To determine whether being both vigorously active and a team sports participant or being vigorously physically active but not a team member is associated with substance use and sexual risk behaviors.
Cross-sectional, using data from the 1999 national Youth Risk Behavior Survey.
A nationally representative sample of 15 349 US high school students.
Main Outcome Measures
Sexual risk behaviors and substance use among those who were both physically active and team sports participants, physically active but not on a sports team, physically nonactive but on a sports team, and physically nonactive and not on a sports team by sex and race/ethnicity.
Nationwide, 41.9% of the students were both physically active and participants on a sports team, 22.1% were physically active but not sports team members, 12.6% were physically nonactive sports team members, and 22.3% were physically nonactive and not sports team members. More female (mean [SD], 29.3% [2.2%]) than male students (15.3% [1.9%]) were nonactive, and more male students were both physically active and participants in team sports (48.9% [3.4%]) than were female students (34.8% [3.2%]). Black students were more likely to be physically nonactive in both the team and nonteam categories than were students overall. Relative to nonactive nonteam female students, physically active female students on sports teams were less likely to be substance users or engage in sexual risk behaviors than were active nonteam and nonactive team female students. Other associations were specific to racial/ethnic subgroups.
Overall, being both physically active and a team sports participant was associated with a lower prevalence of several health risk behaviors.
AMONG THE many long-term health benefits of physical activity (PA) are that it helps prevent chronic conditions such as cardiovascular disease, diabetes mellitus, obesity, colon cancer, and hypertension at later stages in life.1 Because most health behavior patterns are established in childhood,2-4 promotion of PA should begin at an early age.
In addition to its biological benefits, being physically active is associated with a lower prevalence of several health risk behaviors. Cross-sectional studies in the United States have found that adolescent students who participated in sports programs were less likely to use cigarettes,5-9 marijuana,7-9 and other drugs,8 and to engage in risky sexual behaviors.5,8 Similarly, cross-sectional studies of adolescents have reported a negative relationship between being physically active and using cigarettes6,10,11 and marijuana.10,11 In addition, longitudinal studies have found that students who participated in sports were less likely to initiate alcohol use12 and risky sexual behaviors.5,13-15 Another longitudinal study found a negative association between being physically active and using cigarettes.12 Some studies, however, have suggested that being physically active or involved in sports programs is associated with a higher prevalence of substance use and risky sexual behaviors. For example, alcohol use6,12 and sexual activity14 have been reported to be positively associated with PA or sports participation.
Many of the studies just described have grouped the population either by sports participation (eg, athletes vs nonathletes) or PA level (eg, active vs inactive or a gradient of activity levels). Unfortunately, studies that group students by sports participation could include those ranging from the inactive to the vigorously active. Similarly, studies grouping students by PA levels could include youth who participate in team sports as well as those not on teams. Thus, PA level is a confounder when categorizing by sports participation; sports participation is a confounder when grouping by PA level. Unless these variables are controlled during analyses, identifying the factors underlying a relationship between certain risk behaviors and PA levels or sports participation is difficult—is it the PA alone, team membership alone, or a combination?
Few studies that examined substance use and sexual activity controlled for PA level when comparing participants and nonparticipants in team sports.6,8,16 In one such study of tobacco and alcohol use, Rainey et al6 distinguished between team athletes and nonathletes and subdivided the athletes into high and moderate PA groups and the nonathletes into high, moderate, low, and sedentary PA groups. High and moderate PAs were measured differently for athletes and nonathletes, however. Among athletes, the PA level was measured by the number of sports teams to which they belonged; for nonathletes, self-reported PA level was used.6 Certainly, the number of teams on which adolescents participate could be an inadequate predictor of PA levels.
Another concern about available studies is that many have not analyzed data separately for male and female adolescents. We know that female students are generally less likely than male students to participate in vigorous PA and in sports programs.17-21 In addition, studies that have analyzed substance use and sexual behaviors separately for male and female students have shown that the relationship between these behaviors and sports participation5,8,12,13,15 or PA11,12 differs by sex. For example, male sports participants have been found to initiate sex earlier than nonparticipants,5,13 but this same relationship has not been found among female adolescents. Similarly, female students who were more active or more fit were less likely to use cigarettes than female students who were less active or less fit,12 but this same association was not true for male students. Clearly, looking at these behaviors and associations for male and female adolescents separately is appropriate. In addition, the relationship between sports participation and health risk behaviors has been shown to differ by race/ethnicity.8,16
The purpose of the present study is to identify whether being vigorously active alone, a team sports participant alone, or being both vigorously active and a team sports participant is associated with substance use and risky sexual behaviors and whether associations are specific to sex and racial/ethnic subgroups. We controlled for both team sports participation and vigorous PA by stratifying youth into the following 4 categories that considered both factors: (1) vigorously active and involved in team sports, (2) vigorously active but not involved in team sports, (3) not vigorously active but involved in team sports, and (4) neither vigorously active nor involved in team sports. We analyzed data separately for male and female students as well as by race/ethnicity.
We used data from the 1999 national school-based Youth Risk Behavior Survey (YRBS) for this study. The survey sample used a 3-stage cluster sample design, which has been described in detail elsewhere.17 Briefly, the first-stage sampling frame contained 1270 primary sampling units consisting of large counties or groups of smaller adjacent counties. Of the 1270 primary sampling units, 52 were selected with probability proportional to size. At the second sampling stage, schools were selected with probability proportional to their enrollment. Schools with substantial numbers of black and Hispanic students were oversampled to enable separate analysis of these groups. The third stage of sampling was random selection of 1 or 2 classes of a required subject (eg, English or social studies) from grades 9 through 12 at each chosen school. A weighting factor was applied to each student record to adjust for nonresponse and for varying probabilities of selection, including those resulting from oversampling of black and Hispanic students. Data are representative of students in grades 9 through 12 attending public or private schools in the 50 states and the District of Columbia. There were 15 349 respondents from 144 schools with response rates of 77% for the schools and 86% for the students. The overall response rate was 66%.
Survey procedures were designed to protect students' privacy by allowing for anonymous and voluntary participation. Students completed the self-administered questionnaire in their classrooms during a regular class period, recording their responses directly on a computer-scannable booklet.
Students were placed in 1 of 4 categories based on their responses to 2 items: "On how many of the past 7 days did you exercise or participate in physical activities for at least 20 minutes that made you sweat and breathe hard?" (labeled as "vigorous PA") and "During the past 12 months, on how many sports teams did you play (including any teams run by your school or community groups)?" We called those who reported participation on at least 1 sports team and 20 minutes or more of vigorous PA on 3 or more days "active team males or females." We called those reporting no participation on teams but 20 minutes or more of vigorous activity on 3 or more days "active nonteam males or females." We called those reporting at least 1 sports team and fewer than 3 days of vigorous PA for 20 minutes or more "nonactive team males or females." Finally, we called those reporting neither team membership nor 3 or more days of vigorous PA for 20 minutes or more "nonactive nonteam males or females." The term "nonactive" includes students from the inactive to the moderately active. For simplicity, we refer to these groups as "PA categories" although they consider both PA and sports participation.
Of the 15 349 respondents, 15 142 (98.6%) were included in these analyses. Excluded were those missing a response to either of the 2 items used to determine PA categories (n = 207).
The YRBS items on sexual activity and substance use that we used are listed in Table 1. Although it is not a behavior, the YRBS item on forced sex was included in this study because it is an important health risk factor, particularly for female students. Individual items related to lifetime use of cocaine, inhaled substances, heroin, methamphetamines, and injected drugs were combined to create the variable "other drug use." More than 90% of the items on the YRBS that measure self-reported health risk behavior have at least moderate reliabilities (κ ± 41%).22
1999 Youth Risk Behavior Survey Items Selected for This Study
All estimates were based on weighted data, with SUDAAN (Research Triangle Institute, Research Triangle Park, NC) used to compute 95% confidence intervals for prevalence estimates and to conduct logistic regression analyses. Because we expected that responses by male and female students on risk behaviors would differ significantly, we analyzed these behaviors by sex. Risk behaviors that were statistically significantly associated with PA category were used in logistic regression analyses to determine whether the associations remained after controlling for grade and race/ethnicity. A separate series of logistic regression analyses was performed to determine whether the interaction between race/ethnicity and PA category was significantly associated with risk behaviors after controlling for grade. Data are given as frequencies or odds ratios and 95% confidence intervals.
The sample (N = 15 349) was 49.6% female, 60.8% white (non-Hispanic), 14.1% black or African American (non-Hispanic), and 10.4% Hispanic or Latino. Students not reporting any of these racial/ethnic groups (14.7%) are described as "other." The sample was 28.9% 9th-, 26.0% 10th-, 23.6% 11th-, and 21.4% 12th-grade students. Most of the sample was from suburban areas (59.1%) with 29.4% from urban and 11.0% from rural areas.
Overall, most (41.9%) of the respondents were in the active team category, while 22.1% were active nonteam, 12.6% were nonactive team, and 22.3% were nonactive nonteam. Nonactive team students represented 23.1% of all participants in team sports.
The estimated percentage of male students in the physically active team category (48.9%) was higher than the corresponding estimate for female students (34.8%) (Table 2). In both the active nonteam and nonactive team categories, the proportions of male students (22.3% and 11.7%, respectively) and female students (21.9% and 13.4%, respectively) were roughly equal. The percentage of female students in the nonactive nonteam category (29.3%) was almost twice the percentage of male students (15.3%). Finally, the proportion of black students in the nonactive team (15.1%) and nonactive nonteam (28.5%) categories was greater than the overall proportion of students in these categories (12.6% and 22.3%, respectively).
Demographic Characteristics by Physical Activity Categories: 1999 Youth Risk Behavior Survey for 15 349 Participants*
The percentages of students from urban, suburban, and rural areas were similar for each PA category with the exception of the nonactive team category. Fewer rural students were nonactive team (9.3%) than were students overall (12.6%) (Table 2). The percentage of 9th-grade students who were active team members (46.9%) was greater than the percentage of 11th- or 12th-grade students in this category (39.0% and 38.3%, respectively). Conversely, the percentage of 9th-grade students who were nonactive nonteam (16.2%) was lower than the percentage of 11th- (28.5%) or 12th- (25.5%) grade students in this category.
Among male students, the only risk behavior associated with PA category was "other drug use" (data not shown). Active team male students were less likely to use other drugs (18.2% [3.7%]) than were active nonteam male students (25.8% [3.8%]) but did not differ from nonactive team male students (24.1% [5.7%]) and nonactive nonteam male students (22.5% [6.1%]).
Among female students, the prevalence of 6 health risk behaviors varied by PA category (Table 3). Health risk behaviors for which active team female students had a lower prevalence than their active nonteam, nonactive team, and nonactive nonteam counterparts were ever had sexual intercourse, had 4 or more sex partners during lifetime, and had 4 or more sex partners in the last 3 months. Active team female students also had a lower prevalence of having ever been pregnant than did active nonteam and nonactive nonteam female students but did not differ significantly from nonactive team female students. Behaviors for which the active team group had a lower prevalence than the active nonteam group but did not differ from those who were nonactive team and nonactive nonteam were current cigarette use and current marijuana use.
Weighted Percentages for Selected Health Risk Behaviors Among Female High School Students by Physical Activity Categories*
We tested the statistically significant associations between PA category and health risk behaviors given in Table 3 in a series of logistic regression analyses that controlled for grade and race/ethnicity (Table 4). Active team female students had lesser odds of ever having had intercourse, having had 4 or more sexual partners in their lifetime or in the 3 previous months, having ever been pregnant, and currently using cigarettes than nonactive nonteam female students. However, active team female students and nonactive nonteam female students had similar odds of current marijuana use. The odds of having had 4 or more sexual partners in their lifetime or in the 3 previous months, having ever been pregnant, current cigarette use, and current marijuana use were similar for active nonteam, nonactive team, and nonactive nonteam female students.
Adjusted Odds Ratios for Selected Health Risk Behaviors Among High School Females by Physical Activity Category*
In a separate set of logistic regression analyses, the association between health risk behaviors and the interaction between PA category and race/ethnicity was assessed for female students while controlling for grade. With the exception of marijuana use (Wald F, 1.51; P = .09), the interaction between PA category and race/ethnicity was associated significantly with each health risk behavior we examined. The pattern of results varied for each risk behavior, however (Table 5). White nonactive nonteam female students served as the reference group in these analyses. Relative to the reference group, black female students in all categories other than "active team" were significantly more likely to have ever had sex, and white active team female students were significantly less likely to have ever had sex (Wald F, 7.94; P<.001). White, Hispanic, and other active team female students, as well as black active nonteam and Hispanic nonactive team female students were less likely than the reference group to have had had 4 or more lifetime sexual partners. Black nonactive nonteam female students were significantly more likely than their white counterparts to have had 4 or more lifetime partners (Wald F, 4.44; P<.001). Black nonactive team and nonactive nonteam female students were significantly more likely than the reference group to have 4 or more current sexual partners (Wald F, 5.41; P<.001) and to have ever been pregnant (Wald F, 2.89; P = .001), while active team female students who were white or in the other race/ethnicity categories were less likely than the reference group to have 4 or more current sexual partners. Finally, relative to the reference group, all black female students regardless of PA category were significantly less likely to report current cigarette use, as were white active team and Hispanic active nonteam and nonactive nonteam female students (Wald F, 5.53; P<.001).
Adjusted Odds Ratios for the Association Between Selected Health Risk Behaviors Among High School Females and the Interaction Between Physical Activity Group and Race/Ethnicity*
This study found that nearly one fourth (23.1%) of the participants in team sports were not vigorously active. This result may seem surprising, as one might assume that virtually all of these students would exercise vigorously. Several explanations suggest themselves, some of them related to the 7-day time frame for PA on the YRBS. For example, an "out-of-season" athlete could have reported less PA than he or she would have reported had the survey been during the season. An athlete might also have been injured or recently ill. Students who were not starting players might receive little or no playing time during competitions and perhaps less in practice as well. Golfers, bowlers, and other team athletes may not have to engage in vigorous PA. Whatever the reason, this finding suggests that using participation in team sports as a surrogate for PA (or PA levels) may not be prudent.
Unlike the case for female students, only 1 significant association was found for male students between team sports participation and health risk behavior: fewer active team male students were other drug users than were their active nonteam counterparts. Contrary to the results of other studies, active team male students were not more likely than their active nonteam and nonactive peers to have used alcohol,12 to have ever had sex,14 or to have initiated sexual intercourse at an early age.5 They also were no less likely to have used cigarettes or marijuana.8 Our results may differ from previous studies because we categorized students by both PA and team sports participation. Having more categories with fewer students in each may have led to fewer statistically significant findings. In addition, use of a younger population,12 lack of control for PA levels,5,8,14 and geographic variation5 (in the case of age at first sexual intercourse) may explain why results of previous studies differ from our results.
For female students, we did not confirm the findings of Sabo et al5 that athletes reported greater use of condoms and initiated intercourse later than did nonathletes. Again, the differences in these findings could be partially due to our having controlled for vigorous PA. Other possible factors include the smaller sample size (n = 9009) in the study by Sabo et al due, in part, to the exclusion of students belonging to the other race category. Real variability in responses between the 1995 and 1999 YRBS may also explain differences in results.
If PA plus team sports participation modifies health risk behavior, it may do so differently for male vs female students. Perhaps the potential benefits of PA plus sports participation are not enough to offset the higher degree of risk behavior among male students, in general, or the influence of peers and sociocultural norms on male risk behavior. The frequency and nature of media images of male sports figures (and advertising surrounding male sports programming) differ significantly from that of female sports figures. One example of this is the association of sports with alcohol consumption. These media images may play a role in shaping team "cultures" that are different for male and female students.
The motivational factors for team sports participation and the nature of the sports themselves may attract a different subset of male vs female students. For example, male students drawn to more aggressive contact sports or sports that can be perceived as riskier may be more prone to health risk behaviors. Male students may be more motivated to participate in team sports because it is a way to demonstrate physical strength or to achieve status, approval, or reward. Female students may be more motivated by the social interaction or the health benefits of participating in team sports. Whatever the case, it seems that active team female students are more distinct from female students in the other PA categories in terms of health risk behaviors than are their male counterparts.
Overall, exposure to both PA and team sports participation, but neither PA alone nor sports participation alone was associated with less risky behavior among female students. If coaches serve as role models or enforcers of positive behavior, or if the threat of disqualification if caught drinking or using drugs is greater for team participants, why do we not also observe less risky behavior among those in the nonactive team category? The negative effects of risky behavior on athletic performance may be part of the answer. Individuals involved in more vigorously active team sports may avoid alcohol, other drugs, or sexual activity to stay focused and to maintain maximum performance. Individuals involved in team sports that are less physically demanding may place less priority on physical fitness and performance and may, therefore, engage in riskier behavior. Again, the nature of the sport itself may attract those already engaging in certain risk behaviors or create an environment of normative risk behaviors.
Commitment to team sports may leave less time, particularly after school or on weekends, for engaging in risky behavior among active team female students vs active nonteam female students. Dedicated team members may not want to disappoint teammates by performing badly or jeopardizing their team positions. There may be greater peer pressure to be fit and healthy or to outperform other athletes within a team environment vs a nonteam setting. These factors, and many others, may explain the differences between vigorous PA within vs outside the team setting that possibly influence risk behavior.
If PA or team sports participation directly modifies health risk behaviors, results of our analyses by race/ethnicity may indicate that female students in white and "other" race categories potentially benefit from the combination of vigorous PA and team sports participation while black female students may benefit from vigorous PA alone. Hispanic female students may benefit from team sports participation alone for one risk behavior (ie, the number of sexual partners over the lifetime). Overall, however, vigorous PA within a team sports setting could have the greatest potential benefit for the greatest number of female students.
This study may be limited by the fact that students may underreport or overreport their health risk behavior. How much misreporting takes place is unclear, but we should note that most items measuring health risk behavior on the YRBS have acceptable reliability.22 Also, the sample size for the present study was large and thus many significant associations were expected to be found overall. Analyses by race/ethnicity, however, were affected by having fewer participants within PA categories, resulting in wide confidence intervals for reported odds ratios and fewer significant associations. Unfortunately, whether participation in team sports is related causally to any of the health risk behaviors examined cannot be determined because of the cross-sectional nature of this study. In addition, associations between participation in specific team sports and health risk behaviors cannot be determined from this study.
Longitudinal studies are needed to determine the precise nature of the association between participation in team sports and risk behaviors. Regardless of whether team sports attract female students who engage in less risky behavior or whether being on a sports team modifies risky behavior, however, one can hypothesize that a social norm may be established within the team sports environment that could reinforce less risky behavior. If so, participation in team sports would be beneficial regardless of any direct role it might have in causing or modifying behavior.
Barriers to team sports participation by female students,20,23,24 particularly ones from racial and ethnic minority subgroups, need to be eliminated wherever possible. While coordinating efforts to promote PA,25 schools and communities should consider both the additional health benefits of team sports for adolescent females and the known biological benefits associated with PA in both adolescence26-29 and adulthood.30-34
Corresponding author: Kimary Kulig, PhD, MPH, Pfizer Global Pharmaceuticals, US Outcomes Research, 235 E 42nd St, New York, NY 10017 (e-mail: firstname.lastname@example.org).
Accepted for publication May 1, 2003.
Dr Kulig was an Association of Schools of Public Health intern at the time this study was completed.
Numerous studies have shown that PA among adolescents is associated with lower prevalence of several health risk behaviors. In many studies, however, participation in sports was used as a surrogate for being physically active, or members of team sports were presumed to be very physically active. The present study was conducted to distinguish health risk behaviors among adolescents that are specifically associated with participation in team sports rather than associated with PA alone. Results demonstrate that participation in team sports along with being vigorously physically active, but not being on a sports team or vigorously physically active by itself, is associated with lower prevalence of several health risk behaviors among female adolescents. These findings indicate that active team sports participation among female adolescents may afford benefits beyond the known biological health benefits of PA. Barriers to team sports participation by female students must be eliminated, and schools and communities should coordinate efforts to promote such participation.
JM The public health burden of a sedentary lifestyle. Med Sci Sport Exerc.
1992;24(suppl 6)S196- S200Google ScholarCrossref
LL Longitudinal tracking of adolescent smoking, physical activity, and food choice behaviors. Am J Public Health.
1994;841121- 1126PubMedGoogle ScholarCrossref
J Physical activity in childhood and adolescence as predictor of physical activity in young adulthood. Am J Prev Med.
1997;13317- 323PubMedGoogle Scholar
RM Tracking of physical activity and physical fitness across the lifespan. Res Q Exerc Sport.
1996;57(suppl 3)548- 557Google Scholar
M The Women's Sports Foundation Report: Sport and Teen Pregnancy. East Meadow, NY Women's Sports Foundation1998;
RF Patterns of tobacco and alcohol use among sedentary, exercising, nonathletic, and athletic youth. J Sch Health.
1996;6627- 32PubMedGoogle ScholarCrossref
NJ Health risk behaviors of adolescent participants in organized sports. J Adolesc Health.
1998;22460- 465PubMedGoogle ScholarCrossref
M Sports participation and health-related behaviors among US youth. Arch Pediatr Adolesc Med.
2000;154904- 911PubMedGoogle ScholarCrossref
R Risk behavior in adolescence: a psychological framework for understanding and action. J Adolesc Health.
1991;12597- 605PubMedGoogle ScholarCrossref
RR Relationship between physical activity level and cigarette, smokeless tobacco, and marijuana use among public high school adolescents. J Sch Health.
1995;65438- 442PubMedGoogle ScholarCrossref
SG Associations between physical activity and other health behaviors in a representative sample of US adolescents. Am J Public Health.
1996;861577- 1581PubMedGoogle ScholarCrossref
RE Physical activity and the initiation of high-risk health behaviors in adolescents. Med Sci Sports Exerc.
1995;271639- 1645PubMedGoogle ScholarCrossref
GM High school athletic participation, sexual behavior and adolescent pregnancy: a regional study. J Adolesc Health.
1999;25207- 216PubMedGoogle ScholarCrossref
LL Participation in high school sports and adolescent sexual activity. Pediatr Exerc Sci.
1994;669- 74Google Scholar
MJ Athletic participation and sexual behavior in adolescents: the different worlds of boys and girls. J Health Soc Behav.
1998;39108- 123PubMedGoogle ScholarCrossref
MP Gender and racial/ethnic differences in predicting adolescent sexual risk: athletic participation versus exercise. J Health Soc Behav.
2002;43436- 459PubMedGoogle ScholarCrossref
et al. Youth risk behavior surveillance: United States. MMWR Morb Mortal Wkly Rep.
2000;491- 94PubMedGoogle Scholar
L Physical activity patterns in American high school students: results from the 1990 Youth Risk Behavior Survey. Arch Pediatr Adolesc Med.
1994;1481131- 1136PubMedGoogle ScholarCrossref
CR Ethnic, socioeconomic, and sex differences in physical activity among adolescents. J Clin Epidemiol.
1996;49125- 134PubMedGoogle ScholarCrossref
WC A review of correlates of physical activity of children and adolescents. Med Sci Sports Exerc.
2000;32963- 975PubMedGoogle ScholarCrossref
JG Reliability of the 1999 Youth Risk Behavior Survey questionnaire. J Adolesc Health.
2002;31336- 342PubMedGoogle ScholarCrossref
ML Determinants of physical activity among female adolescents. Am J Prev Med.
1997;13115- 122PubMedGoogle Scholar
BM Determinants of adolescent physical activity and inactivity patterns. Pediatrics.
Centers for Disease Control and Prevention, Guidelines for school and community programs to promote lifelong physical activity among young people. MMWR Morb Mortal Wkly Rep.
1997;461- 35PubMedGoogle Scholar
C Physical activity, physical fitness, and coronary heart disease risk factors in youth: the Quebec Family Study. Prev Med.
((6 pt 1))
555- 562PubMedGoogle ScholarCrossref
JM Effects of school-based aerobic exercise on blood pressure in adolescent girls at risk for hypertension. Am J Public Health.
1998;88949- 951PubMedGoogle ScholarCrossref
JH Physical activity and blood pressure in adolescents. Pediatr Exerc Sci.
1994;6361- 380Google Scholar
BR Bone density and bone mineral content of adolescent soccer athletes and competitive swimmers. Pediatr Exerc Sci.
1992;4319- 330Google Scholar
R Adult female hip bone density reflects teenage sports-exercise patterns but not teenage calcium intake. Pediatrics.
((1, pt 1))
40- 44PubMedGoogle ScholarCrossref
U The health benefits of physical activity in children and adolescents: implications for chronic disease prevention. Eur J Pediatr.
1999;158271- 274PubMedGoogle ScholarCrossref
GF Exercise and its role in the prevention and rehabilitation of cardiovascular disease. Ann Behav Med.
1997;19220- 229PubMedGoogle ScholarCrossref
et al. Physical activity and public health: a recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine. JAMA.
1995;273402- 407PubMedGoogle ScholarCrossref
US Department of Health and Human Services, Physical Activity and Health: A Report of the Surgeon General. Atlanta, Ga US Dept of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion1996;234- 235