Psychosomatic Medicine Faster Service from Outside North America
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowRequest Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Motl, R. W.
Right arrow Articles by Dishman, R. K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Motl, R. W.
Right arrow Articles by Dishman, R. K.
Related Collections
Right arrow Exercise
Right arrow Depression
Psychosomatic Medicine 66:336-342 (2004)
© 2004 American Psychosomatic Society

ORIGINAL ARTICLES

Naturally Occurring Changes in Physical Activity Are Inversely Related to Depressive Symptoms During Early Adolescence

Robert W. Motl, PhD, Amanda S. Birnbaum, PhD, Martha Y. Kubik, PhD and Rod K. Dishman, PhD

From the Department of Exercise Science, The University of Georgia, Athens, Georgia (R.W.M., R.K.D.); Department of Public Health, Cornell University, Ithaca, New York (A.S.B); and School of Nursing, University of Minnesota, Minneapolis, Minnesota (M.Y.K.).

Address correspondence and reprint requests to Rod K. Dishman, PhD, Department of Exercise Science, University of Georgia, 300 River Road, Athens, Georgia 30602-6554. E-mail: rdishman{at}coe.uga.edu


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
OBJECTIVE: We examined the relationship between naturally occurring changes in physical activity and depressive symptoms across a 2-year period among adolescent boys and girls.

METHODS: Participants (N = 4594) reported their frequency of physical activity outside of school and completed the Center for Epidemiological Studies Depression scale in the Fall of 1998 (beginning of 7th grade; baseline data), Spring of 1999 (end of 7th grade; interim data), and Spring of 2000 (end of 8th grade; follow-up data).

RESULTS: Latent growth modeling indicated that a 1 SD unit change in the frequency of leisure-time physical activity was inversely related to a .25 SD unit change in depressive symptoms. This relationship was attenuated but remained statistically significant when simultaneously controlling for the confounding variables of sex, socioeconomic status, smoking, alcohol consumption, and the value participants placed on their health, appearance, and achievement.

CONCLUSIONS: Naturally occurring changes in physical activity were negatively related with changes in depressive symptoms. The results encourage randomized controlled trials to experimentally determine whether an increase in physical activity reduces depression risk among adolescent boys and girls.

Key Words: adolescents, • CES-D, • depression, • exercise.

Abbreviations: CES-D = Center for Epidemiological Studies Depression scale;; CFI = Comparative Fit Index;; LGM = latent growth modeling;; NNFI = Non-Normed Fit Index;; RMSEA = root mean square error of approximation;; TEENS = Teens Eating for Energy and Nutrition at School.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The World Health Organization has projected that by the year 2020 depression will be second only to cardiovascular disease as the world’s leading cause of death and disability (1). Depressive disorders now occur earlier in life than in the past (2), and the annual rate of depression among teenagers and young adults in the United States is nearly twice that of adults aged 25 to 44 years (3,4). An estimated 8 to 9% of early adolescents in the US suffer from depression (5,6). Boys and girls have equal rates of depressive disorders during childhood, but girls are twice as likely as boys to develop depression during adolescence (5). Depressive disorders among youth commonly persist, recur, and continue into adulthood and predict more severe depressive illness in the adult years (7). Without successful treatment, depressed adolescents are at an elevated risk for academic failure, social isolation, promiscuity, drug and alcohol abuse, and suicide (7–9).

There is growing, but still limited, evidence from population-based cohort studies (10) and randomized controlled trials (10–13) with adults that exercise is associated with a reduction of depressive symptoms. Much less is known about the potential efficacy of physical activity for the primary and secondary prevention of depression among adolescents. We know of 8 correlational studies with clinical or convenience samples (14–21) and 2 population-based studies (22, 23) that reported that low levels of physical activity or sport participation were negatively associated with depressive symptoms among adolescents. Another population-based study reported higher emotional well-being among physically active youth, independent of social class and health status (24).

The aforementioned studies did not include a longitudinal design that is necessary for examining the association between naturally occurring changes in physical activity and depression across time. Moreover, the studies did not adjust for other risk factors, such as sex, socioeconomic status, smoking, alcohol consumption, and the value participants place on their health, appearance, and achievement that might confound the association between physical activity and depressive symptoms (25–29). Most studies did not use a measure of depressive symptoms that has established validity for depression screening among adolescents. Hence, the absence of studies using longitudinal designs and that control for other risk factors and include a valid measure of depression prevent a conclusion about the population effectiveness of physical activity for reducing depressive symptoms among adolescents.

The present study examined the relationship between naturally occurring changes in leisure-time physical activity and depressive symptoms across a 2-year period among a cohort of participants in the Teens Eating for Energy and Nutrition at School (TEENS) study (30). We initially established the pattern of longitudinal change in leisure-time physical activity and depressive symptoms. We then examined the relationship between changes in leisure-time physical activity and depressive symptoms. Subsequent analyses examined the relationship between changes in physical activity and depressive symptoms simultaneously controlling for the confounding risk factors of sex, socioeconomic status, smoking, alcohol consumption, and the value participants placed on their health, appearance, and achievement. Those variables have been related to both physical activity (27,28) and depressive symptoms (25,26,29).


   METHODS
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Data
Analyses were conducted using data from the TEENS study (30). The TEENS study was a school-based, group-randomized trial designed to alter cancer-related dietary risk behaviors in young adolescents in the Twin Cities, Minnesota. The TEENS study involved the design and evaluation of school-based environmental, classroom, and family interventions to increase fruit and vegetable intake and reduce fat intake in 7th and 8th grade students; the interventions did not target physical activity, sedentary leisure habits, or depression. Though the primary purpose of the TEENS study involved school-based environmental, classroom, and family interventions, the secondary purposes involved an examination of the correlates of change in depression, physical activity, sedentary leisure habits, and other behaviors (28).

School districts located within a 30-mile radius of the St. Paul-Minneapolis, MN, metropolitan area and with a minimum of 20% of students approved for the free/reduced lunch program were eligible to participate in the TEENS study. Schools were required to have both 7th and 8th grades in 1 building and enroll at least 30 students in each of those grades. Fourteen districts (33 schools) were eligible and 9 districts (20 schools) agreed to participate. One of the twenty schools was chosen for pilot evaluation, and 3 were judged ineligible because of scheduling conflicts. The remaining 16 schools formed the school sample for the study. The schools were randomized within pairs matched on enrollment and participation in free and reduced-price lunch programs to intervention and comparison (delayed intervention) conditions.

Data were collected on 3 occasions over a 2-year period. The 3 occasions were Fall of 1998 (beginning of 7th grade; baseline data), Spring of 1999 (end of 7th grade; interim data), and Spring of 2000 (end of 8th grade; follow-up data). The sample comprised 4594 adolescents who provided data on 1 or more of the 3 occasions. There were 3878 adolescents who completed the survey in 1998; 3798 students in 1999; and 3588 students in 2000. The students were balanced between girls (49%) and boys (51%) and primarily were White (62.6%), with additional racial categories of African American (13.8%), Hispanic (3.8%), Asian (7.2%), Native American (1.7%), Multiethnic (6.5%), and other (4.4%). The mean age of the sample at baseline was 12.7 (SD = 0.4) years. The baseline, interim, and follow-up prevalence of elevated depression symptoms based on a Center for Epidemiological Studies Depression Scale (CES-D) cutoff score of 22 (31) was 16.4%, 17.3%, and 19.7%, respectively. The baseline, interim, and follow-up prevalence of elevated depression symptoms based on a CES-D cutoff score of 24 (32) was 12.8%, 14.3%, and 16.5%, respectively. At baseline, approximately 77.9% of the sample reported at least some regular participation in physical activity outside of school, 2.9% reported smoking 1 or more cigarettes during the past 30 days, and 15.2% reported consuming alcohol 1 or more times during the past 30 days.

Measures
Physical activity was measured by a single-item measure of the frequency of regular physical activity outside of school that has been sensitive to detecting intervention effects (33) and effective in tracking of physical activity among youth (34). The item, "Do you get some regular physical activity outside of school? By regular we mean at least 3 times a week for at least 20 minutes at a time." was rated using a 5-point scale with verbal anchors of Most of the time (1), Usually (2), Once in a while (3), Hardly ever (4), and Never (5). The item was reverse-scored before the analyses; higher scores reflected more frequent physical activity outside of school. The single-item measure of physical activity had a test-retest reliability of 0.69 across the 3 time periods. Though other reports on the TEENS data have used a measure of physical activity that combined frequency of physical activity with the intensity of physical activity (28), we selected only the frequency item. Scores from the 2 items were unrelated at the baseline (r = –0.06), interim (r = 0.01), and follow-up measurements (r = 0.02). Hence, the 2 items do not appear to be similarly assessing physical activity, and there is no empirical basis for combining them into a single measure of physical activity. Because of uncertainty about the validity of self-ratings of physical activity intensity by adolescents (35), we believe the frequency question provides the more interpretable measure of physical activity for the purposes of the present analysis.

Depressive symptoms were measured using the 20 item CES-D (36), which has established predictive validity for the screening of adolescent depression (31,32). The 20 items were rated on the basis of frequency of occurrence during the past week using a 4-point scale. The verbal anchors were Rarely or none of the time (less than 1 day), Some or a little of the time (1–2 days), Occasionally or a moderate amount of time (3–4 days), and Most or all of the time (5–7 days). The positively worded items were reverse-scored, and the individual item scores before the analyses were summed to generate a single composite score. Higher CES-D scores reflected elevated depressive symptoms. The summation and comparison of the composite scores across time were supported by previous analyses that confirmed the factorial validity and multi-group and longitudinal invariance of the CES-D, a recommended precursor to the analysis of change over time (37).

Smoking behavior and alcohol consumption were measured by single-item measures. The smoking item, "How frequently have you smoked cigarettes during the past 30 days?" was rated using a 7-point scale with verbal anchors of Not at all (1), Less than 1 cigarette per day (2), 1 to 5 cigarettes per day (3), About .5 pack per day (4), About 1 pack per day (5), About 1.5 packs per day (6), and 2 packs or more per day (7). The alcohol consumption item, "During the last 30 days, how many times have you had alcohol to drink (including beer, wine, and liquor)?" was rated using a 7-point scale with response options of 0 (1), 1 to 2 (2), 3 to 5 (3), 6 to 9 (4), 10 to 19 (5), 20 to 39 (6), and 40 or more (7).

Socioeconomic status was measured as a trichotomous index (28). The index was based on combining 4 variables: participation in free or reduced-price lunch meal program at school, the highest level of education for mother and father, the number of parents the student reported living with, and number of parents who worked full-time (28).

The value participants placed on their health, appearance, and achievement was measured using a 7-item scale (28). An example item is "How well I do in school is very important to me." The 7 items were rated on a 5-point Likert type scale with verbal anchors of Strongly disagree (1) and Strongly agree (5). The items were summed to form a single composite score before the analyses. Higher scores reflected greater value placed on health, appearance, and achievement.

Latent Growth Modeling
We used latent growth modeling (LGM; 37–39) to examine: 1) the patterns of change in physical activity and depressive symptoms separately; 2) the consequence of a change in physical activity on the change in depressive symptoms; and 3) possible confounding influences on the relationship between physical activity and depressive symptoms. The LGM analyses were performed using AMOS 4.0 with full-information maximum likelihood estimation (40), which is an appropriate and optimal method for the treatment of missing data in covariance modeling (40,41), particularly in applications of longitudinal modeling with missing data (42).

Conceptually, LGM is a 2-stage process that invokes a confirmatory factor analytic framework on variables measured longitudinally. In the first stage, individual-level growth models or trajectories are fit to represent change on measures of the same construct obtained on multiple measurement occasions. This within-individual stage models aspects of intraindividual change including the sample mean change trajectory and within sample variability in the mean change trajectory. The second stage, or the between-individual stage, involves an examination of additional variables as consequences and predictors of longitudinal growth trajectories. For example, the effect of a change in physical activity on a change in depressive symptoms can be examined in the second stage of LGM. LGM has a number of advantages over other more commonly adopted approaches used to study change among continuous variables (eg, ANOVA, MANOVA, lagged regression, use of change scores), including the ability to 1) model change at the individual as well as the group-level of analysis; 2) model individual differences in change trajectories (initial status and change functions); 3) model various functional forms of change (eg, linear, quadratic, optimal); 4) model change in several focal variables concomitantly; and 5) directly model important predictors and outcomes of longitudinal change (37–39).

Model Specification
The first stage of LGM for examining the pattern of change on the measure of physical activity is illustrated in Figure 1; this model is identical for the measure of depressive symptoms. Y1, Y2, and Y3 represent measures of physical activity obtained on 3 occasions; {eta}1 and {eta}2 are latent variables (common factors); and the {epsilon}s and {zeta}s are residuals for the Ys and the {eta}s, respectively. The factor loadings of the Ys on the latent {eta}s are presented as fixed parameters: 1) the loadings of the Ys on {eta}1 are fixed equal to 1.0 so that {eta}1 represents initial status on the measure of physical activity measured by the Ys; and 2) factor loadings of the Ys on {eta}2 are fixed equal to 0.0, 1.0, and 2.0 so that {eta}2 represents linear change on the measure of physical activity measured by the Ys. Thus, {eta}1 and {eta}2 together represent the intercept (initial status) and slope (change) components of the linear change function for physical activity measured by the Ys over time. The factor covariance or {psi}12 parameter accounts for relationship between the initial status and change factors. This simple model can be expanded to test for heteroscedastic (ie, {epsilon}11 != {epsilon}22 != {epsilon}33) and homoscedastic (ie, {epsilon}11 = {epsilon}22 = {epsilon}33) residuals, and linear (ie, 0.0, 1.0, 2.0), quadratic (ie, 0.0, 1.0, 4.0), or optimal (ie, 0.0, 1.0, freely estimated) change functions.



View larger version (19K):
[in this window]
[in a new window]
  		Figure 1. Model illustrating the first-stage of latent growth modeling for physical activity and depressive symptoms. {eta}1 = initial status. {eta}2 = change function. Y1 - Y3 = 1 variable measured on 3 occasions. {epsilon}11, {epsilon}22, and {epsilon}33 = item uniquenesses. {zeta}11 and {zeta}22 = factor variances. {psi}12 = factor covariance.

 
The second stage of LGM for examining the relationship between measures of physical activity and depressive symptoms is presented in Figure 2. This model includes the first stage of LGM plus the addition of paths (ßs) between the initial status latent variables and between the change latent variables. In this model, Y1, Y2, Y3, Y4, Y5, and Y6 represent measures of physical activity and depressive symptoms obtained on 3 occasions; {eta}1, {eta}2, {eta}3, and {eta}4 are latent variables (common factors); and the {epsilon}s and {zeta}s represent residuals for the Ys and the {eta}s, respectively. The latent variables {eta}1 and {eta}3 represent the intercept (initial status) components for physical activity and depressive symptoms measured by the Ys over time, respectively. The latent variables {eta}2 and {eta}4 represent the slope (change) components of the linear change function for physical activity and depressive symptoms measured by the Ys over time, respectively.



View larger version (26K):
[in this window]
[in a new window]
  		Figure 2. Model illustrating the second-stage of latent growth modeling for physical activity and depressive symptoms. {eta}1 and {eta}3 = initial status. {eta}2 and {eta}4 = change function. Y1 – Y6 = two variables measured on 3 occasions. {epsilon}11 – {epsilon}66 = item uniquenesses. {zeta}11 – {zeta}44 = factor variances. {psi}12 and {psi}34 = factor covariances.

 
Within the second stage of LGM, ß31 represents the relationship between initial status for physical activity and initial status for depressive symptoms. The ß31 parameter is interpreted as cross-sectional relationship between physical activity and depressive symptoms. The ß42 parameter represents the relationship between change in physical activity and change in depressive symptoms. The parameter is interpreted as a functional change in the relationship between physical activity and depressive symptoms over time. When confounding variables such as sex are included in the second-stage of LGM, the effect is similar to a group by time interaction that represents the degree to which sex-related differences in physical activity or depressive symptoms vary across time.

Model Fit
The chi-square statistic and subjective indices were used to evaluate and compare the fit of the models in LGM analyses. The chi-square statistic assessed absolute fit of the model to the data, but it is sensitive to sample size and assumes the correct model (43,44). Hence, no restrictive model with positive degrees of freedom is able to fit real data, and such models will often be rejected by a formal test of significance with a sufficiently large sample size (43,44). Subjective indices of fit also were employed to judge and compare the fit of the models. The root mean square error of approximation (RMSEA) represents closeness of fit (44). The RMSEA value should approximate or be less that 0.05 to demonstrate close fit of the model (44). The 90% confidence interval (CI) around the RMSEA point estimate should contain 0.05 to indicate the possibility of close model-data fit (44). Both the NNFI and CFI are incremental fit indices that test the proportionate improvement in fit by comparing the target model to a baseline model with no correlations among observed variables (45,46). NNFI and CFI values approximating 0.95 were indicative of good model-data fit (47).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Descriptive Statistics for Physical Activity and Depressive Symptoms
The mean scores and standard deviations for the measures of physical activity and depressive symptoms across the 3 time points for the overall sample, and boys and girls separately are provided in Table 1. The correlation coefficients among the measures of physical activity and depressive symptoms across the 3 time points for the overall sample are provided in Table 2.


View this table:
[in this window]
[in a new window]
  		TABLE 1. Descriptive Statistics for the Measures of Physical Activity and Depressive Symptoms Across Time for the Overall Sample and Males and Females Separately
 

View this table:
[in this window]
[in a new window]
  		TABLE 2. Correlation Coefficients Among the Measures of Physical Activity and Depressive Symptoms Across the 3 Time Points for the Overall Sample
 
Estimating Growth in Physical Activity
The growth in physical activity was best described by an optimal growth function with a homoscedastic residual structure. This model provided a perfect fit for the physical activity data (df = 2; {chi}2 = 4.77; p = .09; RMSEA [90% CI] = 0.017 [0.000–0.038]; NNFI = 0.998; CFI = 0.998). The mean scores provided in Table 1 illustrate the curvilinear change in physical activity across time. There was a slight initial increase in physical activity at the interim measurement, followed by a lack of change in physical activity at the follow-up measurement.

Estimating Growth in Depressive Symptoms
The growth in depressive symptoms was best described by a quadratic growth function with a heteroscedastic residual structure. This model provided a perfect fit for the depressive symptoms data (df = 1; {chi}2 = 0.01; p = .94; RMSEA [90% CI] = 0.000 [0.000–0.014]; NNFI = 1.001; CFI = 1.000). The mean scores provided in Table 1 illustrate the quadratic change in depressive symptoms across time. There was a slight initial increase in depressive symptoms at the interim measurement, followed by a larger increase in depressive symptoms at the follow-up measurement.

Relationship Between Growth in Physical Activity and Depressive Symptoms
The previous analysis provided a good representation of growth in physical activity and depressive symptoms across time. We next tested a model in which initial status and change factors for physical activity were related with initial status and change factors for depressive symptoms. This model provided an excellent fit to the data (df = 10; {chi}2 = 26.14; p = .004; RMSEA [90% CI] = 0.019 [0.010–0.028]; NNFI = 0.995; CFI = 0.997). There were statistically significant and negative relationships between initial status factors (completely standardized ß31 = –0.31) and change factors (completely standardized ß42 = –0.27) for physical activity and depressive symptoms. Thus, initially higher levels of physical activity were associated with initially lower levels of depressive symptoms, and a change in physical activity across time was inversely associated with a change in levels of depressive symptoms across time.

Confounding Influences
We then tested a model that simultaneously controlled for sex, socioeconomic status, smoking, alcohol consumption, and the value participants placed on their health, appearance, and achievement as confounding influences of the relationship between physical activity and depressive symptoms. Sex was included as a correlate of both initial status and change in physical activity and depressive symptoms; initial status and change in socioeconomic status, smoking, alcohol consumption, and the value participants placed on their health, appearance, and achievement were included as correlates of initial status and change, respectively, in physical activity and depressive symptoms. This model provided an excellent fit to the data (df = 117; {chi}2 = 383.56; p<.001; RMSEA [90% CI] = 0.022 [0.020–0.025]; NNFI = 0.977; CFI = 0.984). Though the path coefficients were attenuated, there were still statistically significant and negative relationships between initial status factors (completely standardized ß31 = –0.22) and change factors (completely standardized ß42 = –0.17) for physical activity and depressive symptoms. Thus, initially higher levels of physical activity were associated with initially lower levels of depressive symptoms, and a change in physical activity across time was inversely associated with a change in levels of depressive symptoms across time, even after controlling for the confounding variables. The path coefficients for the relationships of the confounding variables with initial status and change in physical activity and depressive symptoms are provided in Table 3. We tested 1 final model that included the confounders and treatment group, and though the model provided an excellent fit (df = 123; {chi}2 = 390.67; p<.001; RMSEA [90% CI] = 0.022 [0.019–0.024]; NNFI = 0.975; CFI = 0.984) treatment group did not influence the magnitude of the relationships between physical activity and depressive symptoms.


View this table:
[in this window]
[in a new window]
  		TABLE 3. Standardized Path Coefficients for the Relationships of the Confounding Variables with Initial Status and Change in Physical Activity and Depressive Symptoms
 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
To our knowledge, this is the first evidence showing that a change in physical activity is inversely related to a change in depressive symptoms across a 2-year period among community-residing adolescent boys and girls. The primary novel finding was that a 1 SD unit change in the frequency of leisure-time physical activity was inversely related to a .25 SD unit change in depressive symptoms. The magnitude of the relationship between the change in physical activity and the change in depressive symptoms was reduced by approximately one third when simultaneously controlling for the confounding variables of sex, socioeconomic status, smoking, alcohol consumption, and the value participants placed on their health, appearance, and achievement. Our analysis extends beyond previous reports of a cross-sectional association of physical activity and depression symptoms (48). Although we have adopted the perspective that change in physical activity preceded change in depression symptoms, it is not possible to infer the direction of causality from our analysis. It is also possible that increases in depression symptoms contribute to the decline in physical activity observed during adolescence.

The relationship between changes in physical activity and depressive symptoms, though statistically significant, was relatively small in magnitude. The clinical importance of this effect can be partially judged as a binomial effect size (49). The binomial effect size in this observational study approximates an outcome of 4% above a control rate, hypothetically benefitting 135 children in the present sample. Though not directly comparable to our results, a recent 10-week randomized controlled trial of the SSRI sertraline demonstrated a binomial reduction of 8% in physician-rated depression symptoms among 199 adolescents diagnosed with major depression (50). Hence the clinical importance of our results appears small relative to the effect of pharmacotherapy. Nonetheless, we think the present findings are sufficiently strong to encourage a long-term randomized controlled trial of whether physical activity reduces the primary risk of adolescent depression.

When examining growth in physical activity and depression, we observed an optimal change function for physical activity and a quadratic change function for depressive symptoms. The optimal change function was characterized by an initial increase in physical activity followed by a plateau or lack of further changes in physical activity. The magnitude of change in physical activity was small and reflected the relative stability of self-reported leisure-time physical activity among the sample of adolescents. That stable pattern of leisure-time physical activity is consistent with previous research. For example, regular vigorous physical activity defined as 3 or more days per week of running, jogging, or swimming is relatively stable among 12-, 13-, and 14-year-old boys and girls; there is a similar pattern for the prevalence of physical inactivity defined as no participation in vigorous or moderate physical activity (51). There was a quadratic change in depressive symptoms across time. That pattern of change is consistent with previous research among adolescent boys and girls. For example, the onset of depression occurs in early adolescence, and the prevalence of depression continues to increase during late adolescence (26,52). The relative stability of physical activity, combined with the accelerated prevalence of depression during early adolescence, underscore the need for an early intervention to prevent a reduction in physical activity and an increase in depressive symptoms.

Previous researchers have reported an inverse relationship between physical activity and depression among adolescents using data collected with a cross-sectional research design (14–21). In the present study, initial levels of physical activity were negatively related with initial depressive symptoms. This relationship was attenuated by approximately one third when simultaneously controlling for the confounding variables of sex, socioeconomic status, smoking, alcohol consumption, and the value participants placed on their health, appearance, and achievement. Thus, cross-sectional analyses have consistently supported a relationship between physical activity and depression. We now provide stronger evidence for a relationship between changes in physical activity and depressive symptoms across a 2-year period among adolescents.

The inverse relationship between physical activity and depressive symptoms has implications for understanding the primary and secondary prevention of depressive symptoms in adolescents. Though drugs are a common and safe modality for the short-term treatment of adolescent depression (53), medications have unwanted side-effects, and the long-term efficacy and safety of antidepressants have not yet been confirmed by large-scale randomized controlled trials. The most effective form of psychotherapy for treating adolescent depression, cognitive behavior therapy, has an efficacy rate of only 65% (54). Because of the incomplete effectiveness of psychotherapy and the as-yet uncertain safety and efficacy of drug therapy for the long-term treatment of adolescent depression, continued investigation of efficacious, low-risk interventions that can reduce the risk of adolescent depression is warranted. Two quasi-experimental studies with small samples of adolescents have reported small reductions in depressive symptoms after exercise training (20,55). Moreover, recent evidence from a randomized controlled trial indicates that aerobic exercise, antidepressants, or combined exercise and antidepressants had similar effects of reducing depression scores among older patients with major depression after 16-weeks of treatment (56), though medication only resulted in a faster initial response than aerobic exercise alone or in combination with medication in patients with mild symptoms of depression patients (56). Another randomized controlled trial indicates that supervised group exercise, cognitive psychotherapy, or combined exercise and psychotherapy had similar effects of reducing depression scores among adults with mild to moderate depression after 10-weeks of treatment (57). Whether similar effects will be confirmed by a randomized controlled trial or are generalizable to a reduction in the primary risk of developing depression among adolescents has not yet been established. Future studies, therefore, should examine the possible independent and interactive effects of physical activity with pharmacotherapy and psychotherapy among adolescents undergoing treatment for depression.

The present study examined the relationship between naturally occurring changes in physical activity and depression among a large, community-residing sample of adolescent boys and girls. Hence, this was an observational study, not an intervention study. Though the data were collected as part of the TEENS study, the TEENS intervention did not target physical activity, sedentary leisure habits, or depressive symptoms. Moreover, the present results were not affected by whether the participants were in the TEENS intervention group. Although we observed that naturally occurring changes in physical activity were inversely associated with depressive symptoms, we have limited information about the causes of the changes in physical activity or depressive symptoms. The correlates of change in physical activity observed in this study included change in smoking behavior and the value participants placed on their health, appearance, and achievement. The correlates of change in depressive symptoms included change in socioeconomic status, smoking behavior, and alcohol consumption. Future research should identify additional factors that influence natural history changes in physical activity and depression, which will provide a better basis for interventions aimed at increasing physical activity and reducing depression among adolescents.

There were several limitations of this study. As previously noted, the design of the present study did not permit an examination of the causal relationship between physical activity and depressive symptoms. Hence, we are unable to specify whether changes in physical activity are a cause or consequence of changes in depressive symptoms. This only can be evaluated by a randomized controlled trial. Another limitation of this study was the single item measure of physical activity. The single-item measure of physical activity has, however, been sensitive to detecting intervention effects (33) and effective in tracking of physical activity among youth (34). The lack of inclusion of race as a confounder is another limitation. Our sample was racially heterogenous but predominantly White (62.6%), with additional racial categories of African American (13.8%), Hispanic (3.8%), Asian (7.2%), Native American (1.7%), Multiethnic (6.5%), and other (4.4%). The predominance of White adolescents and relatively small samples of other races precluded analysis of racial as a confounding variable, as such an analysis would be heavily biased by the large portion of the sample that is White.

In summary, change in self-reported frequency of leisure-time physical activity was inversely associated with a change in depressive symptoms. The association of physical activity with depressive symptoms was attenuated, but still statistically significant, after controlling for the confounding variables of sex, socioeconomic status, smoking behavior, alcohol consumption, and the value participants placed on their health, appearance, and achievement. Because of the elevated prevalence of depression during adolescence, the incomplete effectiveness of psychotherapy, and the uncertain safety and efficacy of drug therapy for treating adolescent depression, it is important to continue to investigate the efficacy of low-risk interventions for reducing depression symptoms, such as physical activity, that may be more acceptable to youth and families. Early adolescence may be an especially critical time for such intervention, as physical activity begins to decrease, while depression rates continue to increase, during later adolescence.

Received for publication June 17, 2003.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Murray CJL, Lopez AD. Evidence-based health policy – lessons from the Global Burden of Disease Study. Science 1996; 274: 740–3.[Free Full Text]
  2. Cross-National Collaborative Group. The changing rate of major depression. JAMA 1992; 268: 3098–105.[Abstract/Free Full Text]
  3. Kessler RC, McGonagle KA, Zhao S, Nelson CB, Hughes M, Eshleman S, Wittchen H, Kendler KS. Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States: Results from the National Comorbidity Survey. Arch Gen Psychiatry 1994; 51: 8–18.[Abstract/Free Full Text]
  4. Kessler RC, Walters EE. Epidemiology of DSM-III-R major depression and minor depression among adolescents and young adults in the National Comorbidity Survey. Depress Anxiety 1998; 7: 3–14.[CrossRef][Medline]
  5. Birmaher B, Ryan ND, Williamson DE, Brent DA, Kaufman J, Dahl RE, Perel J, Nelson B. Childhood and adolescent depression: A review of the past 10 years. Part I. J Am Acad Child Adolesc Psychiatry 1996; 35: 1427–39.[CrossRef][Medline]
  6. Rushton JL, Forcier M, Schectman RM. Epidemiology of depressive symptoms in the National Longitudinal Study of Adolescent Health. J Am Acad Child Adolesc Psychiatry 2002; 41: 199–205.[CrossRef][Medline]
  7. Weissman MM, Wolk S, Goldstein RB, Moreau B, Adams P, Greenwald S, Klier CM, Ryan ND, Dahl RE, Wickramaratne P. Depressed adolescents grown up. JAMA 1999; 281: 1701–13.[Abstract/Free Full Text]
  8. Ryan ND, Puig-Antich J, Ambrosini P, Rabinovich H, Robinson D, Nelson B, Iyengar S, Twomey J. The clinical picture of major depression in children and adolescents. Arch Gen Psychiatry 1987; 44: 854–61.[Abstract/Free Full Text]
  9. Shaffer, D., Gould, M. S., Fisher, P., Trautman, P., Moreau, D., Kleinman, M., Flory M. Psychiatric diagnosis in child and adolescent suicide. Arch Gen Psychiatry 1996; 53: 339–48.[Abstract/Free Full Text]
  10. Dunn AL, Trivedi MH, O’Neal HA. Physical activity dose-response effects on outcomes of depression and anxiety. Med Sci Sports Exerc 2001; 33: S587–97.[CrossRef][Medline]
  11. Lawlor DA, Hopker SW. The effectiveness of exercise as an intervention in the management of depression: Systematic review and meta-regression analysis of randomised controlled trials. Brit Med J 2001; 322: 1–8.[Abstract/Free Full Text]
  12. Moore KA, Blumenthal JA. Exercise training as an alternative treatment for depression among older adults. Altern Ther Health Med 1998; 4: 48–56.[Medline]
  13. Singh NA, Clements KM, Singh MA. The efficacy of exercise as a long-term antidepressant in elderly subjects: a randomized, controlled trial. J Gerontol A Biol Sci Med Sci 2001; 56: M497–504.[Abstract/Free Full Text]
  14. Brown JD, Lawton M. Stress and well-being in adolescence: The moderating role of physical exercise. J Hum Stress 1986; 12: 125–31.
  15. Field T, Diego M, Sanders CE. Exercise is positively related to adolescents’ relationships and academics. Adolescence 2001; 36: 105–10.[Medline]
  16. Keats MR, Courneya KS, Danielsen S, Whitsett SF. Leisure-time physical activity and psychological well-being in adolescents after cancer diagnosis. J Pediatr Oncol Nurs 1999; 16: 180–8.
  17. McDermott RJ, Hawkins WE, Marty PJ, Littlefield EA, Murray S, Williams TK. Health behavior correlates of depression in a sample of high school students. J School Health 1990; 60: 414–7.[Medline]
  18. Mechanic D, Hansell S. Adolescent competence, psychological well-being, and self-assessed physical health. J Health Soc Behav 1987; 28: 364–74.[CrossRef][Medline]
  19. Murray NG, Kelder SH, Meyers MC. Type of physical activity and level of depression in suburban 9th and 10th grade students. Med Sci Sports Exerc 1998; 30: S117.
  20. Norris R, Carroll D, Cochrane R. The effects of physical activity and exercise training on psychological stress and well-being in an adolescent population. J Psychosom Res 1992; 36: 55–65.[CrossRef][Medline]
  21. Sanders CE, Field TM, Diego M, Kaplan M. Moderate involvement in sports is related to lower depression levels among adolescents. Adolescence 2000; 35: 793–7.[Medline]
  22. Kirkcaldy BD, Shephard RJ, Siefen RG. The relationship between physical activity and self-image and problem behaviour among adolescents. Soc Psychiatry Psychiatr Epidemiol 2002; 37: 544–50.[CrossRef][Medline]
  23. Page RM, Tucker LA. Psychosocial discomfort and exercise frequency: An epidemiological study of adolescents. Adolescence 1994; 29: 183–91.[Medline]
  24. Steptoe A, Butler N. Sports participation and emotional well-being in adolescents. Lancet 1996; 347: 1789–92.[CrossRef][Medline]
  25. Kovacs M. Psychiatric disorders in youths with IDDM: Rates and risk factors. Diabetes Care 1997; 20: 36–44.[Abstract]
  26. Lewinsohn PM, Rohde P, Seeley JR. Major depressive disorder in older adolescents: Prevalence, risk factors, and clinical implications. Clin Psych Rev 1998; 18: 765–94.[CrossRef][Medline]
  27. Pate RR, Heath GW, Dowda M, Trost SG. Associations between physical activity and other health behaviors in a representative sample of US adolescents. Am J Public Health 1996; 86: 1577–81.[Abstract/Free Full Text]
  28. Schmitz KH, Lytle LA, Phillips GA, Murray DM, Birnbaum AS, Kubik MY. Psychosocial correlates of physical activity and sedentary leisure habits in young adolescents: The Teens Eating for Energy and Nutrition study. Prev Med 2002; 34: 266–78.[CrossRef][Medline]
  29. Wells VE, Deykin EY, Klerman GL. Risk factors for depression in adolescence. Psychiatric Devel 1985; 3: 83–108.
  30. Lytle LA, Perry CL. Applying research and theory in program planning: An example from a nutrition education intervention. Health Promotion Practice 2001; 2: 68–80.
  31. Garrison CZ, Addy CL, Jackson KL, McKeown RE, Waller JL. The CES-D as a screen for depression and other psychiatric disorders in adolescents. J Am Acad Child Adolesc Psychiatry 1991; 30: 636–41.[Medline]
  32. Roberts RE, Lewinsohn PM, Seeley JR. Screening for adolescent depression: A comparison of the CES-D and BDI. J Am Acad Child Adolesc Psychiatry 1991; 20: 58–66.
  33. Kelder SH, Perry CL, Klepp KI. Community-wide youth exercise promotion: Long-term outcomes of the Minnesota Heart Health Program and the Class of 1989 Study. J School Health 1993; 63: 218–23.[Medline]
  34. Kelder SH, Perry CL, Klepp KI, Lytle LL. Longitudinal tracking of adolescent smoking, physical activity, and food choice behaviors. Am J Public Health 1994; 84: 1121–6.[Abstract/Free Full Text]
  35. Sirard JR, Pate RR. Physical activity assessment in children and adolescents. Sports Med 2001; 31: 439–54.[CrossRef][Medline]
  36. Radloff LS. The CES-D Scale: A self-report depression scale for research in the general population. Appl Psychol Meas 1997; 1: 385–401.[CrossRef]
  37. Lance CE, Vandenberg RJ, Self RM. Latent growth models of individual change: The case of newcome adjustment. Organ Behav Hum Decis Process 2000; 83: 107–40.[CrossRef][Medline]
  38. Meredith W, Tisak J. Latent curve analysis. Psychometrika 1990; 55: 107–22.
  39. Willett JB, Sayer AG. Using covariance structure analysis to detect correlates and predictors of individual change over time. Psychol Bull 1994; 116: 363–81.[CrossRef]
  40. Arbuckle JL, Wothke W. AMOS 4.0 user’s guide. Chicago, IL: SmallWaters Corporation; 1999.
  41. Enders CK, Bandalos DL. The relative performance of full information maximum likelihood estimation for missing data in structural equation models. Struct Equat Model 2001; 8: 430–57.[CrossRef]
  42. Wothke W. Longitudinal and multigroup modeling with missing data. In: Little TD, Schnabel KU, Baumert J, editors. Modeling longitudinal and multilevel data: Practical issues, applied approaches, and specific examples. Mahwah, NJ: Lawrence Erlbaum; 2000.
  43. Bollen KA, Structural equations with latent variables. New York: Wiley; 1989.
  44. Browne MW, Cudeck R. Alternative ways of assessing model fit. In: Bollen KA, Long JS, editors. Testing structural equation models. Newbury Park, CA: Sage; 1993.
  45. Bentler PM. Comparative fit indices in structural models. Psychol Bull 1990; 107: 238–46.[CrossRef][Medline]
  46. Bentler PM, Bonett DG. Significance tests and goodness of fit in the analysis of covariance structures. Psychol Bull 1980; 88: 588–606.[CrossRef]
  47. Hu L, Bentler PM. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Struct Equat Model 1999; 6: 1–55.
  48. Sallis JF, Prochaska JJ, Taylor WC. A review of correlates of physical activity of children and adolescents. Med Sci Sports Exerc 2000; 32: 963–75.[Medline]
  49. Rosenthal R, Rubin DB. A simple, general purpose display of magnitude of experimental effect. J Ed Psych 1982; 74: 166–9.[CrossRef]
  50. Wagner KD, Ambrosini P, Rynn M, Wohlberg C, Yang R, Greenbaum MS, Childress A, Donnelly C, Deas D; Sertraline Pediatric Depression Study Group. Efficacy of sertraline in the treatment of children and adolescents with major depressive disorder: two randomized controlled trials. JAMA 2003; 290: 1033–41.[Abstract/Free Full Text]
  51. Caspersen CJ, Pereira MA, Curran KM. Changes in physical activity patterns in the United States, by sex and cross-sectional age. Med Sci Sports Exerc 2000; 32: 1601–9.[Medline]
  52. Fleming JE, Offord DR. Epidemiology of childhood depressive disorders: A critical review. J Am Acad Child Adolesc Psychiatry 1990; 29: 571–80.[Medline]
  53. Emslie GJ, Rush AJ, Weinberg WA, Kowatch RA, Hughes CW, Carmody T, Rintelmann J. A double-blind, randomized, placebo-controlled trial of fluoxetine in children and adolescents with depression. Arch Gen Psychiatry 1997; 54: 1031–7.[Abstract/Free Full Text]
  54. Brent DA, Holder D, Kolko D, Birmaher B, Baugher M, Roth C, Iyengar S, Johnson BA. A clinical psychotherapy trial for adolescent depression comparing cognitive, family, and supportive therapy. Arch Gen Psychiatry 1997; 54: 877–85.[Abstract/Free Full Text]
  55. Brown SW, Welsh MC, Labbe EE, Vitulli WF, Kulkarni P. Aerobic exercise in the psychological treatment of adolescents. Percept Mot Skills 1982; 74: 555–60.[CrossRef]
  56. Blumenthal JA, Babyak MA, Moore KA, Craighead WE, Herman S, Khatri P, Waugh R, Napolitano MA, Forman LM, Appelbaum M, Doraiswamy PM, Krishnan KR. Effects of exercise training on older patients with major depression. Arch Intern Med 1999; 159: 2349–56.[Abstract/Free Full Text]
  57. Fremont J, Craighead LW. Aerobic exercise and cognitive therapy in the treatment of dysphoric moods. Cognitve Ther Res 1987; 11: 241–51.



This article has been cited by other articles:


Home page
 J Pediatr PsycholHome page
K. H. Petty, C. L. Davis, J. Tkacz, D. Young-Hyman, and J. L. Waller
Exercise Effects on Depressive Symptoms and Self-Worth in Overweight Children: A Randomized Controlled Trial
J. Pediatr. Psychol., October 1, 2009; 34(9): 929 - 939.
[Abstract] [Full Text] [PDF]

Home page
 J Atten DisordHome page
B. D. Kiluk, S. Weden, and V. P. Culotta
Sport Participation and Anxiety in Children with ADHD
J Atten Disord, May 1, 2009; 12(6): 499 - 506.
[Abstract] [PDF]

Home page
 PediatricsHome page
M. Hamer, E. Stamatakis, and G. Mishra
Psychological Distress, Television Viewing, and Physical Activity in Children Aged 4 to 12 Years
Pediatrics, May 1, 2009; 123(5): 1263 - 1268.
[Abstract] [Full Text] [PDF]

Home page
 Eur J Public HealthHome page
A. Sacker and N. Cable
Do adolescent leisure-time physical activities foster health and well-being in adulthood? Evidence from two British birth cohorts
Eur J Public Health, June 1, 2006; 16(3): 331 - 335.
[Abstract] [Full Text] [PDF]

Home page
 Psychosom. Med.Home page
D. Rodriguez, H. B. Moss, and J. Audrain-McGovern
Developmental Heterogeneity in Adolescent Depressive Symptoms: Associations With Smoking Behavior
Psychosom Med, March 1, 2005; 67(2): 200 - 210.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowRequest Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Motl, R. W.
Right arrow Articles by Dishman, R. K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Motl, R. W.
Right arrow Articles by Dishman, R. K.
Related Collections
Right arrow Exercise
Right arrow Depression

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS