The 2017 national Youth Risk Behavior Survey (YRBS) collected cross-sectional data on a wide range of priority health risk behaviors from a nationally representative sample of public and private high school students in grades 9 through 12. The survey used a three-stage probability sampling methodology that has been previously described (Kann et al., 2018, Brener et al., 2013). A weighting factor was applied to each student record to adjust for the varying probabilities of selection at each stage of sampling, student non-response, and the oversampling of black and Hispanic students. The questionnaire was administered in the classroom during a regular class period by trained data collectors. Responses were recorded directly on computer-scannable questionnaire booklets or answer sheets. Student participation in the survey was anonymous and voluntary, and local procedures were used to obtain parental consent. The school response rate was 75% and the student response rate was 81%, for an overall response rate of 60%, and a sample size of 14,765. The national YRBS has been reviewed and approved by an IRB at the Centers for Disease Control and Prevention (CDC).

Table 1 identifies the questions and analytic coding for each measure included in the analyses. One indicator of mental health, feeling sad and hopeless, was included in the analyses. Two questions assessed suicidal behaviors.Table 1

Question and Analytic Coding for Included Mental Health and Suicidal Outcomes and Physical Activity, Sedentary, and Dietary Behaviors, 2017 YRBS.a

Four measures of physical activity included daily physical activity, muscle-strengthening activity, physical education, and sports team participation. Two measures of sedentary behaviors focused on screen-time use included television viewing and computer or video game use. Six dietary behaviors included breakfast intake, two questions to determine fruit intake, four questions to determine vegetable intake, intake of soda, intake of sports drink, and water consumption.

Demographic variables included sex, race/ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, and non-Hispanic other), and grade in school (9th–12th). Body weight status was based on body mass index (BMI) calculated from self-reported height and weight. Based on reference data from growth charts produced by CDC, students with BMI ≥ 95th percentile for sex and age were considered to have obesity; students with 85th ≤ BMI < 95th percentile were considered to be overweight; students with 5th ≤ BMI < 85th percentile were considered to be normal weight; and students with BMI < 5th percentile were considered to be underweight.

All analyses were conducted using SUDAAN (version 11.0.3) to account for the survey's complex sampling design. Analyses were stratified by sex because previous research shows differences for male and female adolescents by mental health-related outcomes and dietary and physical activity behaviors. Statistical interaction testing results did not indicate a need to stratify by race/ethnicity, grade in school, or BMI percentile. Chi-squared tests were used to assess significant (p < 0.05) bivariate differences in mental health-related outcomes and physical activity, sedentary, and dietary behaviors by male and female students. Sex-stratified multiple logistic regressions were used to separately model each mental health-related outcome variable on categorical variables for physical activity, sedentary, and dietary behaviors, while controlling for race/ethnicity, grade, and body weight status. Results from the multivariable analyses are reported as adjusted prevalence ratios (APRs) with 95% confidence intervals. APRs were considered statistically significant if p < 0.05 or the 95% confidence interval (CI) did not include 1.0.

The overall sample was 14,765 US high school students with approximately half females (50.7%). A total of 53.5% were non-Hispanic white, 13.4% were non-Hispanic black, 22.8% were Hispanic, and 10.3% were non-Hispanic other. The sample was distributed across the grades in school, 27.3% were in 9th grade, 25.6% in 10th, 23.9% in 11th, and 23.0% in 12th. Most US high students had a normal body weight (66.6%), 15.6% were overweight, 14.8% had obesity, and 3.0% were underweight. The analytic sample varied for each model due to missing data from listwise deletion (see Table 3, Table 4 for analytic sample sizes).Table 2

Mental Health and Suicidal Outcomes and Physical Activity, Sedentary, and Dietary Behaviors by Sex Among U.S. High School Students, 2017 YRBS.

Table 2 provides the prevalence estimates and chi-squared p-values of the mental health and suicide-related outcomes and health-related behaviors stratified by sex. Female students compared with male students had significantly higher prevalence rates for feeling sad or hopeless, suicidal thoughts, and suicide attempts. All outcomes and health behaviors were significantly different for males and females except for two of the dietary behaviors and both sedentary behaviors.

Female students compared with male students had significantly higher prevalence of all the insufficient physical activity behaviors. Female students compared to male students had significantly higher prevalence rates for not eating breakfast and for eating fruit or drinking 100% fruit juice. Whereas male students compared to female students had significantly higher prevalence rates for drinking soda or pop and drinking sports drinks.

Male students who did not meet the aerobic physical activity guideline were 23% more likely to feel sad and hopeless than those who did (Table 3). Female and male students who did not play on at least one sports team compared to those who did were 16% and 18%, respectively, more likely to feel sad and hopeless. Female and male students who played video or computer games or used a computer more than two hours per day compared to those who did not were more likely to feel sad and hopeless.

Female and male students who did not meet the aerobic physical activity guideline compared to those who did were 35% and 36%, respectively, more likely to have seriously considered attempting suicide. In addition, female and male students who played video or computer games or used a computer more than two hours per day compared to those who did not were 42% and 58%, respectively, more likely to have seriously considered attempting suicide.

None of the physical activity behaviors were significantly associated with suicide attempts. However, female and male students who watched television more than two hours per day compared to those who did not were 55% and 73%, respectively, more likely to have attempted suicide. Furthermore, male students who played video or computer games or used a computer more than two hours per day compared to those who did not were 56% more likely to have attempted suicide.

Female and male students who did not eat breakfast on all 7 days were 48% and 59%, respectively, more likely to feel sad and hopeless compared to those who did (Table 4). Female students who drank soda or pop one or more times per day in the last 7 days compared to those who did not were 27% more likely to feel sad and hopeless.

Female and male students who did not eat breakfast on all 7 days compared to those who did were 50% and 76%, respectively, more likely to have seriously considered attempting suicide. In addition, female students who drank soda pop one or more times per day compared to those who did not were 36% more likely to have seriously considered attempting suicide.

Female students who did not eat breakfast on all 7 days compared to those who did were more than 2 times as likely to have attempted suicide. Similarly, male students who did not eat breakfast on all 7 days compared to those who did were almost 2 times more likely to have attempted suicide. Both female and male students who drank soda or pop one or more times per day compared to those who did not were 56% and 70%, respectively, more likely to have attempted suicide. In addition, female and male students who drank a sports drink one or more times per day compared to those who did not were 50% and 57%, respectively, more likely to have attempted suicide.

Previous research has found associations between adolescents who participated in daily physical activity and considering attempting suicide (Brosnahan et al., 2004, Brown et al., 2007). The current study found a similar inverse association, showing that male and female students who were not physically active for at least 60 min daily were more likely to seriously consider attempting suicide. Although previous studies have shown a significant association between physical activity and attempted suicide (Brown et al., 2007, Sibold et al., 2015), the current study did not find this association. Studies that have found this association examined vigorous activity among adolescents on various number of days of doing physical activity (Brown et al., 2007, Sibold et al., 2015), while this study examined self-reported moderate and vigorous physical activity combined and on all 7 days, which aligns with national guidance. In addition, the only sex difference found for physical activity was that male students who did not meet the aerobic physical activity guideline were more likely to feel sad and hopeless. This finding might be explained by the support males get from family, friends, and school environments to be more physically active, which in turn might have a stronger influence on their self-efficacy and level of enjoyment (Wenthe et al., 2009), and ultimately, their overall feelings and emotions.

In addition, the current study did not find a significant association between sports participation and suicidal behaviors as other studies have shown (Brown et al., 2007, Gunn and Lester, 2014, Southerland et al., 2016). However, there was a significant association for both male and female students that showed not playing on a sports team increased their likelihood of feeling sad and hopeless, which supports the social benefits of playing on a sports team and demonstrates a possible protective effect. The 2019 release of the National Youth Sports Strategy also underscores the physical activity, psychosocial, and academic achievement benefits of sports team participation (U.S. Department of Health and Human Services, 2019). Although physical education attendance was not significantly associated with mental health outcomes, it is the cornerstone for Comprehensive School Physical Activity Programs, which serves as a national framework for physical education and physical activity in schools (CDC, 2019). However, only 4% of schools nationwide require daily physical education for all students K-12 (CDC, 2015), suggesting opportunities exist for improving adolescents’ attendance in physical education and ultimately increasing daily physical activity.

Adolescents who engage in the recommended levels of daily physical activity can still be at risk for high levels of sedentary behaviors at other times during the day, subsequently leading to increased risk for undesirable health outcomes (Tremblay et al., 2011). Previous studies have found associations between adolescent sedentary behaviors and poor mental health outcomes, including depression, though some findings have been mixed (Benson et al., 2013, Suchert et al., 2015). In this study, playing video games or using the computer for more than two hours per day was significantly associated with greater likelihood of feeling sad and hopeless and seriously considering suicide. Prolonged periods of engaging in sedentary activities can increase social isolation and feelings of loneliness, which are associated with suicidal thoughts and behaviors (Bearman and Moody, 2004). The observed sex differences for the associations between video game and computer usage and suicide attempts highlight the importance of considering other factors such as social media engagement and video game content. Further exploring the types of video games and websites accessed on computers and the ways in which adolescents engage in video gaming and other media might inform differences between female and male adolescents and differential effects on their mental health (Desai et al., 2010).

Similar to findings from previous research, the current study found associations between skipping breakfast and increased risk of mental health related outcomes among males and females. Individuals who feel sad and hopeless or experience other mental health issues may be more likely to skip meals because of low appetite or as a dieting strategy. Alternatively, skipping breakfast could lead to lower calorie intake which may contribute to depressed mood (Fulkerson et al., 2004). Further, natural shifts in sleep patterns among adolescents might also lead to breakfast skipping (Crowley et al., 2007). Efforts to support healthy eating behaviors are needed at multiple levels (US Department of Health and Human Services and US Department of Agriculture, 2015). For example, schools can support healthy eating behaviors, including regular consumption of breakfast by participating in the federal school meal programs (e.g., School Breakfast Program), and offering second chance breakfast during morning break at school.

This study also found associations between consuming soda one or more times per day and all three mental health and suicidal outcomes among females. Interestingly, after adjusting for covariates, among males, consuming soda was only associated with having attempted suicide. It is not clear why these differences between female and male students exist. Other research has shown associations between soda intake and similar mental health related outcomes in adolescents but did not stratify results by sex (Solnick and Hemenway, 2014). Research has also shown associations between caffeine consumption and depressive symptoms in both male and female adolescents (Fulkerson et al., 2004). This study was not able to distinguish between caffeinated and non-caffeinated sodas. The American Academy of Pediatrics states that caffeine use among all children should be discouraged (Committee on Nutrition and the Council on Sports Medicine and Fitness, 2011). Although the Smart Snacks in School nutrition standards allow high schools to sell low calorie caffeinated beverages (e.g., diet soda), school districts can establish nutrition standards that specifically address the sale of caffeinated beverages.

One finding that warrants further research is the association between consuming sports drinks one or more times per day and having attempted suicide among females and males. Sports drinks typically contain added sugar and electrolytes, but do not contain caffeine (Committee on Nutrition and the Council on Sports Medicine and Fitness, 2011). However, it is possible that YRBS participants are including energy drinks, which contain added caffeine and other stimulants, when asked about sports drink consumption. These two types of beverages are often confused (Committee on Nutrition and the Council on Sports Medicine and Fitness, 2011), and could affect the results of the analysis.

A few key limitations should be noted. The cross-sectional data used in these analyses do not allow for conclusions regarding the direction of the modeled associations. Specifically, the correlations in this current study do not show causation. While adolescents who have insufficient physical activity and engage in sedentary and less healthy dietary behaviors may experience poor mental health and suicidal thoughts and behaviors, it is also true that adolescents with poor mental health and suicidal thoughts and behaviors might be less likely to engage in healthy behaviors. Depression also could lead to the social withdrawal, lessened activity, and impact on eating behaviors. However, previous evidence has shown support for a causal association from sedentary behavior to depression (Choi et al., 2019). Longitudinal studies are needed to identify potential cause-and-effect associations between modifiable health risk behaviors and mental health and suicidal behaviors.

While many of the statistically significant effects were small in magnitude, the associations were consistent with previous findings. Additional variables that explain personal (e.g., self-efficacy) and contextual (e.g., peer and parent support) factors of health behaviors might yield greater effects and may be considered in future research. In addition, data came from a single survey so associations could represent a shared response bias. Further, the results are based on student self-report responses to self-administered questionnaires, raising the possibility of reporting error or social-desirability bias, and single-item assessments of suicide attempts seem to be problematic (Hom et al., 2019, Hom et al., 2016). However, YRBS questionnaire items generally demonstrate good test–retest reliability (Brener et al., 2013). CDC also reviewed existing empiric literature to assess cognitive and situational factors that might affect the validity of adolescent self-reporting of behaviors (Brener et al., 2013). Family socioeconomic status was not available but could potentially be a confounder. Results also could be biased by using listwise deletion and not including adolescents who are out of school who may be at higher risk of poor mental health outcomes.

Finally, a greater precision of measurement could be used to assess physical activity, dietary behaviors, sedentary behaviors, and medical record information documenting mental health and suicidal behaviors or more comprehensive scales for these behaviors. Given these limitations, the data should be interpreted cautiously. However, the results of this study reveal potentially important associations between modifiable health risk behaviors and mental health and suicidal behaviors.