Study sample

From 2006 to 2008, all adolescent residents (age 13–19 years) of the Nord-Trøndelag county in Norway (n=10 464) were invited to participate in the third wave of the population-based Nord-Trøndelag Health Study (Young-HUNT3). The attendees completed a comprehensive health-related questionnaire during school hours. Of those invited, 7716 (74%) adolescents responded to the questionnaire and attended a clinical examination. The response rate was 82% among junior high school students and 69% among senior high school students. Participants not enrolled in school (n=84) and participants ≥20 years of age (n=13) were excluded from the analyses in this study. Thus, the study sample comprised 7619 participants (73%) (online supplementary appendix), of whom 4615 (61%) went to junior high school. All participants in senior high school were ≥16 years old.

Exposure variables

Leisure time PA level was assessed by a validated question on frequency of PA from WHO Health Behavior in Schoolchildren (WHO HBSC) Survey Questionnaire21 22: Outside school hours: ‘How often do you usually exercise in your free time so much that you get out of breath or sweat?’. The level of intensity during exercise where you breathe heavily and/or sweat refers to moderate to vigorous activity. Response alternatives were: every day, 4–6 days/week, 2–3 days/week, 1 day/week, less than every week, less than every month and never. Responses were categorised into three levels of PA: ‘low PA’ (≤1 day/week) (reference group), ‘moderate PA’ (2–3 days/week) and ‘high PA’ (≥4 days/week).

Type and frequency of sport participation was assessed by the question: ‘How often have you participated in the following activities/sports in the last 12 months?’; endurance sports (eg, cross-country skiing, swimming, running), team sports (eg, soccer, volleyball, handball), strength sports (eg, weightlifting, body-building), technical sports (eg, track and field, alpine skiing, snowboard), aesthetic sports (eg, dance, gymnastics), martial arts (eg, judo, karate, boxing), extreme sports (eg, rafting, rock climbing, paragliding), jogging/walking and other. Four alternatives were given for describing the frequency of participation in each of the sport categories: never, less than once a week, once a week, several times a week. A frequency of ‘at least once a week’ was defined as active participation. Furthermore, adolescents were classified by their participation in sports into team sports (eg, soccer, volleyball, handball) or individual sports (all other sports). The reference group consisted of those with no or infrequent participation in all the sport categories, as well as those who reported a low level of PA. The groups were mutually exclusive. Responses to ‘jogging/walking’ were not defined as separate sport activities/participation, as they may also be performed in non-sport contexts. The activity ‘jogging/walking’ was, however, included in all exposure categories; ‘jogging/walking’ at least once a week was reported among 61% of those participating in individual sports, among 71% of those participating in team sports and among 33% of those with no/infrequent sport participation or low PA level.

Outcome variables: mental health

Psychological distress was assessed using a validated short version of the Hopkins Symptom Check List Five-item (SCL-5),23 24 including the phrases: ‘During the last 14 days: I have been constantly afraid and anxious; I have felt tense or uneasy; I have felt hopeless about the future; I have felt dejected or sad; I have worried too much about various things’. Responses were scored according to four response alternatives ranging from ‘not at all bothered’ (1) to ‘extremely bothered’ (4). A mean score was calculated, and a cut-off for symptoms of anxiety and depression was set at a mean score above two.23 The SCL-5 version has shown high correlation with the 25-item SCL-25 (r=0.92)24 and good internal consistency (Cronbach’s alpha 0.87).23

Self-esteem/feelings of self-worth was measured using a short version of the Rosenberg Self-Esteem Scale (RSES) (original 10 items),25 including four statements: ‘I have a positive attitude toward myself’, ‘I feel rather useless at times’, ‘I feel that I don’t have much to be proud of’, and ‘I feel that I am a valuable person, at least equal to other people’. Response alternatives were measured on a 4-point scale ranging from ‘I strongly agree’ (1) to ‘I strongly disagree’ (4). For the first and last items, the scores were inversed. A mean score was calculated (range 4–16), with higher scores indicating higher levels of self-esteem. A cut-off was set at a mean score of 10 (midpoint of the scale) to separate low and high self-esteem, corresponding to the recommended cut-off at 25 on the original 10-item RSES (range 10–40).26 The four-item version of the RSES is found to correlate at 0.95 with the full scale and to explain 0.90% of the full-scale variance, and has good internal consistency (Cronbach’s alpha 0.80).27

Life satisfaction was measured with the question: ‘Thinking about your life at the moment, would you say that you by and large are satisfied with life, or are you mostly dissatisfied?’ Response alternatives were measured on a 7-point scale ranging from ‘very satisfied’ (1) to ‘very dissatisfied’ (7), and were coded into a dichotomous outcome variable where adolescents who responded 1–3 were classified as ‘high life satisfaction’, and those who responded 4–7 were classified as ‘low life satisfaction’. A single-item life satisfaction measure is shown to perform almost as well as the multiple-item Satisfaction with Life Scale (SWLS).28 29

Potential confounders

Data on gender and age were obtained from the Norwegian National Population Registry. Socioeconomic status was based on participants’ reports of perceived family economy. Pubertal development was assessed by self-reported pubertal status using the validated Pubertal Development Scale (PDS).30 Participants were asked to rate their own growth and to assess pubic hair growth. Further, boys were asked to assess changes in voice and facial hair growth, while girls were asked about age at menarche and breast development. Pubertal changes were reported on scales ranging from 1 (has not begun) to 4 (development completed). Menarche was dichotomised into yes (coded 4) and no (coded 1). The items were summed up and an average PDS score was calculated.

Exposure to interpersonal violence was assessed with questions derived from The University of California at Los Angeles Post-traumatic Stress Disorder Reaction Index Reaction Index31: ‘Have you ever experienced any of these events?’ (no or yes): (1) Been subjected to violence (beaten or injured), (2) Been subjected to unpleasant/disagreeable sexual acts by someone approximately your own age, (3) Been subjected to unpleasant/disagreeable sexual acts by an adult and (4) Been threatened or physically harassed by fellow students at school over a period of time. Responses were categorised as ‘prior violence’, ‘prior sexual abuse’ (by peer or adult) and ‘prior bullying’.32

Statistical analyses

Categorical variables were described with counts and percentages, and continuous variables with mean and SD. Possible associations between gender and age distribution were assessed with X² tests.

The associations between (1) PA level and (2) type of sport participation and the mental health outcomes were evaluated using multiple logistic regression models. The results are expressed as ORs with 95% CIs. Analyses were stratified by gender and school level (junior and senior high school). To adjust for potential confounding, age, puberty development (PDS), socioeconomic status, prior violence, sexual abuse (by peer or adult) and bullying were all included in the multiple models. All tests were two sided, and the significance level was set to 5%. Analyses were performed using SPSS V.25 (SPSS).

Due to missing data on pubertal development (PDS score) (13% in girls, 15% in boys), a model-based imputation for this variable was performed. Linear regression models stratified by gender were fitted with age and body mass index. Residuals for both models followed standard normal distribution and the model fit was very good. The missing values were replaced with the predicted values. The predicted variables and the original variables for both genders were highly correlated, r=0.75 and r=0.85 for boys and girls, respectively.

Patient and public involvement

When preparing for the Young-HUNT3 survey, the HUNT research centre appointed reference groups consisting of student representatives (13–19 years), as well as representatives from the county school authorities and county doctors. Content of the questionnaires, clinical examination, implementation of results and protection of privacy have been discussed with these representatives. Results from the current study have been presented and discussed with the Youth Panel for Research at the Research and Communication Unit for Musculoskeletal Health, Oslo University Hospital. The HUNT research centre has an active information policy and publishes annual newsletters for participants as well as regular reports on the HUNT website (https://www.ntnu.no/hunt/om).

Characteristics of the study sample

In total, 3785 boys and 3834 girls were included in the analyses. Characteristics of the study sample are summarised in table 1. The mean age was 15.8 years (SD 1.7). About 10% reported living in families with an economy below average, and reports of exposure to interpersonal violence was higher among senior as compared with junior high school students, with physical violence being more common in boys, sexual violence more common in girls, and bullying more evenly distributed between the sexes. Girls reported more mental health problems than boys, with the highest levels of psychological distress found among girls in high school (table 1).

PA level and sport participation

More boys reported a high level of PA (45.2%) compared with girls (35.4%) (p<0.001). For both genders, junior high school students reported significantly higher levels of PA than senior high school students (table 2). Of the various sports, team sports were most commonly reported by both girls and boys, with about 60% of the adolescents participating in such sports at least once a week (table 2). For both genders, participation in team and technical sports was less common among senior high school students as compared with junior high school students (p<0.001). In contrast, a significantly higher proportion of both girls and boys in senior high school were engaged in strength sports compared with those in junior high school. Significantly, more girls than boys participated in aesthetic sports, while a higher proportion of boys participated in strength sports and extreme sports, in both age groups (table 2).

PA and mental health

A high level of PA, compared with a low level of PA, was significantly associated with reduced odds of low self-esteem and low life satisfaction among all students (table 3). In senior high school students, a high level of PA was significantly associated with reduced odds of psychological distress (OR 0.63, 95% CI 0.46 to 0.86 for girls and OR 0.46, 95% CI 0.27 to 0.79 for boys).

Sport participation and mental health

Participation in team sports, compared with no/infrequent sport participation, was significantly associated with reduced odds of low self-esteem for girls, both in junior high school (OR 0.45, 95% CI 0.32 to 0.64) and senior high school (OR 0.57, 95% CI 0.39 to 0.84). A similar non-significant trend was observed in boys (table 4). Participation in individual sports was significantly associated with reduced odds of low self-esteem among senior high school boys (OR 0.37, 95% CI 0.18 to 0.76).

Participation in team sports was associated with reduced odds of low life satisfaction, among all girls and among junior high school boys (table 4). Among girls in senior high school, team sport participation was also significantly associated with reduced odds of psychological distress (OR 0.70, 95% CI 0.49 to 1.00). In boys and junior high school girls, no statistically significant associations between sport participation and psychological distress were revealed in the adjusted models (table 4).

Strengths and limitations

The main strengths of our study include the large sample size of adolescents from an unselected general population, a high participation rate (73%) and the information including frequency of participation in sports. To our knowledge, such comprehensive information about sport participation in relation to gender and age differences has not previously been presented. This information enabled us to examine the relationship between sport participation and mental health outcomes, including the aspect of various types of sports, in contrast to most studies on this topic which mainly focus on general PA. We used validated measures of mental health outcomes. Another strength is that we were able to adjust for a variety of possible confounders, as well as including exposures to interpersonal violence as these exposures may have an impact on both PA behaviours and mental health.33–35

The main limitation of this study is the cross-sectional study design, and the resultant inability to demonstrate the direction of the association. Further, as measures were self-reported, the included variables are susceptible to information bias. We have used a single item measure to assess PA, and the variable used to describe sport participation exposure provides a crude measure of frequency of participation. However, WHO HBSC question of PA used in this study has been found to hold acceptable reliability and validity in adolescent samples.21 22 It should be noted that a proportion of the ‘high PA’ group in this study may not fulfil the recommended levels of daily PA according to WHO guidelines,36 as this group includes all those who played sports or exercised at least 4 days/week. Although measures of psychological distress (SCL-5) and self-esteem (RSES) were shortened versions of the original instruments, the measurement precision of these versions is found to be high and sufficient for use in population-based studies.23 24 27 Furthermore, dichotomisation of the mental health outcomes makes them prone to misclassification. However, the cut-off values to distinguish those with high versus low degree of psychological distress (SCL-5) and low self-esteem (RSES) have both been shown to be clinically relevant cut-points.23 26 In contrast to psychological distress, low self-esteem and low life satisfaction were not more prevalent in the older age group, reflecting the measurement of different phenomena. Psychological distress is found to function as a proxy measure of symptoms of anxiety and depression,23 24 while self-esteem and life satisfaction are more closely related to subjective well-being,28 37 which may be more stable traits.

Although the response rate in Young-HUNT3 was high, the lower response among senior high school students compared with junior high school students may represent a selection bias. In Norway, most adolescents start senior high school the year they turn 16. Differences in school systems in other countries, as well as differences related to opportunities for engagement in sports and the organisation of various youth sports may limit the generalisability of the results from this study. Norwegian society is rooted in egalitarian ideals, with ‘Sport for All’ as a high priority and policy aim38; this may be part of the reason why sport participation found in this study is high by international standards. Lastly, the results of this study should be interpreted with caution due to multiple testing, and replication of results is warranted.

Our results according to previous findings and interpretation of findings

The findings from this study confirmed the gender differences in PA levels found in other population-based studies, with girls being less physically active than boys.11–13 Also in line with reports from other studies, fewer girls were involved in sports.7 13 14 39 The lower levels of PA and sport participation found among senior high school students confirm findings from previous studies reporting a decline in PA during adolescence,10 11 40 41 especially between the ages of 15 and 16 years.11 Drop-out from sports for adolescents at this age may be related, in part, to the difficulty of meeting the increasing demands of participation in both school and sport.7 42 However, a systematic review of factors associated with drop-out from organised sports reported that lack of enjoyment and perceptions of competence are the two most dominant factors related to drop-out from sports among adolescents.42

The current findings are in line with previous studies reporting associations between adolescents’ PA and mental health, including lower likelihood of depressive symptoms,1 11 13 43 44 as well as greater well-being13 and higher self-esteem among those who are physically active.1 44 Longitudinal studies also indicate that PA may protect against the development of depression.45 46 Furthermore, PA may be a helpful intervention for adolescents struggling with depressive symptoms.47 48 Regarding engagement in sports, the results of the present study are in accordance with the literature indicating a positive relationship between sport participation and mental health.9 13 49–51

The findings from this study make several contributions to the body of research on the impact of PA and sports on mental health among adolescents. First, age differences were revealed in these relationships, as a high PA level was associated with reduced psychological distress in high school students only. Explanations for why PA and sport participation may be of greater importance in reducing psychological distress among older adolescents could relate to how peer support and interaction play an increasingly important role during adolescence.19 20 Thus, social and PAs with peers may be particularly beneficial for older adolescents, helping to distract them from depressive thoughts and to reduce the sense of isolation.

Second, this study shows that the potential beneficial effects related to mental health vary across various types of sport. Participation in team sports, compared with no participation, was more strongly related to beneficial mental health outcomes than individual sports compared with no participation, especially in high school girls. Other studies have reported that organised sports had a greater impact on depressive symptoms among girls than boys,11 and that team sports may confer mental health benefits for girls in particular.13 A systematic review of the psychological and social benefits of sport participation argued that team sports seem to be associated with more beneficial outcomes compared with individual sports due to the social aspect of being part of a team.9 In a recent study of a large representative sample of European adolescents, lower levels of anxiety and depression and higher levels of well-being were found among team sport participants.13 Team sport participation during adolescence has also been shown to be associated with lower levels of depressive symptoms in early adulthood.50 52 In line with this, our findings highlight that the type of sport, including the social aspect of participation, should be considered when examining the impact of sports on mental health among adolescents.

Overall, adolescents with a high PA and team sport participation had lower odds of having low self-esteem than of having psychological distress. These results are in accordance with findings from a review study showing that, of the several mental health outcomes associated with PA in young people, the strongest association was with self-esteem.1 It may be that experiences of low self-esteem occur earlier than symptoms of anxiety and depression and that low self-esteem may trigger poor coping behaviour and risk behaviour that subsequently increases the likelihood of mental disorders.53 A Cochrane review reports that exercise interventions have positive short-term effects on self-esteem that may help prevent the development of psychological and behavioural problems.54

This study contributes to our understanding of the implications lower levels of PA may have on mental health in different phases of adolescence. Regardless of the direction of the association between PA and mental health, inactive adolescents may carry a ‘double health burden’, with both physical and mental health challenges. To help more adolescents increase or maintain their levels of PA, interventions could include facilitation of a wider variety of sports activities, and at different skill levels, to reach and engage more adolescents. As lack of enjoyment has previously been found to be the most dominant factor related to drop-out from sports42, increased focus on the joy of sports may be important in order to reduce drop-out rates during adolescence. In efforts to reduce mental health problems, our results suggest that girls in particular should be a target group for promotion of team sport participation.