AMERICAN JOURNAL OF HUMAN BIOLOGY 28:129–137 (2016)
Original Research Article
Quantitative Physical Activity Assessment of Children and Adolescents in a Rural Population from Eastern Nepal KIMBERLY D. WILLIAMS,1,2* JANARDAN SUBEDI,3 BHARAT JHA,4 JOHN BLANGERO,5 SARAH WILLIAMS-BLANGERO,5 AND BRADFORD TOWNE6 1 Department of Anthropology, Temple University, Philadelphia, Pennsylvania 2 Institute for Genomic and Evolutionary Medicine, Temple University, Philadelphia, Pennsylvania 3 Department of Sociology and Gerontology, Miami University, Oxford, Ohio 4 Tribhuvan University Institute of Medicine, Kathmandu, Nepal 5 South Texas Diabetes and Obesity Institute, University of Texas Health Science Center Regional Academic Health Center, Harlingen, Texas 6 Department of Community Health, Lifespan Health Research Center, Boonshoft School of Medicine, Wright State University, Dayton, Ohio
Objectives: We report cross-sectional, objectively measured physical activity data for 399 children and adolescents aged 6 to 18 years. We evaluated physical activity of children and adolescents, considered time spent in each activity intensity category, and explored the impact of growth disruption (stunting and wasting) on physical activity patterns. Methods: Participants wore an Actical (Mini-Mitter, Bend, OR) omnidirectional accelerometer for one week as part of their annual visit to the Jiri Growth Study. The percentage of time spent in standard activity intensities were computed using standard metabolic equivalents (METS) cutpoints and compared by chronological age, sex, and school versus non-school days. Results: Primary findings include (1) children are more active on non-school days and adolescents are more active during the school week; (2) Jirel children do not exhibit the reduction in physical activity that most Western populations experience during the transition from childhood to adolescence; and (3) Jirel children and adolescents routinely meet the suggested one hour/day MVPA threshold; (4) Stunting is prevalent and factors leading to this growth disruption may contribute to the amount of time in sedentary or light physical activity. Conclusions: We report child and adolescent physical activity patterns from the Jirel population of eastern Nepal. In this rural context, children and adolescents are more active than populations reported from Western contexts. This key finding has important biomedical implications for the maintenance of healthy body composition, skeletal health, C 2015 Wiley Periodicals, Inc. V and other health traits. Am. J. Hum. Biol. 28:129–137, 2016. A physically active lifestyle during childhood and adolescence has potential benefits for growth and development as well as adult health. Evidence of these benefits comes from studies that demonstrate positive links between physical activity and body composition (Berkey et al., 2003; Boddy et al., 2014; Guinhouya, 2012; Hay et al., 2012; Lohman et al., 2006; Pizarro et al., 2013; Trost et al., 2013; Ward et al., 2006), and physical activity and cardiovascular health (Hay et al., 2012; Palve et al., 2014; Schaefer et al., 2014). Development of the skeleton (Sayers et al., 2011; Scerpella et al., 2011) and a possible relationship between levels of physical activity and development of motor skills (Cliff et al., 2009; Lopes et al., 2012) are also known. Researchers have worked to elucidate the specific ways in which physical activity may be beneficial. For example, some have compared types of physical activity common to children and adolescents (Duncan et al., 2008) or the amount of time spent engaging in specific intensity levels (Barnes et al., 2013; Martinez-Gomez et al., 2010) in an effort to identify the type of activity that will most often result in positive health outcomes. These inquiries are often complicated by the relationships between body composition, physical activity, nutrition, fitness, and health, which are complex and influenced by a broad range of sociocultural, environmental, and genetic factors. Human biologists seeking to understand how sociocultural practices and barriers, health status, and changing subsistence systems influence physical activity patterns and energetics (e.g., Sherar et al., 2011; Trost et al., 2005) now more routinely use accelerometers, which objectively measure physical activity. The placement of the accelerC 2015 Wiley Periodicals, Inc. V
ometer (e.g. ankle versus hip versus wrist) can influence data collection, and accelerometry may not be able to effectively record specific types of body movements (Snodgrass, 2012; Luzner et al., 2014). On the other hand, recent work has demonstrated that there is good to moderate agreement between accelerometers and ground force plates with regard to the capture of vertical ground reaction forces (Pouliot-Laforte et al., 2014), and others have shown that the Actical accelerometer in particular can discern small changes in physical activity (Montoye et al., 2014b). These recent findings, as well as strengths like freedom from errors in recall memory or incomplete questionnaire information, outweigh the known limitations of this data collection technique. This is especially the case for assessment of physical activity in remote settings where it might be difficult to travel with heavy or awkward equipment or to set up a specialized field laboratory. The ease of use, small size, and sturdy nature of many new accelerometers promotes the consideration of full days of activity assessment, which is especially important in the study of childhood activity. Contract grant sponsor: NIH; Contract grant numbers: HD053206 (KDW); HD40377 (BT); AI370919 (SW-B); AI44406 (SW-B); MH59490 (JB). *Correspondence to: Kimberly D. Williams, Department of Anthropology, 1115 Polett Walk, 210 Gladfelter Hall, Temple University, Philadelphia, PA 19120. E-mail: [email protected] Received 2 February 2015; Revision received 27 April 2015; Accepted 16 June 2015 DOI: 10.1002/ajhb.22762 Published online 16 July 2015 in Wiley Online Library (wileyonlinelibrary.com).
130
K.D. WILLIAMS
This study presents cross-sectional, objectively measured physical activity data for a non-Western population. These data differ from many of the physical activity data published on Western pediatric populations where factors such as TV viewing and active versus passive transportation must be considered. The data presented here are from a population of children and adolescents who engage primarily in active transportation (walking), live in a rugged rural environment, and participate in a variety of subsistence and domestic activities outside of school. We examine how physical activity varies between the sexes, between children and adolescents, and between school and non-school days in this type of free-living context. Finally, these data contribute to the relatively slim literature regarding growth disruption (stunting and wasting) and physical activity levels.
Over the course of this study (2005–2012), the physical activity of 790 individuals (393 boys, 397 girls) was monitored for a total of 1,448 monitoring periods. Data files were excluded from analyses in the following cases: (1) the accelerometer was removed for any portion of the study period, (2) the device suffered mechanical error, (3) operator error (incorrect epoch length, anthropometrics, and/or participant identification), or (4) the participant opened the accelerometer. Data were further limited to the first visit from each participant who had six continuous days of physical activity data, thus creating a crosssectional sample consisting of boys and girls aged 6 to 18 years. School days and non-school days (Saturdays) were compared in order to consider the impact of leisure day activities.
MATERIALS AND METHODS Study population
Z-scores were calculated for height-for-age, weight-forage, and BMI-for-age based on WHO growth standards (de Onis et al., 2007). Descriptive statistics are presented in Table 1 (boys) and Table 2 (girls). The percentage of participants who were stunted (23 SD < z-score heightfor-age < 22 SD) and severely stunted (z-score height-forage < 23 SD) as well as those wasted (23 SD < z-score BMI-for-age < 22 SD) and severely wasted (z-score BMIfor-age < 23SD) were aggregated into the variables “stunted” and “wasted,” respectively. In addition to comparing z-scores for participants by physical activity level, we examined the relationships between the aggregated stunted and wasted categories as well. Data were reduced in the SAS statistical program (version 9.4) using a modified version of the technique described by the Canadian Health Measures Survey (Colley, 2012; www.haloresearch.ca/accel). This important tool provides SAS code to process the large datasets produced specifically by the Actical accelerometer. This is similar to the SAS code available from the National Cancer Institute to process NHANES Actiwatch data (http://appliedresearch.cancer.gov/nhanes_pam). Standardization of data reduction methods aids comparisons across studies that use the same data collection devices. Beyond aiding data reduction, the use of this code ensures the use of standard definitions of physical activity intensity defined by the same MET cut points (Table 3). The choice of MET cut points is important for understanding these types of physical activity, especially when considered as time spent in each category of physical activity rather than total or mean activity counts. Inconsistent definitions of activity intensity may lead to conflicting results (Bailey et al., 2013). We modified the data reduction process to remove the code for analyses of steps. As we did not record steps for all participants, we do not consider those data here. Additionally, our data were collected in 15-second epochs rather than 60-second epochs. The data reduction method allowed for easy aggregation of these epochs in order to consider physical activity count and intensity per minute. The resulting data were analyzed in SPSS (version 22). Mean percentage of the data collection period for each physical activity intensity level: sedentary (SED), light (LPA), moderate (MPA), and vigorous (VPA) were then calculated. MPA and VPA were aggregated to produce the moderate-vigorous activity variable (MVPA). We evaluated school days and non-school days (Saturdays) separately to calculate the (a) the percentage of each age/sex
The Jirels have been the focus of extensive genetic and epidemiological studies for over 25 years (e.g., Blangero, 1988, 1990; Relethford and Blangero, 1990; WilliamsBlangero, 1989a,1989b,1990a, 1993, 1998, 1999, 2002; Williams-Blangero and Blangero, 1989c1990b, WilliamsBlangero et al., 2002). All subjects in this study were participants in the Jiri Growth Study (Williams et al., 2007, 2012, 2013, 2015), a genetic epidemiological study of growth and development. The Jiri Growth Study began in 2000, and from 2005 to 2012, objective physical activity assessment was collected. The Jirel children in this study attend school 6 days a week; Saturday is the only non-school day. The Temple University and the Wright State University Institutional Review Boards for Human Subjects Research approved all informed consent procedures; informed consent was obtained from parents of all subjects prior to participation. Children assented to the data collection protocol. The Nepal Health Research Council also approved the data collection protocol and operation of the Jiri Helminth Project Clinic and Jiri Growth Study. Physical activity assessment Participants were recruited and data were collected by a physician and several assistants who worked for the project in Nepal on a year-round basis. Jiri Growth Study data were collected every Sunday and Thursday morning except during harvest time (late June/early July) and during the Diwali holiday celebration (October). The Actical omnidirectional accelerometer (Mini-Mitter, Bend, OR) was used to collect physical activity data. This device was chosen because of the sturdy, waterproof casing. Data were collected every 15 seconds (15-second epoch). Young children engage in frequent spurts of activity (Bailey et al., 1995; Baquet et al., 2007; Berman et al., 1998); the 15-second data collection epoch was used to capture as much variation in children’s activity as possible. Children were fitted with an accelerometer during their scheduled annual Jiri Growth Study visit. The accelerometer was placed on the left hip just superior to the iliac crest and the children were instructed to wear the accelerometer 24 hours a day, including during times they were bathing or sleeping. Parents were also asked to ensure that the accelerometers were worn at all times unless the belt became uncomfortable. No participants reported that the belt or the device was uncomfortable. American Journal of Human Biology
Anthropometry and growth standards
131
QUANTITATIVE PHYSICAL ACTIVITY ASSESSMENT OF CHILDREN AND ADOLESCENTS TABLE 1. Descriptive Statistics for Male Jirel Children and Adolescents Age 6 yrs (n 5 16) 7 yrs (n 5 26) 8 yrs (n 5 19) 9 yrs (n 5 17) 10 yrs (n 5 16) 11 yrs (n=18) 12 yrs (n=19) 13 yrs (n 5 26) 14 yrs (n 5 17) 15 yrs (n=17) 16 yrs (n 5 13) 17 yrs (n 5 10) 18 yrs (n 5 3)
WT
HT
BMI
zHT/Agea
zBMI/Agea
zWT/Agea,b
16.84 (2.15) 17.77 (1.71) 22.21 (6.79) 21.79 (2.18) 23.25 (3.89) 24.67 (3.09) 29.03 (5.82) 30.65 (6.45) 34.70 (6.86) 37.71 (6.18) 40.46 (8.71) 47.05 (4.95) 44.83 (6.71)
107.04 (5.54) 112.49 (5.57) 121.83 (10.42) 121.52 (5.30) 126.16 (7.28) 130.43 (4.95) 136.35 (7.69) 138.74 (9.18) 143.71 (7.96) 149.02 (9.49) 154.13 (11.25) 158.56 (4.61) 156.30 (2.17)
14.70 (1.40) 14.04 (0.93) 14.68 (1.37) 14.76 (1.05) 14.50 (0.99) 14.45 (1.12) 15.49 (1.91) 15.75 (1.47) 16.72 (2.07) 16.87 (1.19) 16.80 (1.64) 18.68 (1.40) 18.31 (2.21)
21.81 (1.13) 21.75 (1.06) 20.96 (1.84) 21.84 (0.88) 21.80 (1.14) 21.88 (0.74) 21.80 (1.08) 22.33 (1.24) 22.53 (1.04) 22.55 (1.21) 22.41 (1.45) 22.17 (0.60) 22.65 (0.30)
20.59 (1.26) 21.20 (0.83) 20.87 (0.91) 20.97 (0.80) 21.35 (0.79) 21.75 (0.90) 21.39 (1.21) 21.50 (0.87) 21.35 (1.22) 21.50 (0.75) 21.91 (0.98) 21.10 (0.68) 21.53 (1.07)
21.60 (0.99) 22.00 (0.77) 21.28 (1.53) 21.86 (0.70) 22.07 (1.16)
School day wear time
Saturday wear time
55.76 (6.31) 56.38 (15.24) 58.35 (13.96) 57.48 (10.04) 53.53 (7.23) 52.30 (6.45) 51.33 (7.26) 48.56 (8.00) 50.52 (7.05) 49.12 (8.54) 53.69 (20.64) 58.82 (23.59) 45.41 (8.95)
11.62 (1.36) 11.69 (3.13) 11.20 (2.03) 11.79 (1.87) 10.91 (1.19) 10.63 (2.29) 10.57 (2.10) 10.44 (2.10) 11.17 (1.73) 10.52 (1.85) 11.87 (3.81) 12.37 (4.44) 10.53 (1.04)
School day wear time
Saturday wear time
60.09 (11.63) 56.19 (8.19) 66.91 (22.38) 53.99 (4.79) 51.40 (6.66) 50.04 (5.94) 54.09 (8.66) 47.88 (7.03) 48.71 (7.16) 49.92 (9.93) 51.71 (10.51) 50.30 (10.19) 44.59 (11.10)
12.25 (2.42) 11.27 (1.63) 13.42 (4.44) 10.68 (1.38) 10.22 (1.82) 10.58 (2.03) 11.55 (1.99) 9.61 (2.07) 10.14 (2.04) 10.37 (2.49) 11.06 (2.36) 10.58 (2.59) 9.90 (1.98)
Data are presented as mean (SD) in kilograms, centimeters, and hours. a Z-scores are calculated using WHO reference population (de Onis, 2012). b Z weight-for-age up to age 10 years, per WHO guidelines.
TABLE 2. Descriptive Statistics for Female Jirel Children and Adolescents Age 6 yrs (n 5 11) 7 yrs (n 5 17) 8 yrs (n 5 14) 9 yrs (n 5 20) 10 yrs (n 5 19) 11 yrs (n 5 20) 12 yrs (n 5 16) 13 yrs (n 5 11) 14 yrs (n 5 19) 15 yrs (n 5 20) 16 yrs (n 5 12) 17 yrs (n 5 11) 18 yrs (n 5 6)
WT
HT
BMI
zHT/Agea
zBMI/Agea
zWT/Agea,b
16.32 (1.90) 17.03 (2.10) 19.64 (1.77) 22.15 (3.54) 26.11 (7.48) 26.80 (3.83) 29.69 (6.23) 35.05 (9.36) 39.29 (7.41) 39.95 (6.52) 41.75 (4.93) 45.54 (6.99) 49.25 (8.09)
108.05 (4.09) 111.75 (4.91) 117.66 (4.66) 123.33 (5.96) 129.77 (9.07) 132.46 (7.05) 136.49 (8.56) 144.18 (8.23) 148.31 (6.80) 148.19 (6.63) 146.35 (4.40) 151.31 (4.33) 152.08 (8.33)
13.95 (1.05) 13.58 (0.78) 14.18 (0.92) 14.48 (1.34) 15.27 (2.09) 15.21 (1.01) 15.79 (2.02) 16.65 (3.22) 17.72 (2.18) 18.11 (2.08) 19.49 (2.14) 19.90 (2.55) 21.20 (1.81)
21.38 (0.80) 21.66 (0.90) 21.53 (0.80) 21.50 (0.98) 21.39 (1.42) 21.89 (1.06) 22.16 (1.25) 21.76 (1.19) 21.65 (0.98) 21.96 (0.96) 22.38 (0.65) 21.73 (0.65) 21.66 (1.26)
20.99 (0.85) 21.31 (0.65) 21.00 (0.66) 21.07 (0.97) 20.91 (1.15) 21.13 (0.58) 21.24 (1.08) 21.29 (1.59) 20.89 (0.93) 20.95 (1.02) 20.53 (0.85) 20.50 (0.91) 20.06 (0.63)
21.57(0.86) 21.96 (0.91) 21.65(0.63) 21.67(1.12) 21.49 (1.33)
Data are presented as mean (SD) in kilograms, centimeters, and hours. a Z-scores are calculated using WHO reference population (de Onis, 2012). b Z weight-for-age up to age 10 years, per WHO guidelines.
TABLE 3. Physical Activity Variables Used In Data Reduction (After Table 1, Colley, R. 2012. Actical Accelerometer Data Analysis Support Tool: Harmonizing With The Canadian Health Measures Survey (Accel1)) Physical activity variables Wear time 5 24 hours—non-wear time (hours) Non-wear time 5 60 consecutive minutes of zero counts Valid day 5 101 hours of wear time Intensity cut-points: METS Accelerometer count range (cpm) Sedentary (SED) 1
Original Research Article
Quantitative Physical Activity Assessment of Children and Adolescents in a Rural Population from Eastern Nepal KIMBERLY D. WILLIAMS,1,2* JANARDAN SUBEDI,3 BHARAT JHA,4 JOHN BLANGERO,5 SARAH WILLIAMS-BLANGERO,5 AND BRADFORD TOWNE6 1 Department of Anthropology, Temple University, Philadelphia, Pennsylvania 2 Institute for Genomic and Evolutionary Medicine, Temple University, Philadelphia, Pennsylvania 3 Department of Sociology and Gerontology, Miami University, Oxford, Ohio 4 Tribhuvan University Institute of Medicine, Kathmandu, Nepal 5 South Texas Diabetes and Obesity Institute, University of Texas Health Science Center Regional Academic Health Center, Harlingen, Texas 6 Department of Community Health, Lifespan Health Research Center, Boonshoft School of Medicine, Wright State University, Dayton, Ohio
Objectives: We report cross-sectional, objectively measured physical activity data for 399 children and adolescents aged 6 to 18 years. We evaluated physical activity of children and adolescents, considered time spent in each activity intensity category, and explored the impact of growth disruption (stunting and wasting) on physical activity patterns. Methods: Participants wore an Actical (Mini-Mitter, Bend, OR) omnidirectional accelerometer for one week as part of their annual visit to the Jiri Growth Study. The percentage of time spent in standard activity intensities were computed using standard metabolic equivalents (METS) cutpoints and compared by chronological age, sex, and school versus non-school days. Results: Primary findings include (1) children are more active on non-school days and adolescents are more active during the school week; (2) Jirel children do not exhibit the reduction in physical activity that most Western populations experience during the transition from childhood to adolescence; and (3) Jirel children and adolescents routinely meet the suggested one hour/day MVPA threshold; (4) Stunting is prevalent and factors leading to this growth disruption may contribute to the amount of time in sedentary or light physical activity. Conclusions: We report child and adolescent physical activity patterns from the Jirel population of eastern Nepal. In this rural context, children and adolescents are more active than populations reported from Western contexts. This key finding has important biomedical implications for the maintenance of healthy body composition, skeletal health, C 2015 Wiley Periodicals, Inc. V and other health traits. Am. J. Hum. Biol. 28:129–137, 2016. A physically active lifestyle during childhood and adolescence has potential benefits for growth and development as well as adult health. Evidence of these benefits comes from studies that demonstrate positive links between physical activity and body composition (Berkey et al., 2003; Boddy et al., 2014; Guinhouya, 2012; Hay et al., 2012; Lohman et al., 2006; Pizarro et al., 2013; Trost et al., 2013; Ward et al., 2006), and physical activity and cardiovascular health (Hay et al., 2012; Palve et al., 2014; Schaefer et al., 2014). Development of the skeleton (Sayers et al., 2011; Scerpella et al., 2011) and a possible relationship between levels of physical activity and development of motor skills (Cliff et al., 2009; Lopes et al., 2012) are also known. Researchers have worked to elucidate the specific ways in which physical activity may be beneficial. For example, some have compared types of physical activity common to children and adolescents (Duncan et al., 2008) or the amount of time spent engaging in specific intensity levels (Barnes et al., 2013; Martinez-Gomez et al., 2010) in an effort to identify the type of activity that will most often result in positive health outcomes. These inquiries are often complicated by the relationships between body composition, physical activity, nutrition, fitness, and health, which are complex and influenced by a broad range of sociocultural, environmental, and genetic factors. Human biologists seeking to understand how sociocultural practices and barriers, health status, and changing subsistence systems influence physical activity patterns and energetics (e.g., Sherar et al., 2011; Trost et al., 2005) now more routinely use accelerometers, which objectively measure physical activity. The placement of the accelerC 2015 Wiley Periodicals, Inc. V
ometer (e.g. ankle versus hip versus wrist) can influence data collection, and accelerometry may not be able to effectively record specific types of body movements (Snodgrass, 2012; Luzner et al., 2014). On the other hand, recent work has demonstrated that there is good to moderate agreement between accelerometers and ground force plates with regard to the capture of vertical ground reaction forces (Pouliot-Laforte et al., 2014), and others have shown that the Actical accelerometer in particular can discern small changes in physical activity (Montoye et al., 2014b). These recent findings, as well as strengths like freedom from errors in recall memory or incomplete questionnaire information, outweigh the known limitations of this data collection technique. This is especially the case for assessment of physical activity in remote settings where it might be difficult to travel with heavy or awkward equipment or to set up a specialized field laboratory. The ease of use, small size, and sturdy nature of many new accelerometers promotes the consideration of full days of activity assessment, which is especially important in the study of childhood activity. Contract grant sponsor: NIH; Contract grant numbers: HD053206 (KDW); HD40377 (BT); AI370919 (SW-B); AI44406 (SW-B); MH59490 (JB). *Correspondence to: Kimberly D. Williams, Department of Anthropology, 1115 Polett Walk, 210 Gladfelter Hall, Temple University, Philadelphia, PA 19120. E-mail: [email protected] Received 2 February 2015; Revision received 27 April 2015; Accepted 16 June 2015 DOI: 10.1002/ajhb.22762 Published online 16 July 2015 in Wiley Online Library (wileyonlinelibrary.com).
130
K.D. WILLIAMS
This study presents cross-sectional, objectively measured physical activity data for a non-Western population. These data differ from many of the physical activity data published on Western pediatric populations where factors such as TV viewing and active versus passive transportation must be considered. The data presented here are from a population of children and adolescents who engage primarily in active transportation (walking), live in a rugged rural environment, and participate in a variety of subsistence and domestic activities outside of school. We examine how physical activity varies between the sexes, between children and adolescents, and between school and non-school days in this type of free-living context. Finally, these data contribute to the relatively slim literature regarding growth disruption (stunting and wasting) and physical activity levels.
Over the course of this study (2005–2012), the physical activity of 790 individuals (393 boys, 397 girls) was monitored for a total of 1,448 monitoring periods. Data files were excluded from analyses in the following cases: (1) the accelerometer was removed for any portion of the study period, (2) the device suffered mechanical error, (3) operator error (incorrect epoch length, anthropometrics, and/or participant identification), or (4) the participant opened the accelerometer. Data were further limited to the first visit from each participant who had six continuous days of physical activity data, thus creating a crosssectional sample consisting of boys and girls aged 6 to 18 years. School days and non-school days (Saturdays) were compared in order to consider the impact of leisure day activities.
MATERIALS AND METHODS Study population
Z-scores were calculated for height-for-age, weight-forage, and BMI-for-age based on WHO growth standards (de Onis et al., 2007). Descriptive statistics are presented in Table 1 (boys) and Table 2 (girls). The percentage of participants who were stunted (23 SD < z-score heightfor-age < 22 SD) and severely stunted (z-score height-forage < 23 SD) as well as those wasted (23 SD < z-score BMI-for-age < 22 SD) and severely wasted (z-score BMIfor-age < 23SD) were aggregated into the variables “stunted” and “wasted,” respectively. In addition to comparing z-scores for participants by physical activity level, we examined the relationships between the aggregated stunted and wasted categories as well. Data were reduced in the SAS statistical program (version 9.4) using a modified version of the technique described by the Canadian Health Measures Survey (Colley, 2012; www.haloresearch.ca/accel). This important tool provides SAS code to process the large datasets produced specifically by the Actical accelerometer. This is similar to the SAS code available from the National Cancer Institute to process NHANES Actiwatch data (http://appliedresearch.cancer.gov/nhanes_pam). Standardization of data reduction methods aids comparisons across studies that use the same data collection devices. Beyond aiding data reduction, the use of this code ensures the use of standard definitions of physical activity intensity defined by the same MET cut points (Table 3). The choice of MET cut points is important for understanding these types of physical activity, especially when considered as time spent in each category of physical activity rather than total or mean activity counts. Inconsistent definitions of activity intensity may lead to conflicting results (Bailey et al., 2013). We modified the data reduction process to remove the code for analyses of steps. As we did not record steps for all participants, we do not consider those data here. Additionally, our data were collected in 15-second epochs rather than 60-second epochs. The data reduction method allowed for easy aggregation of these epochs in order to consider physical activity count and intensity per minute. The resulting data were analyzed in SPSS (version 22). Mean percentage of the data collection period for each physical activity intensity level: sedentary (SED), light (LPA), moderate (MPA), and vigorous (VPA) were then calculated. MPA and VPA were aggregated to produce the moderate-vigorous activity variable (MVPA). We evaluated school days and non-school days (Saturdays) separately to calculate the (a) the percentage of each age/sex
The Jirels have been the focus of extensive genetic and epidemiological studies for over 25 years (e.g., Blangero, 1988, 1990; Relethford and Blangero, 1990; WilliamsBlangero, 1989a,1989b,1990a, 1993, 1998, 1999, 2002; Williams-Blangero and Blangero, 1989c1990b, WilliamsBlangero et al., 2002). All subjects in this study were participants in the Jiri Growth Study (Williams et al., 2007, 2012, 2013, 2015), a genetic epidemiological study of growth and development. The Jiri Growth Study began in 2000, and from 2005 to 2012, objective physical activity assessment was collected. The Jirel children in this study attend school 6 days a week; Saturday is the only non-school day. The Temple University and the Wright State University Institutional Review Boards for Human Subjects Research approved all informed consent procedures; informed consent was obtained from parents of all subjects prior to participation. Children assented to the data collection protocol. The Nepal Health Research Council also approved the data collection protocol and operation of the Jiri Helminth Project Clinic and Jiri Growth Study. Physical activity assessment Participants were recruited and data were collected by a physician and several assistants who worked for the project in Nepal on a year-round basis. Jiri Growth Study data were collected every Sunday and Thursday morning except during harvest time (late June/early July) and during the Diwali holiday celebration (October). The Actical omnidirectional accelerometer (Mini-Mitter, Bend, OR) was used to collect physical activity data. This device was chosen because of the sturdy, waterproof casing. Data were collected every 15 seconds (15-second epoch). Young children engage in frequent spurts of activity (Bailey et al., 1995; Baquet et al., 2007; Berman et al., 1998); the 15-second data collection epoch was used to capture as much variation in children’s activity as possible. Children were fitted with an accelerometer during their scheduled annual Jiri Growth Study visit. The accelerometer was placed on the left hip just superior to the iliac crest and the children were instructed to wear the accelerometer 24 hours a day, including during times they were bathing or sleeping. Parents were also asked to ensure that the accelerometers were worn at all times unless the belt became uncomfortable. No participants reported that the belt or the device was uncomfortable. American Journal of Human Biology
Anthropometry and growth standards
131
QUANTITATIVE PHYSICAL ACTIVITY ASSESSMENT OF CHILDREN AND ADOLESCENTS TABLE 1. Descriptive Statistics for Male Jirel Children and Adolescents Age 6 yrs (n 5 16) 7 yrs (n 5 26) 8 yrs (n 5 19) 9 yrs (n 5 17) 10 yrs (n 5 16) 11 yrs (n=18) 12 yrs (n=19) 13 yrs (n 5 26) 14 yrs (n 5 17) 15 yrs (n=17) 16 yrs (n 5 13) 17 yrs (n 5 10) 18 yrs (n 5 3)
WT
HT
BMI
zHT/Agea
zBMI/Agea
zWT/Agea,b
16.84 (2.15) 17.77 (1.71) 22.21 (6.79) 21.79 (2.18) 23.25 (3.89) 24.67 (3.09) 29.03 (5.82) 30.65 (6.45) 34.70 (6.86) 37.71 (6.18) 40.46 (8.71) 47.05 (4.95) 44.83 (6.71)
107.04 (5.54) 112.49 (5.57) 121.83 (10.42) 121.52 (5.30) 126.16 (7.28) 130.43 (4.95) 136.35 (7.69) 138.74 (9.18) 143.71 (7.96) 149.02 (9.49) 154.13 (11.25) 158.56 (4.61) 156.30 (2.17)
14.70 (1.40) 14.04 (0.93) 14.68 (1.37) 14.76 (1.05) 14.50 (0.99) 14.45 (1.12) 15.49 (1.91) 15.75 (1.47) 16.72 (2.07) 16.87 (1.19) 16.80 (1.64) 18.68 (1.40) 18.31 (2.21)
21.81 (1.13) 21.75 (1.06) 20.96 (1.84) 21.84 (0.88) 21.80 (1.14) 21.88 (0.74) 21.80 (1.08) 22.33 (1.24) 22.53 (1.04) 22.55 (1.21) 22.41 (1.45) 22.17 (0.60) 22.65 (0.30)
20.59 (1.26) 21.20 (0.83) 20.87 (0.91) 20.97 (0.80) 21.35 (0.79) 21.75 (0.90) 21.39 (1.21) 21.50 (0.87) 21.35 (1.22) 21.50 (0.75) 21.91 (0.98) 21.10 (0.68) 21.53 (1.07)
21.60 (0.99) 22.00 (0.77) 21.28 (1.53) 21.86 (0.70) 22.07 (1.16)
School day wear time
Saturday wear time
55.76 (6.31) 56.38 (15.24) 58.35 (13.96) 57.48 (10.04) 53.53 (7.23) 52.30 (6.45) 51.33 (7.26) 48.56 (8.00) 50.52 (7.05) 49.12 (8.54) 53.69 (20.64) 58.82 (23.59) 45.41 (8.95)
11.62 (1.36) 11.69 (3.13) 11.20 (2.03) 11.79 (1.87) 10.91 (1.19) 10.63 (2.29) 10.57 (2.10) 10.44 (2.10) 11.17 (1.73) 10.52 (1.85) 11.87 (3.81) 12.37 (4.44) 10.53 (1.04)
School day wear time
Saturday wear time
60.09 (11.63) 56.19 (8.19) 66.91 (22.38) 53.99 (4.79) 51.40 (6.66) 50.04 (5.94) 54.09 (8.66) 47.88 (7.03) 48.71 (7.16) 49.92 (9.93) 51.71 (10.51) 50.30 (10.19) 44.59 (11.10)
12.25 (2.42) 11.27 (1.63) 13.42 (4.44) 10.68 (1.38) 10.22 (1.82) 10.58 (2.03) 11.55 (1.99) 9.61 (2.07) 10.14 (2.04) 10.37 (2.49) 11.06 (2.36) 10.58 (2.59) 9.90 (1.98)
Data are presented as mean (SD) in kilograms, centimeters, and hours. a Z-scores are calculated using WHO reference population (de Onis, 2012). b Z weight-for-age up to age 10 years, per WHO guidelines.
TABLE 2. Descriptive Statistics for Female Jirel Children and Adolescents Age 6 yrs (n 5 11) 7 yrs (n 5 17) 8 yrs (n 5 14) 9 yrs (n 5 20) 10 yrs (n 5 19) 11 yrs (n 5 20) 12 yrs (n 5 16) 13 yrs (n 5 11) 14 yrs (n 5 19) 15 yrs (n 5 20) 16 yrs (n 5 12) 17 yrs (n 5 11) 18 yrs (n 5 6)
WT
HT
BMI
zHT/Agea
zBMI/Agea
zWT/Agea,b
16.32 (1.90) 17.03 (2.10) 19.64 (1.77) 22.15 (3.54) 26.11 (7.48) 26.80 (3.83) 29.69 (6.23) 35.05 (9.36) 39.29 (7.41) 39.95 (6.52) 41.75 (4.93) 45.54 (6.99) 49.25 (8.09)
108.05 (4.09) 111.75 (4.91) 117.66 (4.66) 123.33 (5.96) 129.77 (9.07) 132.46 (7.05) 136.49 (8.56) 144.18 (8.23) 148.31 (6.80) 148.19 (6.63) 146.35 (4.40) 151.31 (4.33) 152.08 (8.33)
13.95 (1.05) 13.58 (0.78) 14.18 (0.92) 14.48 (1.34) 15.27 (2.09) 15.21 (1.01) 15.79 (2.02) 16.65 (3.22) 17.72 (2.18) 18.11 (2.08) 19.49 (2.14) 19.90 (2.55) 21.20 (1.81)
21.38 (0.80) 21.66 (0.90) 21.53 (0.80) 21.50 (0.98) 21.39 (1.42) 21.89 (1.06) 22.16 (1.25) 21.76 (1.19) 21.65 (0.98) 21.96 (0.96) 22.38 (0.65) 21.73 (0.65) 21.66 (1.26)
20.99 (0.85) 21.31 (0.65) 21.00 (0.66) 21.07 (0.97) 20.91 (1.15) 21.13 (0.58) 21.24 (1.08) 21.29 (1.59) 20.89 (0.93) 20.95 (1.02) 20.53 (0.85) 20.50 (0.91) 20.06 (0.63)
21.57(0.86) 21.96 (0.91) 21.65(0.63) 21.67(1.12) 21.49 (1.33)
Data are presented as mean (SD) in kilograms, centimeters, and hours. a Z-scores are calculated using WHO reference population (de Onis, 2012). b Z weight-for-age up to age 10 years, per WHO guidelines.
TABLE 3. Physical Activity Variables Used In Data Reduction (After Table 1, Colley, R. 2012. Actical Accelerometer Data Analysis Support Tool: Harmonizing With The Canadian Health Measures Survey (Accel1)) Physical activity variables Wear time 5 24 hours—non-wear time (hours) Non-wear time 5 60 consecutive minutes of zero counts Valid day 5 101 hours of wear time Intensity cut-points: METS Accelerometer count range (cpm) Sedentary (SED) 1
