The study population consisted of 189 children; including 92 (48.7%) boys and ..... Chinapaw, M. J., L. B. Mokkink, M. N. van Poppel, W. van Mechelen, and C. B..
Pediatric Exercise Science, 2012, 24, 229-245 © 2012 Human Kinetics, Inc.
Measuring and Validating Physical Activity and Sedentary Behavior Comparing a Parental Questionnaire to Accelermeter Data and Diaries Bettina Bringolf-Isler Swiss Tropical and Public Health Institute, University of Basel
Urs Mäder and Nicole Ruch Federal Institute of Sport
Susi Kriemler, Leticia Grize, and Charlotte Braun-Fahrländer Swiss Tropical and Public Health Institute, University of Basel Accurately measuring children’s physical activity and their sedentary behavior is challenging. The present study compared 189 parental responses to a questionnaire surveying physical activity and sedentary behavior of children aged 6–14 years, to accelerometer outputs and time activity diaries for the same group. Responses were analyzed taking age, sex and maternal education into account. Correlation coefficients between questionnaire reports and accelerometer-based physical activity across all age groups were acceptable (up to r = .55). Yet, adjustment for age markedly attenuated these associations, suggesting concomitant influences of biological and behavioral processes linked to age. The comparisons of general time indications in the questionnaire with 24h-diary records suggested that parents tended to under- and over-report single activities, possibly due to social desirability. We conclude that physical activity questionnaires need to be designed for specific age groups and be administered in combination with objective measurements.
Background It is widely accepted that regular physical activity is beneficial for children’s health (22), whereas sedentary behavior is a risk factor for chronic diseases (20,27), independent of physical activity levels. However, accurately measuring children’s
Bringolf-Isler, Kriemler, Grize, and Braun-Fahrländer are with the Swiss Tropical and Public Health Institute and the University of Basel, Basel, Switzerland. Mäder and Ruch are with the Federal Institute of Sport, Magglingen, Switzerland. 229
230 Bringolf-Isler et al.
physical activity and their sedentary behavior remains challenging. Compared with adults, children perform more spontaneous movements, resulting in brief bouts of varied activities with frequent rest periods (2), which are difficult to capture with questionnaires. Accelerometers are the preferred instrument for objectively assessing the amount of physical activity, but their use in large samples is timeconsuming and expensive. In addition, they do not provide information about the type of activity that the child engages in (34). Yet, knowledge about these behavioral patterns is important to efficiently promote physical activity. Several physical activity questionnaires for youth have been developed (9), but only a few of them also focus on sedentary behavior (8). Most physical activity questionnaires have shown limited validity, especially in children below the age of ten (9,25). Younger children accumulate high levels of physical activity during unstructured play (4), which is more challenging to capture with a questionnaire than structured sports activities. Questionnaires focusing mainly on structured sports might overlook other relevant sources of physical activity (30). For example, it has been observed that children from low-income families have less access to sports facilities than their higher-income counterparts, but are not less physically active (32). This contributes to discrepancies between self-reported time spent in moderate to vigorous physical activity (MVPA) and accelerometer measurements, as recently reported by a study of Dutch adolescents with different educational backgrounds (29). Sedentary behavior (SB) is distinct from a lack of physical activity (PA) and its assessment is challenging. A recently published systematic review of the validity and reliability of sedentary behavior measures in children and adolescents concluded that most studies found reliable estimates of SB but their validity remained largely untested (16) The aim of the current study was to evaluate the validity and reliability of single and combined questionnaire items for assessing sedentary behavior and physical activity in children by comparing parentally reported SB and PA with accelerometry based time spent in sedentary activity (SA) and MVPA. In addition, we evaluated the relevance of single and combined items for different age groups of children.
Methods Study Population Participants were part of a larger cross-sectional study (3) that included primary school children aged six to ten years (including kindergarten, first grade, and fourth grade), and adolescents aged 13–14 years (eighth grade) living in one of three Swiss communities (Bern, Biel-Bienne and Payerne). A flow-chart of the study design is shown in Figure 1. From a random sample of 19 school classes (at least two classes per grade and community), all children whose parent(s) completed questionnaires (n = 236) were invited to wear an accelerometer device and to complete a time-activity diary (4). The minimum wearing time for accelerometers was set at 10 hr a day for at least three weekdays and one weekend day. 167 (70.7%) children fulfilled these criteria. For diary reports to be included, 80% of the entries into a 24h protocol had to be available and the diary had to be completed for three weekdays and one weekend day. 125 (66.1%) of the children fulfilled these criteria. 147 parents
Measuring Physical Activity and Sedentary Behavior 231
completed the questionnaires again after two months and six months. The personal and social characteristics of the children invited for additional measurements did not differ significantly from those that were not invited. The ethics committee of the canton of Bern approved the study protocol and written informed consent was obtained from the children’s parents.
Physical Activity Questionnaire The inquiry questionnaire was administered in autumn 2004 as a parent proxyreport, mainly using an activity-based approach (18). Questions were structured around a list of activities and asked about activity behavior on a typical weekday and weekend day. Respondents reported how much time (minutes per day) the child performed a specific activity. The following leisure time activities were included to capture sedentary behavior: watching television/video/DVD, sitting at a computer/playing Nintendo/electronic games, doing homework, playing a musical instrument, reading, playing quietly or performing other quiet activities, and to assess active behavior: playing vigorously indoors, playing vigorously outdoors, and time spent sweating or breathing hard. The list of activities was pretested in 50 children and the most commonly reported activities were included in the final questionnaire. Answers were categorized as: never, less than 30 min, 30 min to one hour, one to two hours, two to four hours and more than four hours per day. To obtain an overall estimate of time spent in a specific activity on weekdays and on the weekend, the categorized time-intervals were transformed into a continuous variable using the intervals’ midpoints and an assumed maximum of 8 hr. From a weekly time, a daily average was calculated. A sensitivity analysis indicated that
Figure 1 — Schematic of the study design.
232 Bringolf-Isler et al.
results did not change when the minimum or maximum value of each time-interval was used for the calculations instead of the median of the interval (5). To further assess children’s’ moderate to vigorous physical activities parents were also asked how many hours per day the child spent outdoors, and how many hours of sports training per week the child attended outside of school (in Switzerland, all children attend three physical education classes per week at school, but sports trainings take place in clubs, usually after school). In addition, parents indicated how many days a week they performed physical activities together with their children and evaluated whether their child was “much more active”, “more active”, “equally active”, “less active” or “much less active” than other children of the same age and sex. The categories were subsequently condensed into “more active”, “equally active”, and “less active”, as the distribution was skewed. Finally, the usual mode of transport to school was assessed and categorized as ‘active’ if children walked or biked to school and as ‘non active’ if they used public transport or a car. All questions asking about the usual amount of time spent in leisure time activities were asked again after two months in winter and six months in spring/summer to test the stability of responses.
Accelerometer Measurements Actigraph accelerometers (Model AM7164, formerly Computer Science and Applications (CSA), now Manufacturing Technology Inc. (MTI), Fort Walton Beach, FL) were worn during two seven day periods, one in winter 2004 and one in spring/ summer 2005, to assess average physical activity. The device measures the acceleration of the body on the vertical axis by taking 40 measurements per second and integrating these acceleration signals continuously to the chosen epoch time of one minute. To compare questionnaire responses with accelerometer measurements, metabolic equivalents (MET) were calculated using a published age-dependant equation (14). MET levels greater than or equal to three were defined as MVPA. The equation does not provide a cut-off for sedentary behavior, thus an arbitrary threshold of less than100 counts/min was defined. This threshold has been used in previous studies in children (11,12,31). Minutes of sedentary behavior between 06:00 and 21:00 hr were considered for analysis. To obtain mean minutes of MVPA or SA per day, a weighted average of weekday MVPA and SA (multiplied by five) and of weekend MVPA and SA (multiplied by two) divided by seven was calculated. In this paper the abbreviation MVPA and SA are used exclusively for accelerometer-based activity monitoring.
Time-Activity Diary (Physical Activity Record) The newly developed and validated diary (4) required a record of the same leisure time activities as asked in the questionnaire. For each 15-min interval between 06:00 and 22:00 hr and each hour between 22:00 and 06:00 hr, parents and caregiver respectively had to indicate the child’s activity. It was recommended that this was done together with the child. Adolescents filled in the diary themselves. The diary was completed during two four-day periods, one in winter and one in spring/summer, including two weekdays and the weekend. Diary recording was concomitant with the accelerometer measurements. To make records representa-
Measuring Physical Activity and Sedentary Behavior 233
tive for the whole week, weighting for weekdays and the weekend were carried out as described above.
Data Analyses All analyses were conducted with STATA 9.0. To test the validity of the questionnaire, Spearman correlation coefficients were calculated across all age groups between the reported time spent in each type of physical activity and accelerometerbased MVPA and SA, respectively. Then, correlations adjusting for grade, sex and maternal education were calculated for all children and stratified by age group. For all variables except for times spent outdoors, the variable had to be transformed using the square root or a logarithm. For categorical variables, multivariate linear regression models adjusting for grade, sex and maternal education were calculated. In a next step, we evaluated whether a combination of active or sedentary behaviors as reported in the questionnaire better predicted total MVPA and SA than single activities. The ‘active behavior score’ included the sum of minutes spent vigorously playing indoors and outdoors, cycling and sports training. The ‘sedentary behavior score’ totalled the minutes spent watching television, sitting at a computer, doing homework, reading, playing a musical instrument, other quiet activities and traveling by car/public transport. Again, analyses were stratified by age group. The stability of children’s usual PA and SB measured during different seasons was tested by calculating the mean of the differences and the intraclass correlation coefficients between the weekly number of minutes spent in a given activity reported at baseline, two months later, and six months later. Time indications in the questionnaire were then compared with the diary reports. Separate analyses were run for weekdays and weekends. For each time category of a given activity as reported in the questionnaire, the corresponding mean duration according to the diary was calculated. The results were displayed graphically except when the time category included less than three children.
Results Questionnaire Data The study population consisted of 189 children; including 92 (48.7%) boys and 97 (51.3%) girls. 112 (64.7%) were primary school-aged children (28 in kindergarten, 18 first graders, 70 fourth graders) and 71 (35.3%) adolescents (eighth grade). Table 1 shows the questionnaire reported duration and frequency of different leisure time activities across all age groups and stratified by primary school children and adolescents. In general, primary school children reported spending more time with physical activities, whereas adolescents spent more time in sedentary activities. On average, children reported spending more than three hours per day outdoors, but much less time was devoted to vigorous play outdoors (Tab 1).
Accelerometer Data On average children spent 153.4 min/day (95% CI: 141.5 min/d to 166.2 min/d) in accelerometer-based MVPA and 482.9 min/day (95% CI: 471.5 min/d to 494.3
234
min/day min/day min/day min/day min/day min/day
Sitting at a computer
Doing homework
Reading/looking at books
Playing a musical instrument
Playing quietly
Traveling by car/public transport
min/day min/day min/day
Playing vigorously active indoors
Playing vigorously active outdoors
Cycling min/day min/day
Time spent breathing hard and sweating
Attending sport training (outside school)
Sports activities:
min/day
Time spent outdoors
Active behavior:
min/day
Watching television
Sedentary behavior: 15.0 (4.3–45.0)
8.6 (0.0–17.1)
60.0 (38.6–77.1)
23.6 (15.0–57.9)
57.9 (23.6–90.0)
15.0 (4.3–51.4)
196.3 (136.8–249.7)
12.9 (4.3–23.6)
45.0 (23.6–77.1)
10.7 (0.0–15.0)
23.6 (15.0–45.0)
32.1 (10.7–45.0)
8.6 (0.0–17.1)
66.4 (38.6–77.1)
23.6 (15.0–57.9)
69.6 (45.0–115.7)
23.6 (4.3–57.9)
213.4 (162.4–265.4)
12.9 (4.3–23.6)
45.0 (23.6–77.1)
10.7 (0.0–15.0)
15.0 (15.0–45.0)
15.0 (0.0–36.4)
15.0 (0.0–23.6)
45.0 (23.6–57.9)
Median (p25, p75)
Median (p25, p75) 57.9 (23.6–83.6)
Primary school children (n = 118)
Total (n = 189)
Table 1 Duration and Frequency of Parental Reported Leisure Time Activities in School-Aged Children, Stratified by Age Group
8.6 (0.0–25.7)
55.7 (30.0–77.1)
15.0 (4.3–36.4)
23.6 (15.0–62.1)
15.0 (0.0–30.0)
156.9 (114.4–208.3)
15.0 (4.3–25.7)
45.0 (15.0–57.9)
12.9 (0.0–23.6)
36.4 (15.0–77.1)
45.0 (36.4–77.1)
36.4 (15.0–57.9)
57.9 (45.0–90.0)
Median (p25, p75)
Adolescents (n = 71)
Measuring Physical Activity and Sedentary Behavior 235
min/d) in SA. Boys were more active than girls and physical activity levels decreased by grade. Activity levels did not differ with maternal education levels. Minutes spent in accelerometer based sedentary behavior showed a similar but reciprocal pattern.
Comparison Between Main Questionnaire and Accelerometer Data Table 2a presents the correlations between parents’ reported time spent in sedentary activities and accelerometer based SA. In the unadjusted correlations, 12 out of 16 tested questionnaire items were significantly associated with total accelerometer based SA. However, when the analyses were adjusted for age (four grades), sex, and maternal education these correlations decreased considerably and none of the individual items or the sedentary activity score was significantly associated with SA. Age had the strongest impact on attenuating the associations with accelerometer data, gender and maternal education had almost no effect (data not shown). Questionnaire items such as sitting at a computer, reading, playing a musical instrument or traveling by car/public transport were even negatively correlated with SA. When stratified by age-group crude and adjusted correlations were similar among adolescents covering an age-range of 13–14 yrs. Among primary school children aged 6–10 yrs adjustment for age attenuated the correlations. Table 2b shows the comparison of parents’ reported time spent in active behaviors with accelerometer based MVPA. Time spent vigorously playing outdoors, time spent in sports training (for adolescents), physical activity bouts with the parents, physical activity assessment compared with other children and the active behavior score were the only questionnaire items remaining significantly and positively associated with total MVPA after adjustment. The active behavior score did not substantially improve the correlation between accelerometer and questionnaire data.
Comparison Between the Main Questionnaire and the Repeated Questionnaires The stability of reported usual time spent with specific activities at baseline, two months later in winter, and six months later in spring/summer is shown in Table 3. For most activities, the mean time difference was small and non- significant. Two months after the first survey, children were reported to spend significantly more time watching television and reading, and less time playing vigorously outdoors. Six months after the first survey, no significant differences in reported usual time spent in specific activities were noted, except for time spent vigorously playing indoors. For most activities, the intraclass correlation coefficients were moderate to high and statistically significant. Taken together, the results indicate that parents are fairly consistent in reporting usual physical activity, although some variations due to seasonal influences are to be expected.
Comparison Between the Main Questionnaire and the Diary Figure 2 illustrates the comparison between questionnaire-reported and diaryreported time spent in specific activities. For each category of time indicated in the questionnaire, the mean duration spent in a given activity, according to the diary,
236
0.12 -0.10 0.05
Playing a musical instrument
Playing quietly/other quiet activities
Traveling by car/public transport
0.53***
Doing homework 0.32
0.32***
Sitting at a computer/playing Nintendo/ electronic games
Reading
0.32***
Watching television/video/DVD
-0.03
0.01
-0.01
-0.06
0.03
-0.03
0.00
0.01
0.01
-0.08
-0.03
-0.01
-0.06
0.08
-0.04
-0.10
Spearman Correlation 0.55***
Spearman Correlation Adjusted for school grade, sex and maternal education
Spearman Correlation Adjusted for school grade, sex and maternal education
Sedentary behavior scorea
Sedentary behavior (min/day)
Primary school children (n = 109)
All (n = 167)
All (n = 167)
Table 2 a. Comparison of Parent’s Reported Sedentary Behavior With Accelerometer-Based SA
-0.10
0.20
-0.02
-0.14
0.13
-0.13
0.06
0.19
Spearman Correlation Adjusted for sex and maternal education
Adolescents (n = 58)
237
0.33*** 0.30*** 0.48*** 0.19* 0.15(*) –0.15(*)
Time spent outdoors
Playing vigorously active indoors
Playing vigorously active outdoors
Cycling
Time spent breathing hard and sweating
Attending sports training (outside school)
0.11
0.07
0.11
0.19*
0.08
0.10
0.21*
0.03
–0.01
0.12
0.21*
–0.01
0.06
0.17
Spearman Correlation 0.46**
Spearman Correlation Adjusted for school grade, sex and maternal education
Spearman Correlation Adjusted for school grade, sex and maternal education
Active behavior scoreb
Active behavior (min/day):
Primary school children
All
All
Table 2 b. Comparison of Parents’ Reported Physical Activity With Accelerometer-Based MVPA
(continued)
0.29*
0.18
0.19
0.06
0.20
0.08
0.24
Spearman Correlation Adjusted for sex and maternal education
Adolescents
238
All
Reference 53.4 (24.9; 81.9)*** 17.9 (-8.4–44.3)
Equally active More active
29.3 (-50.6 to -8.0)**
Less active
No
46.4 (15.6–77.1)***
Two or more times a week Reference
7.1 (-18.0–32.2)
Once a week
Yes
Reference
Never
15.7 (3.3–28.0)*
12.2 (-0.2–24.6)(*)
Reference
0.0 (-10.5–10.4)
Reference
13.3 (0.4–26.1)*
5.9 (-4.9–16.7)
Reference
Regression adjusted for school grade, sex Univariate regression and maternal education model (min/day) (min/day)
All
9.6 (-11.0–30.4)
17.3 (-1.2–35.9)(*)
Reference
-8.1 (-21.1–5.3)
Reference
17.5 (0.6–34.4)*
2.2 (-12.6–17.1)
Reference
Regression adjusted for school grade, sex and maternal education (min/day)
Primary school
17.6 (4.7–30.4)**
-4.4 (-8.4–9.7)
Ref
7.1 (-6.2–20.6)
Ref
1.6 (-17.8–21.1)
7.2 (-7.7–22.1)
Ref
Regression adjusted sex and maternal education (min/day)
Adolescents
Accelerometer based MVPA (min/day) associated with each category of activity Mean (95% CI)
of all sedentary activities; bcombination of time spent vigorously playing indoors, vigorously playing outdoors, cycling and sport training; (*)p < 0.1; * p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001
acombination
Comparison with other children of the same age and sex
Active commuting to school
Frequency of physical activity bouts with parents
Categorical variables
Table 2 b. (continued)
239
1.9 (-3.7–7.6) 2.1 (-0.6–10.3) -8.7 (-17.6–0.1)
-10.0 (-19.2 to -0.8)*
Playing a musical instrument
Playing quietly
Playing vigorously indoors
Playing vigorously outdoors
The significance of all intraclass correlations (ICC) was ≤0.001
-1.4 (-7.2–4.5)
6.1 (0.2–11.9)*
Reading
Cycling
-2.3 (-7.3–2.7)
Doing homework
-0.1 (-3.6–3.3)
2.9 (-1.8–7.7)
Sitting at a computer
Traveling by car/public transport
6.8 (0.8–12.7)*
Watching television
Mean difference (min/day)
0.64 (0.54–0.74)
0.49 (0.36–0.62)
0.43 (0.29–0.57)
0.41 (0.27–0.55)
0.42 (0.29–0.55)
0.34 (0.20–0.49)
0.64 (0.54–0.73)
0.56 (0.45–0.67)
0.61 (0.50–0.71)
0.60 (0.49–0.70)
ICC (95% CI)
After two months
2.6 (-4.4–9.7)
3.7 (-0.8–8.3)
-3.1 (-11.3–5.1)
-8.3 (-14.2 to -2.4)*
1.1 (-8.3–10.4)
-0.9 (-4.6–2.9)
-0.6 (-5.5–4.3)
-3.0 (-8.5–2.5)
2.2 (-3.6–8.0)
4.6 (-3.3–12.5)
Mean difference (min/day)
0.45 (0.31–0.58)
0.43 (0.29–0.57)
0.60 (0.49–0.71)
0.67 (0.58–0.75)
0.32 (0.17–0.47)
0.59 (0.48–0.71)
0.54 (0.42–0.66)
0.59 (0.48–0.70)
0.44 (0.30–0.57)
0.49 (0.36–0.62)
ICC (95% CI)
After six months
Table 3 Reliability of Questionnaire Reported Duration of Leisure Time Activities Two and Six Months Later
240 Bringolf-Isler et al.
Figure 2 — Mean minutes/day spent in diary recorded leisure time activities plotted against the respective time indications in the questionnaire The range of minutes covered by the time categories of the questionnaire (x-axis) is shadowed in gray. Results are shown for sedentary and active behavior and stratified by weekday and weekend.
is shown. Analyses were stratified by weekday and weekend and are displayed separately for active and sedentary behavior. For most sedentary activities, an increase in questionnaire reported time was paralleled by an increase in mean time reported in the diary, indicating good agreement in relative terms. Parents reported duration of reading, sitting at a computer and actively playing outdoors was quite similar according to both instruments. Questionnaire and diary reports agreed less for “actively playing indoors” and “cycling”. In absolute terms, questionnaires tended to underestimate time spent watching TV and quiet activities and to overreport the duration of playing a musical instrument, doing homework, actively playing indoors and cycling.
Discussion Accurate assessment of physical activity and sedentary behavior is important to understanding the association between physical activity behavior and health, to monitor secular trends, and to evaluate the effectiveness of interventions. The present study evaluated the validity of parent completed questionnaire items on usual physical activity and sedentary behavior for use in large epidemiological studies. When the questionnaire responses were compared with MVPA and SA, significant associations ranging between r = .19 and r = .53 were found for many of the items when children of all age groups were considered. Yet, adjustment for age markedly
Measuring Physical Activity and Sedentary Behavior 241
attenuated these associations, in particular for items assessing sedentary activities, suggesting concomitant influences of biological and behavioral processes linked to age. The results also indicate that parents’ indications about children’s usual active and sedentary behavior were fairly consistent when tested in different seasons. Comparisons of questionnaire responses to activities reported in the diaries were rather good for reported active play outdoors. Yet socially desirable sedentary activities tended to be over-reported, while TV watching was under-reported. In recent years, many different physical activity questionnaires have been developed and assessed for reliability and validity. Systematic reviews concluded that none showed both acceptable reliability and validity (9) and that even less is known about the validity of sedentary behavior assessment (16). The results of the current study support these findings and additionally highlight the importance of considering age in a validation study. On the one hand several items relating to active behavior were significantly associated with MVPA, however, different items were relevant for different age groups. On the other hand age might act as a confounder. With increasing age a biological decline in physical activity occurs (23) and time spent with sedentary activities such as TV watching or doing homework increases while time spent playing outdoors decreases with age. A high correlation between a specific behavior and SA or MVPA might thus arise because both are linked to age and not necessarily because they are causally related. However, limiting analyses to narrow age-groups might reduce the range of time spent in a given activity within this age-group and thus reduce the power to find a strong correlation with accelerometry based SA or MVPA. Yet, it would be desirable that questionnaire items assessing active and sedentary behavior would be valid within and across age groups. Several Actigraph-specific MVPA cut-points are available for use in children, each producing different estimates of MVPA. The use of any particular cut-off point, though necessary, is a limitation of accelerometers and a possible source of discrepancy when compared with other MVPA estimates (such as self reporting). However, in the present analysis the use of mean counts per minute instead of cutoff points showed comparable results after adjustment for age (data not shown). The study further fond differential reporting of single sedentary activities, which seem to be influenced by social desirability of the specific behavior. Such unbalanced parental reporting might contribute to the low correlation between self-report and objectively measured SA. Screen activities such as watching television or sitting at a computer have been used in several surveys as an indicator for inactive behavior (17). In the current study, the assessment of sedentary behavior with single items or combined as a “sedentary behaviour score” was only correlated with SA across but not within age-groups. These findings suggest that assessing sedentary behavior by questionnaire shows considerable differences to accelerometer based SA and may not be appropriate. The summation of minutes spent in active behavior expressed as an active behavior score did not increase the correlation with accelerometer based MVPA compared with time spent in specific activities. “Vigorously playing outdoors” was the activity that showed the strongest association with MVPA, in primary school children. This is in line with data from previous studies showing that younger children’s MVPA minutes are mainly made up of unstructured play (7,21) if it is of sufficient intensity (4). “Time spent outdoors”, which was a useful physical activity
242 Bringolf-Isler et al.
measure in younger children in American (6,26) and Australian studies (35), was less strongly associated with total MVPA in the current study. Most likely, such questions are context dependent. The duration of leisure time activities seems to differ between Swiss children and American or Australian children. Wen (35) reported that more than one third of the 10–12 year old children from Sydney, Australia, spent less than half an hour per day playing outdoors and more than 40% spent more than two hours watching television or playing computer games after school. In the current study, parents reported that their children spent on average more than three hours playing outdoors (including weekends) and few spent more than one hour watching TV. Thus, asking parents about their children’s time playing outdoors may have a different implication in different countries. Organized sport plays a more important role for overall MVPA in adolescents than for younger children, as has been reported by others (21), yet, as only a short spell of the day is spent with this activity its absolute contribution to average MVPA in adolescents was modest. Active commuting to school has been associated with increased MVPA in some studies (10,28) but not in others (13,19). In the current study, the group of children actively commuting to school walked or biked to school. This is still the main mode of transport for the majority of Swiss school children (3,15). Those using public transport were classified as ‘nonactive’, even though they might have walked partly. As adolescents more often use public transport because of longer distances to school the strict definition of “active transport” might have introduced stronger misclassification among adolescents, and resulted in a non significant and even negative correlation with SA. In general, mode of transport to school did not contribute much to overall MVPA in our sample, as most children live in very close proximity to their schools. Comparing a child’s physical activity level with that of other children as a relative measure of physical activity discriminated between more and less active children. The item has been found to be valid in adults when evaluated against health outcomes (33) and maximal oxygen uptake (1). However, as a relative measure it does not allow quantification of physical activity and it is not suitable for assessing secular trends. The study had some limitations. First, the sample size for stratified analyses was relatively small. Second, the transformation of interval data into a time-score was necessary to obtain an overall measure of physical activity, but it can only be considered an approximation. It was, however, reassuring to note that the calculations were not sensitive to the underlying assumptions. Third, the epoch length of one minute is rather long for research involving children and potentially reduces MVPA estimates (24). Finally the strict inclusion criterion reduced the diary sample considerably.
Conclusions The items and scores tested in this study were designed to measure physical activity and sedentary behavior in a broad age-range of children through proxy reporting by their parents. Results indicated that correlations between questionnaire reports and accelerometer based PA and SA were acceptable when compared across age groups but were markedly attenuated when adjusted for age, suggesting concomitant
Measuring Physical Activity and Sedentary Behavior 243
influences of biological and behavioral processes linked to age. Thus, future studies intending to relate physical activity to health or to monitor secular trends have to be designed specifically for different age groups and cannot rely on questionnaires only, but should combine questionnaires with objective measurements. Acknowledgments The authors are grateful to their many colleagues in the School Health Services of Bern, Biel and Payerne who organized the survey. We would also like to thank the children, parents and teachers for their enthusiastic co-operation, which made this investigation possible. The study was supported by a grant of the Federal Commission of Sport (ESK).
References 1. Ainsworth, B. E., D. R. Jacobs, Jr., and A. S. Leon. Validity and reliability of selfreported physical activity status: the Lipid Research Clinics questionnaire. Med Sci Sports Exerc. 25:92-98.1993. 2. Bailey, R. C., J. Olson, S. L. Pepper, J. Porszasz, T. J. Barstow, and D. M. Cooper. The level and tempo of children’s physical activities: an observational study. Med Sci Sports Exerc. 27:1033-1041.1995. 3. Bringolf-Isler, B., L. Grize, U. Mader, N. Ruch, F. H. Sennhauser, and C. BraunFahrlander. Personal and environmental factors associated with active commuting to school in Switzerland. Prev Med. 46:67-73.2008. 4. Bringolf-Isler, B., L. Grize, U. Mader, N. Ruch, F. H. Sennhauser, and C. BraunFahrlander. Assessment of intensity, prevalence and duration of everyday activities in Swiss school children: a cross-sectional analysis of accelerometer and diary data. Int J Behav Nutr Phys Act. 6:50.2009. 5. Bringolf-Isler, B., L. Grize, U. Mader, N. Ruch, F. H. Sennhauser, and C. BraunFahrlander. Built environment, parents’ perception, and children’s vigorous outdoor play. Prev Med. 50:251-256.2010. 6. Burdette, H. L., R. C. Whitaker, and S. R. Daniels. Parental report of outdoor playtime as a measure of physical activity in preschool-aged children. Arch Pediatr Adolesc Med. 158:353-357.2004. 7. Burdette, H. L., and R. C. Whitaker. Resurrecting free play in young children: looking beyond fitness and fatness to attention, affiliation, and affect. Arch Pediatr Adolesc Med. 159:46-50.2005. 8. Chinapaw, M. J., S. M. Slootmaker, A. J. Schuit, M. van Zuidam, and W. van Mechelen. Reliability and validity of the Activity Questionnaire for Adults and Adolescents (AQuAA). BMC Med Res Methodol. 9:58.2009. 9. Chinapaw, M. J., L. B. Mokkink, M. N. van Poppel, W. van Mechelen, and C. B. Terwee. Physical activity questionnaires for youth: a systematic review of measurement properties. Sports Med. 40:539-563.2010. 10. Cooper, A. R., L. B. Andersen, N. Wedderkopp, A. S. Page, and K. Froberg. Physical activity levels of children who walk, cycle, or are driven to school. Am J Prev Med. 29:179-184.2005. 11. Ekelund, U., S. J. Griffin, and N. J. Wareham. Physical activity and metabolic risk in individuals with a family history of type 2 diabetes. Diabetes Care. 30:337-342.2007. 12. Ekelund, U., S. Brage, S. J. Griffin, and N. J. Wareham. Objectively measured moderate- and vigorous-intensity physical activity but not sedentary time predicts insulin resistance in high-risk individuals. Diabetes Care. 32:1081-1086.2009.
244 Bringolf-Isler et al.
13. Ford, P., R. Bailey, D. Coleman, K. Woolf-May, and I. Swaine. Activity levels, dietary energy intake, and body composition in children who walk to school. Pediatr Exerc Sci. 19:393-407.2007. 14. Freedson, P., D. Pober, and K. F. Janz. Calibration of accelerometer output for children. Med Sci Sports Exerc. 37:S523-530.2005. 15. Grize, L., B. Bringolf-Isler, E. Martin, and C. Braun-Fahrlander. Trend in active transportation to school among Swiss school children and its associated factors: three cross-sectional surveys 1994, 2000 and 2005. Int J Behav Nutr Phys Act. 7:28.2010. 16. Lubans, D.R., K. Hesketh, D.P. Cliff, et al. A systematic review of the validity and reliability of sedentary behaviour measures used with children and adolescents. Obes. Rev. 12:781–799. 17. Marshall, S. J., S. J. Biddle, T. Gorely, N. Cameron, and I. Murdey. Relationships between media use, body fatness and physical activity in children and youth: a metaanalysis. Int J Obes Relat Metab Disord. 28:1238-1246.2004. 18. McMurray, R. G., K. B. Ring, M. S. Treuth, G. J. Welk, R. R. Pate, K. H. Schmitz, J. L. Pickrel, et al. Comparison of two approaches to structured physical activity surveys for adolescents. Med Sci Sports Exerc. 36:2135-2143.2004. 19. Metcalf, B., L. Voss, A. Jeffery, J. Perkins, and T. Wilkin. Physical activity cost of the school run: impact on schoolchildren of being driven to school (EarlyBird 22). Bmj. 329:832-833.2004. 20. Mitchell, J. A., C. Mattocks, A. R. Ness, S. D. Leary, R. R. Pate, M. Dowda, S. N. Blair, et al. Sedentary behavior and obesity in a large cohort of children. Obesity (Silver Spring). 17:1596-1602.2009. 21. Nilsson, A., L. B. Andersen, Y. Ommundsen, K. Froberg, L. B. Sardinha, K. Piehl-Aulin, and U. Ekelund. Correlates of objectively assessed physical activity and sedentary time in children: a cross-sectional study (The European Youth Heart Study). BMC Public Health. 9:322.2009. 22. Rowland, T. Physical Activity, Fitness, and Children. In: Physical Activity and Health. Bouchard, C., S. Blair, andW. Haskell (Ed.). Champaign, Illinois: Human Kinetics, 2007, pp. 424. 23. Rowland, T. W. The biological basis of physical activity. Med Sci Sports Exerc. 30:392399.1997. 24. Rowlands, A. V. Accelerometer assessment of physical activity in children: an update. Pediatr Exerc Sci. 19:252-266.2007. 25. Sallis, J. F., M. J. Buono, J. J. Roby, F. G. Micale, and J. A. Nelson. Seven-day recall and other physical activity self-reports in children and adolescents. Med Sci Sports Exerc. 25:99-108.1993. 26. Sallis, J. F., J. J. Prochaska, and W. C. Taylor. A review of correlates of physical activity of children and adolescents. Med Sci Sports Exerc. 32:963-975.2000. 27. Sardinha, L. B., L. B. Andersen, S. A. Anderssen, A. L. Quiterio, R. Ornelas, K. Froberg, C. J. Riddoch, et al. Objectively measured time spent sedentary is associated with insulin resistance independent of overall and central body fat in 9- to 10-year-old Portuguese children. Diabetes Care. 31:569-575.2008. 28. Sirard, J. R., B. E. Ainsworth, K. L. McIver, and R. R. Pate. Prevalence of active commuting at urban and suburban elementary schools in Columbia, SC. Am J Public Health. 95:236-237.2005. 29. Slootmaker, S. M., A. J. Schuit, M. J. Chinapaw, J. C. Seidell, and W. van Mechelen. Disagreement in physical activity assessed by accelerometer and self-report in subgroups of age, gender, education and weight status. Int J Behav Nutr Phys Act. 6:17.2009. 30. Timperio, A., D. Crawford, A. Telford, and J. Salmon. Perceptions about the local neighborhood and walking and cycling among children. Prev Med. 38:39-47.2004.
Measuring Physical Activity and Sedentary Behavior 245
31. van Sluijs, E.M., N.R. Jones, A.P. Jones, S.J. Sharp, F. Harrison, and S.J. Griffin. School-level correlates of physical activity intensity in 10-year-old children. Int. J. Pediatr. Obes., 2010. 32. Voss, L. D., J. Hosking, B. S. Metcalf, A. N. Jeffery, and T. J. Wilkin. Children from low-income families have less access to sports facilities, but are no less physically active: cross-sectional study (EarlyBird 35). Child Care Health Dev. 34:470-474.2008. 33. Washburn, R. A., L. L. Adams, and G. T. Haile. Physical activity assessment for epidemiologic research: the utility of two simplified approaches. Prev Med. 16:636-646.1987. 34. Welk, G. J., C. B. Corbin, and D. Dale. Measurement issues in the assessment of physical activity in children. Res Q Exerc Sport. 71:S59-73.2000. 35. Wen, L. M., J. Kite, D. Merom, and C. Rissel. Time spent playing outdoors after school and its relationship with independent mobility: a cross-sectional survey of children aged 10-12 years in Sydney, Australia. Int J Behav Nutr Phys Act. 6:15.2009.
