FINAL REPORT OF WORKSHOP PROCEEDINGS AND RECOMMENDATIONS
Compiled by: Amy Branum, M.S.P.H., National Center for Health Statistics, CDC; Janet Fulton, Ph.D., National Center for Chronic Disease Prevention and Health Promotion, CDC; Mary L. Hediger, Ph.D., NICHD, NIH; Richard Troiano, Ph.D., National Cancer Institute, NIH
PURPOSE AND FORMAT OF WORKSHOP
As part of planning for the National Children’s Study (Study), John Himes, Ph.D., M.P.H., University of Minnesota, non-federal co-chair of the Nutrition, Growth, and Pubertal Development Working Group, and Heidi Kalkwarf, Ph.D., R.D., Cincinnati Children’s Hospital Medical Center, core member of the Early Origins of Adult Health Working Group, proposed a white paper to review physical activity measurement across age groups. However, it was decided that a workshop be held to obtain more timely information from a group of experts. Physical activity will be a covariate in testing various hypotheses.
The 2-day workshop had the following objectives:
- Determine the current state of knowledge about methodologies used to assess physical activity during pregnancy, infancy, childhood, and adolescence
- Determine the validity, feasibility, strengths, and limitations of measurement methods during each developmental period
- Recommend measures at specific developmental periods to capture unique information
- Recommend measures to be used in a longitudinal study of children’s health.
Experts were invited to discuss measurement of physical activity in children or adolescents, longitudinal study design, and measurement of physical activity in one of five age groups: pregnancy, infancy (1 to 2 years), preschool (2 to 5 years), elementary school (6 to 11 years), and adolescence (12 to 19 years). Table 1 lists experts and their topics. Speakers were asked to talk about specific measurements that were most appropriate to the Study, particularly as they related to specific Study hypotheses (Table 2). Experts were also asked to consider the option of using standard measurements across age groups.
The second day of the workshop, participants gathered in age-specific breakout sessions to: (1) recommend age-specific measurements and (2) recommend studies to validate measures.
MEASURING PHYSICAL ACTIVITY
The following factors were considered:
- Feasibility (for example, the costs of collecting physical activity data in 100,000 subjects)
- The comparability of methods across ages
- Differences in measurements by ethnicity, culture, or socio-economic status
- The need to anticipate technological advances for use in a longitudinal study.
Measurement tools include questionnaires, motion sensors, direct observation, and biological measures. Various metrics have been used, and their detailed descriptions, as well as strengths and limitations, can be found elsewhere (Montoye, 1996). Questionnaires, motion sensors, and direct observation were considered for use in the Study.
Questionnaires, either self-report or interviewer-administered, are commonly used for adolescent participants to obtain data on the frequency, intensity, duration, and type of physical activity. Information from questionnaires is limited by the participant’s recall of activity; however, they are able to provide a breadth of information and can include information on sedentary behaviors. Several reviews have been written on the reliability and validity of physical activity questionnaires for youth (Baranowski, 1988; Sallis, 1991, 2000).
Motion sensors, such as heart rate monitors, accelerometers, and pedometers, have also been used in studies of physical activity. Comparisons of questionnaire and accelerometer data are inconclusive. Some studies show that youth over-report vigorous activity, yet under-report moderate intensity activity (Hendelman, 2000). Nevertheless, motion sensors yield accurate information about movement, while questionnaires provide information about the type and context of physical activity.
Before deciding on measurement tools, the Study must answer the following questions:
- How much error can be tolerated in a sample of 100,000 participants?
- What are the latest relevant findings on physical activity measurement instruments?
- How can we determine compatibility of methods across age groups and systematize data collection and analysis?
- What new physical activity technologies are available and worthy of rapid development?
MEASUREMENT IN AGE-SPECIFIC GROUPS
Pregnancy
Physical activity measurement during pregnancy presents unique challenges, as metabolic changes affect nearly all organ systems. Metabolic rate and resting cardiac output increases 30-50 percent over non-pregnant levels. Physiologic and anatomical changes during pregnancy can also complicate measurement. It is important to separate non-leisure from leisure-time physical activity, as the effects may differ. Additionally, occupational physical activity is associated with certain adverse pregnancy outcomes.
Self-report questionnaires have been the primary mode used to assess physical activity among pregnant women (Ning, 2003; Hinton, 2001; Stein, 2003). In longitudinal studies, experience shows self-report may be imprecise, with overestimation of activity intensity (Stein, 2003); moreover, error may be affected by age, race/ethnicity, or body mass index (BMI).
A recent study described the pattern of energy expenditure (EE) in pregnant women using heart rate monitors, motion sensors, and questionnaires (Stein, 2003). The study showed that, compared to heart rate monitors, motion sensors underestimated EE by about 400 kcal/day, whereas questionnaires overestimated EE by a similar amount. This suggests that using a combination of motion sensors and questionnaires may provide reasonable estimates of physical activity among pregnant women.
Strategies for physical activity measurement during pregnancy for the Study include:
- Design studies to validate physical activity questionnaires among pregnant women
- Evaluate the determinants of over/underestimation via questionnaires
- Evaluate the applicability of energy cost (MET) and other equations to pregnant women.
Infancy
Activity in infants may relate to their emotions and weight gain. Since infants do not exercise in the traditional manner, their activity is thought to be relatively stable and therefore potentially easier to measure. However, activity increases with age and differs by sex (males are more active than females). In addition, activity is increased with more visual stimuli and is also affected by diet (breastfed infants are more active than formula-fed infants; infants who are malnourished or iron deficient are less active). Another challenge is that questionnaires require report by proxy (parent or caretaker), which may be unreliable. In addition, when motion sensors are used, it is not always clear if the unit is measuring the infant’s activity, the caretaker’s activity, or both.
Methods include maternal reports, observer assessments, and instrument techniques, and temperament questionnaires have been used with maternal reports to adjust for individual behavioral style. Specific maternal report tools are listed in Table 3. The mothers’ proximity to the children, ecologic validation (that is, measures "real" infant behavior), low cost, and ease of training are advantages of questionnaires. A drawback is the potential for maternal subjectivity; for example, mothers who breastfeed may be more sensitive to changes in infant behavior.
Observer assessments include activity subscales from behavioral or developmental assessments, such as the Neonatal Behavioral Assessment Scale (NBAS) or the Bayley Scales of Infant Development (BSID) (see Table 4), and tests specifically designed to capture activity, such as the Laboratory Temperament Assessment Battery (LAB TAB) (Goldsmith, 1996). The LAB TAB measures pre-locomotor and locomotor skills in a laboratory setting using a variety of stimuli. Advantages of these assessments include greater observer objectivity and more standardized behavior sampling. Disadvantages include lack of ecologic validation, the need for trained observers, and the dependency on cooperation of the infant.
Motion sensors, mainly accelerometers, have been used to a lesser extent in infants and could be used in a subsample of participants. Strengths and limitations of use of activity monitors at different ages are listed in Table 5.
In conclusion:
- Accelerometers may be the most objective tools but could be problematic for use in infants less than 2 years old due to the uncertainty of the type of activity that is being captured.
- Accelerometers should be used in conjunction with a diary of infant behavior.
- Maternal rating scales, while subjective, could be the most cost-effective tools.
Preschool-Aged Children (2-5 Years)
Preschool is a critical period for motor pattern development: limb control, body management, and locomotion. There is some debate about tracking motor skills from preschool age through childhood and their relation to physical activity, but motor pattern development is the foundation for physical activity and sport skills later in life. It is estimated that children 1 to 4 years old spend approximately 23 hours a week engaged in physical play, but much more research is needed on this activity, including:
- The relationship and importance of motor skill ability to physical activity
- The duration and intensity of physical activity
- The determinants of physical activity (and/or skillfulness)
- The stability of activity patterns over developmental time
- The relationship of physical activity/skillfulness to health consequences.
Several issues are important when choosing an instrument for preschool children. First, questionnaires have limitations: preschool children cannot reliably report physical activity, parents/caregivers are often not with the child during peak activity, and children’s relatively unstructured days make it difficult for parents/caregivers to recall activity patterns reliably. Also, motion sensors may not be valid with preschool-aged children, and cutoff points for activity intensities are unknown in this age group; additionally, inactivity may be measured as activity, such as being carried or riding (Reilly, 2002).
Alternatives for measuring preschool physical activity include measures of motor skills alone or in combination with tests of "fitness," such as the "get up and go" (time from seated position to upright), one-footed hop, or rolling a ball between pins.
Elementary School-Aged Children (6-11 Years)
Elementary school-aged children participate in a wide variety of activities, which may be highly variable from day to day; may vary by gender, age, race/ethnicity, or culture; may be sporadic; and may be dependent on external factors, such as parents, coaches, seasonality, or environment. Thus, physical activity instruments must be able to capture a variety of physical activity patterns, and maturation issues must be considered to enhance utility in longitudinal research.
The choice of instrument depends on the precision needed for the measurement and its ability to be administered among young children. Instruments used in this age range include questionnaires and motion sensors. Multiple measurement methods may allow more comprehensive assessment of the various components of physical activity.
Self-report questionnaires can be easy to administer, and older children may provide considerable detail on physical activity; however, there are still cognitive limitations among children younger than 8 years, and younger children tend to overestimate time spent in activity. Previous day recall of physical activity is most commonly used and can provide detailed information on activity but requires multiple days to reflect typical activity patterns. Previous day recall measures can either be time-based, such as the PDPAR (Weston, 1997), or activity-based, like the SAPAC (Sallis, 1997). One study showed that the SAPAC overestimated physical activity by four to five times when compared to an activity monitor (McMurray, 1998); however, there was no systematic bias in the error, indicating that a correction factor may be applicable.
Various self-report data instruments can give a more general measure of "typical physical activity" and can provide a quick indicator of general activity but little information about energy expenditure or time spent in the activity. This is not a major limitation, however, and these measures can be useful in classifying children by level of activity. An example of this type of measurement instrument includes the PAQ-C (Crocker, 1997). A relatively new technology for self-reported data is the ACTIVITYGRAM. This instrument is based on the PDPAR but uses a computer interface that can offer prompts to facilitate recall and allows data export and data tracking over time. It is currently being validated and developed for use on the Web.
Activity monitors may be used more reliably and feasibly in children of this age range (see Table 5 for the strengths and limitations of use of activity monitors in children of this age). In addition to limitations listed in Table 5, it has been shown that physical activity is often underestimated during "lifestyle" activities when using only instruments like the CSA monitor (Welk, 2000). This could be a bigger problem when used in children due to their tendency towards more sporadic activity. Reliability studies with accelerometers (MTI, Shalimar, FL) reveal an accelerometer coefficient of variation (CV) of about 10 percent for most monitor types and a CV for participants of about 20 percent (Welk, 2003). Additional work is needed to examine the variability in accelerometer output (by age and size) and to investigate reliability and calibration issues. It may also be useful to examine using a common outcome measure to facilitate comparisons across accelerometer studies.
Again, experts recommended using a combination of methods to measure physical activity in elementary school-aged children. Self-report instruments can enhance data from activity monitors by capturing the time and context of activity. Future pilot studies are needed to determine the effect of age and body size on activity counts, determine the variability between monitors, and assess error models to determine types of bias in self-report measures.
Adolescents (12-18 Years)
Measuring physical activity in adolescents has many of the same challenges as younger school-aged children; however, there are some differences. Adolescents are more likely to participate in organized activities, particularly team sports, and they are also able to report physical activity more accurately.
There are many instruments available for children 9 years and older (Sallis, 2000). Some apply interview methods, while others use self-report techniques, and most separate moderate from vigorous activity. For adolescents, these instruments have acceptable test-retest capabilities (range=0.60 to 0.98), and most show evidence of validity, although none have been validated in more than one study. One validity study reported that overestimation of physical activity is not uniform, either overall or within certain subgroups (Pate, 2002). Accelerometers are commonly used in conjunction with self-report tools to enhance objectivity and indicate types of physical activity among adolescents (please refer to Table 5 for strengths and limitations).
RECOMMENDATIONS FROM GENERAL AND BREAKOUT SESSIONS
Table 6 lists the recommended measures and metrics for the various age groups. For best results, experts also recommended the following:
1. Pregnancy
- Combine accelerometry and questionnaires to capture sufficient information on physical activity during pregnancy.
- Self-report questionnaires should provide information on activity mode (for example, leisure time, occupational, transportation) and context of physical activity (for example, school, home, or work place), sedentary behaviors, social or emotional status (including symptoms of depression), fatigue, and barriers to activity.
- Physical fitness measures (treadmill or step test) should be obtained at entry into the study.
2. Infancy (birth up to 2 years)
- Motor skill assessment should be done to obtain information about fundamental motor movement patterns.
- Accelerometry could be used to assess infant mobility and crawling.
3. Preschool-aged children (2 to 5 years)
- Fundamental motor skills assessment should be done through age 5.
- A parents/caregiver questionnaire on physical activity patterns would supplement accelerometer data and would also provide information on context of physical activity (for example, activity with the family) and sedentary behaviors.
4. Elementary school-aged children (6 to 11 years)
- A combination of accelerometry and questionnaires is recommended.
- A questionnaire administered to the parent or caregiver, will capture information on several physical activity constructs: amount and type of physical activity and sedentary behaviors, context of physical activity (school, home), purpose (transportation), and the environment for physical activity.
- To determine the context of physical activity more accurately, direct observation on a subsample of participants could be done, as direct observation on the entire study sample probably would not be feasible.
5. Adolescence (12+ years)
- Accelerometry and information obtained from questionnaires, which would be completed by the child, are most desirable.
- Questionnaires should capture the following constructs: physical activity preferences, self-efficacy, enjoyment, peer affiliation, and socialization.
- Questionnaires should also collect information on physical activity that is not captured by accelerometry (for example, cycling, swimming).
6. Pilot study recommendations
- Among preschool children, a validation study is recommended to determine the correspondence between accelerometer movement counts and activity patterns.
- To develop systematic collection and analysis methods for accelerometry, several recommendations were made:
- Develop systematic procedures to handle data coding, storage, and interpretation from accelerometers.
- Determine the optimal intensity thresholds and critical periods of data collection for Study age groups.
- Determine the hours per day and the number of days needed for valid estimates of physical activity.
- With advances in technology, calibration studies are regularly needed to ensure comparability of devices over time.
- More expertise on injury (that is types of injury) is needed to better determine the relationship between physical activity and traumatic brain injury.
- Experts strongly recommended collecting information on physical activity as a general exposure variable throughout the lifespan of the Study, particularly as an important predictor of adult chronic disease risk.
Participants
Thomas Baranowski, Ph.D., Baylor College of Medicine
Amy Branum, M.S.P.H., National Center for Health Statistics, CDC, DHHS
Elaine F. Cassidy, Ph.D., Robert Wood Johnson Foundation
William Cameron Chumlea, Ph.D., Wright State University
James F. Clapp, H.D., Case Western University
Jane E. Clark, Ph.D., University of Maryland
Patty Freedson, Ph.D., University of Massachusetts, Amherst
Edward A. Frongillo, Jr., Ph.D., Cornell University
Bernard F. Fuemmeler, Ph.D., M.P.H., National Cancer Institute, NIH, DHHS
Janet E. Fulton, Ph.D., National Center for Chronic Disease Prevention and Health Promotion, CDC, DHHS
Matthew W. Gillman, M.D., S.M., Harvard University Medical School
Eleanor Z. Hanna, Ph.D., Office of Research on Women’s Health, NIH, DHHS
Lynne Haverkos, M.D., M.P.H., NICHD, NIH, DHHS
Mary L. Hediger, Ph.D., NICHD, NIH, DHHS
Heidi J. Kalkwarf, Ph.D., Cincinnati Children’s Hospital Medical Center
Carole A. Kimmel, Ph.D., Office of Research and Development, EPA
Robert J. Kuczmarski, Dr.P.H., R.D., National Institute of Diabetes and Digestive and Kidney Diseases, NIH, DHHS
Robert M. Malina, Ph.D., Tarleton State University
Charles E. Matthews, Ph.D., Vanderbilt University Medical School
Thomas L. McKenzie, Ph.D., San Diego State University
Suzanne B. McMaster, Ph.D., Office of Research and Development, EPA
Deborah H. Olster, Ph.D., Office of Behavioral and Social Sciences Research, NIH, DHHS
Russell R. Pate, Ph.D., University of South Carolina
Mark A. Pereira, Ph.D., University of Minnesota
James M. Pivarnik, Ph.D., Michigan State University
Charlotte A. Pratt, Ph.D., R.D., National Heart, Lung, and Blood Institute, NIH, DHHS
Brian Saelens, Ph.D., Cincinnati Children’s Hospital and Medical Center
Christine G. Spain, M.S., M.A., Office of Disease Prevention and Health Promotion, Office of the Secretary, DHHS
Margarita Treuth, Ph.D., Johns Hopkins Bloomberg School of Public Health
Richard Troiano, Ph.D., National Cancer Institute, NIH, DHHS
Reta Van Orden, M.S., University of Utah
Greg Welk, Ph.D., Iowa State University
Michelle A. Williams, Sc.D., University of Washington
John Worobey, Ph.D., Rutgers University
Table 1. Invited speakers, their affiliations and topics
Speaker name
|
Affiliation |
Topic |
|
Thomas Baranowski, Ph.D. |
Department of Pediatrics, Baylor School of Medicine
|
Physical activity in the context of the Study |
|
Russell R. Pate, Ph.D. |
Dept. of Exercise Science, University of South Carolina
|
Overview of physical activity measures |
|
Edward A. Frongillo, Ph.D. |
Div. of Nutritional Sciences, Cornell University
|
Overview of longitudinal study design measurement issues |
|
Michelle A. Williams, Sc.D. |
Dept. of Epidemiology, University of Washington
|
Physical activity in pregnancy |
|
John Worobey, Ph.D. |
Dept. of Nutritional Sciences, Rutgers University
|
Physical activity during infancy |
|
Jane E. Clark, Ph.D. |
Dept. of Kinesiology, University of Maryland
|
Physical activity in preschool-aged children |
|
Gregory Welk, Ph.D. |
Dept. of Health and Human Performance, Iowa State University
|
Physical activity in elementary school-aged children |
|
Brian E. Saelens, Ph.D. |
Dept. of Pediatrics/Psychology, University of Cincinnati |
Physical activity in adolescents |
Table 2. Specific Study hypotheses and their relevance for physical activity measurement
|
Study Hypothesis
|
Role of Physical Activity |
|
Maternal stress during pregnancy is associated with increased risk of asthma. |
Careful consideration of physical activity, both leisure time and work activity, as a covariate is needed to accurately evaluate the influence of maternal stress during pregnancy on the developing immune system and risk of asthma.
|
|
Social, behavioral, and family factors that affect development of dietary preferences and physical activity patterns early in childhood determine risk of childhood obesity and insulin resistance. |
Assessment of parental dietary and activity practices may be necessary to understand their influence on the acquisition of practices among their children. Childcare and school influences will also need to be evaluated.
|
|
Breast milk feeding, compared with infant formula feeding, and breastfeeding duration are associated with lower rates of obesity and lower risk of insulin resistance.
|
Although the exposure variable will be measured early on, it will be important to assess physical activity level as a covariate in older children when determining obesity and insulin resistance in older children.
|
|
Dietary predictors of obesity and insulin resistance include reduced intake of fiber and whole grains, and high glycemic index.
|
When measuring dietary intake and obesity, physical activity will be an important covariate. |
|
Environmental factors such as distance to parks, availability of walking routes in the neighborhood, and neighborhood safety are associated with risk of obesity and insulin resistance.
|
Physical activity will need to be assessed as an exposure measure. |
|
Repeated head trauma has a cumulative adverse effect on neurocognitive development.
|
Physical activity will need to be measured as an important covariate for these hypotheses. Information on type and frequency of physical activity, particularly participation in sports, will need to be collected. |
Table 3. Examples of maternal report tools to quantify infant activity
|
Tool |
Age Range |
What They Contain |
Reference |
|
Early Infant Temperament Questionnaire
|
1–4 months |
8 items that measure activity |
Medoff-Cooper, 1993 |
|
Revised Infant Temperament Questionnaire
|
4–8 months |
13 items that measure activity |
Carey, 1995 |
|
Toddler Temperament Scale
|
12–24 months |
12 items that measure activity |
Fullard, 1984 |
|
Infant Behavior Questionnaire |
1–12 months |
17 items that measure activity level
|
Rothbart, 1981 |
|
Toddler Behavior Assessment Questionnaire |
18–24 months |
20 items that measure activity level |
Goldsmith, 1996 |
Table 4. Description of NBAS and BSID tools
|
Test
|
Age Range |
Measurement |
|
Neonatal Behavioral Assessment Scale (NBAS) |
0–28 days |
“Activity level” scored from 1 “None” to 9 “Continuous, inconsolable movement”
|
|
Bayley Scales of Infant Development (BSID) |
6–24 months and older |
“Hyperactivity” scored from 1 “Consistently hyperactive; fidgety and agitated in movement” to 5 “Consistently not hyperactive; never fidgety or agitated in movement”
|
Table 5. General advantages and disadvantages of using accelerometers in the Study
|
Advantages
|
Disadvantages |
|
Gives objective measurement of PA, which is also not influenced by ethnicity, culture, or socioeconomic status
|
Cannot assess all forms of activity (for example cycling, swimming) |
|
Allows for continuous measurement |
Can be hard to control individual variability in responses to activity
|
|
Relatively easy to use |
Limitations of prediction equations for estimating EE or time spent in activity
|
|
Typically not burdensome to participant |
Unintended removal |
Table 6. Recommended types and timing of physical activity measures
|
|
|
Entry into study
|
Pregnancy— weeks gestation |
Age group |
|
Recommended measurement |
Construct |
|
16 ±2 |
26 ±1 |
32 ±2 |
1 year |
Preschool |
Elementary
school |
Adolescence |
|
Accelerometry |
Physical activity
|
P |
P |
P |
P |
P a |
P |
P b |
P |
|
Questionnaire |
Physical activity/ sedentary behavior
|
P |
P |
P |
P |
P |
P c |
P c |
P d |
|
Motor skill milestones |
Fundamental motor skills
|
|
|
|
|
Pe |
P |
|
|
|
Direct observation |
Physical activity
|
|
|
|
|
P a |
P f |
P g |
|
|
Cardiorespiratory fitness |
Fitness level |
P |
|
|
|
|
|
|
|
aAccelerometry can be performed in crawling infants
bAccelerometry for this age group can alternate every other year between being done in the whole sample and then half of the sample
cSelf-report/questionnaire data should be collected by the parent for these age groups
dSelf-report/questionnaire data can be collected by the adolescent participant
eDuring the first year, motor milestones should be collected at 3, 6, 9, 12 months of age
fDirect observation can be done on half of study sample
gDirect observation can be done in the whole study sample at ages 6, 8, and 10 years
References
Baranowski T. Validity of self report of physical activity: An information processing approach. Research Quarterly for Exercise and Sport, 59(4):314-327, 1988.
Carey WB, McDevitt SC and Associates (1995). Carey Temperament Scales (Activity subscales). San Antonio, TX: The Psychological Corporation.
Crocker PR, Bailey DA, Faulkner RA, Kowalski KC, McGrath R. Measuring general levels of physical activity: preliminary evidence for the Physical Activity Questionnaire for Older Children. Medicine and Science in Sports and Exercise. 1997;29:1344-9.
Fullard W, McDevitt SC, Carey WB. Assessing temperament in one- to three-year-old children. Journal of Pediatric Psychology.1984;9:205-17.
Goldsmith, HH. Studying temperament via construction of the Toddler Behavior Assessment Questionnaire (Activity subscale). Child Development. 1996;67:218-235.
Goldsmith HH & Rothbart MK (1996). Laboratory Temperament Assessment Battery (LAB-TAB): Prelocomotor and locomotor versions (Activity subscales). Madison, WI: University of Wisconsin.
Hendelman D, Miller K, Baggett C, Debold E, Freedson P. Validity of accelerometry for the assessment of moderate intensity physical activity in the field. Medicine and Science in Sports and Exercise. 2000;32:S442-9.
Hinton PS, Olson CM. Predictors of pregnancy-associated change in physical activity in a rural white population. Maternal Child Health Journal. 2001;5:7-14.
McMurray RG, Harrell JS, Bradley CB, Webb JP, Goodman EM. Comparison of a computerized physical activity recall with a triaxial motion sensor in middle-school youth. Medicine and Science in Sports and Exercise. 1998;30:1238-45.
Medoff-Cooper B, Carey WB, McDevitt SC. The Early Infancy Temperament Questionnaire. Journal of Development and Behavioral Pediatrics. 1993;14:230-5.
Montoye HJ, Kemper HCG, Saris WHM, Washburn RA. Measuring physical activity and energy expenditure. Champaign, IL: Human Kinetics.
Ning Y, Williams MA, Dempsey JC, Sorensen TK, Frederick IO, Luthy DA. Correlates of recreational physical activity in pregnancy. Journal of Maternal, Fetal, and Neonatal Medicine. 2003;13:385-93.
Pate RR, Freedson PS, Sallis JF, Taylor WC, Sirard J, Trost SG, Dowda M. Compliance with physical activity guidelines: prevalence in a population of children and youth. Annals of Epidemiology. 2002;12:303-8.
Reilly JJ, Coyle J, Kelly L, Burke G, Grant S, Paton JY. An objective method for measurement of sedentary behavior in 3- to 4-year olds. Obesity Research. 2003;11:1155-8.
Rothbart MK. Measurement of temperament in infancy. (Activity subscale of the Infant Behavior Questionnaire). Child Development. 1981; 52:569-578.
Sallis JF. Self-report measures of children’s physical activity. Journal of School Health. 1991;61:215-9.
Sallis JF, Saelens BE. Assessment of physical activity by self report: status, limitations, and future directions. Research Quarterly for Exercise and Sport. 71;2000:1-14.
Sallis JF, Strikmiller PK, Harsha DW, Feldman HA, Ehlinger S, Stone EJ, Williston J, Woods S. Validation of interviewer- and self-administered physical activity checklists for fifth grade students. Medicine and Science in Sports and Exercise. 1996;28:840-51.
Stein AD, Rivera JM, Pivarnik JM. Measuring energy expenditure in habitually active and sedentary pregnant women. Medicine and Science in Sports and Exercise.2003;35:1441-6.
Welk GJ, Blair SN, Wood K, Jones S, Thompson RW. A comparative evaluation of three accelerometry-based physical activity monitors. Medicine and Science in Sports and Exercise. 2000;32:S489-97.
Welk GJ, Almeida J, Morss G. Laboratory calibration and validation of the Biotrainer and Actitrac activity monitors. Medicine and Science in Sports and Exercise. 2003;35:1057-64.
Weston AT, Petosa R, Pate RR. Validation of an instrument for measurement of physical activity in youth. Medicine and Science in Sports and Exercise. 1997;29:138-43.
Appendix A-Examples of physical activity studies currently in the field
LEAP: Promotion of Physical Activity in High School Girls
TAAG: The Trial of Activity for Adolescent Girls
http://www.cscc.unc.edu/taag/desc.php
Appendix B-References to instruments
Actical ® Physical Activity Monitor
http://www.minimitter.com
MicroMini-Motionlogger ®
Actigraph Monitor
http://www.mtiactigraph.com
Biotrainer
http://www.imsystems.net
Tritrac
http://www.stayhealthy.com
ACTIVITYGRAM
http://www.fitnessgram.net
Comprehensive list of downloadable measures and surveys
http://www-rohan.sdsu.edu/faculty/sallis/measures.html