, data collection, response rates, and processing is

available in the overview report of this supplement (5). The

prevalence estimates for all physical activity, nutrition, and

body weight questions for the overall study population and by

sex, race/ethnicity, grade, and sexual orientation are available

at https://nccd.cdc.gov/youthonline/App/Default.aspx. The

full YRBS questionnaire is available at https://www.cdc.gov/

healthyyouth/data/yrbs/pdf/2019/2019_YRBS-National-HS-

Questionnaire.pdf.

Measures

The student demographic characteristics analyzed included

sex (female or male) and race/ethnicity. Students were classified

into four racial/ethnic categories: non-Hispanic white (white);

non-Hispanic black (black); Hispanic or Latino of any race

(Hispanic); and other or multiple races. The numbers of

students in the other or multiple racial/ethnic groups were

too small for meaningful analysis; therefore, findings for

those groups are not presented; however, the corresponding

data remain in the analytic sample. This analysis included six

dietary variables and five physical activity variables (Table 1).

The dietary variables included the following: during the 7 days

before the survey, had eaten fruit or drunk 100% fruit juices

<1 time/day, had eaten vegetables <1 time/day, had not eaten

breakfast on all 7 days, had drunk soda or pop ≥1 time/day

(not counting diet soda or diet pop), had drunk a sports drink

≥1 time/day, and had drunk <3 glasses/day of plain water. The

physical activity variables included the following: during the

7 days before the survey, had been physically active for a total

of ≥60 minutes/day on all 7 days, had exercised to strengthen

or tone muscles on ≥3 days, had met both aerobic and muscle-

strengthening physical activity guidelines (defined as being

physically active for a total of ≥60 minutes/day on all 7 days

and doing exercises to strengthen or tone muscles on ≥3 days),

had attended physical education classes on all 5 days in an

average school week, and had played on ≥1 sports team during

the 12 months before the survey.

Analysis

Prevalence estimates and 95% confidence intervals for each

2019 dietary and physical activity behavior were calculated

overall and for each sex and racial/ethnic group. Statistically

significant pairwise differences by sex and race/ethnicity were

determined by t-tests. In addition, prevalence of each dietary

and physical activity behavior was compared for 2017 with

2019 by using t-tests. Differences between prevalence estimates

were considered statistically significant if the t-test p value

was <0.05.

To identify 10-year temporal trends, logistic regression analyses

were used to model linear and quadratic time effects while

controlling for sex, grade (9, 10, 11, and 12), and racial/ethnic

changes over time (6). All variables had data available for

2009–2019, except for did not eat breakfast on all 7 days; were

physically active for a total of ≥60 minutes/day on all 7 days; did

exercises to strengthen or tone muscles on ≥3 days and met both

aerobic and muscle-strengthening physical activity guidelines,

which had data for 2011–2019 only; and drank a sports drink

≥1 time/day and drank <3 glasses/day of plain water, which had

data for 2015–2019 only. Additional information about the

methods used to conduct YRBS trend analyses are provided in

the overview report of this supplement (5).

Results

Dietary Behaviors

Overall

In 2019, nationwide, 41.8% of students had eaten fruit

or drunk 100% fruit juices <1 time/day; 40.7% had eaten

vegetables <1 time/day; 16.7% had not eaten breakfast on

all 7 days; 15.1% had drunk sugar-sweetened soda or pop

≥1 time/day (not counting diet soda or diet pop); 10.6%

had drunk a sports drink ≥1 time/day; and 44.6% had drunk

<3 glasses/day of plain water (Table 2). A higher percentage of

male students than female students had drunk sugar-sweetened

soda or pop ≥1 time/day (18.2% versus 11.7%) and had drunk

a sports drink ≥1 time/day (14.0% versus 7.1%). A higher

percentage of black students than white and Hispanic students

had eaten fruit or drunk 100% fruit juices <1 time/day (47.8%

versus 42.1% and 39.5%, respectively), had eaten vegetables

<1 time/day (54.8% versus 35.5% and 46.8%, respectively),

had drunk a sports drink ≥1 time/day (15.6% versus 9.3%

Supplement

66 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 1. Question wording and details for included dietary and physical activity behavior variables — Youth Risk Behavior Survey, United

States, 2019

Variable Question Response options

Years of data

available for 10-year

trend analysis Coding for analysis

Dietary behaviors

Ate fruit or drank

100% fruit juices

<1 time/day

During the past 7 days, how many times did you . . .

• drink 100% fruit juices such as orange juice,

apple juice, or grape juice? (Do not count

punch, Kool-Aid, sports drinks, or other

fruit-flavored drinks.)

• eat fruit? (Do not count fruit juice.)

I did not [drink 100% fruit juice]/[eat

fruit] during the past 7 days, 1–3 times

during the past 7 days, 4–6 times

during the past 7 days,1 time/day,

2 times/day, 3 times/day, or

≥4 times/day

2009–2019 <1 time/day versus

≥1 time/day

Ate vegetables

<1 time/day

During the past 7 days, how many times

did you eat . . .

• green salad?

• potatoes? (Do not count French fries, fried

potatoes, or potato chips.)

• carrots?

• other vegetables? (Do not count green salad,

potatoes, or carrots.)

I did not eat [green salad]/[potatoes]/

[carrots]/[other vegetables] during

the past 7 days, 1–3 times during the

past 7 days, 4–6 times during the past

7 days, 1 time/day, 2 times/day,

3 times/day, or ≥4 times/day

2009–2019 <1 time/day versus

≥1 time/day

Did not eat breakfast

on all 7 days

During the past 7 days, on how many days did

you eat breakfast?

0 days, 1 day, 2 days, 3 days, 4 days,

5 days, 6 days, or 7 days

2011–2019 <7 days versus 7 days

Drank soda or pop

≥1 time/day

During the past 7 days, how many times did you

drink a can, bottle, or glass of soda or pop, such

as Coke, Pepsi, or Sprite? (Do not count diet

soda or diet pop.)

I did not drink soda or pop during the

past 7 days, 1–3 times during the past

7 days, 4–6 times during the past

7 days, 1 time/day, 2 times/day,

3 times/day, or ≥4 times/day

2009–2019 ≥1 time/day versus

<1 time/day

Drank a sports drink

≥1 time/day

During the past 7 days, how many times did you

drink a can, bottle, or glass of a sports drink,

such as Gatorade or Powerade? (Do not count

low-calorie sports drinks such as Propel or G2.)

I did not drink sports drinks during the

past 7 days, 1–3 times during the past

7 days, 4–6 times during the past

7 days, 1 time/day, 2 times/day,

3 times/day, or ≥4 times/day

2015–2019 ≥1 time/day versus

<1 time/day

Drank <3 glasses/day

of plain water

During the past 7 days, how many times did you

drink a bottle or glass of plain water? (Count

tap, bottled, and unflavored sparkling water.)

I did not drink water during the past

7 days, 1–3 times during the past

7 days, 4–6 times during the past

7 days, 1 time per day, 2 times per day,

3 times/day, or ≥4 times/day

2015–2019 ≥3 times/day versus

<3 times/day

Physical activity behaviors

Were physically

active for a total of

≥60 minutes/day

on all 7 days

During the past 7 days, on how many days were

you physically active for a total of at least

60 minutes per day? (Add up all the time you

spent in any kind of physical activity that

increased your heart rate and made you

breathe hard some of the time.)

0 days, 1 day, 2 days, 3 days, 4 days,

5 days, 6 days, or 7 days

2011–2019 7 days versus <7 days

Did exercises to

strengthen or tone

muscles on ≥3 days

During the past 7 days, on how many days did you

do exercises to strengthen or tone your muscles,

such as push-ups, sit-ups, or weightlifting?

0 days, 1 day, 2 days, 3 days, 4 days,

5 days, 6 days, or 7 days

2011–2019 ≥3 days versus <3 days

Met both aerobic

and muscle-

strengthening

physical activity

guidelines

[See “were physically active for a total of

≥60 minutes/day on all 7 days” and “did

exercises to strengthen or tone muscles on

≥3 days.”]

Not applicable 2011–2019 Physically active for

≥60 minutes/day on all

7 days

and did exercises

to strengthen or tone

muscles on ≥3 days

versus physically active

for <60 minutes/day on

all 7 days or did exercises

to strengthen or tone

muscles on <3 days

Attended physical

education classes

on all 5 days

In an average week when you are in school, on

how many days do you go to physical

education (PE) classes?

0 days, 1 day, 2 days, 3 days, 4 days, or

5 days

2009–2019 ≥5 days versus <5 days

Played on ≥1 sports

team

During the past 12 months, on how many sports

teams did you play? (Count any teams run by

your school or community groups.)

0 teams, 1 team, 2 teams, or ≥3 teams 2009–2019 ≥1 team versus <1 team

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 67

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 2. Percentage of high school students who engaged in selected dietary and physical activity behaviors, by sex and race/ethnicity —

Youth Risk Behavior Survey, United States, 2019

Variable

Total

% (95% CI)

Sex Race/Ethnicity

Female

% (95% CI)

Male

% (95% CI)

White,

non-Hispanic

% (95% CI)

Black,

non-Hispanic

% (95% CI)

Hispanic

% (95% CI)

Dietary behaviors

Ate fruit or drank 100% fruit juices

<1 time/day*

41.8 (39.8–43.8) 43.0 (40.7–45.4) 40.6 (38.2–43.1) 42.1 (39.2–45.1) 47.8

†,§

(43.6–51.9) 39.5 (36.7–42.3)

Ate vegetables <1 time/day

40.7 (38.0–43.4) 40.4 (37.2–43.6) 41.1 (38.1–44.3) 35.5 (33.2–37.8) 54.8

†,§

(50.1–59.4) 46.8

†

(41.8–52.0)

Did not eat breakfast on all 7 days during the

7 days before the survey

16.7 (15.3–18.1) 16.7 (15.2–18.3) 16.6 (14.9–18.4) 15.3 (13.9–16.8) 21.1

†

(17.3–25.6) 16.9 (14.1–20.0)

Drank sugar-sweetened soda or pop ≥1

time/day**

15.1 (13.1–17.2) 11.7 (9.9–13.8) 18.2

††

(15.9–20.8) 15.2 (12.7–18.0) 16.9 (13.5–21.0) 16.1 (13.1–19.6)

Drank a sports drink ≥1 time/day

§§

10.6 (9.2–12.3) 7.1 (5.7–8.8) 14.0

††

(11.9–16.4) 9.3 (7.7–11.2) 15.6

†,§

(12.9–18.8) 11.9

†

(10.2–13.8)

Drank <3 glasses/day of plain water

¶¶

44.6 (42.7–46.5) 44.1 (42.0–46.1) 45.0 (42.3–47.6) 44.2 (41.7–46.7) 54.8

†,§

(49.0–60.4) 44.2 (41.8–46.7)

Physical activity behaviors

Were physically active for a total of

≥60 minutes/day on all 7 days***

23.2 (21.9–24.6) 15.4 (14.2–16.6) 30.9

††

(28.9–33.1) 25.6 (24.1–27.2) 21.1

†

(17.6–25.2) 20.9

†

(18.6–23.5)

Did exercises to strengthen or tone muscles

on ≥3 days

†††

49.5 (47.6–51.3) 39.7 (37.2–42.4) 59.0

††

(56.8–61.0) 50.8 (48.2–53.4) 47.0 (42.7–51.2) 48.1 (44.5–51.9)

Met both aerobic and muscle-strengthening

physical activity guidelines

§§§

16.5 (14.6–18.6) 10.1 (8.7–11.6) 23.1

††

(20.4–26.0) 18.4 (15.8–21.4) 13.4

†

(9.5–18.4) 16.0 (13.7–18.6)

Went to physical education classes on

all 5 days

¶¶¶

25.9 (21.5–31.0) 22.8 (17.9–28.5) 28.9

††

(24.6–33.7) 24.3 (18.8–30.7) 23.8 (17.4–31.7) 29.9 (24.5–36.0)

Played on ≥1 sports team**** 57.4 (54.3–60.4) 54.6 (51.1–58.0) 60.2

††

(56.9–63.4) 62.0 (58.1–65.7) 56.1

†

(51.4–60.7) 51.6

†

(46.5–56.6)

Abbreviation: CI = confidence interval.

* Such as orange juice, apple juice, or grape juice, not counting punch, Kool-Aid, sports drinks, or other fruit-flavored drinks during the 7 days before the survey.

†

Significantly different than white students based on t-test analysis (p<0.05).

Significantly different than Hispanic students based on t-test analysis (p<0.05).

Green salad, potatoes (not counting French fries, fried potatoes, or potato chips), carrots, or other vegetables during the 7 days before the survey.

** Such as Coke, Pepsi, or Sprite, not counting diet soda or diet pop, during the 7 days before the survey.

††

Significantly different than female students based on t-test analysis (p<0.05).

§§

Such as Gatorade or PowerAde, not counting low-calorie sports drinks such as Propel water or G2, during the 7 days before the survey.

¶¶

Counting tap, bottled, and unflavored sparkling water during the 7 days before the survey.

*** Adding up time spent in any kind of physical activity that increased their heart rate and made them breathe hard some of the time during the 7 days before the survey.

†††

Such as push-ups, sit-ups, or weightlifting during the 7 days before the survey.

§§§

Were physically active for ≥60 minutes/day on all 7 days and did exercises to strengthen or tone muscles on ≥3 of the 7 days before the survey.

¶¶¶

In an average week when the student was in school.

**** Counting any teams run by their school or community groups during the 12 months before the survey.

and 11.9%, respectively), and had drunk <3 glasses/day of

plain water (54.8% versus 44.2% and 44.2%, respectively).

In addition, a higher percentage of Hispanic students than

white students had eaten vegetables <1 time/day (46.8% versus

35.3%) and had drunk a sports drink ≥1 time/day (11.9%

versus 9.3%), and a higher percentage of black students than

white students had not eaten breakfast on all 7 days (21.1%

versus 15.3%).

Trends

Trend analyses indicated that, during 2009–2019, a

significant linear increase occurred in the percentage of students

who had eaten fruit or drunk 100% fruit juices <1 time/day

overall and among female, male, white, black, and Hispanic

students (Table 3). Significant quadratic trends were not

identified except among black students. The percentage of

black students who had eaten fruit or drunk 100% fruit

juices <1 time/day did not change during 2009–2015 and

then increased during 2015–2019. During 2017–2019, the

percentage of students who had eaten fruit or drunk 100%

fruit juices <1 time/day increased among male students and

black students.

During 2009–2019, a significant linear increase occurred

in the percentage of students who had eaten vegetables

<1 time/day overall and among male, white, and black students.

Significant quadratic trends were not identified, except among

black students. The percentage of black students who had eaten

vegetables <1 time/day did not change during 2009–2015 and

then increased during 2015–2019.

During 2011–2019, a significant linear increase occurred in

the percentage of students who had not eaten breakfast on all

7 days overall and among female, male, and white students.

During 2017–2019, the percentage of students who had not

eaten breakfast on all 7 days increased among students overall

and among female, male, white, and black students.

Supplement

68 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 3. Percentage of high school students who engaged in selected dietary behaviors, by sex, race/ethnicity, and survey year — Youth Risk

Behavior Survey, United States, 2009–2019

Behavior

Prevalence (%)

Linear change* Quadratic change*

Change during

2017–2019

†

2009 2011 2013 2015 2017 2019

Ate fruit or drank 100% fruit juices <1 time/day

Total 35.2 36.0 37.4 36.7 39.2 41.8 Increased during

2009–2019

None None

Female 37.6 38.4 40.0 37.9 41.8 43.0 Increased during

2009–2019

None None

Male 33.0 33.9 34.7 35.4 36.7 40.6 Increased during

2009–2019

None Increased

White, non-Hispanic 34.4 35.8 39.3 37.0 40.4 42.1 Increased during

2009–2019

None None

Black, non-Hispanic 39.2 36.4 36.5 37.8 39.3 47.8 Increased during

2009–2019

None during

2009–2015

Increased during

2015–2019

Increased

Hispanic 35.6 35.3 35.0 35.9 37.6 39.5 Increased during

2009–2019

None None

Ate vegetables <1 time/day

Total 37.3 37.7 38.5 39.0 40.6 40.7 Increased during

2009–2019

None None

Female 38.4 38.4 38.7 40.0 40.7 40.4 None None None

Male 36.3 37.2 38.5 38.0 40.6 41.1 Increased during

2009–2019

None None

White, non-Hispanic 32.7 34.3 35.2 35.8 37.2 35.5 Increased during

2009–2019

None None

Black, non-Hispanic 48.8

45.7 48.1 47.5 50.6 54.8 Increased during

2009–2019

None during

2009–2015

Increased during

2015–2019

None

Hispanic 45.9 43.6 43.1 43.5 43.9 46.8 None None None

Did not eat breakfast on all 7 days during the 7 days before the survey

Total —** 13.1 13.7 13.8 14.1 16.7 Increased during

2011–2019

—

††

Increased

Female —** 13.9 13.8 14.2 14.5 16.7 Increased during

2011–2019

—

††

Increased

Male —** 12.3 13.5 13.3 13.6 16.6 Increased during

2011–2019

—

††

Increased

White, non-Hispanic —** 12.0 11.5 12.0 12.8 15.3 Increased during

2011–2019

—

††

Increased

Black, non-Hispanic —** 16.1 16.0 18.0 15.2 21.1 None —

††

Increased

Hispanic —** 14.4 17.4 14.7 16.0 16.9 None —

††

None

Drank sugar-sweetened soda or pop ≥1 time/day

§§

Total 29.2 27.8 27.0 20.4 18.7 15.1 Decreased during

2009–2019

None during

2009–2013

Decreased during

2013–2019

Decreased

Female 23.3 24.0 24.1 16.4 15.4 11.7 Decreased during

2009–2019

None during

2009–2013

Decreased during

2013–2019

Decreased

Male 34.6 31.4 29.9 24.3 22.3 18.2 Decreased during

2009–2019

None Decreased

White, non-Hispanic 29.0 28.8 29.0 19.7 19.6 15.2 Decreased during

2009–2019

None Decreased

Black, non-Hispanic 33.7 28.0 30.2 22.7 21.5 16.9 Decreased during

2009–2019

None None

Hispanic 28.1 27.0 22.6 21.7 17.0 16.1 Decreased during

2009–2019

None None

See table footnotes on the next page.

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 69

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 3. (Continued) Percentage of high school students who engaged in selected dietary behaviors, by sex, race/ethnicity, and survey year —

Youth Risk Behavior Survey, United States, 2009–2019

Behavior

Prevalence (%)

Linear change* Quadratic change*

Change during

2017–2019

†

2009 2011 2013 2015 2017 2019

Drank a sports drink ≥1 time/day

¶¶

Total —** —** —** 13.8 12.4 10.6 Decreased during

2015–2019

—

††

None

Female —** —** —** 8.8 8.2 7.1 None —

††

None

Male —** —** —** 18.7 16.9 14.0 Decreased during

2015–2019

—

††

Decreased

White, non-Hispanic —** —** —** 12.4 10.7 9.3 Decreased during

2015–2019

—

††

None

Black, non-Hispanic —** —** —** 19.7 21.1 15.6 None —

††

Decreased

Hispanic —** —** —** 15.7 13.5 11.9 Decreased during

2015–2019

—

††

None

Drank <3 glasses/day of plain water***

Total —** —** —** 50.5 48.7 44.6 Decreased during

2015–2019

—

††

Decreased

Female —** —** —** 51.9 48.8 44.1 Decreased during

2015–2019

—

††

Decreased

Male —** —** —** 49.0 48.6 45.0 Decreased during

2015–2019

—

††

Decreased

White, non-Hispanic —** —** —** 50.1 48.8 44.2 Decreased during

2015–2019

—

††

Decreased

Black, non-Hispanic —** —** —** 60.9 52.7 54.8 None —

††

None

Hispanic —** —** —** 49.7 47.5 44.2 Decreased during

2015–2019

—

††

Decreased

* Based on trend analyses by using a logistic regression model controlling for sex, race/ethnicity, and grade (p<0.05).

†

Based on t-test analysis (p<0.05).

Such as orange juice, apple juice, or grape juice, not counting punch, Kool-Aid, sports drinks, or other fruit-flavored drinks, during the 7 days before the survey.

Green salad, potatoes (not counting French fries, fried potatoes, or potato chips), carrots, or other vegetables during the 7 days before the survey.

** Data not available. Question not asked in that year.

††

Insufficient years of data to assess quadratic trends.

§§

Such as Coke, Pepsi, or Sprite, not counting diet soda or diet pop, during the 7 days before the survey.

¶¶

Such as Gatorade or PowerAde, not counting low-calorie sports drinks such as Propel water or G2, during the 7 days before the survey.

*** Counting tap, bottled, and unflavored sparkling water during the 7 days before the survey.

During 2009–2019, a significant linear decrease occurred

in the percentage of students who had drunk sugar-sweetened

soda or pop ≥1 time/day overall and among female, male, white,

black, and Hispanic students (Figure 1). Significant quadratic

trends were identified overall and among female students.

Overall and among female students, the percentage of students

who had drunk sugar-sweetened soda or pop ≥1 time/day did

not change during 2009–2013 and then decreased during

2013–2019. During 2017–2019, the percentage of students who

had drunk sugar-sweetened soda or pop ≥1 time/day decreased

overall and among female, male, and white students.

During 2015–2019, a significant linear decrease occurred

in the percentage of students who had drunk a sports drink

≥1 time/day overall and among male, white, and Hispanic

students (Figure 2). During 2017–2019, the percentage of

students who had drunk a sports drink ≥1 time/day decreased

among male students and black students.

During 2015–2019, a significant linear decrease occurred

in the percentage of students who had drunk <3 glasses/day

of plain water overall and among female, male, white, and

Hispanic students. During 2017–2019, the percentage of

students who had drunk <3 glasses/day of plain water decreased

overall and among female, male, white, and Hispanic students.

Physical Activity Behaviors

Overall

In 2019, nationwide, 23.2% of students had been physically

active for ≥60 minutes/day on all 7 days; 49.5% had exercised

to strengthen or tone their muscles on ≥3 days/week; 16.5%

had met both aerobic and muscle-strengthening physical

activity guidelines; 25.9% had attended physical education

classes on all 5 days in an average school week; and 57.4% had

played on ≥1 sports team (Table 2). A higher percentage of

male students than female students had been physically active

for ≥60 minutes/day on all 7 days (30.9% versus 15.4%), had

exercised to strengthen or tone muscles on ≥3 days (59.0%

versus 39.7%), had met both aerobic and muscle-strengthening

physical activity guidelines (23.1% versus 10.1%), had

Supplement

70 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

FIGURE 1. Percentage of high school students who had drunk sugar-sweetened soda or pop ≥1 time per day during the 7 days before the

survey, overall and by sex and race/ethnicity* — Youth Risk Behavior Survey, United States, 2009–2019

Overall

Female

Male

2009 2011 2013 2015 2017 2019

Percentage

2009 2011 2013 2015 2017 2019

White, non-Hispanic

Black, non-Hispanic

Hispanic

100

Year

* During 2009–2019, a significant linear decrease was observed in the percentage of students who had drunk sugar-sweetened soda or pop ≥1 time/day overall and among

female, male, white, black, and Hispanic students. Based on trend analyses by using a logistic regression model controlling for sex, race/ethnicity, and grade (p<0.05).

attended physical education classes on all 5 days in an average

school week (28.9% versus 22.8%), and had played on

≥1 sports team (60.2% versus 54.6%) (Figure 3). A higher

percentage of white students than black students had been

physically active for ≥60 minutes/day on all 7 days (25.6%

versus 21.1%), had met both aerobic and muscle-strengthening

physical activity guidelines (18.4% versus 13.4%), and had

played on ≥1 sports team (62.0% versus 56.1%). In addition,

a higher percentage of white students than Hispanic students

had been physically active for ≥60 minutes/day on all 7 days

(25.6% versus 20.9%) and had played on ≥1 sports team

(62.0% versus 51.6%) (Table 2).

Trends

During 2011–2019, a significant linear decrease occurred

in the percentage of students who had been physically active

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 71

US Department of Health and Human Services/Centers for Disease Control and Prevention

FIGURE 2. Percentage of high school students who had drunk a sports drink ≥1 time per day during the 7 days before the survey, overall and

by sex and race/ethnicity* — Youth Risk Behavior Survey, United States, 2015–2019

Overall

Female

Male

2015 2017 2019

Percentage

100

White, non-Hispanic

Black, non-Hispanic

Hispanic

Year Year

* During 2015–2019, a significant linear decrease was observed in the percentage of students who had drunk a sports drink ≥1 time/day overall and among male,

white, and Hispanic students. Based on trend analysis by using a logistic regression model controlling for sex, race/ethnicity, and grade (p<0.05).

for ≥60 minutes/day on all 7 days overall and among female,

male, white, black, and Hispanic students (Table 4). During

2017–2019, the percentage of students who had been

physically active for ≥60 minutes/day on all 7 days decreased

overall and among male students and Hispanic students.

During 2011–2019, a significant linear decrease occurred in

the percentage of students who had exercised to strengthen or

tone their muscles on ≥3 days/week overall and among male,

white, black, and Hispanic students. During 2017–2019, no

significant changes occurred in the percentage of students

who had exercised to strengthen or tone their muscles on

≥3 days/week overall or among the sex or racial/ethnic groups.

During 2011–2019, a significant linear decrease occurred

in the percentage of students who had met both aerobic and

muscle-strengthening physical activity guidelines overall

and among female, male, white, and black students. During

2017–2019, the percentage of students who had met both

aerobic and muscle-strengthening physical activity guidelines

did not significantly change overall but decreased among male

students and Hispanic students.

During 2009–2019 and during 2017–2019, no significant

linear changes occurred in the percentage of students who had

attended physical education classes on all 5 days in an average

school week or had played on ≥1 sports team overall or among

the sex and racial/ethnic groups, except among female students.

Among female students, a significant linear decrease occurred

in the percentage who had attended physical education classes

on all 5 days in an average school week.

Discussion

With the exception of decreases in the percentages of

students who had consumed soda ≥1 time/day, sports drinks

≥1 time/day, and <3 glasses/day of plain water, high school

students’ dietary and physical activity behaviors have not

improved during the previous 10 years and, in certain cases,

have worsened. This is cause for concern because healthy

dietary and physical activity behaviors are important for

growth and development, academic outcomes, and prevention

of chronic diseases, including type 2 diabetes, heart disease,

hypertension, and obesity (1,7). Recent data demonstrate

that approximately one in five adolescents have prediabetes,

which increases the risk for type 2 diabetes and cardiovascular

diseases (8). In addition, data from the National Health and

Nutrition Examination Survey reveal that, in the United States

during 2007–2008, approximately 18.1% of youths aged

12–19 years had obesity and this increased to 20.6% during

2015–2016 (9). In this analysis, in which differences by race/

ethnicity exist, black and Hispanic high school students have

poorer dietary and physical activity behaviors, compared

Supplement

72 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

FIGURE 3. Percentage* of high school students who had engaged in physical activity

†

and physical education during the 7 days before the

survey, overall and by sex

— Youth Risk Behavior Survey, United States, 2019

Overall

Female

Male

Were physically active

for a total of

≥60 minutes/day

on all 7 days

Did exercises to

strengthen or tone

muscles on ≥3 days

Met both aerobic

and muscle-strengthening

physical activity guidelines

Attended physical

education classes

on all 5 days

Played on ≥1

sports team

during the past

12 months

100

Percentage

Activity

* Bars represent the percentage of respondents with a “yes” response, overall and by sex.

†

The “met both aerobic and muscle-strengthening physical activity guidelines” variable is defined as being physically active for a total of ≥60 minutes/day on all

7 days and doing exercises to strengthen or tone muscles on ≥3 days during the 7 days before the survey (Source: U.S. Department of Health and Human Services.

Physical activity guidelines for Americans. 2nd ed. Washington, DC: US Department of Health and Human Services; 2018. https://www.hhs.gov/fitness/be-active/

physical-activity-guidelines-for-americans/index.html).

In 2019, a significantly higher percentage of male than female students had been physically active for ≥60 minutes/day on all 7 days during the 7 days before the

survey, had exercised to strengthen or tone muscles on ≥3 days during the 7 days before the survey, had met the aerobic and muscle-strengthening physical activity

guidelines during the 7 days before the survey, had attended physical education classes on all 5 days in an average school week when the student was in school,

and had played on ≥1 sports team during the past 12 months. Based on t-test analysis (p<0.05).

with white high school students. These findings also indicate

that male students have poorer dietary behaviors but better

physical activity behaviors than do female students. Addressing

dietary and physical activity behaviors can benefit all students

and is especially important for those with increased risk for

chronic diseases (e.g., students from low-income families and

racial/ethnic minorities).

Dietary Behaviors

No improvements occurred in fruit or vegetable consumption

during 2009–2019 and, in many cases, have worsened. Overall,

consumption of fruits and vegetables remained low in 2019.

For example, four of 10 high school students had eaten fruit

or drunk 100% fruit juices <1 time/day. Similarly, four of 10

had eaten vegetables <1 time/day. Although the prevalence of

having eaten fruit or drunk 100% fruit juice <1 time/day and

having eaten vegetables <1 time/day is similar for male students

and female students, recommended daily intakes differ by age

and sex. Females and males aged 14–18 years need 1.5 cups

and 2 cups, respectively, of fruits, and 2.5 cups and 3 cups,

respectively, of vegetables (https://www.choosemyplate.gov/

resources/MyPlatePlan). Although YRBS measures frequency

of intake and not the amount consumed, children and

adolescents who meet the recommended amounts typically

consume fruits and vegetables multiple times throughout the

day (10); therefore, consuming fruits or vegetables <1 time/day

is likely insufficient. Strategies that encourage adolescents to

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 73

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 4. Percentage of high school students who engaged in selected physical activity behaviors, by sex, race/ethnicity, and survey year —

Youth Risk Behavior Survey, United States, 2009–2019

Behavior

Prevalence (%)

Linear change* Quadratic change*

Change during

2017–2019

†

2009 2011 2013 2015 2017 2019

Were physically active for a total of ≥60 minutes/day on all 7 days

Total —

28.7 27.1 27.1 26.1 23.2 Decreased during

2011–2019

—** Decreased during

2017–2019

Female —

18.5 17.7 17.7 17.5 15.4 Decreased during

2011–2019

—** None

Male —

38.3 36.6 36.0 35.3 30.9 Decreased during

2011–2019

—** Decreased during

2017–2019

White, non-Hispanic —

30.4 28.2 29.0 27.1 25.6 Decreased during

2011–2019

—** None

Black, non-Hispanic —

26.0 26.3 24.2 24.5 21.1 Decreased during

2011–2019

—** None

Hispanic —

26.5 25.5 24.6 25.8 20.9 Decreased during

2011–2019

—** Decreased during

2017–2019

Did exercises to strengthen or tone muscles on ≥3 days

††

Total —

55.6 51.7 53.4 51.1 49.5 Decreased during

2011–2019

—** None

Female —

43.8 41.6 42.7 40.8 39.7 None —** None

Male —

66.7 61.8 63.7 62.1 59.0 Decreased during

2011–2019

—** None

White, non-Hispanic —

55.7 52.4 54.5 50.6 50.8 Decreased during

2011–2019

—** None

Black, non-Hispanic —

54.0 48.8 52.3 51.0 47.0 Decreased during

2011–2019

—** None

Hispanic —

56.6 53.3 52.4 52.3 48.1 Decreased during

2011–2019

—** None

Met guidelines for aerobic and muscle-strengthening physical activity

§§

Total —

21.9 21.6 20.5 20.0 16.5 Decreased during

2011–2019

—** None

Female —

12.7 13.0 12.2 12.1 10.1 Decreased during

2011–2019

—** None

Male —

30.7 30.3 28.6 28.5 23.1 Decreased during

2011–2019

—** Decreased during

2017–2019

White, non-Hispanic —

23.9 22.6 22.7 20.8 18.4 Decreased during

2011–2019

—** None

Black, non-Hispanic —

18.4 20.6 15.7 17.7 13.4 Decreased during

2011–2019

—** None

Hispanic —

18.9 20.5 18.7 20.0 16.0 None —** Decreased during

2017–2019

See table footnotes on the next page.

increase the quantity of fruits and vegetables each time they

consume them are likely needed to help them meet the daily

recommendations (10). For example, schools can offer students

multiple fruit and vegetable choices each day through school

meal programs, including through grab-and-go salads (11).

Sugar-sweetened beverages (SSBs) are the primary source

of added sugars in U.S. youths’ diets (1). Frequently

drinking SSBs is associated with health conditions, including

obesity, type 2 diabetes, heart disease, and tooth decay (12).

Alternatively, drinking enough water every day is good for

overall health and is associated with higher Healthy Eating

Index scores among adolescents (13). (More information about

the Healthy Eating Index is available at https://www.fns.usda.

gov/resource/healthy-eating-index-hei.) YRBS asks about two

specific types of SSBs, soda or pop and sports drinks. This study

identified substantial decreases in the percentage of students

who had drunk soda or pop ≥1 time/day overall and among all

sex and racial/ethnic groups. In addition, decreases occurred

in the percentage of students who had drunk a sports drink

≥1 time/day overall and among female, white, and Hispanic

students. Despite these improvements in soda and sports drink

consumption, consumption of these beverages is common.

Differences also existed by sex and race/ethnicity. Similar to

this study, previous studies reported that SSB intake was higher

among males than among females (14) and among black and

Hispanic adolescents than among white adolescents (15). One

possible explanation for the differences between racial/ethnic

groups is that beverage companies disproportionately market

SSBs to black and Hispanic youths (16).

Supplement

74 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 4. (Continued) Percentage of high school students who engaged in selected physical activity behaviors, by sex, race/ethnicity, and survey

year — Youth Risk Behavior Survey, United States, 2009–2019

Behavior

Prevalence (%)

Linear change* Quadratic change*

Change during

2017–2019

†

2009 2011 2013 2015 2017 2019

Went to physical education classes on all 5 days

¶¶

Total 33.3 31.5 29.4 29.8 29.9 25.9 None None None

Female 31.9 27.2 24.0 25.5 25.3 22.8 Decreased during

2009–2019

None None

Male 34.6 35.5 34.9 33.8 34.7 28.9 None None None

White, non-Hispanic 30.6 33.0 27.1 25.4 27.2 24.3 None None None

Black, non-Hispanic 37.0 27.6 26.6 35.8 28.5 23.8 None None None

Hispanic 40.5 30.0 37.7 37.7 37.4 29.9 None None None

Played on ≥1 sports team***

Total 58.3 58.4 54.0 57.6 54.3 57.4 None None None

Female 52.3 52.6 48.5 53.0 49.3 54.6 None None None

Male 63.8 64.0 59.6 62.2 59.7 60.2 None None None

White, non-Hispanic 61.1 60.9 55.2 62.4 54.5 62.0 None None None

Black, non-Hispanic 57.3 57.0 55.2 57.6 59.1 56.1 None None None

Hispanic 53.2 54.1 51.2 48.5

52.2 51.6 None None None

* Based on trend analyses by using a logistic regression model controlling for sex, race/ethnicity, and grade (p<0.05).

†

Based on t-test analysis (p<0.05).

Adding up time spent in any kind of physical activity that increased their heart rate and made them breathe hard some of the time during the 7 days before the survey.

Data not available. Question not asked in that year.

** Insufficient years of data to assess quadratic trends.

††

Such as push-ups, sit-ups, or weightlifting during the 7 days before the survey.

§§

Were physically active for ≥60 minutes/day on all 7 days and did exercises to strengthen or tone muscles on ≥3 of the 7 days before the survey.

¶¶

In an average week when the student was in school.

*** Counting any teams run by their school or community groups during the 12 months before the survey.

During the 2014–15 school year, the Smart Snacks in School

nutrition standards were implemented, which decreased

students’ access to SSBs at school. (More information about

the Smart Snacks in School nutrition standards is available at

https://www.gpo.gov/fdsys/pkg/FR-2013-06-28/pdf/2013-

15249.pdf.) Additional policy and educational approaches

(e.g., health education classes or communitywide campaigns)

might help further reduce SSB access in schools and other

settings and help adolescents choose healthier beverage options,

including plain water.

Physical Activity Behaviors

Overall, prevalence of health-promoting physical activity

behaviors was low in 2019 and either decreased or did not

change during the previous 10 years. Healthy People 2020

monitors four of the five physical activity behaviors included

in this study (https://www.healthypeople.gov/), and these

behaviors will continue to be monitored with Healthy People

2030. Healthy People 2020 objective PA-3 aims to increase the

proportion of adolescents who meet federal physical activity

guidelines for aerobic physical activity to ≥31.6% (PA-3.1),

muscle-strengthening activity to ≥61.2% (PA-3.2), and both

aerobic physical activity and muscle-strengthening activity to

≥24.1% (PA-3.3). The proportions of students meeting the

aerobic, muscle-strengthening, or both guidelines decreased

during 2011–2019, and 2019 data indicate that adolescents

continue to fall short of achieving these targets.

One of the Healthy People 2020 objectives (PA-5) is to increase

the proportion of adolescents who participate in daily school

physical education to ≥36.6%. Given no increase in this behavior

during 2009–2019 and that only 25.9% of high school students

attended daily physical education class during 2019, the target

for this objective is unlikely to be met in 2020. Students can

accumulate approximately 40% of their daily physical activity

through participation in physical education (17), demonstrating

that physical education at school is an effective strategy for helping

high school students meet the federal physical activity guidelines.

During 2015–2016, although the majority of U.S. states required

public high schools to provide physical education, few states

mandated a time requirement for high school students, and many

states permitted students to substitute other activities for their

physical education requirement. (More information about the

status of physical education in the United States is available at

https://www.shapeamerica.org/MemberPortal/SHAPE_Sign_I.

aspx?WebsiteKey=c03f2b51-3ee7-46fa-b587-de18213dcae5&Lo

ginRedirect=true&returnurl=%2fadvocacy%2fson%2f.)

The 2019 release of the National Youth Sports Strategy

highlighted youth sports participation for its physical activity,

psychosocial, and academic achievement benefits. (More

information about the National Youth Sports Strategy is

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 75

US Department of Health and Human Services/Centers for Disease Control and Prevention

available at https://health.gov/our-work/physical-activity/

national-youth-sports-strategy.) Despite these benefits, only

57.4% of high school students reported participating in

sports. The National Survey of Children’s Health also assesses

participation in youth sports, with similar estimates to YRBS

for youths aged 14–17 years (3).

Across all the physical activity behaviors, a higher percentage

of males than females met aerobic, muscle-strengthening, or

both guidelines, participated in daily physical education, and

played on ≥1 sports team. These differences might be caused

by gender stereotypes, self-efficacy, self-consciousness, or

social influences (18). When overcoming barriers to physical

activity, particularly for adolescent females, strategies that

span the Social-Ecological Model by addressing individual,

interpersonal, organizational, community, and societal

components might need to be considered.

Addressing Dietary and

Physical Activity Behaviors

Improving dietary and physical activity behaviors among

adolescents requires efforts across multiple settings. For

example, schools can implement policies and practices (e.g.,

local school wellness policies) (https://www.fns.usda.gov/tn/

local-school-wellness-policy) that support healthy eating and

physical activity, including ensuring the following: 1) that

foods and beverages sold during the school day meet Smart

Snacks in School nutrition standards, 2) that school meals are

appealing and include menu items that students enjoy, and

3) that students have access to free drinking water during the

school day (11). Schools can also help students meet the federal

physical activity guidelines by providing physical activity

opportunities before, during, and after the school day. This

can be achieved by developing, implementing, and evaluating

a comprehensive school physical activity program, which serves

as a national framework for physical education and physical

activity in schools. (More guidance on comprehensive school

physical activity programs is available at https://www.cdc.gov/

healthyschools/physicalactivity/pdf/13_242620-A_CSPAP_

SchoolPhysActivityPrograms_Final_508_12192013.pdf.)

Health education is another way that schools can help students

develop the knowledge and skills needed for making health-

enhancing decisions. These school efforts can be addressed and

coordinated through the Whole School, Whole Community,

Whole Child Model, which highlights the interconnectedness

of multiple health behaviors and outcomes and promotes

collaboration among diverse partners, including mental health

professionals, school leaders, school nurses, physical and health

educators, and parents for promoting health and well-being for

all students. (More information about the Whole School, Whole

Community, Whole Child approach is available at https://www.

cdc.gov/healthyschools/wscc/index.htm.)

Community members and parents can reinforce the

messages promoted within the school and can participate on

the school wellness or school health teams that are addressing

healthy eating, physical education, and physical activity. (More

information about parent engagement in school health is

available at https://www.cdc.gov/healthyyouth/protective/

pdf/parent_engagement_strategies.pdf.) In addition, parents

and community members can engage in physical activity

with adolescents, provide social supports for adolescents that

increase physical activity while decreasing sedentary behaviors,

and make choices that support healthy eating.

Community-based interventions that address healthy eating

and physical activity through policy and environmental

changes can improve dietary and physical activity behaviors

and weight-status outcomes among youths (19–21). These

kinds of community-based approaches often adopt multiple

strategies, including providing information (e.g., messaging

campaigns and healthy recipe demonstrations), providing

incentives, and improving access to opportunities for practicing

healthy behaviors through policy and systems changes. Having

multiple activities that target specific behaviors and using a

mix of behavioral change strategies appear to be important

for making health behavior changes (19). Community-

based interventions that also include the school setting are

more effective in influencing outcomes among youths than

interventions that occur only in the community (20).

Limitations

General limitations for the YRBS are available in the

overview report of this supplement (5). The findings in this

report are subject to at least one additional limitation. Certain

questions about dietary behaviors (e.g., fruit consumption)

ask about frequency rather than portion size; therefore, these

data cannot directly determine whether students are meeting

specific recommendations for age and sex (22).

Conclusion

Because of the limited progress in increasing the prevalence

of healthy dietary and physical activity behaviors among U.S.

high school students, multicomponent approaches, including

policy and environmental changes and opportunities for

adolescents to learn about and practice making healthy choices,

are needed to facilitate healthy dietary and physical activity

patterns. Schools, communities, and families can work together

in creating healthy environments where adolescents thrive.

Supplement

76 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

Conicts of Interest

All authors have completed and submitted the International

Committee of Medical Journal Editors form for disclosure of

potential conflicts of interest. No potential conflicts of interest

were disclosed.

References

1. US Department of Health and Human Services; US Department of

Agriculture. 2015–2020 dietary guidelines for Americans. 8th ed.

Washington, DC: US Department of Health and Human Services; US

Department of Agriculture; 2015. http://health.gov/

dietaryguidelines/2015/guidelines

2. US Department of Health and Human Services. Physical activity

guidelines for Americans, 2nd ed. Washington, DC: US Department of

Health and Human Services; 2018. https://health.gov/sites/default/

files/2019-09/Physical_Activity_Guidelines_2nd_edition.pdf

3. The Child and Adolescent Health Measurement Initiative. National

Survey of Children’s Health. Baltimore, MD: Johns Hopkins University,

Bloomberg School of Public Health; 2016. https://www.childhealthdata.

org/learn-about-the-nsch/NSCH

4. Kann L, McManus T, Harris WA, et al. Youth risk behavior surveillance—

United States, 2015. MMWR Surveill Summ 2016;65(No. SS-6).

5. Underwood JM, Brener N, Thornton J, et al. Overview and methods

for the Youth Risk Behavior Surveillance System—United States, 2019.

In: Youth Risk Behavior Surveillance—United States, 2019. MMWR

Suppl 2020;69(No. Suppl 1).

6. Hinkle DE, Wiersma W, Jurs SG. Applied statistics for the behavioral

sciences. 5th ed. Boston, MA: Houghton Mifflin; 2003.

7. Rasberry CN, Tiu GF, Kann L, et al. Health-related behaviors and

academic achievement among high school students—United States,

2015. MMWR Morb Mortal Wkly Rep 2017;66:921–27. http://dx.doi.

org/10.15585/mmwr.mm6635a1

8. Andes LJ, Cheng YJ, Rolka DB, Gregg EW, Imperatore G. Prevalence

of prediabetes among adolescents and young adults in the United States,

2005–2016. JAMA Pediatr 2019;174:e194498. https://doi.org/10.1001/

jamapediatrics.2019.4498

9. Fryar CD, Carroll MD, Ogden CL. Prevalence of overweight, obesity,

and severe obesity among children and adolescents aged 2–19 years:

United States, 1963–1965 through 2015–2016. Hyattsville, MD: US

Department of Health and Human Services, CDC, National Center for

Health Statistics; 2018. https://www.cdc.gov/nchs/data/hestat/obesity_

child_15_16/obesity_child_15_16.htm

10. Moore LV, Hamner HC, Kim SA, Dalenius K. Common ways Americans

are incorporating fruits and vegetables into their diet: intake patterns

by meal, source and form, National Health and Nutrition Examination

Survey 2007–2010. Public Health Nutr 2016;19:2535–9. https://doi.

org/10.1017/S1368980016000586

11. CDC. School health guidelines to promote healthy eating and physical

activity. MMWR Recomm Rep 2011;60(No. RR-5).

12. Dietary Guidelines Advisory Committee. Scientific report of the 2015

Dietary Guidelines Advisory Committee: advisory report to the Secretary

of Health and Human Services and the Secretary of Agriculture.

Washington, DC: US Department of Agriculture, Agricultural Research

Service; 2015. https://health.gov/our-work/food-nutrition/2015-2020-

dietary-guidelines/advisory-report

13. Leung CW, DiMatteo SG, Gosliner WA, Ritchie LD. Sugar sweetened

beverage and water intake in relation to diet quality in U.S. children. Am J

Prev Med 2018;54:394–402. https://doi.org/10.1016/j.amepre.2017.11.005

14. Rosinger A, Herrick K, Gahche J, Park S. Sugar-sweetened beverage

consumption among U.S. youth, 2011–2014. NCHS Data Brief

2017;(271):1–8.

15. Bleich SN, Vercammen KA, Koma JW, Li Z. Trends in beverage

consumption among children and adults, 2003–2014. Obesity (Silver

Spring) 2018;26:432–41. https://doi.org/10.1002/oby.22056

16. Harris JL, Schwartz MB, LoDolce M, et al; Sugary Drink FACTS 2014.

Some progress but much room for improvement in marketing to youth.

Hartford, CT: University of Connecticut Rudd Center for Food Policy

and Obesity; 2014. http://sugarydrinkfacts.org/resources/

SugaryDrinkFACTS_Report.pdf

17. Bassett DR, Fitzhugh EC, Heath GW, et al. Estimated energy

expenditures for school-based policies and active living. Am J Prev Med

2013;44:108–13. https://doi.org/10.1016/j.amepre.2012.10.017

18. Spencer RA, Rehman L, Kirk SF. Understanding gender norms, nutrition,

and physical activity in adolescent girls: a scoping review. Int J Behav

Nutr Phys Act 2015;12:6. https://doi.org/10.1186/s12966-015-0166-8

19. Ritchie LD, Woodward-Lopez G, Au LE, et al; Healthy Communities

Study Team. Associations of community programs and policies with

children’s dietary intakes: the Healthy Communities Study. Pediatr Obes

2018;13(Suppl 1):14–26. https://doi.org/10.1111/ijpo.12440

20. Wang Y, Cai L, Wu Y, et al. What childhood obesity prevention

programmes work? A systematic review and meta-analysis. Obes Rev

2015;16:547–65. https://doi.org/10.1111/obr.12277

21. Pate RR, Frongillo EA, McIver KL, et al; Healthy Communities Study

Team. Associations between community programmes and policies and

children’s physical activity: the Healthy Communities Study. Pediatr

Obes 2018;13(Suppl 1):72–81. https://doi.org/10.1111/ijpo.12426

22. Eaton DK, Olsen EO, Brener ND, et al. A comparison of fruit and

vegetable intake estimates from three survey question sets to estimates

from 24-hour dietary recall interviews. J Acad Nutr Diet

2013;113:1165–74. https://doi.org/10.1016/j.jand.2013.05.013

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 77

US Department of Health and Human Services/Centers for Disease Control and Prevention

Transportation Risk Behaviors Among High School Students —

Youth Risk Behavior Survey, United States, 2019

Merissa A. Yellman, MPH

; Leah Bryan, MPH

; Erin K. Sauber-Schatz, PhD

; Nancy Brener, PhD

Division of Injury Prevention, National Center for Injury Prevention and Control, CDC;

Division of Adolescent and School Health, National Center for

HIV/AIDS, Viral Hepatitis, STD, and TB Prevention, CDC

Abstract

Motor-vehicle crashes are a leading cause of death and nonfatal injury among U.S. adolescents, resulting in approximately 2,500

deaths and 300,000 nonfatal injuries each year. Risk for motor-vehicle crashes and resulting injuries and deaths varies, depending

on such behaviors as seat belt use or impaired or distracted driving. Improved understanding of adolescents’ transportation risk

behaviors can guide prevention efforts.

Therefore, data from the 2019 Youth Risk Behavior Survey were analyzed to determine

prevalence of transportation risk behaviors, including not always wearing a seat belt, riding with a driver who had been drinking

alcohol (riding with a drinking driver), driving after drinking alcohol, and texting or e-mailing while driving. Differences by

student characteristics (age, sex, race/ethnicity, academic grades in school, and sexual identity) were calculated. Multivariable

analyses controlling for student characteristics examined associations between risk behaviors. Approximately 43.1% of U.S. high

school students did not always wear a seat belt and 16.7% rode with a drinking driver during the 30 days before the survey.

Approximately 59.9% of students had driven a car during the 30 days before the survey. Among students who drove, 5.4% had

driven after drinking alcohol and 39.0% had texted or e-mailed while driving. Prevalence of not always wearing a seat belt was

higher among students who were younger, black, or had lower grades. Riding with a drinking driver was higher among Hispanic

students or students with lower grades. Driving after drinking alcohol was higher among students who were older, male, Hispanic,

or had lower grades. Texting while driving was higher among older students or white students. Few differences existed by sexual

identity. Multivariable analyses revealed that students engaging in one transportation risk behavior were more likely to engage in

other transportation risk behaviors. Traffic safety and public health professionals can use these findings to reduce transportation

risk behaviors by selecting, implementing, and contextualizing the most appropriate and effective strategies for specific populations

and for the environment.

Introduction

Motor-vehicle crashes are predictable and preventable.

However, in the United States, they remain the second leading

cause of death among adolescents and the fourth leading

cause of nonfatal injury. During 2018, approximately 2,500

adolescents (persons aged 12–19 years) died in motor-vehicle

crashes; of those deaths, >75% were occupants of passenger

vehicles (i.e., cars, pickup trucks, vans, or sport utility vehicles)

(1). Motor-vehicle crashes also resulted in approximately

297,000 nonfatal injuries among adolescents during 2018.

Moreover, fatal and nonfatal motor-vehicle–crash injuries

among adolescents resulted in approximately $12 billion in

medical and work-loss costs during 2018 (https://www.cdc.

gov/injury/wisqars).

Passenger-related transportation risk behaviors (e.g., nonuse

of seat belts or riding with a driver who had been drinking

Corresponding author: Merissa A. Yellman, MPH, Division of Injury

Prevention, National Center for Injury Prevention and Control.

Telephone: 404-498-5299; E-mail: [email protected].

alcohol) increase the risk for injury or death in a crash or

risk for a crash itself. Seat belt use among adolescents and

young adults is typically lower than among adults of other

age groups (1) (https://crashstats.nhtsa.dot.gov/Api/Public/

ViewPublication/812781). For instance, the National

Occupant Protection Use Survey Controlled Intersection

Study uses a probability-based sample of observational surveys

conducted on an annual basis to produce estimates of seat belt

use nationwide at a typical daylight moment. Results during

2016–2018 indicate that seat belt use among adolescents and

young adults aged 16–24 years was approximately 87% each

year, whereas seat belt use among adults aged ≥25 years was

90% or higher (https://crashstats.nhtsa.dot.gov/Api/Public/

ViewPublication/812781). Previous research also demonstrates

that high school students put themselves at risk by riding with

drivers who have been drinking alcohol (2).

Per mile driven, drivers aged 16–19 years have crash rates

approximately four times greater than those of drivers aged

≥20 years (1); a leading contributor is driver inexperience

(1,3). Because of this elevated crash risk, engagement in

driver-related transportation risk behaviors (e.g., driving after

Supplement

78 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

drinking alcohol or texting or e-mailing while driving) puts

adolescents at even higher risk. For example, drinking alcohol

negatively affects a person’s ability to drive safely regardless of

age. However, even at the same blood alcohol concentration

(BAC), drivers aged 16–20 years have a much higher risk for

being involved in a crash than older drivers (1,4). Similarly, the

negative effects of driver inexperience on driving performance

are worsened by cell phone–related driver distraction (5).

For this report, 2019 data from the Youth Risk Behavior

Survey (YRBS) were analyzed by student characteristics to

determine the prevalence of four transportation risk behaviors

among U.S. high school students. Associations between

engagement in multiple transportation risk behaviors also were

calculated. This study provides an update on which adolescent

groups have an elevated prevalence of engaging in transportation

risk behaviors and reveals the extent to which adolescents engage

in multiple transportation risk behaviors. The findings can help

traffic safety and public health professionals appropriately select,

tailor, and implement effective strategies to have a greater impact

on reducing risk behaviors, thereby preventing crashes, injuries,

and deaths among adolescents.

Methods

Data Source

This report includes data from CDC’s 2019 YRBS, a cross-

sectional, school-based survey conducted biennially since

1991. Each survey year, CDC collects data from a nationally

representative sample of public and private school students in

grades 9–12 in the 50 U.S. states and the District of Columbia.

Additional information about YRBS sampling, data collection,

response rates, and processing is available in the overview

report of this supplement (6). The prevalence estimates for all

unintentional injury questions for the overall study population

and by sex, race/ethnicity, grade, and sexual orientation are

available at https://nccd.cdc.gov/youthonline/App/Default.

aspx. The full YRBS questionnaire is available at https://

www.cdc.gov/healthyyouth/data/yrbs/pdf/2019/2019_YRBS-

National-HS-Questionnaire.pdf.

Measures

This study examined two passenger- and two driver-related

transportation risk behaviors among U.S. high school students.

The overall analytic sample was used for the passenger-related

risk behaviors, which included not always wearing a seat belt

when riding in a car driven by someone else and riding with a

driver who had been drinking alcohol (riding with a drinking

driver). Not always wearing a seat belt was assessed with the

question, “How often do you wear a seat belt when riding in

a car driven by someone else?” Response options included

“always,” “most of the time,” “sometimes,” “rarely,” or “never,”

with any response other than “always” being defined as not

always wearing a seat belt. Riding with a drinking driver was

assessed with the question, “During the past 30 days, how many

times did you ride in a car or other vehicle driven by someone

who had been drinking alcohol?” Responses were dichotomized

(0 times versus ≥1 time). Students who reported riding with a

drinking driver at least once during the previous 30 days were

classified as having engaged in the behavior.

Driver-related transportation risk behaviors included driving

when they had been drinking alcohol (driving after drinking

alcohol) and texting or e-mailing while driving (texting while

driving). Driving after drinking alcohol was assessed with

the question, “During the past 30 days, how many times did

you drive a car or other vehicle when you had been drinking

alcohol?” Texting while driving was assessed with the question,

“During the past 30 days, on how many days did you text or

e-mail while driving a car or other vehicle?” Students who

indicated they had not driven a car or other vehicle during the

past 30 days on each respective question were excluded from

the analysis for these questions. Responses among drivers were

categorized as 0 times or days versus ≥1 time or day.

An approximation of driving prevalence among students is

presented to provide context for the driver-related behaviors.

However, driving prevalence is not directly captured in the 2019

YRBS. For this approximation, students who chose a response

other than “I did not drive a car or other vehicle during the past

30 days” for both driver-related questions (driving after drinking

alcohol and texting while driving) were classified as drivers, and

students who indicated that they did not drive a car or other

vehicle during the past 30 days were classified as nondrivers.

Driver classification was independent of students’ responses to

the two questions about passenger-related transportation risk

behaviors because students who drove during the past 30 days

could also be passengers when they were not driving during the

same 30-day period.

All transportation risk behaviors were analyzed by self-

reported student characteristics, including age (14, 15,

16, 17, or ≥18 years), sex (male or female), race/ethnicity

(non-Hispanic white [white]; non-Hispanic black [black]; or

Hispanic or Latino of any race [Hispanic]), academic grades in

school (mostly As or Bs versus mostly Cs, Ds, or Fs), and sexual

identity (heterosexual; lesbian, gay, or bisexual; or not sure).

Although data from students in other or multiple racial/ethnic

groups were collected, the numbers were too small to produce

statistically stable estimates specific to other or multiple racial/

ethnic groups; therefore, these data are not presented as a

separate group in this report but were retained in the analytic

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 79

US Department of Health and Human Services/Centers for Disease Control and Prevention

sample. In addition, students aged <14 years (n = 87) were not

included in the analysis by age because the sample of students

in this age category was too small for meaningful analysis and

because these students cannot legally drive anywhere in the

United States (1).

Analysis

For this report, unadjusted weighted prevalence estimates and

corresponding 95% confidence intervals were calculated, and

posthoc t-tests were used to assess between-group differences.

Differences between prevalence estimates were considered

statistically significant if the t-test p value was <0.05. In the

results, only statistically significant differences in prevalence

estimates are reported.

Logistic regression models that controlled for age, sex,

race/ethnicity, academic grades in school, and sexual

identity produced adjusted prevalence ratios and examined

the associations between transportation risk behaviors. For

passenger-related transportation risk behaviors, students who

did not engage in the risk behaviors were designated as the

referent group. For driver-related transportation risk behaviors,

students who drove but did not engage in the risk behaviors

were designated as the referent group. Adjusted prevalence

ratios were considered statistically significant if their pairwise

comparison between groups (risk versus referent) was p <0.05.

Results

In 2019, a total of 43.1% of U.S. high school students had

not always worn a seat belt and 16.7% had ridden with a

drinking driver during the 30 days before the survey (Table 1).

Among the 59.9% of respondents who had driven a car or other

vehicle during the 30 days before the survey, 5.4% had driven

after drinking alcohol and 39.0% had texted while driving.

Both driving after drinking alcohol and texting while driving

usually increased with age. Specifically, prevalence of driving

after drinking alcohol was higher among students aged ≥18

years (8.9%) than among students aged 16 (4.0%), 15 (2.6%),

or 14 (2.7%) years (Table 1). In addition, prevalence was higher

among students aged 17 (5.9%) years than among those aged

15 (2.6%) years. For texting while driving, prevalence was

higher among students aged ≥18 (59.5%) years than among

students aged 17 (50.9%), 16 (30.5%), 15 (15.5%), or

14 (15.5%) years. Prevalence also was higher among students

aged 17 years than among those aged 16, 15, or 14 years and

higher among students aged 16 years than among those aged

15 or 14 years.

Conversely, not always wearing a seat belt usually decreased

with age. Prevalence of not always wearing a seat belt was lower

among students aged ≥18 years (39.4%) than among students

aged 16 (43.5%), 15 (46.9%), or 14 (45.7%) years. Similarly,

prevalence was lower among students aged 17 (38.9%) years

than among all younger students. For riding with a drinking

driver, no differences occurred by age.

Differences by race/ethnicity were detected for all four

transportation risk behaviors but did not demonstrate a

consistent pattern. Prevalence of not always wearing a seat

belt was higher among black students (61.7%) than among

Hispanic students (48.2%) or white students (36.6%). In

addition, prevalence among Hispanic students was higher than

among white students. For the alcohol-related transportation

risk behaviors, Hispanic students (20.8%) had a higher

prevalence of riding with a drinking driver than black students

(15.9%) or white students (15.1%), and Hispanic students

(6.6%) had a higher prevalence of driving after drinking alcohol

than black students (4.1%). In contrast, prevalence of texting

while driving was higher among white students (43.9%) than

among black students (29.5%) or Hispanic students (35.2%).

Students whose academic grades in school were mostly Cs, Ds,

or Fs had a higher prevalence of not always wearing a seat belt

(57.0%), riding with a drinking driver (20.1%), and driving

after drinking alcohol (7.4%) than students whose academic

grades in school were mostly As or Bs (38.8%, 15.3%, and

4.7%, respectively); however, prevalence of texting while

driving did not differ by this characteristic.

Few differences were identified when examining behaviors by

sex and by sexual identity. Only alcohol-related transportation

risk behaviors demonstrated differences. Among students who

had driven during the 30 days before the survey, male students

(7.0%) had a higher prevalence of driving after drinking alcohol

than female students (3.6%). By sexual identity, students who

were not sure of their sexual identity (21.9%) had a higher

prevalence of riding with a drinking driver than heterosexual

students (15.7%); however, the prevalence was not different

from lesbian, gay, or bisexual students (19.2%).

Multivariable analyses indicated that, for each transportation

risk behavior, students engaging in that behavior were more likely

to engage in each of the other transportation risk behaviors,

after controlling for age, sex, race/ethnicity, academic grades

in school, and sexual identity (Table 2). For passenger-related

transportation risk behaviors, students who did not always

wear a seat belt were 1.80 times as likely to have ridden with a

drinking driver, 2.73 times as likely to have driven after drinking

alcohol, and 1.29 times as likely to have texted while driving than

students who always wore a seat belt. Students who had ridden

with a drinking driver during the 30 days before the survey were

1.42 times as likely to not always wear a seat belt, 9.87 times as

likely to have driven after drinking alcohol, and 1.50 times as

likely to have texted while driving than students who had not

Supplement

80 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 1. Unweighted number and unadjusted weighted prevalence estimates of high school students* who engaged in transportation risk

behaviors, by selected characteristics — Youth Risk Behavior Survey, United States, 2019

Characteristic

Total

Did not always wear a

seat belt

Rode with a driver who

had been drinking

alcohol**

Drove when they had

been drinking

alcohol**

,††

Texted or e-mailed while

driving

††,§§

No.

†

% (95% CI) No.

% (95% CI)

Total 13,677 NA 4,852 43.1 (40.2–45.9) 2,214 16.7 (15.2–18.2) 423 5.4 (4.5–6.5) 2,784 39.0 (36.4–41.7)

Age (yrs)

¶¶

14 1,699 11.9 (10.9–13.0) 573 45.7 (40.9–50.5) 276 16.4 (13.9–19.1) 14 2.7 (0.9–7.5) 51 15.5 (11.2–21.0)

15 3,473 24.8 (23.5–26.0) 1,283 46.9 (42.7–51.1) 557 16.7 (14.6–19.1) 49 2.6 (1.7–3.9) 211 15.5 (11.8–20.2)

16 3,628 25.6 (24.5–26.7) 1,318 43.5 (39.5–47.6) 564 16.0 (13.8–18.5) 112 4.0 (2.8–5.6) 730 30.5 (25.8–35.5)

17 3,102 23.7 (22.5–24.8) 1,045 38.9 (35.6–42.4) 481 16.0 (13.8–18.5) 138 5.9 (4.3–7.9) 1,072 50.9 (46.5–55.3)

≥18 1,616 13.7 (12.6–14.9) 574 39.4 (36.6–42.4) 279 18.4 (15.4–21.7) 91 8.9 (6.4–12.4) 672 59.5 (54.9–63.9)

Sex

Male 6,641 50.6 (49.1–52.1) 2,369 43.3 (40.0–46.7) 1,015 15.6 (14.1–17.2) 257 7.0 (5.6–8.8) 1,434 39.6 (36.6–42.6)

Female 6,885 49.4 (47.9–50.9) 2,440 42.7 (39.7–45.7) 1,141 17.5 (15.6–19.5) 149 3.6 (2.8–4.6) 1,311 38.4 (35.5–41.4)

Race/Ethnicity***

White,

non-Hispanic

6,668 51.2 (46.4–56.0) 2,079 36.6 (33.8–39.6) 986 15.1 (13.5–16.8) 207 5.1 (3.9–6.5) 1,608 43.9 (40.4–47.5)

Black,

non-Hispanic

2,040 12.2 (10.2–14.6) 901 61.7 (56.3–66.8) 325 15.9 (13.3–18.7) 47 4.1 (2.6–6.4) 312 29.5 (24.3–35.2)

Hispanic 3,038 26.1 (21.8–30.9) 1,237 48.2 (45.0–51.4) 605 20.8 (18.7–23.1) 107 6.6 (5.2–8.5) 562 35.2 (30.8–39.8)

Academic grades

†††

Mostly As or Bs 9,785 75.1 (72.2–77.8) 3,152 38.8 (36.0–41.6) 1,449 15.3 (13.8–17.0) 248 4.7 (3.8–5.9) 2,070 40.4 (37.8–43.1)

Mostly Cs, Ds, or Fs 2,677 20.6 (18.3–23.2) 1,226 57.0 (53.4–60.5) 547 20.1 (17.7–22.8) 133 7.4 (5.7–9.6) 548 37.1 (32.2–42.4)

Sexual identity

Heterosexual 10,853 84.4 (83.4–85.3) 3,741 42.1 (39.1–45.2) 1,656 15.7 (14.1–17.4) 322 5.2 (4.2–6.4) 2,268 39.6 (36.6–42.6)

Lesbian, gay, or

bisexual

1,531 11.2 (10.4–12.0) 564 44.7 (39.4–50.1) 283 19.2 (16.0–22.9) 39 4.7 (2.4–9.0) 257 34.7 (28.4–41.7)

Not sure 591 4.5 (3.9–5.0) 208 43.3 (37.6–49.2) 125 21.9 (16.8–28.1) 24 9.5 (4.8–17.7) 93 31.7 (22.0–43.4)

Abbreviations: CI = confidence interval; NA = not applicable.

* Unadjusted weighted prevalence estimates and corresponding 95% CIs were calculated and are presented in the table. Posthoc t-tests were used to assess

between-group differences. Differences were considered statistically significant if the t-test p value was <0.05. Statistical significance is not indicated in the table

due to the large number of different pairwise comparisons; however, all significant differences are described in the results.

†

The unweighted number of students for each characteristic only includes students who selected a response on the survey question pertaining to that characteristic.

Students who did not select a response were not included in the analysis for that characteristic but were retained in the analytic sample for every question on

which they provided a response.

Students who selected any response on the survey question pertaining to a risk behavior were included in the analysis for that behavior; however, only the

unweighted numbers of students who engaged in that behavior are presented in the table. Students who did not select a response were not included in the

analysis for that behavior but were retained in the analytic sample for every question on which they provided a response.

Most of the time, sometimes, rarely, or never wore a seat belt when riding in a car driven by someone else.

** ≥1 time during the 30 days before the survey.

††

Among students who had driven a car or other vehicle during the 30 days before the survey.

§§

On ≥1 day during the 30 days before the survey.

¶¶

The total column percentages for age do not add up to 100% because students aged <14 years are not presented because they cannot drive legally in any U.S. state.

*** The total column percentages for race/ethnicity do not add up to 100% because other non-Hispanic race categories are not presented.

†††

The total column percentages for academic grades do not add up to 100% because students who were not sure about their grades or who responded “none of

these grades” are not presented.

ridden with a drinking driver. For driver-related transportation

risk behaviors, students who had driven after drinking alcohol

at least once during the 30 days before the survey were 1.65

times as likely to not always wear a seat belt, 4.91 times as likely

to have ridden with a drinking driver, and 2.38 times as likely

to have texted while driving than students who had not driven

after drinking alcohol. Students who had texted while driving

on at least one day during the 30 days before the survey were

1.32 times as likely to not always wear a seat belt, 1.96 times as

likely to have ridden with a drinking driver, and 12.64* times

* Estimate should be interpreted with caution because the 95% confidence

interval is wide.

as likely to have driven after drinking alcohol than students who

had not texted while driving.

Discussion

Transportation risk behaviors varied by student characteristics,

with age, race/ethnicity, and academic grades demonstrating

the most differences. Increased engagement in driver-related

transportation risk behaviors as students become older has been

reported in other studies (7–9). This finding is not surprising

because adolescents engage in certain risky driver–related

behaviors less often when an adult supervisor is present in the

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 81

US Department of Health and Human Services/Centers for Disease Control and Prevention

TABLE 2. Adjusted prevalence ratios* for high school students who engaged in multiple transportation risk behaviors — Youth Risk Behavior

Survey, United States, 2019

Transportation risk behavior

†

Did not always wear

a seat belt

Rode with a driver who

had been drinking

alcohol

Drove when they had been

drinking alcohol

¶,

Texted or

e-mailed while

driving**

,††

aPR (95% CI) aPR (95% CI) aPR (95% CI) aPR (95% CI)

Did not always wear a seat belt

NA 1.80 (1.59–2.04) 2.73 (1.81–4.11) 1.29 (1.19–1.41)

Rode with a driver who had been drinking alcohol

1.42 (1.32–1.53) NA 9.87 (7.14–13.64) 1.50 (1.37–1.65)

Drove when they had been drinking alcohol

¶,

** 1.65 (1.40–1.95) 4.91 (4.17–5.77) NA 2.38 (2.15–2.63)

Texted or e-mailed while driving**

,††

1.32 (1.20–1.44) 1.96 (1.69–2.27) 12.64 (8.45–18.91)

§§

Abbreviations: aPR = adjusted prevalence ratio; CI = confidence interval; NA = not applicable.

* Multivariable logistic regression models that controlled for age, sex, race/ethnicity, academic grades, and sexual identity were used to produce the aPRs and

corresponding 95% CIs presented in the table. The aPRs were considered statistically significant if the p value of their pairwise comparison between groups (risk

versus referent) was <0.05. All aPRs in the table are significant.

†

Students who engaged in protective behaviors (i.e., always wearing a seat belt) or did not engage in risk behaviors (i.e., riding with a driver who had been drinking

alcohol, driving when they had been drinking alcohol among students who had driven, or texting or e-mailing while driving among students who had driven)

were the referent group.

Most of the time, sometimes, rarely, or never wore a seat belt when riding in a car driven by someone else.

≥1 time during the 30 days before the survey.

** Among students who had driven a car or other vehicle during the 30 days before the survey.

††

On ≥1 day during the 30 days before the survey.

§§

Estimate should be interpreted with caution because the 95% CI is wide.

vehicle, as is required when adolescents possess a driver’s permit

(https://aaafoundation.org/distracted-driving-among-newly-

licensed-teen-drivers). As adolescents age, begin to drive without

adult supervision, and gain driving experience, driver-related

risk behaviors can be more common (9) (https://aaafoundation.

org/distracted-driving-among-newly-licensed-teen-drivers).

The positive association between age and texting while driving

illustrates the need to sustain attention to preventing the behavior

throughout adolescence (9). On the other hand, the prevalence

of not always wearing a seat belt decreased by age, possibly

indicating that although adolescents are typically more willing

to engage in risky transportation behaviors as they become older,

they still maintain a sense of self-preservation and risk perception

and therefore take precautions by wearing seat belts.

This study demonstrated that Hispanic students had a higher

prevalence of riding with a drinking driver and driving after

drinking alcohol than white students or black students. One

study described similar findings about drinking and driving

among Hispanics in the literature (10). Additional research

to explore which Hispanic populations might be at higher

risk found that U.S.-born Hispanic youths were more likely

to initiate drinking and driving behavior compared with first-

generation immigrant Hispanic youths, even after adjusting

for demographic variables (10). Additional research is needed

to determine whether different strategies to reduce alcohol-

impaired driving should be selected for or tailored to specific

Hispanic populations based on nativity status.

Other studies have reported that students with lower

academic grades were more likely to engage in other health-

related risk behaviors (e.g., risky sexual behaviors or substance

use) (11). The 2019 YRBS illustrates that this association

extends to engagement in transportation risk behaviors. Lower

academic achievement might be indicative of an underlying

tendency to make riskier decisions, or risky behaviors

themselves might lead to lower academic achievement. More

research into a potential causal association and the temporality

of that association is warranted. Of note, texting while driving

was the one transportation risk behavior that did not differ

by academic achievement. One potential explanation is that

although adolescents understand that texting while driving

is unsafe, the perceived benefits of texting while driving and

the motivations for engaging in the behavior often differ

from other transportation risk behaviors and can outweigh

the perceived risks for adolescents at the moment when they

choose to do it (8,9).

In this study, students engaging in any given transportation

risk behavior were more likely to engage in each of the other

measured transportation risk behaviors, even after controlling

for student characteristics. Associations with alcohol-

related behaviors were highest, particularly for driving after

drinking alcohol. Students who engaged in any of the other

transportation risk behaviors were approximately 3–13 times

as likely to have also engaged in driving after drinking alcohol

at least once during the 30 days before the survey. This might

signify a general willingness to engage in risky behaviors among

students who choose to drink and drive. This finding is also

concerning because of the potential additive effects of these

transportation risk behaviors. For example, adolescents who

drive after drinking alcohol, thus increasing their risk for a

crash, are also more likely to not always wear a seat belt, which

increases their risk for injury or death during a crash.

Because students engaged in multiple transportation

risk behaviors, interventions designed to address multiple

transportation risk behaviors might concurrently help reduce

Supplement

82 MMWR / August 21, 2020 / Vol. 69 / No. 1

US Department of Health and Human Services/Centers for Disease Control and Prevention

those behaviors. Existing infrastructure and resources for

comprehensive school and community programs designed

to address different health behaviors could be leveraged to

expand the benefits of these programs to transportation risk

behaviors. For example, programs that already rely on family

engagement could incorporate safe driving, because parental

involvement is crucial for teaching adolescents how to drive

by providing varied practice opportunities, promulgating

safe driver behaviors, and instilling the importance of

avoiding transportation risk behaviors (https://www.cdc.

gov/parentsarethekey/parents/index.html). Programs that

provide counseling and social services for adolescents could

incorporate brief alcohol interventions, which are promising

for reducing drinking and driving among adolescents at high

risk for engaging in the behavior (12,13).

Engagement in all of these transportation risk behaviors

across the United States remains high. Considering that

adolescent drivers (16–19 years of age) have the highest crash

rates (1), the fact that only six of 10 adolescents in this study

always wore seat belts is concerning. Measures that are effective

for increasing seat belt use, such as primary enforcement seat

belt laws that allow police to ticket drivers or passengers for

being unrestrained even in the absence of other violations (13),

also can be beneficial for preventing crashes or crash injuries

involving other contributing factors. For example, evidence

indicates that primary enforcement seat belt laws are effective

for reducing fatal alcohol-related crashes among underage

drivers aged 15–20 years (14).

Although this study did not find many differences in riding

with a drinking driver by student characteristics, approximately

one of every five students engaged in the behavior

. Riding with

a drinking driver is intrinsically unsafe and also is associated

with adolescent drinking and driving

(15). Longitudinal

research has revealed that adolescent passengers who are

exposed to drinking and driving at a young age are more likely

to engage in drinking and driving themselves as they become

older and begin to drive (16). Additional research about

the drinking drivers with whom adolescents ride and their

relationships with the drinking drivers (e.g., parents, other

family members, or peers) might be useful for designing and

implementing targeted interventions.

In every U.S. state, minimum legal drinking age (MLDA)

laws stipulate that drinking alcohol is illegal for anyone aged

<21 years, as is driving after drinking any amount of alcohol

(zero tolerance laws) (1,13). Despite these laws, approximately

one fifth of drivers aged 16–20 years killed in crashes during

2018 had BACs of ≥0.08% (1). This study found that in 2019,

a total of 5.4% of students who drove did so after drinking

alcohol at least once in the previous 30 days. Driving after

drinking alcohol is risky and unacceptable at any age; however,

the risk is even higher among adolescent drivers aged 16–20

years, even at BACs below the legal limit for adults (4). Zero

tolerance laws (7,13,14), graduated driver licensing systems

(7), and MLDA laws (7,13,14) are effective in helping reduce

drinking and driving and alcohol-related crashes and injuries

among adolescents, and they should continue to remain

universally implemented. Other general population deterrent

approaches that are effective for preventing alcohol-impaired

driving overall also can be beneficial for specifically preventing

adolescent drinking and driving. For example, publicized

sobriety checkpoints are highly effective for reducing drinking

and driving overall (13), and evidence indicates that they can

reduce alcohol-impaired driving (17) and alcohol-related

crashes among underage drivers (14).

Consistent with two recent studies, this analysis determined

that texting while driving among adolescents remains high,

increases with age, and is more common among white students

than students of other races/ethnicities (8,9). Similar to the

other studies, this analysis also determined that adolescents

who engage in texting while driving are more likely to engage

in other transportation risk behaviors (8,9). Awareness

campaigns, education, and changes in policy related to texting

while driving have had mixed effectiveness (9,13). Because of

this, such technologic interventions as in-vehicle cell phone

blocking technologies can serve as potential solutions; however,

the effectiveness and acceptability of such solutions require

more research (9,13).

Lack of parental monitoring and supervision is a common

underlying contributor to many health risk behaviors, and

parental involvement can be especially important for reducing

transportation risk behaviors. For example, one study found

that adolescents with supportive parents who monitor their

behavior were less likely to engage in multiple passenger- and

driver-related transportation risk behaviors, including seat belt

nonuse, cell phone use while driving, and drinking and driving,

than adolescents with uninvolved parents (18). Parents/guardians

also can play a vital role in teaching adolescents to drive by

helping ensure they gain valuable driving experience and by

setting rules and expectations for adolescent drivers, including

rules and expectations for not engaging in transportation

risk behaviors. Parent-teen driving agreements (https://www.

cdc.gov/parentsarethekey/parents/index.html) can formalize

those expectations and demonstrate a commitment between

parents and adolescents to adhere to safe driving practices while

adolescents gain new driving privileges over time.

Supplement

MMWR / August 21, 2020 / Vol. 69 / No. 1 83

US Department of Health and Human Services/Centers for Disease Control and Prevention

Limitations

General limitations for the YRBS are available in the overview

report of this supplement (6). The findings in this report are

subject to at least two additional limitations. First, YRBS does

not quantify driving or riding exposure in general or during the

30 days before the survey. How many trips each student takes as

a driver or as a passenger and the amount of time each student

spends on the road are unknown. High school students who

take more frequent trips or drive for longer times or distances

might have more opportunity to engage in transportation risk

behaviors because of a higher exposure that is not captured by the

survey. Second, for riding with a driver who had been drinking

alcohol, the relationship between the student and the drinking

driver (e.g., parent/guardian, other family member, a peer, or

someone else) is unknown. The nature of this relationship

might have implications for designing potential strategies and

prevention messages for empowering adolescents so that they

can intervene (15).

Conclusion

Motor-vehicle–crash injuries remain a leading cause of

death among adolescents. Despite this, passenger- and driver-

related transportation risk behaviors that increase the risk for

crashes, injuries, and deaths remain too common. Reducing

transportation risk behaviors among adolescents by using proven

strategies, especially those that can target multiple transportation

risk behaviors, can help prevent crashes, reduce injuries, and

save lives. Because driver-related transportation risk behaviors

increased with age, continued emphasis on implementation

of effective strategies for preventing these behaviors with high

school juniors and seniors should be considered.

Conicts of Interest

All authors have completed and submitted the International

Committee of Medical Journal Editors form for disclosure of

potential conflicts of interest. No potential conflicts of interest

were disclosed.

References

1. Insurance Institute for Highway Safety. Teenagers. Arlington, VA:

Insurance Institute for Highway Safety, Highway Loss Data Institute;

2019. https://www.iihs.org/topics/teenagers

2. O’Malley PM, Johnston LD. Driving after drug or alcohol use by US

high school seniors, 2001-2011. Am J Public Health 2013;103:2027–34.

https://doi.org/10.2105/AJPH.2013.301246

3. McCartt AT, Shabanova VI, Leaf WA. Driving experience, crashes and

traffic citations of teenage beginning drivers. Accid Anal Prev

2003;35:311–20. https://doi.org/10.1016/S0001-4575(02)00006-4

4. Voas RB, Torres P, Romano E, Lacey JH. Alcohol-related risk of driver

fatalities: an update using 2007 data. J Stud Alcohol Drugs

2012;73:341–50. https://doi.org/10.15288/jsad.2012.73.341

5. Llerena LE, Aronow KV, Macleod J, et al. An evidence-based review:

distracted driver. J Trauma Acute Care Surg 2015;78:147–52. https://

doi.org/10.1097/TA.0000000000000487

6. Underwood JM, Brener N, Thornton J, et al. Overview and methods

for the Youth Risk Behavior Surveillance System—United States, 2019.

In: Youth Risk Behavior Surveillance—United States, 2019. MMWR

Suppl 2020;69(No. Suppl 1).

7. CDC. Vital signs: drinking and driving among high school students

aged ≥16 years—United States, 1991-2011. MMWR Morb Mortal

Wkly Rep 2012;61:796–800.

8. Olsen EO, Shults RA, Eaton DK. Texting while driving and other risky

motor vehicle behaviors among US high school students. Pediatrics

2013;131:e1708–15. https://doi.org/10.1542/peds.2012-3462

9. Li L, Shults RA, Andridge RR, Yellman MA, Xiang H, Zhu M. Texting/

emailing while driving among high school students in 35 states, United

States, 2015. J Adolesc Health 2018;63:701–8. https://doi.org/10.1016/j.

jadohealth.2018.06.010

10. Maldonado-Molina MM, Reingle JM, Jennings WG, Prado G. Drinking

and driving among immigrant and US-born Hispanic young adults:

results from a longitudinal and nationally representative study. Addict

Behav 2011;36:381–8. https://doi.org/10.1016/j.addbeh.2010.12.017

11. Rasberry CN, Tiu GF, Kann L, et al. Health-related behaviors and

academic achievement among high school students—United States,

2015. MMWR Morb Mortal Wkly Rep 2017;66:921–7. https://doi.

org/10.15585/mmwr.mm6635a1

12. Steinka-Fry KT, Tanner-Smith EE, Hennessy EA. Effects of brief alcohol

interventions on drinking and driving among youth: a systematic review

and meta-analysis. J Addict Prev 2015;3:1–22.

13. Richard CM, Magee K, Bacon-Abdelmoteleb P, Brown JL. Countermeasures

that work: a highway safety countermeasure guide for state highway safety

offices. 9th ed. Washington, DC: National Highway Traffic Safety

Administration (NHTSA); 2018. Report no. DOT HS 812 478. https://

www.nhtsa.gov/sites/nhtsa.dot.gov/files/documents/812478_

countermeasures-that-work-a-highway-safety-countermeasures-guide-.pdf

14. Romano E, Scherer M, Fell J, Taylor E. A comprehensive examination of

U.S. laws enacted to reduce alcohol-related crashes among underage drivers.

J Safety Res 2015;55:213–21. https://doi.org/10.1016/j.jsr.2015.08.001

15. Leadbeater BJ, Foran K, Grove-White A. How much can you drink before

driving? The influence of riding with impaired adults and peers on the

driving behaviors of urban and rural youth. Addiction 2008;103:629–37.

https://doi.org/10.1111/j.1360-0443.2008.02139.x

16. Li K, Simons-Morton BG, Vaca FE, Hingson R. Association between

riding with an impaired driver and driving while impaired. Pediatrics

2014;133:620–6. https://doi.org/10.1542/peds.2013-2786

17. Nelson TF, Xuan Z, Babor TF, et al. Efficacy and the strength of evidence

of U.S. alcohol control policies. Am J Prev Med 2013;45:19–28. https://

doi.org/10.1016/j.amepre.2013.03.008

18. Ginsburg KR, Durbin DR, García-España JF, Kalicka EA, Winston FK.

Associations between parenting styles and teen driving, safety-related

behaviors and attitudes. Pediatrics 2009;124:1040–51. https://doi.

org/10.1542/peds.2008-3037

ISSN: 2380-8950 (Print)

The Morbidity and Mortality Weekly Report (MMWR) Series is prepared by the Centers for Disease Control and Prevention (CDC) and is available free

of charge in electronic format. To receive an electronic copy each week, visit MMWR at https://www.cdc.gov/mmwr/index.html.

Readers who have difficulty accessing this PDF file may access the HTML file at https://www.cdc.gov/mmwr/ind2020_su.html. Address all inquiries about

the MMWR Series, including material to be considered for publication, to Executive Editor, MMWR Series, Mailstop E-90, CDC, 1600 Clifton Rd., N.E.,

Atlanta, GA 30329-4027 or to [email protected]v.

All material in the MMWR Series is in the public domain and may be used and reprinted without permission; citation as to source, however, is appreciated.

MMWR and Morbidity and Mortality Weekly Report are service marks of the U.S. Department of Health and Human Services.

Use of trade names and commercial sources is for identification only and does not imply endorsement by the U.S. Department of Health and Human Services.

References to non-CDC sites on the Internet are provided as a service to MMWR readers and do not constitute or imply endorsement of these organizations

or their programs by CDC or the U.S. Department of Health and Human Services. CDC is not responsible for the content of these sites. URL addresses

listed in MMWR were current as of the date of publication.