Journal of Physical Activity and Health, 2015, 12, 1184 -1191 http://dx.doi.org/10.1123/jpah.2014-0202 © 2015 Human Kinetics, Inc.
REVIEW
Step-Count Guidelines for Children and Adolescents: A Systematic Review Michael Pereira da Silva, Fabio Eduardo Fontana, Eric Callahan, Oldemar Mazzardo, and Wagner De Campos Background: The aim of this systematic review was to identify the most optimal step-count cutoff for children and adolescents (5–19 years old) among guidelines currently available in the literature. Methods: The databases searched were PubMed, SportDiscus, Science Direct, Web of Science and LILACS. Studies were categorized into Health Cohort studies or Physical Activity (PA) Cohort studies according to the reference standard used. The quality of the studies was assessed using the QUADAS-2 instrument. Results: Six Health and 3 PA Cohort studies were included in the final pool of papers after Full Text reading. With the exception of a single study, studies demonstrated a high risk of methodological bias in at least 1 of the QUADAS-2 domains. Guidelines ranged from 10,000 to 16,000 steps/day for the Health studies (5–16 years old), and from 9,000 to 14,000 steps/day for PA studies (6–19 years old). Due to the high risk of methodological bias, none of the Health Cohort guidelines were endorsed. The PA Cohort study with the lowest risk of methodological bias suggested 12,000 steps/day for children and adolescents irrespective of gender. Conclusion: PA Cohort studies demonstrated lower risk of methodological bias than Health Cohort studies. The optimal youth step-count guideline of 12,000 steps/day was endorsed. Keywords: youth, QUADAS-2
The detection of insufficient levels of physical activity is essential for the identification of children and adolescents more prone to health problems such as diabetes, high blood pressure, and obesity,1,2 but it is currently difficult to accurately measure physical activity habits. Subjective measures, such as self-reported questionnaires, are compromised by the ability of children and adolescents to recall all relevant details about their daily physical activity behaviors and quantify physical activity intensity.3,4 This problem is amplified in children due to their less developed cognitive capacity and the more intermittent nature of their physical activity behavior.4 Accelerometers objectively measure duration, intensity, and frequency of physical activity,5,6 but they are expensive and lack a gold standard procedure to convert acceleration counts to exercise intensity.7 Pedometers may be a viable alternative to objectively measure physical activity by providing an overall measure of daily steps taken. They are inexpensive and conveniently sized,7–9 and its output is reliable and simple to comprehend.7 Pedometer step-counts correlate well with accelerometry and measures of energy expenditure.10 Pedometer daily step-counts can assist the screening and surveillance of physical activity behavior in pediatric populations. Several previous studies have proposed step-count guidelines,11–19 but the great variability in the recommendations may deter their use in field interventions (ie, 9000 to 16,000 steps/day). A recent comprehensive narrative review of literature20 has recommended a minimum of 13,000 to 15,000 steps/day for elementary school boys, 11,000 to 12,000 steps/day for elementary school girls, and 10,000 to 11,700 for adolescent boys and girls, but the guidelines were arbitrarily selected and the range may be confusing to practitioners. Thus, the aim of this study is to systematically assess
Silva ([email protected]), Mazzardo, and De Campos are with the Dept of Physical Education, Federal University of Parana, Curitiba, Parana, Brazil. Fontana and Callahan are with the School of Health, Physical Education, and Leisure Services, University of Northern Iowa, Cedar Falls, IA. 1184
the literature developing youth step-count guidelines to identify the most optimal cutoff point among step-count guidelines for children and adolescents (5–19 years old) currently available in the literature. For such, this review systematically extracted data from eligible studies and evaluated the risk of methodological bias and concerns over the applicability of findings. Based on the reference standard, studies were assigned to 2 groups: Health and Physical Activity (PA) Review Cohorts. The Health Review Cohort encompassed studies referencing step-count guidelines on Body Mass Index (BMI) or body fat percentage.12,14–17,19 Guidelines proposed by these studies represent the minimum count of daily steps necessary for the prevention of obesity. The PA Review Cohort included studies referencing guidelines on the attendance of physical activity recommendations.11,13,18 Step-count guidelines proposed by these studies indicate whether PA recommendations are being met.
Methods Literature Search A systematic search of peer reviewed articles published between January 1, 2000 and September 30, 2013 was performed using 5 databases (PubMed, SportsDiscus, Science Direct, Web of Science, and LILACS). Key words were used as MeSH terms and text words as appropriate for each database: step-counts (walking, pedometer), physical activity (motor activity, physical activity assessment, physical activity surveillance), health outcomes (body mass index, body composition, body fat), cutoff (cut-off points, step-count recommendations, step-count guidelines), and children (child, adolescent, adolescents, teenager). Search statements were truncated using Boolean terms. A new search was made between November 1, 2013 and March 31, 2014, but no additional studies were included. Screening of relevant papers followed a 4-step process: 1) assessment of titles, 2) assessment of abstracts, 3) full article review, and 4) screening for additional references among selected papers.
Step-Count Guidelines: A Systematic Review 1185
Two reviewers independently reviewed the papers during all 4 steps. Disagreements between the 2 reviewers were resolved by discussion at consensus meetings. Duplicate papers were excluded. According to the reference standard, eligible studies were assigned to Health (BMI or body fat percentage) or PA Review Cohorts (attendance to physical activity recommendations).
Demographic Characteristics
Inclusion Criteria To be included in the review, studies had to meet the following criteria: 1) development of step-count guidelines or testing existing step-count guidelines; 2) target group consisting of children and/ or adolescents 5 to 19 years of age; 3) basing step-count cutoffs on either a physical activity standard (ie, attainment of 60 minutes of moderate to vigorous physical activity daily) or a health standard (ie, BMI, body fat percentage); and 4) articles written in English, Portuguese, or Spanish. Downloaded by R B Draughon Library on 09/18/16, Volume 12, Article Number 8
studies were found after cross-referencing the selected studies. After reading the full text of all 16 studies, 9 studies fully met the inclusion criteria. Six of these studies were classified into Health Review Cohort and 3 in the PA Review Cohort. A summary of the review process is provided in Figure 1.
Data Extraction One reviewer extracted the following data from relevant articles: study identifiers (authors, year, country, journal), study design, study population (sample size, age), index text, reference standard, and proposed step-count cutoffs. A second reviewer double checked the extracted data, and discrepancies were resolved by discussion at consensus meetings.
Methodological Quality Assessment (QUADAS-2) The QUADAS-2 instrument assesses the methodological quality of diagnostic studies.21 It evaluates the risk of methodological bias and concerns over the applicability of findings. Risk of methodological bias is evaluated across 4 domains: participant selection, index text, reference standard, and flow and timing of data collection (ie, high, low, or unclear risk). Participant selection refers to the quality of the procedures used to select participants. For example, convenient sampling is a source of bias in a study measuring the quality of a diagnostic test. The index test pertains to the test being evaluated. In this systematic review, the index test was daily step-counts. Problems in the interpretation or procedures used to conduct the index test may introduce bias to the study. The reference standard refers to the test used to classify the target condition. BMI, body fat percentage, and attainment to physical activity guidelines measured via accelerometry have been used as reference standards to develop step-count guidelines. The reference standard itself, its conduct or interpretation may bias the results of a study. Flow and timing of data collection assesses the possibility that the sequence of study procedures bias the results of a study. QUADAS-2 also evaluates concerns over the applicability of findings across the aforementioned first 3 domains (ie, high, low, or unclear concern). As recommended,21 minor adaptations to the QUADAS-2 instrument were made to specifically assess research regarding step-count guidelines. QUADAS-2 was piloted on 4 studies before quality assessment (The modified instrument is available upon request). Two independent reviewers conducted the quality appraisal, and disagreements were resolved by discussion at consensus meetings.
Results Included Studies The initial search identified 452 references. Screening of Titles and Abstracts led to 13 potentially pertinent references. Three other
Studies in the Health Review Cohort ranged from 665 to 2698 participants, and studies in the PA Review Cohort ranged from 34 to 2217 participants. Participants in the Health Review Cohort were 5 to 17 years old and those in the PA Review Cohort were 6 to 19 years old. Studies in the Health Review Cohort were conducted in 4 countries [Australia (2 studies, N = 3373), United States of America (2 studies, N = 1879), New Zealand (1 study, N = 969), Multicenter (1 study, NUnited States of America = 711, NAustralia = 563; NSweden = 680)]. Studies in the PA Review Cohort were conducted in 3 countries [United States of America (1 study, N = 2217), Canada (1 study, N = 1619), United Kingdom (1 study, N = 34)]. All studies were written in English.
Methodological Quality Table 1 provides a summary of the methodological quality appraisal of the studies. Overall, 8 of the 9 studies had methodological limitations leading to a high risk of bias in at least 1 of the QUADAS-2 domains. Regarding the applicability of step-count guidelines to interventions, 6 of the 9 studies demonstrated a low concern.
Risk of Bias Patient Selection. Several studies (Health Review Cohort: 3 studies; PA Review Cohort: 1 study) showed high risk of bias during the selection of participants. The most common risk was convenient sampling of participants. Procedures used to select participants were unclear in 1 study.16 Index Test. Seven studies demonstrated high risk of bias during the handling of the index test (Health Review Cohort: 5 studies; PA Review Cohort: 2 studies). The most common limitations were the selection of a single cutoff point for a broad age range and use of 1000 steps/day increments to guide the statistical selection of optimal step-count cutoffs. One study did not provide sufficient information for the interpretation of the index test.16 Reference Standard. Five studies in the Health Review Cohort showed high risk of bias for the reference standard. They used BMI as the reference, even though the limitations of BMI have been extensively reported.22–24 The other Health study was rated unclear for risk of bias.15 Bioimpedance assessed body fat percentage, but the study lacked descriptions of preassessment procedures and information about the validity of the equipment used. All 3 studies in the PA Review Cohort demonstrated low risk of bias. Although there is no gold standard to measure physical activity habits, the accelerometer procedures used in the studies are considered highly acceptable in the literature.4,25–27 Flow and Timing. Two studies in the Health Review Cohort revealed high risk of bias because they did not include all subjects in the analysis.15,17 Four Health studies lacked a description of the interval between index testing and reference standard assessment,12,14,16,19 and the risk of bias during flow and timing of data collection was rated unclear. All studies in the PA Review
JPAH Vol. 12, No. 8, 2015
Downloaded by R B Draughon Library on 09/18/16, Volume 12, Article Number 8
1186 Silva et al
Figure 1 — Review process.
Cohort demonstrated low risk of bias for flow and timing of assessment.
study.13 In this study, steps were measured via accelerometry without a corresponding pedometer adapted step/day cutoff. Reference Standards. In Health and PA Review Cohorts, con-
Concerns of Applicability Patient Selection. Two studies in the Health Review Cohort
demonstrated high concern for participant selection.12,19 They referenced studies describing the representativeness of original samples but did not provide information about the representativeness of the subsamples used for developing step-count guidelines. All studies in the PA Review Cohort demonstrated low concern for the handling of participant selection. Index Test. All studies in the Health Review Cohort demonstrated
low concern for the applicability of the index test. Concern about the applicability of the index test was high in 1 PA Review Cohort
cerns about applicability of reference standard to assess the target conditions were low.
Optimal Step-Counts Guidelines Health Review Cohort Studies.
Table 2 provides a summary of the characteristics and suggested step-count guidelines for studies in the Health Review Cohort. Pedometers were used to measure step-counts in all studies. Five studies used BMI12,14,16,17,19 and 1 study used body fat percentage as the reference standard.15 With the exception of 2 studies that used ROC Curves,12,14 the contrasting groups method was the
JPAH Vol. 12, No. 8, 2015
Step-Count Guidelines: A Systematic Review 1187
Table 1 QUADAS-2 Results Risk of bias Participant selection
Beets et al (2008) Dollman et al (2010) Duncan et al (2007)
Study
Applicability concerns
Index test
Reference standard
Flow and timing
Patient selection
Index test
Reference standard
↑
↑
↑
?
↓
↓
↓
↓
↑
↑
?
?
↓
↓
↑
↑
?
↑
↓
↓
↓
BMI and body fat outcomes
Laurson et al (2008)
?
?
↑
?
↓
↓
↓
McComack et al (2011)
↓
↑
↑
↑
?
↓
↓
Tudor-Locke et al (2004)
↑
↑
↑
?
↓
↓
↓
Adams, Johnson and Tudor-Locke (2013)
↓
↑
↓
↓
↓
↓
↓
Colley, Jassen and Tremblay (2012)
↓
↓
↓
↓
↓
↑
↓
Rowlands and Eston (2005)
↑
↑
↓
↓
↓
↓
↓
Downloaded by R B Draughon Library on 09/18/16, Volume 12, Article Number 8
Physical activity outcome
Note. ↓ = Low; ↑ = High; ? = Unclear.
most common analysis used in the Health Cohort Review studies to identify optimal step-count cutoff points. Beets et al12 tested step-count cutoffs previously developed by Tudor-Locke et al.19 They concluded that step-count cutoffs were unable to distinguish between healthy and unhealthy BMI classified weight.12 Similarly, Dollman et al14 indicated a low capacity of step-count cutoff (AUC range: .52 to .60) to discriminate between youth individuals with healthy and overweight BMI classifications . Step-count guidelines proposed by studies in the Health Review Cohort were generally gender specific. For boys, step-count guidelines ranged from 11,000 to 16,000 steps/day. For girls, step-count guidelines ranged from 10,000 to 16,000 steps/day. McCormack et al17 suggested a single cutoff for both sexes, and Dollman et al14 proposed age specific cutoff points. Out of 4 QUADAS-2 methodological domains, studies in the Health Review Cohort demonstrated a high risk of methodological bias in at least 2 domains. Laurson et al16 demonstrated a high risk of bias in 1 domain, but there was insufficient information for the quality rating of the other 3 domains. Due to the high risk of methodological bias, this systematic review does not endorse any of the step-count guidelines proposed by Health Review Cohort studies. PA Review Cohort Studies. Table 3 provides a summary of the characteristics and suggested step-count guidelines for studies in the PA Review Cohort. The index test, daily step-counts, was measured by accelerometry in 2 studies13,28 and by pedometry in the other study.18 Sixty minutes of moderate to vigorous physical activity, assessed via accelerometry, was the reference standard in all PA Review Cohort studies. ROC curves were used to identify optimal step-count cutoffs in all studies in the PA Review Cohort. Two studies computed separate step-count guidelines by gender and developmental period, but they each suggested a single guideline due to the similarities among the cutoffs for each subgroup.13,28 Rowland and Eston18 suggested separate step-count guidelines for boys and girls 8 to 10 years old. Step-count guidelines proposed by the PA Review Cohort studies ranged from 9000 to 14,000 steps/day. Colley et al13 demonstrated low risk of bias across all 4 QUADAS-2 domains but a high concern over the applicability of the index test. This study suggested a cutoff of 12,000 steps/day. Adams et al11 demonstrated
low concern across all 3 applicability domains but demonstrated a high risk of bias in the index test domain. This study suggested 9000 steps/day.
Discussion Pedometer daily step-counts are a convenient and reliable measure of physical activity behavior among children and adolescents.7–9 However, previously proposed step-count guidelines vary greatly (ie, 9000–16,000 steps/day).11–19 Tudor-Locke et al20 comprehensively reviewed the literature developing youth step-count guidelines, but arbitrarily recommended steps/day ranges for children and adolescents. Thus, the purpose of this study was to systematically assess the literature developing step-count guidelines for children and adolescents. To reduce bias during the review process, systematic procedures and protocols were used to identify, select, and extract data from relevant articles, and selected articles were appraised based on methodological quality and applicability of step-count guidelines. Instead of a range, the use of systematic procedures also permitted the recommendation of a single most optimal stepcount cutoff point. This systematic review also benefited from the publication of 2 recent step-count guideline papers with excellent research quality appraisal.
Health Review Cohort Studies In general, studies in the Health Review Cohort demonstrated a high risk of methodological bias. Sources of methodological limitations included the selection of participants, interpretation of the index test, choice of the reference standard, and flow and timing of data collection. Participant selection is an important step in diagnostic studies. Random selection of participants is necessary to avoid the risk of methodological bias,21 but 3 studies conveniently sampled participants. The most common risk of bias in the interpretation of the index test was the use of 1,000-step increments to select optimal steps/day cutoffs. Such a procedure decreases the precision of the selected steps/day guideline. For example, choosing 12,000 steps/ day may lead to less desirable values of sensitivity and specificity than a cutoff of 12,450 steps/day.
JPAH Vol. 12, No. 8, 2015
1188 Silva et al
Downloaded by R B Draughon Library on 09/18/16, Volume 12, Article Number 8
Table 2 Summary of Health Review Cohort Studies Results Cutoff point suggested/ Conclusion
Study
Sample
Step-count assessment
Reference standard
Beets et al (2008)
United States, 1067 participants, Age: 6–12 years
Pedometer: Walk4Life MLS 2525 and Yamax SW 200. The pedometer was worn for 4 consecutive days. Tudor-Locke et al (2004) cutoff points were tested.
BMI: international cutpoint for youth was used (Cole et al, 2000).
None: Tudor-Locke cutoffs (15,000 steps/day for boys and 12,000 for girls) were unable to distinguish between healthy and unhealthy weight.
Dollman et al (2010)
Australia, 2698 participants, Age: 5–16 year. (Children: 5–12 years; Adolescents: 13–16 years)
Pedometer: New Lifestyles 1000. The pedometer was worn for 7 consecutive days. Four days (3 weekdays and 1 weekend) were necessary for analyses.
BMI: international cutpoint for youth was used (Cole et al, 2000).
Suggested cutoffs were 12,000 for younger males, 11,000 for older males, and 10,000 for younger females. The AUC for older females was nonsignificant.
Duncan et al (2007)
New Zealand, 969 participants, Age: 5–12 years
Pedometer: Multidaymemory NL-2000. The pedometer was worn for 3 weekdays and 2 weekend days.
% Body Fat: A hand-to-foot Suggested cutoffs were 16,000 steps/day for boys bioelectrical impedance and 13,000 for girls. analyzer (Model BIM4). Excess of %BF was classified as % Body Fat ≥85th Percentile.
Laurson et al (2008)
United States, 812 participants, Age: 6–12 years
Pedometer: DigiwalkerSW-200. The pedometer was worn for 7 consecutive days. Four days (3 weekdays and 1 weekend) were necessary for analyses.
BMI: international cutpoint for youth was used (Cole et al, 2000).
Suggested cutoffs for obese were 11,500 steps/day for boys and 10,000 for girls. The cutoffs for overweight were 12,000 steps/day for boys and 10,500 for girls.
McComack et al (2011)
Australia, 675 participants, Age: 7–16 years
Pedometer: Digiwalk SW-700. The pedometer was worn for 7 consecutive days. Four days (3 weekdays and 1 weekend) were necessary for analyses.
BMI: international cutpoint for youth was used (Cole et al, 2000).
The suggested cutoff was 16,000 steps/day for both boys and girls.
Tudor-Locke et al (2004)
Multicenter (United States, Australia & Sweden), 1954 participants, Age: 6–12 years
Pedometer: My Life Stepper, MLS-2000. The pedometer was worn for 4 consecutive school days.
BMI: international cutpoint for youth was used (Cole et al, 2000).
Cutoffs ranged from 14,000 to 17,000 steps/day for boys and from 10,000 to 13,000 for girls. Suggested cutoffs were 15,000 steps/ day for boys and 12,000 for girls.
Most Health Review Cohort studies used BMI as the reference standard. BMI often misclassifies individuals due to its inability to account for lean mass weight and differences in biological maturation.29–31 The low association between PA and BMI may affect the ability of step-counts to discriminate obesity levels.32,33 In fact, Dollman et al14 indicated a poor diagnostic ability of step-counts to differentiate between BMI classified healthy and overweight children and adolescents (AUC range: .52 to .60). All Health Review Cohort studies demonstrated the same limitation for flow and timing of data collection. They lacked a description of the
interval between index test and reference standard measurements. If too much time elapses between the 2 measurements, variations in the index test (steps/day) or reference standard (obesity classification) may bias the results. Developmental and gender differences in BMI have been widely reported. Health review cohort studies generally took gender into account for the development of step-count guidelines. With the exception of a single study, step-count guidelines were offered separately by gender.17 However, only Dollman et al14 developed optimal step-count guidelines by gender and developmental stage.
JPAH Vol. 12, No. 8, 2015
Step-Count Guidelines: A Systematic Review 1189
Downloaded by R B Draughon Library on 09/18/16, Volume 12, Article Number 8
Table 3 Summary of PA Review Cohort Studies Results Cutoff point suggested/ Conclusion
Study
Sample
Step-count assessment
Reference standard
Adams, Johnson and Tudor-Locke (2013)
United States, 2217 participants, Age: 6–17 years (Children: 6–11 years; Adolescents: 12–17 years)
Accelerometer: ActiGraph 7164. Average daily accelerometer steps was computed and adapted to pedometer steps.
Accelerometer: ActiGraph 7164. Participants with at least 1 valid day (≥ 10 hours) were included in the analyses. Freedson 3 and 4 METs and Evenson equations were used.
A range of 11,500–13,500 accelerometer steps/day for children and 11,500–14,000 accelerometer steps/ day for adolescents was suggested for boys and girls. A pedometer-friendly adaptation of 9000 steps/day was recommended for boys and girls.
Colley, Jassen and Tremblay (2012)
Canada, 1619 participants, Age: 6–19 years
Accelerometer: Actical. Pedometer friendly adaptations were not offered.
Accelerometer: Actical. Each of 4+ valid days (≥ 10 hours) was assessed separately. The cutoff point for MVPA was >1500 counts/min.
The cutoff point of 12,000 steps/day was recommended for boys and girls 6–19 years of age.
Rowlands and Eston (2005)
United Kingdom, 34 participants, Age: 8.3–10.8 years
Pedometer: Yamax DigiWalker DW-200. Previous cutoff points were tested (Tudor-Locke et al, 2004; Vincent & Pangrazi, 2002).
Accelerometer: Tritrac-T303 and T303A. A single measurement day was used in analyzes. Counts/min used to specify physical activity intensities was based in Oxygen uptake studies (Eston et al, 1998; Rowlands et al, 1999).
Tudor-Locke et al (2004) cutoff point for boys may be too high to identify 60 min of MVPA. It was recommended 13,000 steps/ day for boys and 12,000 for girls.
Ignoring age-related differences in BMI was a source of bias in 5 studies12,14,16,17,19. Methodological limitations prevent the endorsement of stepcount guidelines proposed by the Health Review Cohort studies. Future studies should choose health standards with stronger validity evidence and more highly associated with physical activity (ie, metabolic syndrome risk factors). It is also recommended that future studies assess the sensitivity and specificity across the entire continuum of possible steps/day cutoffs.
PA Review Cohort Studies PA Review Cohort studies took into consideration gender and developmental differences during the development of step-count guidelines. All 3 studies13,18,28 established gender and age appropriate step-count guidelines. Rowlands and Eston18 restricted the sample to participants between 8 and 10 years of age suggesting distinct step-count cutoffs by gender. Adams et al28 developed separate guidelines by gender for children (6–11 years old) and adolescents (12–17 years old), but the similarity in step-count cutoffs among the different subgroups led to the recommendation of a single pedometer scaled step-count cutoff. Colley et al13 developed separate guidelines by gender for 3 age groups: children (6–10 years old), early adolescence (11–14 years old), and late adolescence (15–19 years old). Colley et al13 also recommended a single step-count guideline for girls and boys 6 to 19 years of age due to the similarity among the cut-points developed for the subgroups. Quality assessment of studies guided the determination of the most optimal step-count guidelines currently available in the literature. Guidelines offered by studies with a high risk of
methodological bias are likely to be imprecise for the identification of children not reaching physical activity recommendations. Rowlands and Eston18 had a high risk of methodological bias during the selection of participants and interpretation of the index test. In addition, inferences made from this study are limited by the small sample power of the study (N = 17 participants/group). Adams et al28 used a large and representative sample of US children and adolescents to develop step-count guidelines. Although steps were measured via accelerometry, a pedometer adapted 9000 steps/day guideline was suggested. The applicability of the guidelines was improved by converting accelerometer to pedometer steps, but 9000 steps/day may be too low to identify individuals reaching 60 minutes of moderate to vigorous activity. In fact, the cutoff is the lowest cutoff among all the studies developing step-count guidelines. This study was also limited by the selection of step-count cutoffs using 500-step increments. Colley et al13 demonstrated a high concern over the applicability of the index test selected. The study measured steps using Actical accelerometers, but it did not provide an estimate of corresponding pedometer steps/day cutoff. It is possible that the guidelines developed using accelerometer counted steps are not appropriate for pedometer based interventions. Accelerometers are considered to be more sensitive at identifying steps during low frequency acceleration than pedometers.34 While concerns of the applicability of the step-count guidelines were high, this study was methodologically well designed. It was the only study to demonstrate low risk of bias in all 4 methodological domains. Thus, this review endorses the guidelines suggested by Colley et al13 of 12,000 steps/day. Interestingly, this cutoff was commonly proposed by many of the Health and PA Review Cohort studies. In the Health Review Cohort studies,
JPAH Vol. 12, No. 8, 2015
1190 Silva et al
12,000 steps/day was suggested twice for girls12,19 and twice for boys.14,16 Among studies in the PA Review Cohort, Rowlands and Eston18 suggested it for 8- to 10-year-old girls, and Adams, Johnson and Tudor-Locke28 included 12,000 steps/day within the suggested accelerometer steps/day guidelines. It is strongly recommended that pedometry can be used as the index test of future studies. The use of pedometers to count steps eliminates possible errors related to the accelerometer-pedometer conversion process. Low-cost and easiness-of-use make pedometers friendly devices to count daily steps for the discrimination between sufficiently and insufficiently active individuals.
Downloaded by R B Draughon Library on 09/18/16, Volume 12, Article Number 8
Limitations, Strengths, and Recommendations for Future Studies This systematic review is not without limitations. Meta-analytic procedures were not possible due to the limited number of studies on the topic (N = 9), and divergent use of methodological (Health vs. PA Review Cohorts) and statistical procedures (Contrasting groups method vs. ROC curves). The performance of a meta-analysis would have strengthened the choice of a step-count guideline. There are also several strengths to this systematic review. Systematic review procedures were strictly followed. Two reviewers were used throughout the review process, and QUADAS-2 is considered an appropriate instrument to assess methodological bias of diagnostic studies.21 Six out of nine studies included in this systematic review were published in the last 6 years.12–14,16,17,28 Finally, recommendations made in this systematic review should assist the design of future studies. It is recommended that future research developing step-count guidelines base the index test on pedometry, develop guidelines for a narrow developmental span, assess the sensitivity and specificity of the entire steps/day continuum, use ROC curves to determine optimal cutoff points,35,36 clearly describe the timeline between the index test and reference standard, randomly sample participants, and use accelerometry output as the reference standard. In addition to values of sensitivity and specificity of each proposed steps/day cutoff, ROC curves also provide an indication of the discriminatory power of the index test based on the area under the curve statistic. Although accelerometry was suggested as the reference standard, it is important to acknowledge the lack of a gold standard to measure PA behavior in field research. Consequently, studies developing step-count guidelines using PA reference standards should check the utility of proposed step-count guidelines for the identification of health conditions such as the metabolic syndrome.
Conclusion The quality of studies developing step-count guidelines was systematically assessed to determine the most optimal cutoff point among currently available guidelines. Out of the nine studies assessed, Colley et al13 had the lowest risk of methodological bias. Thus, this systematic review endorses the step-count guidelines proposed by Colley et al13 of 12,000 steps/day for children and adolescents irrespective of gender. Acknowledgments The authors would like to acknowledge the Federal University of Parana and the University of Northern Iowa.
References 1. Luke A, Dugas LR, Durazo-Arvizu RA, Cao G, Cooper RS. Assessing physical activity and its relationship to cardiovascular risk factors: NHANES 2003-2006. BMC Public Health. 2011;11(1):387. PubMed doi:10.1186/1471-2458-11-387 2. Goodman E, Daniels SR, Meigs JB, Dolan LM. Instability in the diagnosis of metabolic syndrome in adolescents. Circulation. 2007;115(17):2316–2322. PubMed doi:10.1161/CIRCULATIONAHA.106.669994 3. Sallis JF, Buono MJ, Roby JJ, Micale FG, Nelson JA. Sevenday recall and other physical activity self-reports in children and adolescents. Med Sci Sports Exerc. 1993;25(1):99–108. PubMed doi:10.1249/00005768-199301000-00014 4. Corder K, Ekelund U, Steele RM, Wareham NJ, Brage S. Assessment of physical activity in youth. J Appl Physiol. 2008;105(3):977–987. PubMed doi:10.1152/japplphysiol.00094.2008 5. Nader PR, Bradley RH, Houts RM, McRitchie SL, O’Brien M. Moderate-to-vigorous physical activity from ages 9 to 15 years. JAMA. 2008;300(3):295–305. PubMed doi:10.1001/jama.300.3.295 6. Troiano RP, Berrigan D, Dodd KW, Mâsse LC, Tilert T, McDowell M. Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc. 2008;40(1):181–188. PubMed doi:10.1249/ mss.0b013e31815a51b3 7. Duncan JS, Schofield G, Duncan EK. Pedometer-determined physical activity and body composition in New Zealand children. Med Sci Sports Exerc. 2006;38(8):1402–1409. PubMed doi:10.1249/01. mss.0000227535.36046.97 8. Duncan MJ, Al-Nakeeb Y, Woodfield L, Lyons M. Pedometer determined physical activity levels in primary school children from central England. Prev Med. 2007;44(5):416–420. PubMed doi:10.1016/j. ypmed.2006.11.019 9. Gabel L, Proudfoot N a, Obeid J, et al. Step count targets corresponding to new physical activity guidelines for the early years. Med Sci Sports Exerc. 2013;45(2):314–318. PubMed doi:10.1249/ MSS.0b013e318271765a 10. Tudor-Locke C, Williams JE, Reis JP, Pluto D. Utility of pedometers for assessing physical activity: convergent validity. Sports Med. 2002;32(12):795–808. PubMed doi:10.2165/00007256-20023212000004 11. Adams MA, Caparosa S, Thompson S, Norman GJ. Translating physical activity recommendations for overweight adolescents to steps per day. Am J Prev Med. 2009;37(2):137–140. PubMed doi:10.1016/j. amepre.2009.03.016 12. Beets MW, Le Masurier GC, Beighle A, et al. Are current body mass index referenced pedometer step-count recommendations applicable to US youth? J Phys Act Health. 2008;5(5):665–674. PubMed 13. Colley RC, Janssen I, Tremblay MS. Daily step target to measure adherence to physical activity guidelines in children. Med Sci Sports Exerc. 2012;44(5):977–982. PubMed doi:10.1249/MSS.0b013e31823f23b1 14. Dollman J, Olds TS, Esterman A, Kupke T. Pedometer step guidelines in relation to weight status among 5- to 16-year-old Australians. Pediatr Exerc Sci. 2010;22(2):288–300. PubMed 15. Duncan JS, Schofield G, Duncan EK. Step count recommendations for children based on body fat. Prev Med. 2007;44(1):42–44. PubMed doi:10.1016/j.ypmed.2006.08.009 16. Laurson KR, Eisenmann JC, Welk GJ, Wickel EE, Gentile DA, Walsh DA. Evaluation of youth pedometer-determined physical activity guidelines using receiver operator characteristic curves. Prev Med. 2008;46(5):419–424. PubMed doi:10.1016/j.ypmed.2007.12.017 17. McCormack GR, Rutherford J, Giles-Corti B, Tudor-Locke C, Bull F. BMI-referenced cut-points for recommended daily pedometerdetermined steps in Australian children and adolescents. Res Q Exerc Sport. 2011;82(2):162–167. PubMed doi:10.1080/02701367.2011.10 599743 18. Rowlands AV, Eston RG. Comparison of accelerometer and pedometer measures of physical activity in boys and girls, ages 8–10 years. Res Q Exerc Sport. 2005;76(3):251–257. PubMed doi:10.1080/0270136 7.2005.10599296
JPAH Vol. 12, No. 8, 2015
Downloaded by R B Draughon Library on 09/18/16, Volume 12, Article Number 8
Step-Count Guidelines: A Systematic Review 1191
19. Tudor-Locke C, Pangrazi RP, Corbin CB, et al. BMI-referenced standards for recommended pedometer-determined steps/day in children. Prev Med. 2004;38(6):857–864. PubMed doi:10.1016/j. ypmed.2003.12.018 20. Tudor-Locke C, Craig CL, Beets MW, et al. How many steps/day are enough? for children and adolescents. Int J Behav Nutr Phys Act. 2011;8(1):78. PubMed doi:10.1186/1479-5868-8-78 21. Whiting PF, Rutjes AWS, Westwood ME, et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med. 2011;155(8):529–536. PubMed doi:10.7326/0003-4819155-8-201110180-00009 22. Must A, Anderson SE. Body mass index in children and adolescents: considerations for population-based applications. Int J Obes. 2006;30(4):590–594. PubMed doi:10.1038/sj.ijo.0803300 23. Burkhauser RV, Cawley J. Beyond BMI: the value of more accurate measures of fatness and obesity in social science research. J Health Econ. 2008;27(2):519–529. PubMed doi:10.1016/j. jhealeco.2007.05.005 24. Ljungvall A, Gerdtham UG, Lindblad U. Misreporting and misclassification: implications for socioeconomic disparities in body-mass index and obesity. Eur J Health Econ. 2015;16(1):5–20. PubMed 25. Trost SG, Pate RR, Freedson PS, Sallis JF, Taylor WC. Using objective physical activity measures with youth: how many days of monitoring are needed? Med Sci Sports Exerc. 2000;32(2):426–431. PubMed doi:10.1097/00005768-200002000-00025 26. Trost SG, McIver KL, Pate RR. Conducting accelerometer-based activity assessments in field-based research. Med Sci Sports Exerc. 2005;37(11, Suppl):S531–S543. PubMed doi:10.1249/01. mss.0000185657.86065.98 27. Trost SG. State of the art reviews: measurement of physical activity in children and adolescents. Am J Lifestyle Med. 2007;1(4):299–314. doi:10.1177/1559827607301686
28. Adams MA, Johnson WD, Tudor-Locke C. Steps/day translation of the moderate-to-vigorous physical activity guideline for children and adolescents. Int J Behav Nutr Phys Act. 2013;10(10):49. PubMed doi:10.1186/1479-5868-10-49 29. Nihiser AJ, Lee SM, Wechsler H, et al. Body mass index measurement in schools. J Sch Heal. 2007;77(10):651–71; quiz 722–4. 30. Bonthuis M, Jager KJ, Abu-Hanna A, Verrina E, Schaefer F, van Stralen KJ. Application of body mass index according to height-age in short and tall children. PLoS One. 2013;8(8):e72068. PubMed doi:10.1371/ journal.pone.0072068 31. Wickramasinghe VP, Cleghorn GJ, Edmiston KA, Murphy AJ, Abbott RA, Davies PSW. Validity of BMI as a measure of obesity in Australian white Caucasian and Australian Sri Lankan children. Ann Hum Biol. 2005;32(1):60–71. PubMed doi:10.1080/03014460400027805 32. Wittmeier KDM, Mollard RC, Kriellaars DJ. Physical activity intensity and risk of overweight and adiposity in children. Obes (Silver Spring). 2008;16(2):415–20. 33. Ribeiro J, Guerra S, Pinto A, Oliveira J, Duarte J, Mota J. Overweight and obesity in children and adolescents: relationship with blood pressure, and physical activity. Ann Hum Biol. 2003;30(2):203–213. PubMed doi:10.1080/0301446021000054587 34. Tudor-Locke C, Ainsworth BE, Thompson RW, Matthews CE. Comparison of pedometer and accelerometer measures of free-living physical activity. Med Sci Sports Exerc. 2002;34(12):2045–2051. PubMed doi:10.1097/00005768-200212000-00027 35. Hajian-Tilaki K. Receiver Operating Characteristic (ROC) curve analysis for medical diagnostic test evaluation. Caspian J Intern Med. 2013;4(2):627–635. PubMed 36. Schulzer M. Diagnostic tests: a statistical review. Muscle Nerve. 1994;17(7):815–819. PubMed doi:10.1002/mus.880170719
JPAH Vol. 12, No. 8, 2015
REVIEW
Step-Count Guidelines for Children and Adolescents: A Systematic Review Michael Pereira da Silva, Fabio Eduardo Fontana, Eric Callahan, Oldemar Mazzardo, and Wagner De Campos Background: The aim of this systematic review was to identify the most optimal step-count cutoff for children and adolescents (5–19 years old) among guidelines currently available in the literature. Methods: The databases searched were PubMed, SportDiscus, Science Direct, Web of Science and LILACS. Studies were categorized into Health Cohort studies or Physical Activity (PA) Cohort studies according to the reference standard used. The quality of the studies was assessed using the QUADAS-2 instrument. Results: Six Health and 3 PA Cohort studies were included in the final pool of papers after Full Text reading. With the exception of a single study, studies demonstrated a high risk of methodological bias in at least 1 of the QUADAS-2 domains. Guidelines ranged from 10,000 to 16,000 steps/day for the Health studies (5–16 years old), and from 9,000 to 14,000 steps/day for PA studies (6–19 years old). Due to the high risk of methodological bias, none of the Health Cohort guidelines were endorsed. The PA Cohort study with the lowest risk of methodological bias suggested 12,000 steps/day for children and adolescents irrespective of gender. Conclusion: PA Cohort studies demonstrated lower risk of methodological bias than Health Cohort studies. The optimal youth step-count guideline of 12,000 steps/day was endorsed. Keywords: youth, QUADAS-2
The detection of insufficient levels of physical activity is essential for the identification of children and adolescents more prone to health problems such as diabetes, high blood pressure, and obesity,1,2 but it is currently difficult to accurately measure physical activity habits. Subjective measures, such as self-reported questionnaires, are compromised by the ability of children and adolescents to recall all relevant details about their daily physical activity behaviors and quantify physical activity intensity.3,4 This problem is amplified in children due to their less developed cognitive capacity and the more intermittent nature of their physical activity behavior.4 Accelerometers objectively measure duration, intensity, and frequency of physical activity,5,6 but they are expensive and lack a gold standard procedure to convert acceleration counts to exercise intensity.7 Pedometers may be a viable alternative to objectively measure physical activity by providing an overall measure of daily steps taken. They are inexpensive and conveniently sized,7–9 and its output is reliable and simple to comprehend.7 Pedometer step-counts correlate well with accelerometry and measures of energy expenditure.10 Pedometer daily step-counts can assist the screening and surveillance of physical activity behavior in pediatric populations. Several previous studies have proposed step-count guidelines,11–19 but the great variability in the recommendations may deter their use in field interventions (ie, 9000 to 16,000 steps/day). A recent comprehensive narrative review of literature20 has recommended a minimum of 13,000 to 15,000 steps/day for elementary school boys, 11,000 to 12,000 steps/day for elementary school girls, and 10,000 to 11,700 for adolescent boys and girls, but the guidelines were arbitrarily selected and the range may be confusing to practitioners. Thus, the aim of this study is to systematically assess
Silva ([email protected]), Mazzardo, and De Campos are with the Dept of Physical Education, Federal University of Parana, Curitiba, Parana, Brazil. Fontana and Callahan are with the School of Health, Physical Education, and Leisure Services, University of Northern Iowa, Cedar Falls, IA. 1184
the literature developing youth step-count guidelines to identify the most optimal cutoff point among step-count guidelines for children and adolescents (5–19 years old) currently available in the literature. For such, this review systematically extracted data from eligible studies and evaluated the risk of methodological bias and concerns over the applicability of findings. Based on the reference standard, studies were assigned to 2 groups: Health and Physical Activity (PA) Review Cohorts. The Health Review Cohort encompassed studies referencing step-count guidelines on Body Mass Index (BMI) or body fat percentage.12,14–17,19 Guidelines proposed by these studies represent the minimum count of daily steps necessary for the prevention of obesity. The PA Review Cohort included studies referencing guidelines on the attendance of physical activity recommendations.11,13,18 Step-count guidelines proposed by these studies indicate whether PA recommendations are being met.
Methods Literature Search A systematic search of peer reviewed articles published between January 1, 2000 and September 30, 2013 was performed using 5 databases (PubMed, SportsDiscus, Science Direct, Web of Science, and LILACS). Key words were used as MeSH terms and text words as appropriate for each database: step-counts (walking, pedometer), physical activity (motor activity, physical activity assessment, physical activity surveillance), health outcomes (body mass index, body composition, body fat), cutoff (cut-off points, step-count recommendations, step-count guidelines), and children (child, adolescent, adolescents, teenager). Search statements were truncated using Boolean terms. A new search was made between November 1, 2013 and March 31, 2014, but no additional studies were included. Screening of relevant papers followed a 4-step process: 1) assessment of titles, 2) assessment of abstracts, 3) full article review, and 4) screening for additional references among selected papers.
Step-Count Guidelines: A Systematic Review 1185
Two reviewers independently reviewed the papers during all 4 steps. Disagreements between the 2 reviewers were resolved by discussion at consensus meetings. Duplicate papers were excluded. According to the reference standard, eligible studies were assigned to Health (BMI or body fat percentage) or PA Review Cohorts (attendance to physical activity recommendations).
Demographic Characteristics
Inclusion Criteria To be included in the review, studies had to meet the following criteria: 1) development of step-count guidelines or testing existing step-count guidelines; 2) target group consisting of children and/ or adolescents 5 to 19 years of age; 3) basing step-count cutoffs on either a physical activity standard (ie, attainment of 60 minutes of moderate to vigorous physical activity daily) or a health standard (ie, BMI, body fat percentage); and 4) articles written in English, Portuguese, or Spanish. Downloaded by R B Draughon Library on 09/18/16, Volume 12, Article Number 8
studies were found after cross-referencing the selected studies. After reading the full text of all 16 studies, 9 studies fully met the inclusion criteria. Six of these studies were classified into Health Review Cohort and 3 in the PA Review Cohort. A summary of the review process is provided in Figure 1.
Data Extraction One reviewer extracted the following data from relevant articles: study identifiers (authors, year, country, journal), study design, study population (sample size, age), index text, reference standard, and proposed step-count cutoffs. A second reviewer double checked the extracted data, and discrepancies were resolved by discussion at consensus meetings.
Methodological Quality Assessment (QUADAS-2) The QUADAS-2 instrument assesses the methodological quality of diagnostic studies.21 It evaluates the risk of methodological bias and concerns over the applicability of findings. Risk of methodological bias is evaluated across 4 domains: participant selection, index text, reference standard, and flow and timing of data collection (ie, high, low, or unclear risk). Participant selection refers to the quality of the procedures used to select participants. For example, convenient sampling is a source of bias in a study measuring the quality of a diagnostic test. The index test pertains to the test being evaluated. In this systematic review, the index test was daily step-counts. Problems in the interpretation or procedures used to conduct the index test may introduce bias to the study. The reference standard refers to the test used to classify the target condition. BMI, body fat percentage, and attainment to physical activity guidelines measured via accelerometry have been used as reference standards to develop step-count guidelines. The reference standard itself, its conduct or interpretation may bias the results of a study. Flow and timing of data collection assesses the possibility that the sequence of study procedures bias the results of a study. QUADAS-2 also evaluates concerns over the applicability of findings across the aforementioned first 3 domains (ie, high, low, or unclear concern). As recommended,21 minor adaptations to the QUADAS-2 instrument were made to specifically assess research regarding step-count guidelines. QUADAS-2 was piloted on 4 studies before quality assessment (The modified instrument is available upon request). Two independent reviewers conducted the quality appraisal, and disagreements were resolved by discussion at consensus meetings.
Results Included Studies The initial search identified 452 references. Screening of Titles and Abstracts led to 13 potentially pertinent references. Three other
Studies in the Health Review Cohort ranged from 665 to 2698 participants, and studies in the PA Review Cohort ranged from 34 to 2217 participants. Participants in the Health Review Cohort were 5 to 17 years old and those in the PA Review Cohort were 6 to 19 years old. Studies in the Health Review Cohort were conducted in 4 countries [Australia (2 studies, N = 3373), United States of America (2 studies, N = 1879), New Zealand (1 study, N = 969), Multicenter (1 study, NUnited States of America = 711, NAustralia = 563; NSweden = 680)]. Studies in the PA Review Cohort were conducted in 3 countries [United States of America (1 study, N = 2217), Canada (1 study, N = 1619), United Kingdom (1 study, N = 34)]. All studies were written in English.
Methodological Quality Table 1 provides a summary of the methodological quality appraisal of the studies. Overall, 8 of the 9 studies had methodological limitations leading to a high risk of bias in at least 1 of the QUADAS-2 domains. Regarding the applicability of step-count guidelines to interventions, 6 of the 9 studies demonstrated a low concern.
Risk of Bias Patient Selection. Several studies (Health Review Cohort: 3 studies; PA Review Cohort: 1 study) showed high risk of bias during the selection of participants. The most common risk was convenient sampling of participants. Procedures used to select participants were unclear in 1 study.16 Index Test. Seven studies demonstrated high risk of bias during the handling of the index test (Health Review Cohort: 5 studies; PA Review Cohort: 2 studies). The most common limitations were the selection of a single cutoff point for a broad age range and use of 1000 steps/day increments to guide the statistical selection of optimal step-count cutoffs. One study did not provide sufficient information for the interpretation of the index test.16 Reference Standard. Five studies in the Health Review Cohort showed high risk of bias for the reference standard. They used BMI as the reference, even though the limitations of BMI have been extensively reported.22–24 The other Health study was rated unclear for risk of bias.15 Bioimpedance assessed body fat percentage, but the study lacked descriptions of preassessment procedures and information about the validity of the equipment used. All 3 studies in the PA Review Cohort demonstrated low risk of bias. Although there is no gold standard to measure physical activity habits, the accelerometer procedures used in the studies are considered highly acceptable in the literature.4,25–27 Flow and Timing. Two studies in the Health Review Cohort revealed high risk of bias because they did not include all subjects in the analysis.15,17 Four Health studies lacked a description of the interval between index testing and reference standard assessment,12,14,16,19 and the risk of bias during flow and timing of data collection was rated unclear. All studies in the PA Review
JPAH Vol. 12, No. 8, 2015
Downloaded by R B Draughon Library on 09/18/16, Volume 12, Article Number 8
1186 Silva et al
Figure 1 — Review process.
Cohort demonstrated low risk of bias for flow and timing of assessment.
study.13 In this study, steps were measured via accelerometry without a corresponding pedometer adapted step/day cutoff. Reference Standards. In Health and PA Review Cohorts, con-
Concerns of Applicability Patient Selection. Two studies in the Health Review Cohort
demonstrated high concern for participant selection.12,19 They referenced studies describing the representativeness of original samples but did not provide information about the representativeness of the subsamples used for developing step-count guidelines. All studies in the PA Review Cohort demonstrated low concern for the handling of participant selection. Index Test. All studies in the Health Review Cohort demonstrated
low concern for the applicability of the index test. Concern about the applicability of the index test was high in 1 PA Review Cohort
cerns about applicability of reference standard to assess the target conditions were low.
Optimal Step-Counts Guidelines Health Review Cohort Studies.
Table 2 provides a summary of the characteristics and suggested step-count guidelines for studies in the Health Review Cohort. Pedometers were used to measure step-counts in all studies. Five studies used BMI12,14,16,17,19 and 1 study used body fat percentage as the reference standard.15 With the exception of 2 studies that used ROC Curves,12,14 the contrasting groups method was the
JPAH Vol. 12, No. 8, 2015
Step-Count Guidelines: A Systematic Review 1187
Table 1 QUADAS-2 Results Risk of bias Participant selection
Beets et al (2008) Dollman et al (2010) Duncan et al (2007)
Study
Applicability concerns
Index test
Reference standard
Flow and timing
Patient selection
Index test
Reference standard
↑
↑
↑
?
↓
↓
↓
↓
↑
↑
?
?
↓
↓
↑
↑
?
↑
↓
↓
↓
BMI and body fat outcomes
Laurson et al (2008)
?
?
↑
?
↓
↓
↓
McComack et al (2011)
↓
↑
↑
↑
?
↓
↓
Tudor-Locke et al (2004)
↑
↑
↑
?
↓
↓
↓
Adams, Johnson and Tudor-Locke (2013)
↓
↑
↓
↓
↓
↓
↓
Colley, Jassen and Tremblay (2012)
↓
↓
↓
↓
↓
↑
↓
Rowlands and Eston (2005)
↑
↑
↓
↓
↓
↓
↓
Downloaded by R B Draughon Library on 09/18/16, Volume 12, Article Number 8
Physical activity outcome
Note. ↓ = Low; ↑ = High; ? = Unclear.
most common analysis used in the Health Cohort Review studies to identify optimal step-count cutoff points. Beets et al12 tested step-count cutoffs previously developed by Tudor-Locke et al.19 They concluded that step-count cutoffs were unable to distinguish between healthy and unhealthy BMI classified weight.12 Similarly, Dollman et al14 indicated a low capacity of step-count cutoff (AUC range: .52 to .60) to discriminate between youth individuals with healthy and overweight BMI classifications . Step-count guidelines proposed by studies in the Health Review Cohort were generally gender specific. For boys, step-count guidelines ranged from 11,000 to 16,000 steps/day. For girls, step-count guidelines ranged from 10,000 to 16,000 steps/day. McCormack et al17 suggested a single cutoff for both sexes, and Dollman et al14 proposed age specific cutoff points. Out of 4 QUADAS-2 methodological domains, studies in the Health Review Cohort demonstrated a high risk of methodological bias in at least 2 domains. Laurson et al16 demonstrated a high risk of bias in 1 domain, but there was insufficient information for the quality rating of the other 3 domains. Due to the high risk of methodological bias, this systematic review does not endorse any of the step-count guidelines proposed by Health Review Cohort studies. PA Review Cohort Studies. Table 3 provides a summary of the characteristics and suggested step-count guidelines for studies in the PA Review Cohort. The index test, daily step-counts, was measured by accelerometry in 2 studies13,28 and by pedometry in the other study.18 Sixty minutes of moderate to vigorous physical activity, assessed via accelerometry, was the reference standard in all PA Review Cohort studies. ROC curves were used to identify optimal step-count cutoffs in all studies in the PA Review Cohort. Two studies computed separate step-count guidelines by gender and developmental period, but they each suggested a single guideline due to the similarities among the cutoffs for each subgroup.13,28 Rowland and Eston18 suggested separate step-count guidelines for boys and girls 8 to 10 years old. Step-count guidelines proposed by the PA Review Cohort studies ranged from 9000 to 14,000 steps/day. Colley et al13 demonstrated low risk of bias across all 4 QUADAS-2 domains but a high concern over the applicability of the index test. This study suggested a cutoff of 12,000 steps/day. Adams et al11 demonstrated
low concern across all 3 applicability domains but demonstrated a high risk of bias in the index test domain. This study suggested 9000 steps/day.
Discussion Pedometer daily step-counts are a convenient and reliable measure of physical activity behavior among children and adolescents.7–9 However, previously proposed step-count guidelines vary greatly (ie, 9000–16,000 steps/day).11–19 Tudor-Locke et al20 comprehensively reviewed the literature developing youth step-count guidelines, but arbitrarily recommended steps/day ranges for children and adolescents. Thus, the purpose of this study was to systematically assess the literature developing step-count guidelines for children and adolescents. To reduce bias during the review process, systematic procedures and protocols were used to identify, select, and extract data from relevant articles, and selected articles were appraised based on methodological quality and applicability of step-count guidelines. Instead of a range, the use of systematic procedures also permitted the recommendation of a single most optimal stepcount cutoff point. This systematic review also benefited from the publication of 2 recent step-count guideline papers with excellent research quality appraisal.
Health Review Cohort Studies In general, studies in the Health Review Cohort demonstrated a high risk of methodological bias. Sources of methodological limitations included the selection of participants, interpretation of the index test, choice of the reference standard, and flow and timing of data collection. Participant selection is an important step in diagnostic studies. Random selection of participants is necessary to avoid the risk of methodological bias,21 but 3 studies conveniently sampled participants. The most common risk of bias in the interpretation of the index test was the use of 1,000-step increments to select optimal steps/day cutoffs. Such a procedure decreases the precision of the selected steps/day guideline. For example, choosing 12,000 steps/ day may lead to less desirable values of sensitivity and specificity than a cutoff of 12,450 steps/day.
JPAH Vol. 12, No. 8, 2015
1188 Silva et al
Downloaded by R B Draughon Library on 09/18/16, Volume 12, Article Number 8
Table 2 Summary of Health Review Cohort Studies Results Cutoff point suggested/ Conclusion
Study
Sample
Step-count assessment
Reference standard
Beets et al (2008)
United States, 1067 participants, Age: 6–12 years
Pedometer: Walk4Life MLS 2525 and Yamax SW 200. The pedometer was worn for 4 consecutive days. Tudor-Locke et al (2004) cutoff points were tested.
BMI: international cutpoint for youth was used (Cole et al, 2000).
None: Tudor-Locke cutoffs (15,000 steps/day for boys and 12,000 for girls) were unable to distinguish between healthy and unhealthy weight.
Dollman et al (2010)
Australia, 2698 participants, Age: 5–16 year. (Children: 5–12 years; Adolescents: 13–16 years)
Pedometer: New Lifestyles 1000. The pedometer was worn for 7 consecutive days. Four days (3 weekdays and 1 weekend) were necessary for analyses.
BMI: international cutpoint for youth was used (Cole et al, 2000).
Suggested cutoffs were 12,000 for younger males, 11,000 for older males, and 10,000 for younger females. The AUC for older females was nonsignificant.
Duncan et al (2007)
New Zealand, 969 participants, Age: 5–12 years
Pedometer: Multidaymemory NL-2000. The pedometer was worn for 3 weekdays and 2 weekend days.
% Body Fat: A hand-to-foot Suggested cutoffs were 16,000 steps/day for boys bioelectrical impedance and 13,000 for girls. analyzer (Model BIM4). Excess of %BF was classified as % Body Fat ≥85th Percentile.
Laurson et al (2008)
United States, 812 participants, Age: 6–12 years
Pedometer: DigiwalkerSW-200. The pedometer was worn for 7 consecutive days. Four days (3 weekdays and 1 weekend) were necessary for analyses.
BMI: international cutpoint for youth was used (Cole et al, 2000).
Suggested cutoffs for obese were 11,500 steps/day for boys and 10,000 for girls. The cutoffs for overweight were 12,000 steps/day for boys and 10,500 for girls.
McComack et al (2011)
Australia, 675 participants, Age: 7–16 years
Pedometer: Digiwalk SW-700. The pedometer was worn for 7 consecutive days. Four days (3 weekdays and 1 weekend) were necessary for analyses.
BMI: international cutpoint for youth was used (Cole et al, 2000).
The suggested cutoff was 16,000 steps/day for both boys and girls.
Tudor-Locke et al (2004)
Multicenter (United States, Australia & Sweden), 1954 participants, Age: 6–12 years
Pedometer: My Life Stepper, MLS-2000. The pedometer was worn for 4 consecutive school days.
BMI: international cutpoint for youth was used (Cole et al, 2000).
Cutoffs ranged from 14,000 to 17,000 steps/day for boys and from 10,000 to 13,000 for girls. Suggested cutoffs were 15,000 steps/ day for boys and 12,000 for girls.
Most Health Review Cohort studies used BMI as the reference standard. BMI often misclassifies individuals due to its inability to account for lean mass weight and differences in biological maturation.29–31 The low association between PA and BMI may affect the ability of step-counts to discriminate obesity levels.32,33 In fact, Dollman et al14 indicated a poor diagnostic ability of step-counts to differentiate between BMI classified healthy and overweight children and adolescents (AUC range: .52 to .60). All Health Review Cohort studies demonstrated the same limitation for flow and timing of data collection. They lacked a description of the
interval between index test and reference standard measurements. If too much time elapses between the 2 measurements, variations in the index test (steps/day) or reference standard (obesity classification) may bias the results. Developmental and gender differences in BMI have been widely reported. Health review cohort studies generally took gender into account for the development of step-count guidelines. With the exception of a single study, step-count guidelines were offered separately by gender.17 However, only Dollman et al14 developed optimal step-count guidelines by gender and developmental stage.
JPAH Vol. 12, No. 8, 2015
Step-Count Guidelines: A Systematic Review 1189
Downloaded by R B Draughon Library on 09/18/16, Volume 12, Article Number 8
Table 3 Summary of PA Review Cohort Studies Results Cutoff point suggested/ Conclusion
Study
Sample
Step-count assessment
Reference standard
Adams, Johnson and Tudor-Locke (2013)
United States, 2217 participants, Age: 6–17 years (Children: 6–11 years; Adolescents: 12–17 years)
Accelerometer: ActiGraph 7164. Average daily accelerometer steps was computed and adapted to pedometer steps.
Accelerometer: ActiGraph 7164. Participants with at least 1 valid day (≥ 10 hours) were included in the analyses. Freedson 3 and 4 METs and Evenson equations were used.
A range of 11,500–13,500 accelerometer steps/day for children and 11,500–14,000 accelerometer steps/ day for adolescents was suggested for boys and girls. A pedometer-friendly adaptation of 9000 steps/day was recommended for boys and girls.
Colley, Jassen and Tremblay (2012)
Canada, 1619 participants, Age: 6–19 years
Accelerometer: Actical. Pedometer friendly adaptations were not offered.
Accelerometer: Actical. Each of 4+ valid days (≥ 10 hours) was assessed separately. The cutoff point for MVPA was >1500 counts/min.
The cutoff point of 12,000 steps/day was recommended for boys and girls 6–19 years of age.
Rowlands and Eston (2005)
United Kingdom, 34 participants, Age: 8.3–10.8 years
Pedometer: Yamax DigiWalker DW-200. Previous cutoff points were tested (Tudor-Locke et al, 2004; Vincent & Pangrazi, 2002).
Accelerometer: Tritrac-T303 and T303A. A single measurement day was used in analyzes. Counts/min used to specify physical activity intensities was based in Oxygen uptake studies (Eston et al, 1998; Rowlands et al, 1999).
Tudor-Locke et al (2004) cutoff point for boys may be too high to identify 60 min of MVPA. It was recommended 13,000 steps/ day for boys and 12,000 for girls.
Ignoring age-related differences in BMI was a source of bias in 5 studies12,14,16,17,19. Methodological limitations prevent the endorsement of stepcount guidelines proposed by the Health Review Cohort studies. Future studies should choose health standards with stronger validity evidence and more highly associated with physical activity (ie, metabolic syndrome risk factors). It is also recommended that future studies assess the sensitivity and specificity across the entire continuum of possible steps/day cutoffs.
PA Review Cohort Studies PA Review Cohort studies took into consideration gender and developmental differences during the development of step-count guidelines. All 3 studies13,18,28 established gender and age appropriate step-count guidelines. Rowlands and Eston18 restricted the sample to participants between 8 and 10 years of age suggesting distinct step-count cutoffs by gender. Adams et al28 developed separate guidelines by gender for children (6–11 years old) and adolescents (12–17 years old), but the similarity in step-count cutoffs among the different subgroups led to the recommendation of a single pedometer scaled step-count cutoff. Colley et al13 developed separate guidelines by gender for 3 age groups: children (6–10 years old), early adolescence (11–14 years old), and late adolescence (15–19 years old). Colley et al13 also recommended a single step-count guideline for girls and boys 6 to 19 years of age due to the similarity among the cut-points developed for the subgroups. Quality assessment of studies guided the determination of the most optimal step-count guidelines currently available in the literature. Guidelines offered by studies with a high risk of
methodological bias are likely to be imprecise for the identification of children not reaching physical activity recommendations. Rowlands and Eston18 had a high risk of methodological bias during the selection of participants and interpretation of the index test. In addition, inferences made from this study are limited by the small sample power of the study (N = 17 participants/group). Adams et al28 used a large and representative sample of US children and adolescents to develop step-count guidelines. Although steps were measured via accelerometry, a pedometer adapted 9000 steps/day guideline was suggested. The applicability of the guidelines was improved by converting accelerometer to pedometer steps, but 9000 steps/day may be too low to identify individuals reaching 60 minutes of moderate to vigorous activity. In fact, the cutoff is the lowest cutoff among all the studies developing step-count guidelines. This study was also limited by the selection of step-count cutoffs using 500-step increments. Colley et al13 demonstrated a high concern over the applicability of the index test selected. The study measured steps using Actical accelerometers, but it did not provide an estimate of corresponding pedometer steps/day cutoff. It is possible that the guidelines developed using accelerometer counted steps are not appropriate for pedometer based interventions. Accelerometers are considered to be more sensitive at identifying steps during low frequency acceleration than pedometers.34 While concerns of the applicability of the step-count guidelines were high, this study was methodologically well designed. It was the only study to demonstrate low risk of bias in all 4 methodological domains. Thus, this review endorses the guidelines suggested by Colley et al13 of 12,000 steps/day. Interestingly, this cutoff was commonly proposed by many of the Health and PA Review Cohort studies. In the Health Review Cohort studies,
JPAH Vol. 12, No. 8, 2015
1190 Silva et al
12,000 steps/day was suggested twice for girls12,19 and twice for boys.14,16 Among studies in the PA Review Cohort, Rowlands and Eston18 suggested it for 8- to 10-year-old girls, and Adams, Johnson and Tudor-Locke28 included 12,000 steps/day within the suggested accelerometer steps/day guidelines. It is strongly recommended that pedometry can be used as the index test of future studies. The use of pedometers to count steps eliminates possible errors related to the accelerometer-pedometer conversion process. Low-cost and easiness-of-use make pedometers friendly devices to count daily steps for the discrimination between sufficiently and insufficiently active individuals.
Downloaded by R B Draughon Library on 09/18/16, Volume 12, Article Number 8
Limitations, Strengths, and Recommendations for Future Studies This systematic review is not without limitations. Meta-analytic procedures were not possible due to the limited number of studies on the topic (N = 9), and divergent use of methodological (Health vs. PA Review Cohorts) and statistical procedures (Contrasting groups method vs. ROC curves). The performance of a meta-analysis would have strengthened the choice of a step-count guideline. There are also several strengths to this systematic review. Systematic review procedures were strictly followed. Two reviewers were used throughout the review process, and QUADAS-2 is considered an appropriate instrument to assess methodological bias of diagnostic studies.21 Six out of nine studies included in this systematic review were published in the last 6 years.12–14,16,17,28 Finally, recommendations made in this systematic review should assist the design of future studies. It is recommended that future research developing step-count guidelines base the index test on pedometry, develop guidelines for a narrow developmental span, assess the sensitivity and specificity of the entire steps/day continuum, use ROC curves to determine optimal cutoff points,35,36 clearly describe the timeline between the index test and reference standard, randomly sample participants, and use accelerometry output as the reference standard. In addition to values of sensitivity and specificity of each proposed steps/day cutoff, ROC curves also provide an indication of the discriminatory power of the index test based on the area under the curve statistic. Although accelerometry was suggested as the reference standard, it is important to acknowledge the lack of a gold standard to measure PA behavior in field research. Consequently, studies developing step-count guidelines using PA reference standards should check the utility of proposed step-count guidelines for the identification of health conditions such as the metabolic syndrome.
Conclusion The quality of studies developing step-count guidelines was systematically assessed to determine the most optimal cutoff point among currently available guidelines. Out of the nine studies assessed, Colley et al13 had the lowest risk of methodological bias. Thus, this systematic review endorses the step-count guidelines proposed by Colley et al13 of 12,000 steps/day for children and adolescents irrespective of gender. Acknowledgments The authors would like to acknowledge the Federal University of Parana and the University of Northern Iowa.
References 1. Luke A, Dugas LR, Durazo-Arvizu RA, Cao G, Cooper RS. Assessing physical activity and its relationship to cardiovascular risk factors: NHANES 2003-2006. BMC Public Health. 2011;11(1):387. PubMed doi:10.1186/1471-2458-11-387 2. Goodman E, Daniels SR, Meigs JB, Dolan LM. Instability in the diagnosis of metabolic syndrome in adolescents. Circulation. 2007;115(17):2316–2322. PubMed doi:10.1161/CIRCULATIONAHA.106.669994 3. Sallis JF, Buono MJ, Roby JJ, Micale FG, Nelson JA. Sevenday recall and other physical activity self-reports in children and adolescents. Med Sci Sports Exerc. 1993;25(1):99–108. PubMed doi:10.1249/00005768-199301000-00014 4. Corder K, Ekelund U, Steele RM, Wareham NJ, Brage S. Assessment of physical activity in youth. J Appl Physiol. 2008;105(3):977–987. PubMed doi:10.1152/japplphysiol.00094.2008 5. Nader PR, Bradley RH, Houts RM, McRitchie SL, O’Brien M. Moderate-to-vigorous physical activity from ages 9 to 15 years. JAMA. 2008;300(3):295–305. PubMed doi:10.1001/jama.300.3.295 6. Troiano RP, Berrigan D, Dodd KW, Mâsse LC, Tilert T, McDowell M. Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc. 2008;40(1):181–188. PubMed doi:10.1249/ mss.0b013e31815a51b3 7. Duncan JS, Schofield G, Duncan EK. Pedometer-determined physical activity and body composition in New Zealand children. Med Sci Sports Exerc. 2006;38(8):1402–1409. PubMed doi:10.1249/01. mss.0000227535.36046.97 8. Duncan MJ, Al-Nakeeb Y, Woodfield L, Lyons M. Pedometer determined physical activity levels in primary school children from central England. Prev Med. 2007;44(5):416–420. PubMed doi:10.1016/j. ypmed.2006.11.019 9. Gabel L, Proudfoot N a, Obeid J, et al. Step count targets corresponding to new physical activity guidelines for the early years. Med Sci Sports Exerc. 2013;45(2):314–318. PubMed doi:10.1249/ MSS.0b013e318271765a 10. Tudor-Locke C, Williams JE, Reis JP, Pluto D. Utility of pedometers for assessing physical activity: convergent validity. Sports Med. 2002;32(12):795–808. PubMed doi:10.2165/00007256-20023212000004 11. Adams MA, Caparosa S, Thompson S, Norman GJ. Translating physical activity recommendations for overweight adolescents to steps per day. Am J Prev Med. 2009;37(2):137–140. PubMed doi:10.1016/j. amepre.2009.03.016 12. Beets MW, Le Masurier GC, Beighle A, et al. Are current body mass index referenced pedometer step-count recommendations applicable to US youth? J Phys Act Health. 2008;5(5):665–674. PubMed 13. Colley RC, Janssen I, Tremblay MS. Daily step target to measure adherence to physical activity guidelines in children. Med Sci Sports Exerc. 2012;44(5):977–982. PubMed doi:10.1249/MSS.0b013e31823f23b1 14. Dollman J, Olds TS, Esterman A, Kupke T. Pedometer step guidelines in relation to weight status among 5- to 16-year-old Australians. Pediatr Exerc Sci. 2010;22(2):288–300. PubMed 15. Duncan JS, Schofield G, Duncan EK. Step count recommendations for children based on body fat. Prev Med. 2007;44(1):42–44. PubMed doi:10.1016/j.ypmed.2006.08.009 16. Laurson KR, Eisenmann JC, Welk GJ, Wickel EE, Gentile DA, Walsh DA. Evaluation of youth pedometer-determined physical activity guidelines using receiver operator characteristic curves. Prev Med. 2008;46(5):419–424. PubMed doi:10.1016/j.ypmed.2007.12.017 17. McCormack GR, Rutherford J, Giles-Corti B, Tudor-Locke C, Bull F. BMI-referenced cut-points for recommended daily pedometerdetermined steps in Australian children and adolescents. Res Q Exerc Sport. 2011;82(2):162–167. PubMed doi:10.1080/02701367.2011.10 599743 18. Rowlands AV, Eston RG. Comparison of accelerometer and pedometer measures of physical activity in boys and girls, ages 8–10 years. Res Q Exerc Sport. 2005;76(3):251–257. PubMed doi:10.1080/0270136 7.2005.10599296
JPAH Vol. 12, No. 8, 2015
Downloaded by R B Draughon Library on 09/18/16, Volume 12, Article Number 8
Step-Count Guidelines: A Systematic Review 1191
19. Tudor-Locke C, Pangrazi RP, Corbin CB, et al. BMI-referenced standards for recommended pedometer-determined steps/day in children. Prev Med. 2004;38(6):857–864. PubMed doi:10.1016/j. ypmed.2003.12.018 20. Tudor-Locke C, Craig CL, Beets MW, et al. How many steps/day are enough? for children and adolescents. Int J Behav Nutr Phys Act. 2011;8(1):78. PubMed doi:10.1186/1479-5868-8-78 21. Whiting PF, Rutjes AWS, Westwood ME, et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med. 2011;155(8):529–536. PubMed doi:10.7326/0003-4819155-8-201110180-00009 22. Must A, Anderson SE. Body mass index in children and adolescents: considerations for population-based applications. Int J Obes. 2006;30(4):590–594. PubMed doi:10.1038/sj.ijo.0803300 23. Burkhauser RV, Cawley J. Beyond BMI: the value of more accurate measures of fatness and obesity in social science research. J Health Econ. 2008;27(2):519–529. PubMed doi:10.1016/j. jhealeco.2007.05.005 24. Ljungvall A, Gerdtham UG, Lindblad U. Misreporting and misclassification: implications for socioeconomic disparities in body-mass index and obesity. Eur J Health Econ. 2015;16(1):5–20. PubMed 25. Trost SG, Pate RR, Freedson PS, Sallis JF, Taylor WC. Using objective physical activity measures with youth: how many days of monitoring are needed? Med Sci Sports Exerc. 2000;32(2):426–431. PubMed doi:10.1097/00005768-200002000-00025 26. Trost SG, McIver KL, Pate RR. Conducting accelerometer-based activity assessments in field-based research. Med Sci Sports Exerc. 2005;37(11, Suppl):S531–S543. PubMed doi:10.1249/01. mss.0000185657.86065.98 27. Trost SG. State of the art reviews: measurement of physical activity in children and adolescents. Am J Lifestyle Med. 2007;1(4):299–314. doi:10.1177/1559827607301686
28. Adams MA, Johnson WD, Tudor-Locke C. Steps/day translation of the moderate-to-vigorous physical activity guideline for children and adolescents. Int J Behav Nutr Phys Act. 2013;10(10):49. PubMed doi:10.1186/1479-5868-10-49 29. Nihiser AJ, Lee SM, Wechsler H, et al. Body mass index measurement in schools. J Sch Heal. 2007;77(10):651–71; quiz 722–4. 30. Bonthuis M, Jager KJ, Abu-Hanna A, Verrina E, Schaefer F, van Stralen KJ. Application of body mass index according to height-age in short and tall children. PLoS One. 2013;8(8):e72068. PubMed doi:10.1371/ journal.pone.0072068 31. Wickramasinghe VP, Cleghorn GJ, Edmiston KA, Murphy AJ, Abbott RA, Davies PSW. Validity of BMI as a measure of obesity in Australian white Caucasian and Australian Sri Lankan children. Ann Hum Biol. 2005;32(1):60–71. PubMed doi:10.1080/03014460400027805 32. Wittmeier KDM, Mollard RC, Kriellaars DJ. Physical activity intensity and risk of overweight and adiposity in children. Obes (Silver Spring). 2008;16(2):415–20. 33. Ribeiro J, Guerra S, Pinto A, Oliveira J, Duarte J, Mota J. Overweight and obesity in children and adolescents: relationship with blood pressure, and physical activity. Ann Hum Biol. 2003;30(2):203–213. PubMed doi:10.1080/0301446021000054587 34. Tudor-Locke C, Ainsworth BE, Thompson RW, Matthews CE. Comparison of pedometer and accelerometer measures of free-living physical activity. Med Sci Sports Exerc. 2002;34(12):2045–2051. PubMed doi:10.1097/00005768-200212000-00027 35. Hajian-Tilaki K. Receiver Operating Characteristic (ROC) curve analysis for medical diagnostic test evaluation. Caspian J Intern Med. 2013;4(2):627–635. PubMed 36. Schulzer M. Diagnostic tests: a statistical review. Muscle Nerve. 1994;17(7):815–819. PubMed doi:10.1002/mus.880170719
JPAH Vol. 12, No. 8, 2015
