Share this article on:

Tracking of physical activity in young children


Medicine & Science in Sports & Exercise: January 1996 - Volume 28 - Issue 1 - p 92-96

The purpose of this study was to determine whether physical activity behavior tracks during early childhood. Forty-seven children (22 males, 25 females) aged 3-4 yr at the beginning of the study were followed over a 3-yr period. Heart rates were measured at least 2 and up to 4 d·yr-1 with a Quantum XL Telemetry heart rate monitor. Physical activity was quantified as the percentage of observed minutes between 3:00 and 6:00 p.m. during which heart rate was 50% or more above individual resting heart rate(PAHR-50 Index). Tracking of physical activity was analyzed using Pearson and Spearman correlations. Yearly PAHR-50 index tertiles were created and examined for percent agreement and Cohen's kappa. Repeated measures ANOVA was used to calculate the intraclass correlation coefficient across the 3 yr of the study. Spearman rank order correlations ranged from 0.57 to 0.66 (P < 0.0001). Percent agreement ranged from 49% to 62%. The intraclass R for the 3 yr was 0.81. It was concluded that physical activity behavior tends to track during early childhood.

Department of Exercise Science, University of South Carolina, Columbia, SC 29208; and Department of Behavioral Science, UT-M.D. Anderson Cancer Center Houston, TX 77030

Submitted for publication September 1994.

Accepted for publication April 1995.

The collection of this data was supported with a grant from the National Heart Lung and Blood Institute (HL38131). The authors appreciate the efforts of J. Puhl and K. Greaves for collecting the physical activity data and J. Baranowski for managing the data collection, and C. Spelman for her assistance with preparation of this manuscript.

Address for correspondence: Russell R. Pate, Ph.D., Department of Exercise Science, University of South Carolina, Columbia, SC 29208.

It is well understood that certain behaviors observed during adulthood are associated with increased risk of chronic disease morbidity and mortality (4,23). Among these risk factors are cigarette smoking (7), consumption of a high-fat diet(16), and physical inactivity(3,24,27). In recent years, numerous public health initiatives have been directed toward reducing the prevalence of these behaviors among adults (21). However, it is now known that health behavior change programs targeted at adults are expensive and typically show high rates of recidivism(11,14,17). Accordingly, many experts have recommended that primary prevention of chronic disease be pursued through interventions that are directed at children(28,32).

An assumption that underlies health promotion programs in childhood and youth is that behavioral risk factors for chronic disease exhibit stability over time. That is, it has been assumed that children who adopt high risk behaviors tend to maintain those behaviors through childhood and into adulthood. There is evidence that this is the case for cigarette smoking(13,29) and dietary behaviors(6). However, very little is known about the tendency of physical activity to track during childhood. Accordingly, the purpose of this study was to determine whether physical activity behavior tracks during childhood.

Back to Top | Article Outline



The subjects for this study were 47 children who, at the beginning of the study, were 3-4 yr of age. These children were a subgroup of the participants of one of the eight sites of the Study of Children's Activity and Nutrition(SCAN), a multicenter longitudinal investigation of the development of cardiovascular disease risk factors and related behaviors in children(2). Informed consent was obtained from the guardian of each child and the study was reviewed and approved by the Institutional Review Board of the University of Texas Medical Branch, Galveston.

A total of 263 youngsters participated in SCAN at the Galveston site; various techniques were used to recruit participants(15). Families whose immediate members had a history of chronic illness (e.g., hypertension), families without at least one parent residing in the household, and children with a disability that could affect participation (e.g., mental retardation) were excluded to avoid cultural and other differences within this group.

The 47 children, upon whom the analyses presented in this study were based, were those whose physical activity had been observed on a minimum of two (and up to four) occasions from 3:00 to 6:00 p.m. during each of the 3 yr of the study. Table 1 provides descriptive data on the total sample of 263 subjects and on the subgroup included in this study. The subgroup included a somewhat lower proportion of African-American youngsters and a greater fraction of non-Hispanic white youngsters than in the total sample. The physical characteristics and mean resting heart rate of the 263 subjects and the subgroup included in this study are shown inTable 2. Statistical analysis using t-tests revealed no significant differences between the two samples. Thus, the subjects in this study appear to have been representative of the larger group.

Back to Top | Article Outline


Families participated in annual summer clinics at which a variety of physiological and self-report measures were collected. Resting heart rates were collected at the first annual summer clinic in 1986. Children were observed for up to 4 d between annual clinics, with an intended approximate 3 months' delay between observation days. Field heart rates for this study were collected from August 1986, through July 1989.

Back to Top | Article Outline


Resting heart rate was measured using an automatic Dinamap Adult/Pediatric Vital Signs Monitor (Model 845 XT/XT-IEC). All subjects were measured in the early morning, without prior exercise, having fasted overnight, lying at rest for 15 min, and using the right arm. Appropriate cuff size was determined by measuring the circumference of the upper arm at its largest point using a standardized tape measure. Because of normal variability in resting heart rate, five heart rate readings were taken, 1 min apart, and the last four were recorded. The mean of the four values was used in the analyses. For the analyses performed in this study, resting heart rate was calculated as the average of resting heart rates measured during the three clinic visits.

Physical activity was assessed by measuring heart rate using a portable monitor, the Quantum XL Telemetry heart rate monitor, which consists of a transmitter and a receiver. The transmitter (43 g) measures 137 × 30× 12 mm, and was attached to the chest by self-adhering electrodes placed approximately 2 inches below the nipples. Hypoallergenic micropore tape was applied to prevent the electrodes from becoming disconnected. The receiver(47 g, 51 × 45 × 15 mm) is a microcomputer that was programmed to record and store heart rate data once every minute. Leger and Thivierge(20) determined that under laboratory conditions, the Quantum XL was the most accurate and reliable of 13 commercially available portable heart rate monitors and was found by Treiber et al.(30) to be a valid measure of physical activity in field settings.

Heart rate monitors were worn between the approximate hours of 7:00 a.m. and 7:00 p.m. However, for the analyses presented in this study, only the data collected between 3:00 and 6:00 p.m. were used. This time period was selected because it corresponds to the after-school period when children have the most discretion over their level of activity. This was important during the third year of the study, as many of the children were in school for much of the day. On observation days parents were instructed to allow the child to engage in his/her normal activities. All monitoring occurred on Monday through Thursday. Due to some technical difficulties with the heart rate monitors and parents curtailing the measurement sessions, fewer than 180 min of data were obtained from some children. Subjects were excluded if heart rate had been recorded for less than 120 min during this 3-h period. To exclude erroneous data, recorded heart rates less than 55 bpm or greater than 215 bpm were deleted.

Physical activity was quantified as the percentage of observed minutes during which heart rate was 50% or more above individual resting heart rate(PAHR-50 index). For a given year, a child's physical activity was taken as the mean PAHR-50 index for the 2-4 observation days. DuRant et al.(12) recently reported the PAHR-50 index to be a reliable indicator of physical activity in young children. The validity of the PAHR-50 index has not been formally evaluated.

Back to Top | Article Outline

Statistical Analyses

The tendency of physical activity behavior to track across the 3 yr of the study was assessed by calculating Pearson product-moment and Spearman rank order correlation coefficients. Also, yearly PAHR-50 index tertiles were created and examined for percent agreement and Cohen's kappa across years. Repeated measures ANOVA was used to calculate the intraclass correlation coefficient across the 3 yr of data. Significance was set at the 0.05 level.

Back to Top | Article Outline


Means and standard deviations for the PAHR-50 index and the Pearson and Spearman correlation coefficients for each year are shown inTable 3. The mean PAHR-50 index changed little from year to year, ranging from 12 to 15% during all 3 yr. This indicated that in the group as a whole, the overall level of physical activity was relatively constant throughout the study. Repeated-measures ANOVA revealed no significant differences among ethnic or gender groups with respect to mean PAHR-50 values.

Pearson correlations coefficients for the PAHR-50 index between years ranged from 0.53 to 0.63. The large standard deviations in comparison with the means revealed the data were not normally distributed, suggesting that the Spearman correlation might be more appropriate. Spearman rank order correlations were significant (P <.0001) and ranged from 0.57 to 0.66. This indicated that youngsters generally maintained their ranking with respect to physical activity behavior over the 3-yr study period. Since tracking is often quantified in terms of percentiles, tertiles for the mean PAHR-50 index were constructed and examined for consistency over the 3 yr of the study. Tertile means and standard deviations are shown inTable 4. The intraclass correlation coefficient across the 3 yr of physical activity data was 0.81. This result, along with the percentage of agreement and Cohen's kappa between the yearly tertiles(Table 5) demonstrated that physical activity levels within the group exhibited stability across years and that physical activity tended to track from year to year.

Back to Top | Article Outline


Previous studies have demonstrated that serum lipids(18,22,33), blood pressure(8,19), body composition (9), and physical fitness (10,26) tend to track during childhood. Importantly, each of these factors is known to be affected by physical activity behavior. The present study is the first to examine the tracking of physical activity behavior per se in young children. Our results provide strong evidence that physical activity behavior tends to track during early childhood and that less active children tend to remain less active than the majority of their peers.

A strength of the present study was our use of heart rate monitoring as an objective measure of physical activity. Importantly, heart rate monitoring avoids problems with recall and subjectivity, and is relatively inexpensive to administer in large-scale studies (25). We consider our use of the PAHR-50 index a further strength, as this measure is specific to more vigorous activity in children, is less likely to be affected by factors such as emotional state and climate, and controls for differences in fitness levels and/or age (12).

In keeping with previous studies, both Pearson and Spearman correlation coefficients were used to assess the extent of tracking. Pearson and Spearman correlations ranged from 0.53 to 0.63 and 0.57 to 0.66, respectively. In absolute terms, these correlations are no more than moderate, however, when compared with other tracking studies of risk factors such as body composition(0.53-0.64)(10), serum lipids(0.38-0.66)(33), and systolic blood pressure(0.25-0.65)(8), it appears that these correlations are typical and supportive of our conclusion that physical activity behavior tends to track during early childhood. In support of our conclusions, Bloom(5) defined a stable characteristic as one that exhibits a correlation of greater than 0.50 for two measures obtained at least 1 yr apart. Furthermore, it is possible that our correlation coefficients underestimated the true level of tracking within the group, given that we used only two to four observations of physical activity per year. A larger number of observations would be expected to produce a more reliable index and higher tracking coefficients.

It is important to note that this study was conducted with several important limitations. Physical activity was measured only in young children, only for 3 yr, with only two to four observations per year. Furthermore, because we chose to examine physical activity between the hours of 3:00 and 6:00 p.m., it is possible that some children may have exhibited different activity patterns at other times during the day. However, within the limitations of the study design, our findings provide support for the hypothesis that physical activity behavior has stability over time. This was evidenced by the 62% agreement for the PAHR-50 index tertiles between year 1 and year 3. Therefore, although we know relatively little about how to promote increased physical activity in children, our results strongly suggest that there is a somewhat stable group of children that should be the focus of such interventions.

The results of this investigation clearly highlight the need for physicians and other primary health care providers to become actively involved in the promotion of physical activity and fitness in children and youth. Given their standing within the community and frequent contact with parents and children, health care professionals are uniquely qualified to deliver a broad range of primary and secondary prevention services designed to reduce the prevalence of negative health behaviors such as inactivity. Along these lines, the American Medical Association recently developed Guidelines for Adolescent Preventive Services (GAPS) (1). This document challenges primary health care providers to make preventive services a greater component of their clinical practice. Specifically, GAPS recommends that physicians regularly counsel adolescents about the benefits of exercise and encourage them to engage in safe exercise on a regular basis (1). Our observation of a relatively stable group of children with low levels of physical activity strongly suggests that these guidelines be extended to include children under the age of 10 yr, and that health care professionals, in coordination with schools and other community organizations, become actively involved in the promotion and assessment of physical activity in young children.

Finally, our conclusion, that physical activity in young children tracks during early childhood, indicates that further research in this area is warranted. Among the key unanswered questions are (a) for how many years is tracking evident? (b) Does physical activity track from childhood to adulthood? (c) What are the genetic, social, and environmental factors that explain the tracking of physical activity behavior? and (d) Does the tracking in physical activity account for the tracking in the physiological risk factors? Furthermore, our conclusion indicates that interventions targeted at promotion of physical activity in low active children are needed and worthy of exploration.











Back to Top | Article Outline


1. American Medical Association. Guidelines for Adolescent Preventive Services. Chicago: American Medical Association, 1992.
2. Baranowski, T., E. Stone, R. Klesges, et al. Studies of child activity and nutrition (SCAN): longitudinal research on CVD risk factors and CVH behaviors in young children. Cardiovasc. Risk Factors 2:4-16, 1993.
3. Berlin, J. A. and G. A. Colditz. A meta-analysis of physical activity in the prevention of coronary heart disease. Am. J. Epidemiol. 132:612-628, 1990.
4. Blair, S. N., H. W. Kohl, R. S. Paffenbarger, D. G. Clark, K. H. Cooper, and L. W. Gibbons. Physical fitness and all-cause mortality. J.A.M.A. 262:2395-2401, 1989.
5. Bloom, B. S. Stability and Change in Human Characteristics. New York: John Wiley & Sons, 1964, pp. 1-16.
6. Burns, T. L., P. P. Moll, and R. M. Lauer. Genetic models of human obesity-family studies. Crit. Rev. Food Sci. Nutr. 33:339-343, 1993.
7. Center For Disease Control. Smoking attributable mortality and years of potential life lost-United States, 1988. Morbid. Mortal. Wkly. Rep. 40:62-63, 1991.
8. Clarke, W. R., H. G. Schrott, P. E. Leaverton, W. E. Connor, and R. M. Lauer. Tracking of blood lipids and blood pressure in children: the Muscatine study. Circulation 58:626-634, 1978.
9. Clarke, W. R. and R. M. Lauer. Does obesity track into childhood? Crit. Rev. Food Sci. Nutr. 33:423-430, 1993.
10. Dennison, B. A., J. H. Straus, E. D. Mellits, and E. Charney. Childhood fitness tests: predictor of adult physical activity levels?Pediatrics 82:324-330, 1988.
11. Dishman, R. K. Exercise Adherence. Champaign, IL: Human Kinetics, 1988, pp. 279-396.
12. Durant, R. H., T. Baranowski, H. Davis, et al. Reliability and variability of indicators of heart-rate monitoring in children. Med. Sci. Sports Exerc. 25:389-395, 1993.
13. Ershler, J., H. Levanthal, R. Fleming, K. Glynn. The quitting experience for smokers in sixth through twelfth grade. Addict. Behav. 14:365-378, 1989.
14. Haynes, R. B. Compliance with health advice: an overview with special reference to exercise programs. J. Cardiac Rehab. 4:120-123, 1984.
15. Hooks, P. C., Y. Tsong, T. Baranowski, J. Henske, P. R. Nader, and J. S. Levin. Recruitment strategies for multi-ethnic family and community research: a primer and model. Community Fam. Health 11:48-59, 1988.
16. Keys, A. Serum cholesterol response to dietary cholesterol. Am. J. Clin. Nutr. 40:351-359, 1984.
17. King, A. C. Community interventions for promotion of physical activity and fitness. Exerc. Sports Sci. Rev. 19:211-259, 1991.
18. Laskarzewski, P., J. A. Morrison, I. De Groot, et al. Lipid and lipoprotein tracking in 108 children over a four-year period.Pediatrics 64:584-591, 1979.
19. Lauer, R. M., A. R. Andersen, R. Beaglehole, and T. L. Burns. Factors related to tracking of blood pressure in children. U.S. National Center for Health Statistics Health Examination Survey Cycles II and III. Hypertension 6:307-314, 1984.
20. Leger, L. and M. Thivierge. Heart rate monitors: validity, stability, and functionality. Physician Sportsmed. 116:143-151, 1988.
21. McGinnis, J. M. The public burden of sedentary lifestyles. Med. Sci. Sports Exerc. 24(Suppl.):S196-S200, 1992.
22. Orchard, T. J., R. P. Donahue, L. H. Kuller, et al. Cholesterol screening in childhood: does it predict adult hypercholesterolemia? The Beaver County experience. J. Pediatr. 103:687-691, 1983.
23. Paffenbarger, R. S., R. T. Hyde, A. L. Wing, and C.-C. Hsieh. Physical activity, all-cause mortality and longevity of college alumni.N. Engl. J. Med. 314:605-613, 1986.
24. Paffenbarger, R. S., R. T. Hyde, and A. L. Wing. Physical activity and physical fitness as determinants of health and longevity. In: Exercise, Fitness, and Health. C. Bouchard, R. J. Shephard, T. Stephens, J. R. Sutton, and B. D. McPherson (Eds.). Champaign, IL: Human Kinetics, 1990, pp. 33-48.
25. Pate, R. R. Physical activity assessment in children and adolescents. Crit. Rev. Food Sci. Nutr. 33:321-326, 1993.
26. Powell, K. E. and W. Dysinger. Childhood participation in organized school sports and physical education as precursors of adult physical activity. Am. J. Prev. Med. 3:276-281, 1987.
27. Powell, K. E., P. D. Thompson, C. J. Caspersen, and J. S. Kendrick. Physical activity and the incidence of coronary heart disease.Annu. Rev. Public Health 8:253-287, 1987.
28. Simons-Morton, B. G., N. M. O'Hara, D. G. Simons-Morton, and G. S. Parcel. Children and fitness: a public health perspective. Res. Q. Exerc. Sport 58:295-303, 1987.
29. Taoli, E. and E. L. Wynder. Effect of the age at which smoking begins on frequency of smoking in adulthood. N. Engl. J. Med. 325:968-969, 1991.
30. Treiber, F. A., L. Musante, S. Hartdagan, H. Davis, M. Levy, and W. B. Strong. Validation of a heart rate monitor with children in laboratory and field settings. Med. Sci. Sports Exerc. 21:338-342, 1989.
31. U.S. Department Of Health And Human Services.Surgeon General's Report on Nutrition and Health. Washington, DC: U.S. Department of Health and Human Services, 1988, pp. 83-135.
    32. U.S. Department Of Health And Human Services.Healthy People 2000: National Health and Disease Prevention Objectives (full report, with commentary). Washington, DC: Department of Health and Human Services. Publication 91:50212, 1991, pp. 91-125.
    33. Webber, L. S., S. R. Srinivasan, W. A. Wattigney, and G. S. Berensen. Tracking of serum lipids and lipoproteins from childhood to adulthood. Am. J. Epidemiol. 133:884-899, 1991.


    ©1996The American College of Sports Medicine