Sample weights have been applied. Overlapping 95% CIs demonstrate no statistically significant difference between the effect sizes for males vs females. Data were adjusted for age in wave 1, race/ethnicity (non-Hispanic white, black, Hispanic, and Asian, Native American, or other), sex, parental education (high school or less, some college, and college graduate or more), family structure (2 parents, single mom, and single dad or other), gap in health insurance between wave 1 and wave 4, region, and school characteristics (public or private and mean daily attendance). Depression was defined as having received a diagnosis of depression; current depressive symptoms, a score of 10 or higher on the 10-item subscale of the Center for Epidemiologic Studies Depression scale; and anxiety, having received a diagnosis of anxiety.
eTable 1. Association between Team Sports and Depression, Current Depressive Symptoms, and Anxiety in Full Sample after Accounting for Baseline Mental Health at Wave I without and with Propensity Weighting
eTable 2. Association between Other Activities and Depression, Current Depressive Symptoms, and Anxiety in Full Sample without and with Propensity Weighting
eTable 3. Association between Number of Sports Participated in and Depression, Current Depressive Symptoms, and Anxiety in Full Sample without Propensity Weighting
eTable 4. Association between Specific Sports and Depression, Current Depressive Symptoms, and Anxiety in Full Sample without Propensity Weighting
Customize your JAMA Network experience by selecting one or more topics from the list below.
Easterlin MC, Chung PJ, Leng M, Dudovitz R. Association of Team Sports Participation With Long-term Mental Health Outcomes Among Individuals Exposed to Adverse Childhood Experiences. JAMA Pediatr. 2019;173(7):681–688. doi:10.1001/jamapediatrics.2019.1212
Is team sports participation during adolescence associated with better long-term mental health outcomes among individuals exposed to adverse childhood experiences?
In this study of 9668 individuals from a nationally representative database, among those with adverse childhood experiences, team sports participation during adolescence was significantly associated with better adult mental health outcomes, especially for males, including lower likelihood of having ever received a diagnosis of depression or anxiety and having current depressive symptoms.
The findings suggest that participation in team sports is associated with better adult mental health outcomes among individuals exposed to adverse childhood experiences.
Adverse childhood experiences (ACEs) are associated with long-term poor mental health. Less is known about factors that improve long-term mental health among those with ACEs.
To evaluate, among those exposed to ACEs, whether team sports participation during adolescence is associated with better mental health in adulthood and whether the association between team sports participation and mental health varies by sex.
Design, Setting, and Participants
This study used data from 9668 individuals who participated in waves 1 (1994-1995) and 4 (2008) of the National Longitudinal Study of Adolescent to Adult Health. Individuals were included if they had complete data on exposure to ACEs (physical and sexual abuse, emotional neglect, parental alcohol misuse, parental incarceration, and living with a single parent) and a valid sample weight. Statistical analysis was performed from November 6, 2017, to January 4, 2019.
Main Outcomes and Measures
The association between team sports participation in grades 7 to 12 (wave 1) and diagnosis of depression and/or anxiety and current depressive symptoms (determined by Center for Epidemiologic Studies Depression scale–10 scores) at ages 24 to 32 years (wave 4) among individuals exposed to ACEs. Multivariable logistic regression models were weighted based on propensity scores for factors associated with team sports participation and controlled for individual, family, and school characteristics. Interaction terms tested whether associations between team sports participation and mental health varied by sex.
Of 9668 individuals included in the study (4470 male [50.0%]; mean [SD] age, 15.2 [1.75] years), 4888 (49.3%) reported 1 or more ACE and 2084 (21.3%) reported 2 or more ACEs. Among those with ACEs, team sports participation during adolescence was significantly associated with lower odds of receiving a diagnosis of depression (unadjusted rate, 16.8% vs 22.0%; propensity score–weighted [PSW] adjusted odds ratio [aOR], 0.76; 95% CI, 0.59-0.97) or anxiety (11.8% vs 16.8%; PSW aOR, 0.70; 95% CI, 0.56-0.89) and having current depressive symptoms (21.9% vs 27.5%; PSW aOR, 0.85; 95% CI, 0.71-1.01). There were no significant differences in associations between team sports participation and mental health by sex. Stratified analyses showed significant associations for all outcomes among males (depression: PSW aOR, 0.67 [95% CI, 0.46-0.99]; anxiety: PSW aOR, 0.66 [95% CI, 0.45-0.96]; depressive symptoms: PSW aOR, 0.75 [95% CI 0.56-0.99]) but only 1 outcome among females (anxiety: PSW aOR, 0.73; 95% CI, 0.54-0.98).
Conclusions and Relevance
Among individuals affected by ACEs, team sports participation in adolescence was associated with better adult mental health. Team sports may be an important and scalable resilience builder.
Adverse childhood experiences (ACEs) classically include physical or emotional neglect or abuse, sexual abuse, domestic abuse, exposure to household substance misuse or mental illness, parental separation or divorce, and parental incarceration.1 Adverse childhood experiences are common, with about half of children experiencing 1 and one-quarter experiencing 2 or more.2,3 Physical and mental health problems are more prevalent among those with ACEs throughout the life course.3-6 Risk increases in a graded fashion for multiple mental health conditions (including depression and anxiety)2,7 that are associated with worse life and health outcomes.8,9 Furthermore, the association between ACEs and mental health is thought to be one pathway through which parental ACEs might influence child health.10 Therefore, interventions to prevent ACEs or ameliorate their effects have the potential to improve health throughout the life course.11-15
Little is known, however, about what factors improve long-term outcomes for those exposed to ACEs. Physical activity in childhood and adolescence has been shown to have benefits for mental health, including improvements in depressive symptoms,16-19 stress,16 and self-esteem.18-20 Team sports in particular are thought to lead to improved mental health because of their social nature and resulting social support.20,21 However, it is not known whether team sports participation among the subgroup of youths exposed to ACEs is associated with improved mental health outcomes. For example, ACEs may affect children in ways that could alter or limit the benefits of sports participation.
We sought to examine whether, among those affected by ACEs, team sports participation during adolescence is associated with improved adult mental health. We hypothesized that team sports participation in the context of ACEs may be associated with improved mental health outcomes through direct physiologic alterations, promotion of resilient characteristics, and/or improved social support. Because studies have shown differential effects of ACEs by sex22 and differential benefits of physical activity on mental health by sex,20,23,24 we also examined whether the association between team sports and mental health varies by sex.
We used the National Longitudinal Study of Adolescent to Adult Health (Add Health)25 to compare the emergence of 3 different adult mental health outcomes (diagnosis of depression, diagnosis of anxiety, and depressive symptoms) among those with ACEs during childhood who did and did not participate in team sports during adolescence. We used survey data from wave 1 (collected during the 1994-1995 school year, when respondents were in grades 7-12) to examine participation in team sports during adolescence and data from wave 4 (collected in 2008 when respondents were 24-32 years of age) to examine adult mental health. The study was approved by the UCLA Institutional Review Board, with a waiver of informed consent because the study was a secondary analysis of deidentified data.
We included individuals with data from the wave 1 in-school interview (n = 90 118), wave 1 in-home interview (n = 15 355), and wave 4 in-home interview (n = 11 681) who had complete data on exposure to ACEs (9887), had been exposed to 1 or more ACE (n = 5009), and had a valid sample weight (n = 4888). To provide context for the analytic sample, the demographics are given with 4870 participants with complete data and a valid sample weight who did not report ACEs and were thus excluded from our analytic sample (Table 1). Adverse childhood experiences included sexual abuse, physical abuse (more than 2 incidents), emotional neglect (more than 10 incidents), parental alcohol misuse (5 or more drinks on 1 occasion in the past month), parental incarceration, and living with a single parent. Physical abuse was defined as more than 2 incidents and emotional neglect as more than 10 incidents according to definitions in a previous Add Health analysis.26
Three self-reported mental health outcomes were examined in wave 4: having ever received a diagnosis of depression, having ever received a diagnosis of anxiety, and screening positive for current depressive symptoms according to the 10-item subscale of the Center for Epidemiologic Studies Depression scale (CES-D–10). A score of 10 or more on the CES-D–10 was considered to be a positive screen result for depression, which corresponds with the clinical cutoff.27
The primary independent variable was participation in team sports during adolescence. Participation was determined based on the answer to the following wave 1 in-school questionnaire item: “Are you participating/do you plan to participate in the following teams (check all that apply): baseball/softball, basketball, cheerleading/dance team, field hockey, football, ice hockey, soccer, swimming, tennis, track, volleyball, wrestling, other sport.”25 Adolescents were designated as participating in team sports if they selected 1 or more sport.
Covariates were selected based on their potential to be associated with both sports participation and mental health, including age, race/ethnicity, sex, parental education, family structure, an indicator for having a gap in health insurance at any wave, and school characteristics (region, public or private, and average daily attendance).
Potential mediators included variables that are associated with changes in psychosocial support (school connectedness, self-esteem, feeling socially accepted, and feeling that friends care)28,29 and possible physiologic changes (level of physical activity over the past week, times exercised hard over the past week).
Univariate analyses examined distributions of the 3 outcome variables, primary regressor, and covariates. Bivariate analyses included χ2 tests and analysis of variance. Separate multivariable logistic regressions examined the association between participating in team sports in adolescence and each of the 3 adult mental health outcomes, controlling for the covariates listed above. All statistical analyses were performed in Stata, version 15.1 (StataCorp). The svy and subpop option was used to adjust for survey design elements including stratification, clustering, and weighting. We included a gender × sports participation interaction term in the models and conducted sex-stratified analyses to explore whether the association between team sports and mental health varies by sex.
To account for children who participate in team sports potentially being inherently different from those who do not, we weighted the analyses for participation in team sports based on propensity scores. Candidate variables for the propensity score–weighted model included factors thought to be associated with participation in extracurricular activities based on a review of the literature; these factors were mapped to variables in Add Health.30 First, associations were examined among potential variables; if the associations were moderate or strong, an index was created when possible, and when not possible, the variable with the strongest bivariate association with team sports participation was used. The components of the propensity score were sex, race/ethnicity, parental education, whether the adolescent played a sport with his or her father in the past month, region, urbanicity, school size, and neighborhood connectedness. The propensity score was developed according to published methods for propensity score–weighted analyses with complex surveys.31 The method involved using the survey weight as a covariate in the propensity score model, calculating the propensity score, assessing the score’s balance across the 2 groups, weighting the 2 groups by the score, and multiplying the propensity score by the survey weight to obtain a new weight. Each outcome was then modeled using svyset in Stata with the new weight.
In sensitivity analyses, we compared the propensity score–weighted analysis with models without propensity score weighting. We assessed for reverse causality by accounting for baseline mental health in wave 1, excluding those with a diagnosis of the relevant condition at baseline when modeling diagnosis of depression or anxiety and controlling for baseline CES-D–10 score when modeling depression symptoms. To examine whether team sports participation might be a proxy for participation in any school-related activities, we modeled whether participation in other activities (such as orchestra, band, chorus, and drama club) was associated with each mental health outcome. We also assessed for a dose-response relationship, with the dose being number of sports in which individuals participated, and examined the association of specific sport types with each mental health outcome.
We tested whether each mediating variable was significantly associated with team sports participation and each mental health outcome and whether the associations between team sports participation and mental health diminished when the variable was included in the model. We used the Karlson-Holm-Breen method to determine the degree to which the mediating variable affected the association between team sports participation and each adult mental health outcome.32 Statistical significance was set at P ≤ .05, and 2-sided tests were used. Missing data represented less than 1% for all variables in our primary analyses. Statistical analysis was performed from November 6, 2017, to January 4, 2019.
Of 9668 individuals included in the study (4470 male [50.0%]; mean [SD] age, 15.2 [1.75] years), 4888 (49.3%) individuals reported 1 or more ACE and 2084 (21.3%) reported 2 or more ACEs. The most common ACE was having a single parent (2670; 27.2%), followed by parental incarceration (1657; 16.7%) and parental alcohol misuse (1177; 13.5%). Sexual abuse was least common (488; 5.2%), followed by physical abuse (925; 9.1%) and emotional neglect (1097; 11.2%). Table 1 shows the descriptive statistics for the study population (those with 1 or more ACEs) and compares it with the Add Health population who did not have ACEs (excluded from the analytic sample). The population affected by ACEs included significantly fewer non-Hispanic white individuals, more black individuals, lower parental education, and higher rates of gaps in health insurance. Individuals exposed to ACEs also had lower levels of participation in team sports and other school-related activities.
At baseline (wave 1), adolescents with ACEs were significantly more likely to have received a diagnosis of depression (126 [2.4%] vs 54 [1.2%], P = .006) and have current depressive symptoms (1372 [26.6%] vs 943 [18.1%], P < .001) compared with adolescents without ACEs (Table 1). Baseline rates of having received a diagnosis of anxiety were not significantly different between the 2 groups (43 [0.9%] vs 24 [0.5%], P = .07). In adulthood, those with a history of ACEs had high rates of all 3 mental health outcomes, with 876 (19.1%) having received a diagnosis of depression, 628 (14.0%) having received a diagnosis of anxiety, and 1132 (24.4%) having current depressive symptoms. The mean (SD) age at diagnosis was 20.5 (5.5) years for depression and 22.6 (4.9) years for anxiety.
Team sports participation during adolescence was associated with significantly lower odds of having ever received a diagnosis of depression (unadjusted rate, 16.8% vs 22.0%; propensity score–weighted adjusted odds ratio [aOR], 0.76; 95% CI, 0.59-0.97) and having ever received a diagnosis of anxiety (unadjusted rate, 11.8% vs 16.8%; propensity score–weighted aOR, 0.70; 95% CI, 0.56-0.89). Team sports participation was not significantly associated with current depressive symptoms after full adjustment (unadjusted rate, 21.9% vs 27.5%; propensity score–weighted aOR, 0.85; 95% CI, 0.71-1.01).
The association between team sports and better mental health was not significantly different for males and females (a team sports participation × gender interaction term was not significant for any of the 3 mental health outcomes) (depression: propensity score–weighted aOR, 0.86 [95% CI, 0.57-1.30]; anxiety: propensity score–weighted aOR, 0.89 [95% CI, 0.55-1.48]; current depressive symptoms: propensity score–weighted aOR, 0.83 [95% CI, 0.59-1.18]). However, sex-stratified analyses revealed a greater number of significant associations between team sports participation and mental health for boys than for girls (Figure). Among boys with ACEs, team sports participation during adolescence remained significantly associated with lower odds of each mental health outcome (depression: propensity score–weighted aOR, 0.67 [95% CI, 0.46-0.99]; anxiety: propensity score–weighted aOR, 0.66 [95% CI, 0.45-0.96]; depressive symptoms: propensity score–weighted aOR, 0.75 [95% CI, 0.56-0.99]), whereas for girls, team sports participation was only significantly associated with lower odds of having ever received a diagnosis of anxiety (propensity score–weighted aOR, 0.73; 95% CI, 0.54-0.98) (Figure).
To assess for reverse associations, we accounted for baseline mental health. The association between participation in team sports and mental health remained significant for anxiety and was nearly significant for depression diagnosis and current depressive symptoms (eTable 1 in the Supplement). To examine whether team sports participation might be associated with participation in any school-related activities, we modeled whether participation in other activities was associated with each mental health outcome. Unlike participation in team sports, participation in other school-based activities was not significantly associated with lower odds of any of the mental health outcomes in adulthood (eTable 2 in the Supplement). We also assessed for a dose-response relationship between number of sports in which individuals participated and each mental health outcome (eTable 3 in the Supplement). Although there was a consistent association for those participating in 1 or 2 sports, there was not consistent evidence of a dose-response relationship. Furthermore, we examined the association between specific sport types and each mental health outcome (eTable 4 in the Supplement). Although limited by cell sizes, the 4 largest sports had negative associations with each mental health outcome, with basketball and football having significant associations. Thus, the results do not seem to be driven by a particular sport.
To better identify the pathways though which team sports participation might improve mental health, we tested whether the associations between sports participation and the study outcomes were mediated by increased psychosocial support, physiologic changes, or both. Candidate mediators included school connectedness, self-esteem, feeling socially accepted, feeling as if friends care, level of physical activity over the past week, and times exercised hard over the past week. We examined each candidate mediator in each model and used the Karlson-Holm-Breen method32 to quantify the proportion of the association between sports participation and mental health that was mediated (Table 2). School connectedness (depression, 14.9%; current depressive symptoms, 16.7%; and anxiety, 10.3%), self-esteem (depression, 11.0%; current depressive symptoms, 29.0%; and anxiety, 7.8%), and feeling socially accepted (depression, 14.6%; current depressive symptoms, 20.6%; and anxiety, 10.0%) significantly mediated the association between team sports and all 3 mental health outcomes; together, they mediated 15.8%-35.8% of the association.
In this analysis of a large nationally representative database, 49% of adolescents experienced 1 or more ACEs, which is similar to rates found in other large-scale analyses.2,3 Children affected by ACEs had poor mental health outcomes, with greater numbers of ACEs being associated with higher frequencies of ever having received a diagnosis of depression or anxiety and depressive symptoms in adulthood. Despite this risk profile, team sports participation during adolescence was significantly associated with improved adult mental health, particularly for boys, suggesting that exposure to protective factors in childhood might be associated with improved long-term mental health in this vulnerable population.
These findings suggest that children affected by ACEs might benefit from participating in team sports and other programs that provide psychosocial support. Given the infrastructure that exists for youth sports in the United States already, augmenting and expanding sports programs may be a relatively achievable and inexpensive strategy to mitigate the association of ACEs with health. Ensuring that such programs maintain access for youths at high risk for exposure to ACEs is an important consideration for policymakers and child health advocates. For example, sports programs that require high fees for participation or that are limited to high-resource neighborhoods might exacerbate, rather than mitigate, health disparities associated with ACEs. This is a growing concern given school budget challenges and the emergence and expansion of pay-to-play policies.33,34
Clinicians might consider specifically encouraging patients with ACEs to participate in team sports or other activities that build social support, particularly those with a low likelihood for physical injury or concussion. The American Academy of Pediatrics13 suggests that pediatricians routinely screen for ACEs. Thus, clinicians may soon be tasked with providing recommendations to families affected by ACEs regarding how best to support the long-term health of their children. Identifying interventions that increase resilience for patients with ACEs may be critical for moving from merely documenting the associations between ACEs and negative health outcomes to preventing and/or improving those outcomes.
Our mediation analysis suggests that active variables associated with resilience building that could be targeted by other interventions include school connectedness, self-esteem, and feeling socially accepted. These same resilience factors might be enhanced by programs at the school- and community-level in line with the Centers for Disease Control and Prevention’s Whole School, Whole Community, Whole Child model.35 Given the population health and intergenerational health consequences of ACEs, identifying practical strategies to build resilience, rigorously evaluating such interventions, and then investing in evidence-based strategies to enhance resilience might yield significant health benefits for a substantial group of children and adults.
This study is limited by the study design and variables available in Add Health. There was some loss to follow-up over the course of this study and incomplete data on certain variables. However, the proportion of missing data among all variables in the primary analyses was less than 1%. The team sports participation survey question asked about current and planned participation, which may not have occurred. This likely biases our results toward the null hypothesis. Similarly, the sensitivity analyses accounting for baseline mental health may be overly conservative because adolescent mental health may be an intermediate outcome. We attempted to assess for dose-response using number of sports because the survey did not ask about intensity of participation, but this is an inadequate proxy. Number of sports often does not indicate intensity of participation. Involvement in 3 or more sports, for instance, may indicate a different nature of involvement (less intensive, less committed), limiting its use in a dose-response analysis. Alternatively, participating in more sports may increase stress owing to the large time commitment and could signal a more fragmented social support network. Regardless, there is little reason to expect a dose-response relationship is this context. In addition, components of the ACE indicator could be subject to recall and social desirability bias because they required recall of events before 18 years of age and address sensitive issues. These biases would tend to lead to underreporting of ACEs. However, it is reassuring that the rates of ACEs found are similar to those reported in other large-scale studies.2,3 Furthermore, although we controlled for contextual factors and conducted propensity score weighting for sports participation, unmeasured differences between adolescents who participate in sports and those who do not may remain and confound results. For example, an unmeasured confounder would need to have a propensity score–weighted aOR of 1.56 (95% CI, 1.21-2.78) for depression, 1.40 (95% CI, 1.00-2.19) for depressive symptoms, and 1.67 (95% CI, 1.51-3.00) for anxiety to account for our findings.
Furthermore, causality could not be determined with our study design. Social determinants of health are difficult to isolate, even with comprehensive modeling. It is impossible to fully capture the social environment of a teenager, and this environment is constantly changing. Although team sports may help mitigate the effects of ACEs for some individuals, other factors likely play a role in mitigating these effects also, such as change in economic status and income level.36
These findings appear to have important implications for pediatricians and parents of children exposed to ACEs. Given that participation in team sports was associated with improved adult mental health among those with ACEs, pediatricians might consider recommending team sports participation for patients with ACEs and parents might consider enrolling their children with ACEs in team sports. Similarly, policymakers and child health advocates may consider advocating to make these programs accessible, equitable, and strong.
Accepted for Publication: January 23, 2019.
Corresponding Author: Molly C. Easterlin, MD, UCLA National Clinician Scholars Program, 10940 Wilshire Blvd, Ste 710, Los Angeles, CA 90024 (email@example.com).
Published Online: May 28, 2019. doi:10.1001/jamapediatrics.2019.1212
Author Contributions: Dr Easterlin had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Easterlin, Chung, Dudovitz.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Easterlin.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: All authors.
Supervision: Chung, Dudovitz.
Conflict of Interest Disclosures: None reported.
Funding/Support: Dr Easterlin is supported by the Cedars-Sinai Medical Center via the UCLA National Clinician Scholars Program in the David Geffen School of Medicine at UCLA.
Role of the Funder/Sponsor: The supporting institution had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Create a personal account or sign in to: