This article was downloaded by: [University of Sussex Library] On: 25 March 2015, At: 07:36 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
European Journal of Sport Science Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/tejs20
Cardiovascular fitness is associated with bias between self-reported and objectively measured physical activity a
a
a
S. A. Tomaz , E. V. Lambert , D. Karpul & T. L. Kolbe-Alexander
ab
a
Department of Human Biology, Faculty of Health Sciences, MRC/UCT Research Unit for Exercise Science and Sports Medicine, University of Cape Town, Cape Town, South Africa b
Centre for Research on Exercise, Physical Activity and Health, School of Human Movement Studies, University of Queensland, Brisbane, QLD, Australia Published online: 24 Dec 2014.
Click for updates To cite this article: S. A. Tomaz, E. V. Lambert, D. Karpul & T. L. Kolbe-Alexander (2014): Cardiovascular fitness is associated with bias between self-reported and objectively measured physical activity, European Journal of Sport Science, DOI: 10.1080/17461391.2014.987323 To link to this article: http://dx.doi.org/10.1080/17461391.2014.987323
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
European Journal of Sport Science, 2014 http://dx.doi.org/10.1080/17461391.2014.987323
ORIGINAL ARTICLE
Cardiovascular fitness is associated with bias between self-reported and objectively measured physical activity
S. A. TOMAZ1, E. V. LAMBERT1, D. KARPUL1, & T. L. KOLBE-ALEXANDER1,2 1
Downloaded by [University of Sussex Library] at 07:36 25 March 2015
Department of Human Biology, Faculty of Health Sciences, MRC/UCT Research Unit for Exercise Science and Sports Medicine, University of Cape Town, Cape Town, South Africa, 2Centre for Research on Exercise, Physical Activity and Health, School of Human Movement Studies, University of Queensland, Brisbane, QLD, Australia
Abstract The aim of this research study was to determine whether the level of agreement between self-reported and objective measures of physical activity (PA) is influenced by cardiovascular fitness. Participants (n = 113) completed the Global Physical Activity Questionnaire (GPAQ), a health risk assessment and a sub-maximal 12-minute step test. Age-predicted V_ O2 max was used to classify participants as lower fit and higher fit (HF). ActiGraph (GT3X) accelerometers were worn for 7 consecutive days. Matthews cut points were used to calculate minutes of moderate and vigorous PA (MVPA) per week. Bland–Altman plots were used to measure limits of agreement between GPAQ and ActiGraph MVPA. The participants’ mean age was 37.9 ± 12.7 years and more than 60% were categorised as HF (n = 71). Moderate PA was over-reported in 39% of all participants. Most of the over-reporters for moderate PA were in the HF group (64.1%). Vigorous PA was overreported by 72.6% of all participants. The discrepancy between self-reported and objective measures of vigorous PA increased with increasing self-reported time spent in vigorous PA. Fitter individuals appear to over-report PA more than lesser fit participants, suggesting that fitness could influence the level of agreement between self-reported and objective measures of PA. Keywords: Accelerometer, ActiGraph GT3X, GPAQ, physical activity
Introduction Physical inactivity has been identified as the fourth leading cause of global mortality related to noncommunicable diseases (NCDs), resulting in an estimated 3.2 million deaths, and is therefore a global concern (WHO, 2010b). The current guideline for physical activity (PA) for adults is 150 minutes of moderate intensity activity, or 60 minutes of vigorous activity per week (American College of Sports Medicine [ACSM], 2010). Therefore, those individuals who are not meeting these guidelines are most often referred to as physically inactive in the literature (WHO, 2010a). Based on these guidelines, 46.4% of South African males and 55.7% of South African females are insufficiently active based on self-report data obtained from the Global Status Report (WHO, 2010b). Consequently, the prevalence of inactivity is higher among the South African
adult population than the global prevalence which is 28% in men and 34% in women (WHO, 2010b). PA can be quantified by both subjective and objective measures using self-report questionnaires and accelerometers, respectively. Accelerometers are small devices used to measure movement using counts from which moderate and vigorous intensity PA is estimated. A review by Prince et al. (2008) noted that accelerometry is the most common method used for measuring PA objectively (Prince et al., 2008). The ActiGraph brand of accelerometers has been used extensively in research and well validated (Le Masurier & Tudor-Locke, 2003; Lyden, Kozey, Staudenmeyer, & Freedson, 2011; McClain, Sisson, & Tudor-Locke, 2007; Plasqui & Westerterp, 2006, 2007).The GT3X has been found to have good inter-instrument reliability (SantosLozano, Torres-Luque, et al., 2012), as well as being
Correspondence: S. A. Tomaz, Department of Human Biology, Faculty of Health Sciences, MRC/UCT Research Unit for Exercise Science and Sports Medicine, University of Cape Town, Rondebosch, Cape Town 7700, South Africa. E-mail: [email protected] © 2014 European College of Sport Science
Downloaded by [University of Sussex Library] at 07:36 25 March 2015
2
S. A. Tomaz et al.
an accurate tool in measuring free-living PA and in clinical studies (Santos-Lozano, Marín, et al., 2012; Vanhelst et al., 2012). The Global Physical Activity Questionnaire (GPAQ) and International Physical Activity Questionnaire (IPAQ) are examples of questionnaires that have been used to measure self-reported PA. The GPAQ was developed as a global surveillance tool for PA and has been used successfully for population surveillance in African countries (Guthold et al., 2011), including South Africa (Armstrong & Bull, 2006). Both questionnaires require participants to estimate the amount of moderate and vigorous intensity PA performed, for at least 10 minutes at a time, in the various domains of PA. The domains include transport, occupation, leisure time and domestic. There have been several comparative studies evaluating the discrepancies between objective and self-reported measurement of PA (Downs, Van Hoomissen, Lafrenz & Julka, 2013; Tucker, Welk & Beyler, 2011). Tucker et al. (2011) compared selfreport and accelerometer PA outcomes and determined the relationship between these two measurement tools using the NHANES sample (n = 3082). Results of this study showed large discrepancies between the accelerometer (ActiGraph model 7164) and self-report, with 62% of the sample meeting guidelines according to self-report, compared with objective measurement showing only 9.6% of the sample meeting guidelines. A more recent study conducted in college students compared the ActiGraph GT3X with the short IPAQ (Downs et al., 2013), and showed that self-reported moderate and vigorous PA (MVPA) averaged 66.14 min/day, as opposed to objectively measured MVPA which averaged 19.90 min/day. Furthermore, these measures were not shown to be significantly correlated (r(77) = 0.21; p = 0.08). Characteristics of participants who over-report have been identified in studies comparing self-report and objectively measured activity. These participants tend to be older, less than well educated (the majority of the participants in this particular sample were said to have completed a minimum of 11 years of formal education), are more likely to be smokers, and have central obesity (Craig et al., 2003; Fillipas, Cicuttini, Holland, & Cherry, 2010; Fogelholm et al., 2006; Gerrard, 2012; Irwin, Ainsworth, & Conway, 2001a; Tehard et al., 2005; Timperio, Salmon, & Crawford, 2003;). Habitual levels of PA and cardiovascular fitness could also influence self-reported PA. A recent study by Schuna, Johnson, and TudorLocke (2013) investigated, among other factors, the difference in self-reported and objectively measured PA (Schuna et al., 2013). This study showed that individuals who self-report compliance with MVPA
guidelines do accumulate more objectively measured PA (for all PA indicators, including light, MVPA). This study was performed on the NHANES cohort and investigated the level of agreement based on levels of PA, and more specifically, meeting PA guidelines and how it may influence an individual’s ability and accuracy when reporting PA (Schuna et al., 2013). Thus, there has been limited research investigating the relationship between cardiovascular fitness levels and the accuracy of self-reported PA. Identifying factors, which may influence selfreported PA, in relation to objectively measured PA, may assist researchers with data interpretation and application of self-reported activity, as well as practitioners, for exercise prescription. Therefore, the purpose of this study was to determine the extent to which cardiovascular fitness influences self-reported moderate and vigorous intensity PA, in relation to objectively measured PA. We hypothesised that persons with higher fitness (HF) levels would report PA with greater accuracy than individuals who present with lower fitness (LF) levels. Methods Participants A convenience sample of 113 men (n = 52) and women (n = 61) were recruited from individuals who volunteered to participate in an annual health and risk assessment (HRA) and fitness test, which formed part of a health insurance wellness programme. The goal of the HRA with respect to the health wellness programme is to promote healthy lifestyle behaviour and thus play a role in reducing the risk of NCDs. All participants provided signed informed consent prior to participating in the study. The research study was approved by the Human Research Ethics committee of the University of Cape Town (Ethics reference number 348/2008). Measures Questionnaires Health risk assessment. The HRA was comprised of questions pertaining to the participant’s general demographic information, personal medical history, and family history of disease, medication, injuries sustained and exercise-induced symptoms. Clinical measures and anthropometry. Blood pressure was measured one time following 5 minutes of sitting using a manual blood pressure cuff (ACSM, 2010). Height was measured in meters using an upright Seca stadiometer. Participants stood barefoot with their heels and head in contact with the wall and arms at their side. Weight was measured in
Self-reported and objectively measured PA
Downloaded by [University of Sussex Library] at 07:36 25 March 2015
kilograms using a Clover Scale (Model TCS-A300), with the value rounded to the nearest 100 g. Body mass index (BMI) was calculated using the equation: BMI = weight (kg)/height (m)2. Waist circumference was measured narrowest point of the torso, between the xiphoid process and umbilicus, using a standard tape measure on the skin (ACSM, 2010). Body fat (BF) percentage was estimated, using measures from four different skinfolds sites, including the biceps, triceps, subscapular and suprailiac regions using Harpenden skinfold callipers, as described by Durnin and Womersly (1974). Fitness assessment. A 25 cm Reebok step was used to complete a 12-minute sub-maximal step test that has been validated for use in measuring exercise capacity for HRA purposes (Dugas et al., 2006). The step test consisted of four, 2-minute stepping stages, where each stage was followed by a 1-minute rest period prior to commencing the next stage (Pillay, KolbeAlexander, van, & Lambert, 2012). Participants were required to step up and down to the sound of a metronome, of which the cadence increased as the step test progressed from stage one to stage four (80, 96, 112 and 120 steps/min, respectively). A Suunto heart rate monitor was used to record the heart rate response during the step test. The heart rate file was log analysed using the Suunto Training Manager and Suunto Monitor software (Suunto Oy, Finland). The heart rate response to exercise was then regressed to predict peak metabolic equivalents (METS) at age-predicted maximum heart rate. An estimated maximal oxygen uptake (ml/kg/min) was calculated, based on heart rate recovery at the end of the last stage, using the following equation: 44.891 – (age × 0.262) – (gender × 0.855) + (peak METS × 0.994) + (maximum reported MET hrs/wk of activity × 0.163) (Dugas, van der Merwe, Odendaal, Noakes, & Lambert, 2005; Keytel et al., 2005). The normative values for maximal aerobic power (referred to as predicted V_ O2 max in this paper) according to the ACSM Guidelines for Exercise Testing and Prescription were used to categorise participants in two groups, namely ‘LF’ and ‘HF’ (ACSM, 2010). Participants who fell below the 50th percentile (according to the ACSM (2010) genderbased norms for age-predicted V_ O2 max ) were categorised as LF participants. Global Physical Activity Questionnaire. The GPAQ is based on self-reported recall of moderate and vigorous intensity activities. The GPAQ has been validated as a reliable and valid measure in population surveillance (Bull, Maslin, & Armstrong, 2009; Craig et al., 2003; Hagstromer, Oja, & Sjostrom, 2006). Activities included in the report should be performed for at least 10 consecutive
3
minutes, in a typical week, in each of the four domains, including transport, occupational, leisure and domestic activities (Armstrong & Bull, 2006). The GPAQ was administered to each of the participants individually. Accelerometry. The tri-axial ActiGraph GT3X accelerometer (ActiGraph, Shalimar, FL, USA) was used in this research study. The device was set to measure motion in all three axes and was initialised to record data in 60-second epochs (Trost, McIver, & Pate, 2005). Accelerometer data were captured and cleaned using the ActiLife Software Version 5 (ActiLife software; Pensacola, FL, USA). Data from all three axes were captured and used in the analysis. The accelerometer was given to the participants after the GPAQ was administered, and participants were requested to maintain their usual activity levels while wearing the device. Participants were requested to wear the accelerometer on their right hip for seven consecutive days during their waking hours, but not while bathing, showering or swimming. A minimum of 4 days of wear, with 600 minutes per day were required for data analysis. Non-wear time was defined as 60 continuous minutes of no counts (zeros). Four days of monitoring moderate and vigorous intensity activity has been found to provide 80% reliability (Matthews, Ainsworth, Thompson, & Bassett, 2002; Trost et al., 2005). Moderate and vigorous intensity PA levels were estimated based on accelerometer counts using the cut-off points according to Matthews (2005). The Matthews equation stipulates that counts between 760 and 5998 (inclusive) represents moderate intensity PA and counts above 5999 is indicative of vigorous activity (Matthews, 2005). The Matthews and Troiano et al. (2008) cut points were recently compared to indirect caloriometry and free-living conditions (Crouter, Dellavalle, Haas, Frongillo, & Bassett, 2013). The Matthews cut point over-estimated moderate PA by 33.4% in comparison to freeliving conditions, whereas the Troiano’s cut point under-estimated moderate PA by 50.4% (Crouter et al., 2013). Of the cut points evaluated in this study by Crouter et al., Matthews cut points had lowest mean bias for time spent in MVPA, overestimating measured time spent in MVPA in comparison to indirect calirometry by 13.6 minutes (26.3%) (Crouter et al., 2013). Objective PA from the GT3X was analysed for any moderate-to-vigorous activity which took place for 10-minute bouts/intervals. MVPA was calculated by summing moderate and vigorous activity (Downs et al., 2013). Within each 10-minute bout of moderate and/or vigorous PA, 1 or 2 minutes of ‘dropped’ counts were allowed (Masse et al., 2005;
4
S. A. Tomaz et al.
Ward, Evenson, Vaughn, Rodgers, & Troiano, 2005). An exercise bout was therefore excluded if the programme detected a drop in counts for greater than 2 minutes within the 10-minute period. The term ‘delta’ was used to describe the difference between the objective and self-report measurements of PA. Delta values were recorded as minutes. Minutes spent in PA reported in the GPAQ were subtracted from the minutes measured by the accelerometer (GT3X). Therefore, a positive value would mean that the participant under-reported and negative value would indicate over-reporting. A zerovalue represents accurate reporting.
compare self-reported moderate PA to objectively measured moderate activity, as well as for selfreported vigorous activity and objectively measured vigorous activity. Bland–Altman plots were used to calculate the limits of agreement between the accelerometer counts (objective) and the GPAQ (self-report) for both the HF and LF groups. Intra-class correlations (ICCs) were performed, in addition to the use of the Spearman–Brown Prophecy formula to determine the number of days of monitor wear required to achieve a desired ICC of r = 0.8 (Baranowski & de Moor, 2000).
Downloaded by [University of Sussex Library] at 07:36 25 March 2015
Statistical analysis
Results
STATISTICA 9 software (Stasoft Inc., Tulsa, OK, USA) was used for all statistical analyses. Descriptive statistics were applied to calculate the mean values and standard deviations of all demographic, physical fitness, GPAQ and ActiGraph GT3X accelerometer results. A custom Matlab (R2011b The Mathworks Inc.) program was used to calculate the total amount of time spent doing moderate and vigorous intensity PA using 10-minute bouts. Shapiro–Wilk tests for normality were performed to check whether the data were normally distributed. T-tests for independent groups were used to determine significant differences between age, weight, BMI, BF percentage, systolic and diastolic blood pressure (SBP and DBP, respectively) for the HF and LF groups, as well as between the male and female participants. Spearman’s (non-parametric) correlations were used to measure the relationship between the delta (discrepancy between objective and self-reported PA) and age, V_ O2 max and body composition variables. Furthermore, Spearman’s correlations were used to determine how well the self-reported and objectively measured values for moderate intensity PA and vigorous intensity PA were correlated. Analyses of co-variances were used to measure delta between the HF and LF groups. T-tests for independent variables were performed to
Demographic characteristics The mean age of the participants was 38.8 ± 13.1 and 37.2 ± 12.3 years, for the men (n = 52) and women (n = 61), respectively. The age group of 20–30 years was the most represented age group. All of the participants were members of the health insurance programme, and 95.5% (n = 108) of the sample were employed at the time of testing. The mean BMI for all participants was 25.7 kg m−2. The men were significantly heavier (p = 0.000; t = 7.24), had lower BF percentage (p = 0.000; t = 7.55) and had greater waist circumferences than the women (p = 0.000; t = 6.65) (Table I). Subsequent General Linear Model analyses therefore co-varied for gender due to the significant differences between the men and women. Results of the Shapiro–Wilk tests suggested that the data were not normally distributed (p < 0.001). Fitness test results The predicted V_ O2 max scores for the total sample ranged between 20.9 ml/kg/min and 69.4 ml/kg/min. This score was significantly higher for the men than for the women, 43.9 ml/kg/min vs. 35.0 ml/kg/min, respectively, p < 0.00001; t = 4.73. Participants achieving a V_ O2 max below the 50th percentile (based
Table I. Participant characteristics: demographic and clinical results All (n = 113) Age (years) Waist (cm) BMI (kg m2) Body fat (%) SBP (mmHg) DBP (mmHg)
37.9 82.0 25.7 25.4 118.7 70.2
(12.7) (13.3)a*** (4.4)a*** (7.3)a*** (12.9)a* (10.1)**
Men (n = 52) 38.8 89.6 26.7 20.9 123.2 71.0
(13.1) (12.3) (4.2) (6.0) (12.9) (10.6)
Women (n = 61) 37.2 75.4 24.9 29.3 114.8 69.5
(12.3) (10.3) (4.4) (5.8) (11.7) (9.6)
Higher fitness (HF; n = 71) 36.9 78.6 24.1 22.6 116.6 67.9
(12.5) (10.9) (3.4) (6.7) (9.9) (8.9)
Lower fitness (LF; n = 42) 39.7 87.4 28.3 30.3 122.2 74.1
(12.9) (15.1) (4.7) (5.4) (16.4) (10.8)
Values are presented as mean ± SD. a For differences in men and women. *p < 0.05 for differences between HF and LF; **p < 0.005 for differences between HF and LF; ***p < 0.0005 for differences HF and LF.
Self-reported and objectively measured PA
5
Table II. HF vs. LF for GPAQ data
Domain Vigorous occupational activity Moderate occupational activity MVPA occupational Travelling activity Vigorous leisure time activity Moderate leisure time activity MVPA leisure time activity Total vigorous activity (all domains) Total moderate activity (all domains) Total MVPA (all domains)
All valid (n = 113) 6.90 11.68 18.58 11.72 42.61 50.71 93.32 184.25 241.33 425.58
(49.26) (48.68) (90.39) (23.79) (34.04) (45.35) (63.51) (237.35) (328.76) (500.77)
High fit (HF; n = 71) 10.99 16.06 27.04 15.35 52.11 57.11 109.22 247.75 304.86 552.61
(61.95) (57.33) (111.46) (27.27) (33.54) (46.79) (63.13) (271.38) (382.76) (578.69)
Low fit (LF; n = 42) 0.00 4.29 4.29 5.60 26.55 39.88 66.43 76.90 133.93 210.83
(0.00) (27.77) (27.77) (14.66) (28.70) (41.08) (55.08) (96.75) (162.22) (195.20)
Diff between HF and LF group (p) 0.254 0.216 0.197 0.035* 0.00007** 0.0504 0.0004** 0.0001** 0.007* 0.0003**
Downloaded by [University of Sussex Library] at 07:36 25 March 2015
HF, higher fitness; LF, lower fitness; MVPA, moderate-to-vigorous physical activity. Values are represented as a 7-day average in minutes; mean ± SD. *p < 0.05 for differences HF and LF groups; **p < 0.0005 for differences HF and LF groups.
on ACSM criteria; ACSM, 2010) were placed in the LF group (n = 42). Seventy-one participants (62.8%) were categorised as HF, with 69.2% (n = 36) of the men and 57.4% (n = 35) of the women falling in the HF category, respectively. Therefore, by design, mean V_ O2 max for the HF and LF group were 43.9 ml/kg/min and 30.8 ml/kg/min, respectively (p < 0.001).
Self-reported PA results: GPAQ Table II illustrates all differences between the HF and LF groups for self-reported PA. Seven participants referred to themselves as currently ‘inactive’, and recalled no moderate or vigorous activity in the leisure domain. Average time spent in self-reported vigorous activity in the leisure domain for the total group per day was 43 ± 34 minutes, 95% confidence interval (CI) = 36; 49 minutes, median = 45 minutes. The HF group reported significantly more minutes of leisure time vigorous PA (52 ± 34 minutes) per day than the LF group (27 ± 29 minutes; p < 0.001; t = −4.13). Similarly, total time spent in vigorous activity, which includes all domains for a 7-day week, was significantly different between groups, with the HF group reporting more minutes (248 ± 271) than the LF group (77 ± 97); p = 0.0001; t = −3.93). Average time spent in self-reported moderate intensity activity per day for the total sample was 51 ± 45 minutes, 95% CI = 42; 59 minutes, median = 45 minutes. The HF group reported significantly more total moderate intensity PA, measured for all domains for a 7-day week, than the LF group (p < 0.001). There were no significant differences between the men and women for any self-reported moderate or vigorous intensity activity.
Objective measures of PA Participants wore the accelerometers for an average of 6.1 ± 1.5 days. All participants (n = 113) had 4 or more valid days of wear. In cases where participants wore the accelerometer for more than 7 days, only the first 7 days of data were used in the statistical analyses. There were no significant differences between men and women for objectively measured vigorous intensity or moderate intensity PA. However, there was a significant difference in objectively measured moderate intensity PA between the HF and LF groups (p = 0.003; t = −2.96) (Figure 1). The difference between the HF and LF groups for objectively measured vigorous intensity activity approached significance (p = 0.08; t = −1.71). The
Figure 1. Differences in objectively measured PA between the HF and LF groups.
6
S. A. Tomaz et al.
HF group did significantly more objectively measured MVPA (measured as total PA) than the LF group (p = 0.001; t = −3.31; shown in Figure 1).
Downloaded by [University of Sussex Library] at 07:36 25 March 2015
Bland–Altman plots Figure 2A represents the difference between selfreported and objectively measured moderate intensity PA. Only two participants accurately reported moderate intensity activity (δ = 0). Just over a third of the total sample (34.5%, n = 39) over-reported moderate PA, with more over-reporters being in the HF group (n = 25). Figure 2B represents the difference in reported vigorous PA values from the GPAQ and objectively measured vigorous minutes. Few participants (14.2%, n = 16) reported vigorous intensity activity accurately. However, these participants did not do any vigorous intensity activity according to the accelerometer, and because they had reported no
vigorous activity in the GPAQ, their report was accurate. There were 82 participants (72.6% of the total sample) who over-reported vigorous PA. Figure 2B represents a homoscedastic distribution. As the duration of reported vigorous activity increased, so did the discrepancy between reported and measured vigorous activity increase. Objective vs. self-reported PA results Spearman’s correlations showed that the selfreported moderate PA correlated significantly with objectively measured moderate PA (p < 0.05, r = 0.20). Correlations for vigorous PA were significant (p < 0.05, r = 0.31). Delta minutes were not significantly different between the HF and LF groups for moderate intensity PA (37 minutes vs. −19 minutes; p = 0.304; t = 1.03). The LF group over-reported vigorous activity to a significantly lesser extent than the HF group (−63 minutes vs. −211 minutes; p = 0.001; t = 3.32). Similarly, delta for MVPA was significantly different between the HF and LF groups (−26 minutes vs. −230 minutes; p = 0.012; t = 2.54). The mean ICC value between 2 and 7 days for this study was calculated as r = 0.28. The Spearman– Brown Prophecy formula was used to determine the amount of days required to achieve an ICC of 0.8 (being a ‘desirable ICC’) (Baranowski & de Moor, 2000). The Spearman–Brown Prophecy formula suggests that we needed 10 days of monitoring of MVPA data. This is unusually high because 4 days of MVPA monitoring is described as sufficient for the analysis of MVPA data (Matthews et al., 2002). Furthermore, it suggests that the day-to-day variability is quite high; hence, the need for more days to get a better idea of activity. Discussion
Figure 2. Bland–Altman plot showing the difference the HF and LF groups for (A) bouted moderate intensity and (B) vigorous intensity PA.
The first important finding of this research study is that the level of agreement between self-reported activity and objectively measured vigorous activity was weaker in the HF group. This suggests that the HF group over-estimates their time spent in vigorous intensity PA, when compared with accelerometry. Alternatively, another explanation might be that the HF group misclassify moderate activity to be of a more vigorous nature. Our findings are different to previous research which has shown that self-reported PA measured using the IPAQ is well correlated with objectively measured vigorous intensity PA (Hagstromer et al., 2006). With respect to moderate PA, the same study by Hagströmer et al. supported our findings although they used the IPAQ, which was found to correlate poorly with moderate intensity PA
Downloaded by [University of Sussex Library] at 07:36 25 March 2015
Self-reported and objectively measured PA (Hagstromer et al., 2006). However, there is limited data describing the extent to which fitness may impact on difference in self-reported and objectively measured activity. Irwin et al. concluded that PA may influence estimation of energy expenditure. Their protocol used doubly labelled water to measure energy expenditure and a 7-day PA recall, and lacked an objective measure of cardiorespiratory fitness (V_ O2 max ) (Irwin, Ainsworth, & Conway, 2001b). These researchers’ findings also suggested that more active individuals were more likely to under-report activity, specifically vigorous PA. Thus, these findings are contrary to those observed in our study. These differences may be due to the fact that doubly labelled water is considered the criterion method for energy expenditure, and therefore will yield different results for activity than accelerometry. Furthermore, the Irwin study collected data from adult males (Irwin et al., 2001b), whereas our sample included women. Another reason for the HF groups’ over-reporting could relate to the manner in which the data were analysed. This is evident when we compare our findings to that of a recent study done in the USA NHANES cohort. Our sample self-reported ±60.79 min/day of MVPA, which is similar to that reported in the NHANES study (54.8 ± 1.9 min/day) (Atienza et al., 2011). However, the objectively measured MVPA in our study was far higher than what was observed in the NHANES sample (±38.74 vs. 6.7 min/day) (Atienza et al., 2011). This may be due to the fact that USA NHANES researchers used the Troiano’s cut points, whereas we used the Matthews cut points. The Matthews cut points for moderate PA is lower than that of Troiano, 760–5998 counts per minute vs. 2020 and 5998 counts per minute, respectively. Consequently, our data analysis would have resulted in more PA being classified as moderate intensity. There are many methods and equations used to analyse and making accelerometer data meaningful, as well as many different sets of cut points. Currently, there are no specific equations or data reduction methods used for different ActiGraph models. Therefore, it may be necessary to use validated cut points for the newer models. Equations may need to be more attentive to the new technology found in the newer generation ActiGraph models. Another possible explanation for the over-estimation of vigorous intensity PA may be that the HF group participate in more PA than LF individuals, creating more room for recall error. Field notes and general comments from participants suggested that attending gym, jogging and doing activities such as aerobics were classified as vigorous. However, it is possible that the accelerometer may not capture these activities as vigorous intensity. Many of our participants reported engaging in weight and/or
7
strength training, but because this mode of training does not commonly involve continuous ambulation, an accelerometer attached to one’s hip would not be sensitive to weight training activities such as a bench press or a bicep curl (Matthews, 2005). Another limitation of using accelerometers includes the inability to wear the accelerometer under water (Martinez-Gomez et al., 2010); however, this limitation may have impacted on only a small number of participants in this study. Limitations and strengths One of the limitations of this research study was that the participants comprised of individuals who were completing the health and fitness assessment for their medical insurance company. This could have resulted in some bias as people who are of a HF level or more physically active being more likely to volunteer compared with a LF individual. In an attempt to address the limitations of the small sample size as well as large difference in numbers in the HF and LF groups, we co-varied for confounding variables in the statistical analysis of the data. To the knowledge of the researchers, this is the first research study to look at the impact of cardiovascular fitness on the difference between selfreported and objective measures of PA. Summary Cardiovascular fitness appears to play a role in the accuracy of self-reported total, moderate and vigorous intensity PA levels. The findings of this research study did not support our original hypothesis that the HF group would have better agreement than the LF group. This may be due to HF individuals doing more PA and therefore having more room for reporting error. Additional research is warranted to validate and measure the extent to which fitness can influence self-reported activity in a larger group of people which includes a range of fitness levels. Acknowledgements The funding for this project was received from an internal fund of author T.-L. Kolbe-Alexander, from the University of Cape Town. We appreciate the assistance with recruitment from the Sports Science Institute of South Africa and Lander and Pursad Biokineticists. A special thank you to Mike Lambert for his assistance with the statistical methods. References ACSM. (2010). ACSM’s Guidelines for Exercise Testing and Prescription (8th ed.). Philadelphia: Lippincott Williams & Wilkins.
Downloaded by [University of Sussex Library] at 07:36 25 March 2015
8
S. A. Tomaz et al.
Armstrong, T., & Bull, F. C. (2006). Development of the World Health Organization global physical activity questionnaire. Journal of Public Health, 14(2), 66–70. doi:10.1007/s10389006-0024-x Atienza, A. A., Moser, R. P., Perna, F., Dodd, K., BallardBarbash, R., Troiano, R. P., & Berrigan, D. (2011). Self-reported and objectively measured activity related to biomarkers using NHANES. Medicine & Science in Sports & Exercise, 43, 815–821. doi:10.1249/MSS.0b013e3181fdfc32 Baranowski, T., & de Moor, C. (2000). How many days was that? Intra-individual variability and physical activity assessment. Research Quarterly for Exercise & Sport, 71(2 Suppl.), S74–S78. Bull, F. C., Maslin, T. S., & Armstrong, T. (2009). Global physical activity questionnaire (GPAQ): Nine country reliability and validity study. Journal of Physical Activity and Health, 6, 790–804. Craig, C. L., Marshall, A. L., Sjostrom, M., Bauman, A. E., Booth, M. L., Ainsworth, B. E., … Oja, P. (2003). International physical activity questionnaire: 12-country reliability and validity. Medicine & Science in Sports & Exercise, 35, 1381–1395. doi:10.1249/01.MSS.0000078924.61453.FB Crouter, S. E., Dellavalle, D. M., Haas, J. D., Frongillo, E. A., & Bassett, D. R. (2013). Validity of actiGraph 2-regression model, Matthews cut-points, and NHANES cut-points for assessing free-living physical activity. Journal of Physical Activity and Health, 10, 504–514. Downs, A., Van Hoomissen, H. J., Lafrenz, A., & Julka, D. L. (2013). Accelerometer-measured versus self-reported physical activity in college students: Implications for research and practice. Journal of American College Health. Dugas, L. R., Lambert, E. V., Plattner, T., Gay, J., Lambert, M. I., & Noakes, T. D. (2006). A new step test-derived fitness index may be used to predict maximal functional capacity in healthy persons. Medicine & Science in Sports & Exercise, 38 (5), S510. Dugas, L. R., van der Merwe, L., Odendaal, H., Noakes, T. D., & Lambert, E. V. (2005). A novel energy expenditure prediction equation for intermittent physical activity. Medicine & Science in Sports & Exercise, 37(12), 2154–2161. doi:10.1249/01.mss. 0000177742.12931.50 Durnin, J., & Womersley, J. (1974). Body fat assessed from total body density and its estimation from skinfold thickness: Measurements on 481 men and women aged from 16 to 72 Years. British Journal of Nutrition, 32(1), 77–97. Fillipas, S., Cicuttini, F., Holland, A. E., & Cherry, C. L. (2010). The international physical activity questionnaire overestimates moderate and vigorous physical activity in HIV-infected individuals compared with accelerometry. Journal of the Association of Nurses in AIDS Care, 21(2), 173–181. doi:10.1016/j. jana.2009.11.003 Fogelholm, M., Malmberg, J., Suni, J., Santtila, M., Kyrolainen, H., Mantysaari, M., & Oja, P. (2006). International physical activity questionnaire: Validity against fitness. Medicine & Science in Sports & Exercise, 38, 753–760. doi:10.1249/01. mss.0000194075.16960.20 Gerrard, P. (2012). Accuracy of self-reported physical activity as an indicator of cardiovascular fitness depends on education level. Archives of Physical Medicine and Rehabilitation, 93, 1872– 1874. doi:10.1016/j.apmr.2012.04.007 Guthold, R., Louazani, S. A., Riley, L. M., Cowan, M. J., Bovet, P., Damasceno, A., … Armstrong, T. P. (2011). Physical activity in 22 African countries: Results from the World Health Organization STEPwise approach to chronic disease risk factor surveillance. American Journal of Preventive Medicine, 41(1), 52–60. doi:10.1016/j.amepre.2011.03.008 Hagströmer, M., Oja, P., & Sjöström, M. (2006). The international physical activity questionnaire (IPAQ): A study of
concurrent and construct validity. Public Health Nutrition, 9, 755–762. doi:10.1079/PHN2005898 Irwin, M. L., Ainsworth, B. E., & Conway, J. M. (2001). Estimation of energy expenditure from physical activity measures: Determinants of accuracy. Obesity Research, 9, 517–525. doi:10.1038/oby.2001.68 Keytel, L., Goedecke, J., Noakes, T., Hiiloskorpi, H., Laukkanen, R., van der Merwe, L., & Lambert, E. V. (2005). Prediction of energy expenditure from heart rate monitoring during submaximal exercise. Journal of Sports Sciences, 23, 289–297. doi:10.1080/02640410470001730089 Le Masurier, G. C., & Tudor-Locke, C. (2003). Comparison of pedometer and accelerometer accuracy under controlled conditions. Medicine & Science in Sports & Exercise, 35, 867–871. doi:10.1249/01.MSS.0000064996.63632.10 Lyden, K., Kozey, S. L., Staudenmeyer, J. W., & Freedson, P. S. (2011). A comprehensive evaluation of commonly used accelerometer energy expenditure and MET prediction equations. European Journal of Applied Physiology, 111(2), 187–201. doi:10.1007/s00421-010-1639-8 Martinez-Gomez, D., Puertollano, A., Warnberg, J., Calabro. M. A., Welk, G. J., Sjostrom, M., … Marcos, A. (2010). Comparison of the ActiGraph accelerometer and Bouchard diary to estimate energy expenditure in Spanish adolescents. Nutricion Hospitalaria, 24, 701–710. Masse, L. C., Fuemmeler, B. F., Anderson, C. B., Matthews, C. E., Trost, S. G., Catellier, D. J., & Treuth, M. (2005). Accelerometer data reduction: A comparison of four reduction algorithms on select outcome variables. Medicine & Science in Sports & Exercise, 37(Suppl.), S544–S554. doi:10.1249/01.mss. 0000185674.09066.8a Matthews, C. E. (2005). Calibration of accelerometer output for adults. Medicine & Science in Sports & Exercise, 37(Suppl.), S512–S522. doi:10.1249/01.mss.0000185659.11982.3d Matthews, C. E., Ainsworth, B. E., Thompson, R. W., & Bassett, D. R., Jr. (2002). Sources of variance in daily physical activity levels as measured by an accelerometer. Medicine & Science in Sports & Exercise, 34, 1376–1381. doi:10.1097/00005768200208000-00021 McClain, J. J., Sisson, S. B., & Tudor-Locke, C. (2007). Actigraph accelerometer interinstrument reliability during free-living in adults. Medicine & Science in Sports & Exercise, 39, 1509–1514. doi:10.1249/mss.0b013e3180dc9954 Pillay, J. D., Kolbe-Alexander, T. L., van Mechelen, W., & Lambert, E. V. (2012). Steps that count – the association between the number and intensity of steps accumulated and fitness and health measures. Journal of Physical Activity and Health, 11, 10–17. Plasqui, G., & Westerterp, K. R. (2006). Accelerometry and heart rate as a measure of physical fitness: Cross-validation. Medicine & Science in Sports & Exercise, 38, 1510–1514. doi:10.1249/01. mss.0000228942.55152.84 Plasqui, G., & Westerterp, K. R. (2007). Physical activity assessment with accelerometers: An evaluation against doubly labeled water. Obesity, 15, 2371–2379. doi:10.1038/oby. 2007.281 Prince, S. A., Adamo, K. B., Hamel, M., Hardt, J., Gorber, S. C., & Tremblay, M. (2008). A comparison of direct versus self-report measures for assessing physical activity in adults: A systematic review. The International Journal of Behavioural Nutrition and Physical Activity, 5(1), 56. doi:10.1186/1479-5868-5-56 Santos-Lozano, A., Marín, P. J., Torres-Luque, G., Ruiz, J. R., Lucía, A., & Garatachea, N. (2012). Technical variability of the GT3X accelerometer. Medical Engineering & Physics, 34, 787– 790. doi:10.1016/j.medengphy.2012.02.005 Santos-Lozano, A., Torres-Luque, G., Marín, P., Ruiz, J., Lucia, A., & Garatachea, N. (2012). Intermonitor variability of GT3X
Downloaded by [University of Sussex Library] at 07:36 25 March 2015
Self-reported and objectively measured PA accelerometer. International Journal of Sports Medicine, 33, 994– 999. doi:10.1055/s-0032-1312580 Schuna, J. M., Jr., Johnson, W. D., & Tudor-Locke, C. (2013). Adult self-reported and objectively monitored physical activity and sedentary behavior: NHANES 2005–2006. International Journal of Behavior and Nutrition Physical Activity, 10, 126. Tehard, B., Saris, W. H. M., Astrup, A., Martinez, J. A., Taylor, M. A., Barbe, P., Richterova, B., Guy-Grand, B., Sørensen, T. I., & Oppert, J. M. (2005). Comparison of two physical activity questionnaires in obese subjects: The NUGENOB study. Medicine & Science in Sports & Exercise, 37, 1535–1541. doi:10.1249/01.mss.0000177464.68521.3b Timperio, A., Salmon, J., & Crawford, D. (2003). Validity and reliability of a physical activity recall instrument among overweight and non-overweight men and women. Journal of Science and Medicine in Sport, 6, 477–491. doi:10.1016/S1440-2440 (03)80273-6 Troiano, R. P., Berrigan, D., Dodd, K. W., Mâsse, L. C., Tilert, T., & McDowell, M. (2008). Physical activity in the United States measured by accelerometer. Medicine & Science in Sports & Exercise, 40(1), 181–188. doi:10.1249/mss.0b013e 31815a51b3 Trost, S. G., McIver, K. L., & Pate, R. R. (2005). Conducting accelerometer-based activity assessments in field-based research.
9
Medicine & Science in Sports & Exercise, 37(Suppl.), S531–S543. doi:10.1249/01.mss.0000185657.86065.98 Tucker, J. M., Welk, G. J., & Beyler, N. K. (2011). Physical activity in U.S.: Adults compliance with the physical activity guidelines for Americans. American Journal of Preventive Medicine, 40, 454–461. doi:10.1016/j.amepre.2010.12.016 Vanhelst, J., Mikulovic, J., Bui-Xuan, G., Dieu, O., Blondeau, T., Fardy, P., & Béghin, L. (2012). Comparison of two ActiGraph accelerometer generations in the assessment of physical activity in free living conditions. BMC Research Notes, 5(1), 187. doi:10.1249/MSS.0b013e3181c29e90 Ward, D. S., Evenson, K. R., Vaughn, A., Rodgers, A. B., & Troiano, R. P. (2005). Accelerometer use in physical activity: Best practices and research recommendations. Medicine & Science in Sports & Exercise, 37(Suppl.), S582–S588. doi:10.1249/01.mss.0000185292.71933.91 WHO. (2010a). Global recommendations on physical activity for health. Retrieved from http://whqlibdoc.who.int/publications/ 2010/9789241599979_eng.pdf WHO (2010b). Global status report on noncommunicable diseases. Retrieved from http://www.who.int/nmh/publications/ncd_report_full_en.pdf
European Journal of Sport Science Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/tejs20
Cardiovascular fitness is associated with bias between self-reported and objectively measured physical activity a
a
a
S. A. Tomaz , E. V. Lambert , D. Karpul & T. L. Kolbe-Alexander
ab
a
Department of Human Biology, Faculty of Health Sciences, MRC/UCT Research Unit for Exercise Science and Sports Medicine, University of Cape Town, Cape Town, South Africa b
Centre for Research on Exercise, Physical Activity and Health, School of Human Movement Studies, University of Queensland, Brisbane, QLD, Australia Published online: 24 Dec 2014.
Click for updates To cite this article: S. A. Tomaz, E. V. Lambert, D. Karpul & T. L. Kolbe-Alexander (2014): Cardiovascular fitness is associated with bias between self-reported and objectively measured physical activity, European Journal of Sport Science, DOI: 10.1080/17461391.2014.987323 To link to this article: http://dx.doi.org/10.1080/17461391.2014.987323
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
European Journal of Sport Science, 2014 http://dx.doi.org/10.1080/17461391.2014.987323
ORIGINAL ARTICLE
Cardiovascular fitness is associated with bias between self-reported and objectively measured physical activity
S. A. TOMAZ1, E. V. LAMBERT1, D. KARPUL1, & T. L. KOLBE-ALEXANDER1,2 1
Downloaded by [University of Sussex Library] at 07:36 25 March 2015
Department of Human Biology, Faculty of Health Sciences, MRC/UCT Research Unit for Exercise Science and Sports Medicine, University of Cape Town, Cape Town, South Africa, 2Centre for Research on Exercise, Physical Activity and Health, School of Human Movement Studies, University of Queensland, Brisbane, QLD, Australia
Abstract The aim of this research study was to determine whether the level of agreement between self-reported and objective measures of physical activity (PA) is influenced by cardiovascular fitness. Participants (n = 113) completed the Global Physical Activity Questionnaire (GPAQ), a health risk assessment and a sub-maximal 12-minute step test. Age-predicted V_ O2 max was used to classify participants as lower fit and higher fit (HF). ActiGraph (GT3X) accelerometers were worn for 7 consecutive days. Matthews cut points were used to calculate minutes of moderate and vigorous PA (MVPA) per week. Bland–Altman plots were used to measure limits of agreement between GPAQ and ActiGraph MVPA. The participants’ mean age was 37.9 ± 12.7 years and more than 60% were categorised as HF (n = 71). Moderate PA was over-reported in 39% of all participants. Most of the over-reporters for moderate PA were in the HF group (64.1%). Vigorous PA was overreported by 72.6% of all participants. The discrepancy between self-reported and objective measures of vigorous PA increased with increasing self-reported time spent in vigorous PA. Fitter individuals appear to over-report PA more than lesser fit participants, suggesting that fitness could influence the level of agreement between self-reported and objective measures of PA. Keywords: Accelerometer, ActiGraph GT3X, GPAQ, physical activity
Introduction Physical inactivity has been identified as the fourth leading cause of global mortality related to noncommunicable diseases (NCDs), resulting in an estimated 3.2 million deaths, and is therefore a global concern (WHO, 2010b). The current guideline for physical activity (PA) for adults is 150 minutes of moderate intensity activity, or 60 minutes of vigorous activity per week (American College of Sports Medicine [ACSM], 2010). Therefore, those individuals who are not meeting these guidelines are most often referred to as physically inactive in the literature (WHO, 2010a). Based on these guidelines, 46.4% of South African males and 55.7% of South African females are insufficiently active based on self-report data obtained from the Global Status Report (WHO, 2010b). Consequently, the prevalence of inactivity is higher among the South African
adult population than the global prevalence which is 28% in men and 34% in women (WHO, 2010b). PA can be quantified by both subjective and objective measures using self-report questionnaires and accelerometers, respectively. Accelerometers are small devices used to measure movement using counts from which moderate and vigorous intensity PA is estimated. A review by Prince et al. (2008) noted that accelerometry is the most common method used for measuring PA objectively (Prince et al., 2008). The ActiGraph brand of accelerometers has been used extensively in research and well validated (Le Masurier & Tudor-Locke, 2003; Lyden, Kozey, Staudenmeyer, & Freedson, 2011; McClain, Sisson, & Tudor-Locke, 2007; Plasqui & Westerterp, 2006, 2007).The GT3X has been found to have good inter-instrument reliability (SantosLozano, Torres-Luque, et al., 2012), as well as being
Correspondence: S. A. Tomaz, Department of Human Biology, Faculty of Health Sciences, MRC/UCT Research Unit for Exercise Science and Sports Medicine, University of Cape Town, Rondebosch, Cape Town 7700, South Africa. E-mail: [email protected] © 2014 European College of Sport Science
Downloaded by [University of Sussex Library] at 07:36 25 March 2015
2
S. A. Tomaz et al.
an accurate tool in measuring free-living PA and in clinical studies (Santos-Lozano, Marín, et al., 2012; Vanhelst et al., 2012). The Global Physical Activity Questionnaire (GPAQ) and International Physical Activity Questionnaire (IPAQ) are examples of questionnaires that have been used to measure self-reported PA. The GPAQ was developed as a global surveillance tool for PA and has been used successfully for population surveillance in African countries (Guthold et al., 2011), including South Africa (Armstrong & Bull, 2006). Both questionnaires require participants to estimate the amount of moderate and vigorous intensity PA performed, for at least 10 minutes at a time, in the various domains of PA. The domains include transport, occupation, leisure time and domestic. There have been several comparative studies evaluating the discrepancies between objective and self-reported measurement of PA (Downs, Van Hoomissen, Lafrenz & Julka, 2013; Tucker, Welk & Beyler, 2011). Tucker et al. (2011) compared selfreport and accelerometer PA outcomes and determined the relationship between these two measurement tools using the NHANES sample (n = 3082). Results of this study showed large discrepancies between the accelerometer (ActiGraph model 7164) and self-report, with 62% of the sample meeting guidelines according to self-report, compared with objective measurement showing only 9.6% of the sample meeting guidelines. A more recent study conducted in college students compared the ActiGraph GT3X with the short IPAQ (Downs et al., 2013), and showed that self-reported moderate and vigorous PA (MVPA) averaged 66.14 min/day, as opposed to objectively measured MVPA which averaged 19.90 min/day. Furthermore, these measures were not shown to be significantly correlated (r(77) = 0.21; p = 0.08). Characteristics of participants who over-report have been identified in studies comparing self-report and objectively measured activity. These participants tend to be older, less than well educated (the majority of the participants in this particular sample were said to have completed a minimum of 11 years of formal education), are more likely to be smokers, and have central obesity (Craig et al., 2003; Fillipas, Cicuttini, Holland, & Cherry, 2010; Fogelholm et al., 2006; Gerrard, 2012; Irwin, Ainsworth, & Conway, 2001a; Tehard et al., 2005; Timperio, Salmon, & Crawford, 2003;). Habitual levels of PA and cardiovascular fitness could also influence self-reported PA. A recent study by Schuna, Johnson, and TudorLocke (2013) investigated, among other factors, the difference in self-reported and objectively measured PA (Schuna et al., 2013). This study showed that individuals who self-report compliance with MVPA
guidelines do accumulate more objectively measured PA (for all PA indicators, including light, MVPA). This study was performed on the NHANES cohort and investigated the level of agreement based on levels of PA, and more specifically, meeting PA guidelines and how it may influence an individual’s ability and accuracy when reporting PA (Schuna et al., 2013). Thus, there has been limited research investigating the relationship between cardiovascular fitness levels and the accuracy of self-reported PA. Identifying factors, which may influence selfreported PA, in relation to objectively measured PA, may assist researchers with data interpretation and application of self-reported activity, as well as practitioners, for exercise prescription. Therefore, the purpose of this study was to determine the extent to which cardiovascular fitness influences self-reported moderate and vigorous intensity PA, in relation to objectively measured PA. We hypothesised that persons with higher fitness (HF) levels would report PA with greater accuracy than individuals who present with lower fitness (LF) levels. Methods Participants A convenience sample of 113 men (n = 52) and women (n = 61) were recruited from individuals who volunteered to participate in an annual health and risk assessment (HRA) and fitness test, which formed part of a health insurance wellness programme. The goal of the HRA with respect to the health wellness programme is to promote healthy lifestyle behaviour and thus play a role in reducing the risk of NCDs. All participants provided signed informed consent prior to participating in the study. The research study was approved by the Human Research Ethics committee of the University of Cape Town (Ethics reference number 348/2008). Measures Questionnaires Health risk assessment. The HRA was comprised of questions pertaining to the participant’s general demographic information, personal medical history, and family history of disease, medication, injuries sustained and exercise-induced symptoms. Clinical measures and anthropometry. Blood pressure was measured one time following 5 minutes of sitting using a manual blood pressure cuff (ACSM, 2010). Height was measured in meters using an upright Seca stadiometer. Participants stood barefoot with their heels and head in contact with the wall and arms at their side. Weight was measured in
Self-reported and objectively measured PA
Downloaded by [University of Sussex Library] at 07:36 25 March 2015
kilograms using a Clover Scale (Model TCS-A300), with the value rounded to the nearest 100 g. Body mass index (BMI) was calculated using the equation: BMI = weight (kg)/height (m)2. Waist circumference was measured narrowest point of the torso, between the xiphoid process and umbilicus, using a standard tape measure on the skin (ACSM, 2010). Body fat (BF) percentage was estimated, using measures from four different skinfolds sites, including the biceps, triceps, subscapular and suprailiac regions using Harpenden skinfold callipers, as described by Durnin and Womersly (1974). Fitness assessment. A 25 cm Reebok step was used to complete a 12-minute sub-maximal step test that has been validated for use in measuring exercise capacity for HRA purposes (Dugas et al., 2006). The step test consisted of four, 2-minute stepping stages, where each stage was followed by a 1-minute rest period prior to commencing the next stage (Pillay, KolbeAlexander, van, & Lambert, 2012). Participants were required to step up and down to the sound of a metronome, of which the cadence increased as the step test progressed from stage one to stage four (80, 96, 112 and 120 steps/min, respectively). A Suunto heart rate monitor was used to record the heart rate response during the step test. The heart rate file was log analysed using the Suunto Training Manager and Suunto Monitor software (Suunto Oy, Finland). The heart rate response to exercise was then regressed to predict peak metabolic equivalents (METS) at age-predicted maximum heart rate. An estimated maximal oxygen uptake (ml/kg/min) was calculated, based on heart rate recovery at the end of the last stage, using the following equation: 44.891 – (age × 0.262) – (gender × 0.855) + (peak METS × 0.994) + (maximum reported MET hrs/wk of activity × 0.163) (Dugas, van der Merwe, Odendaal, Noakes, & Lambert, 2005; Keytel et al., 2005). The normative values for maximal aerobic power (referred to as predicted V_ O2 max in this paper) according to the ACSM Guidelines for Exercise Testing and Prescription were used to categorise participants in two groups, namely ‘LF’ and ‘HF’ (ACSM, 2010). Participants who fell below the 50th percentile (according to the ACSM (2010) genderbased norms for age-predicted V_ O2 max ) were categorised as LF participants. Global Physical Activity Questionnaire. The GPAQ is based on self-reported recall of moderate and vigorous intensity activities. The GPAQ has been validated as a reliable and valid measure in population surveillance (Bull, Maslin, & Armstrong, 2009; Craig et al., 2003; Hagstromer, Oja, & Sjostrom, 2006). Activities included in the report should be performed for at least 10 consecutive
3
minutes, in a typical week, in each of the four domains, including transport, occupational, leisure and domestic activities (Armstrong & Bull, 2006). The GPAQ was administered to each of the participants individually. Accelerometry. The tri-axial ActiGraph GT3X accelerometer (ActiGraph, Shalimar, FL, USA) was used in this research study. The device was set to measure motion in all three axes and was initialised to record data in 60-second epochs (Trost, McIver, & Pate, 2005). Accelerometer data were captured and cleaned using the ActiLife Software Version 5 (ActiLife software; Pensacola, FL, USA). Data from all three axes were captured and used in the analysis. The accelerometer was given to the participants after the GPAQ was administered, and participants were requested to maintain their usual activity levels while wearing the device. Participants were requested to wear the accelerometer on their right hip for seven consecutive days during their waking hours, but not while bathing, showering or swimming. A minimum of 4 days of wear, with 600 minutes per day were required for data analysis. Non-wear time was defined as 60 continuous minutes of no counts (zeros). Four days of monitoring moderate and vigorous intensity activity has been found to provide 80% reliability (Matthews, Ainsworth, Thompson, & Bassett, 2002; Trost et al., 2005). Moderate and vigorous intensity PA levels were estimated based on accelerometer counts using the cut-off points according to Matthews (2005). The Matthews equation stipulates that counts between 760 and 5998 (inclusive) represents moderate intensity PA and counts above 5999 is indicative of vigorous activity (Matthews, 2005). The Matthews and Troiano et al. (2008) cut points were recently compared to indirect caloriometry and free-living conditions (Crouter, Dellavalle, Haas, Frongillo, & Bassett, 2013). The Matthews cut point over-estimated moderate PA by 33.4% in comparison to freeliving conditions, whereas the Troiano’s cut point under-estimated moderate PA by 50.4% (Crouter et al., 2013). Of the cut points evaluated in this study by Crouter et al., Matthews cut points had lowest mean bias for time spent in MVPA, overestimating measured time spent in MVPA in comparison to indirect calirometry by 13.6 minutes (26.3%) (Crouter et al., 2013). Objective PA from the GT3X was analysed for any moderate-to-vigorous activity which took place for 10-minute bouts/intervals. MVPA was calculated by summing moderate and vigorous activity (Downs et al., 2013). Within each 10-minute bout of moderate and/or vigorous PA, 1 or 2 minutes of ‘dropped’ counts were allowed (Masse et al., 2005;
4
S. A. Tomaz et al.
Ward, Evenson, Vaughn, Rodgers, & Troiano, 2005). An exercise bout was therefore excluded if the programme detected a drop in counts for greater than 2 minutes within the 10-minute period. The term ‘delta’ was used to describe the difference between the objective and self-report measurements of PA. Delta values were recorded as minutes. Minutes spent in PA reported in the GPAQ were subtracted from the minutes measured by the accelerometer (GT3X). Therefore, a positive value would mean that the participant under-reported and negative value would indicate over-reporting. A zerovalue represents accurate reporting.
compare self-reported moderate PA to objectively measured moderate activity, as well as for selfreported vigorous activity and objectively measured vigorous activity. Bland–Altman plots were used to calculate the limits of agreement between the accelerometer counts (objective) and the GPAQ (self-report) for both the HF and LF groups. Intra-class correlations (ICCs) were performed, in addition to the use of the Spearman–Brown Prophecy formula to determine the number of days of monitor wear required to achieve a desired ICC of r = 0.8 (Baranowski & de Moor, 2000).
Downloaded by [University of Sussex Library] at 07:36 25 March 2015
Statistical analysis
Results
STATISTICA 9 software (Stasoft Inc., Tulsa, OK, USA) was used for all statistical analyses. Descriptive statistics were applied to calculate the mean values and standard deviations of all demographic, physical fitness, GPAQ and ActiGraph GT3X accelerometer results. A custom Matlab (R2011b The Mathworks Inc.) program was used to calculate the total amount of time spent doing moderate and vigorous intensity PA using 10-minute bouts. Shapiro–Wilk tests for normality were performed to check whether the data were normally distributed. T-tests for independent groups were used to determine significant differences between age, weight, BMI, BF percentage, systolic and diastolic blood pressure (SBP and DBP, respectively) for the HF and LF groups, as well as between the male and female participants. Spearman’s (non-parametric) correlations were used to measure the relationship between the delta (discrepancy between objective and self-reported PA) and age, V_ O2 max and body composition variables. Furthermore, Spearman’s correlations were used to determine how well the self-reported and objectively measured values for moderate intensity PA and vigorous intensity PA were correlated. Analyses of co-variances were used to measure delta between the HF and LF groups. T-tests for independent variables were performed to
Demographic characteristics The mean age of the participants was 38.8 ± 13.1 and 37.2 ± 12.3 years, for the men (n = 52) and women (n = 61), respectively. The age group of 20–30 years was the most represented age group. All of the participants were members of the health insurance programme, and 95.5% (n = 108) of the sample were employed at the time of testing. The mean BMI for all participants was 25.7 kg m−2. The men were significantly heavier (p = 0.000; t = 7.24), had lower BF percentage (p = 0.000; t = 7.55) and had greater waist circumferences than the women (p = 0.000; t = 6.65) (Table I). Subsequent General Linear Model analyses therefore co-varied for gender due to the significant differences between the men and women. Results of the Shapiro–Wilk tests suggested that the data were not normally distributed (p < 0.001). Fitness test results The predicted V_ O2 max scores for the total sample ranged between 20.9 ml/kg/min and 69.4 ml/kg/min. This score was significantly higher for the men than for the women, 43.9 ml/kg/min vs. 35.0 ml/kg/min, respectively, p < 0.00001; t = 4.73. Participants achieving a V_ O2 max below the 50th percentile (based
Table I. Participant characteristics: demographic and clinical results All (n = 113) Age (years) Waist (cm) BMI (kg m2) Body fat (%) SBP (mmHg) DBP (mmHg)
37.9 82.0 25.7 25.4 118.7 70.2
(12.7) (13.3)a*** (4.4)a*** (7.3)a*** (12.9)a* (10.1)**
Men (n = 52) 38.8 89.6 26.7 20.9 123.2 71.0
(13.1) (12.3) (4.2) (6.0) (12.9) (10.6)
Women (n = 61) 37.2 75.4 24.9 29.3 114.8 69.5
(12.3) (10.3) (4.4) (5.8) (11.7) (9.6)
Higher fitness (HF; n = 71) 36.9 78.6 24.1 22.6 116.6 67.9
(12.5) (10.9) (3.4) (6.7) (9.9) (8.9)
Lower fitness (LF; n = 42) 39.7 87.4 28.3 30.3 122.2 74.1
(12.9) (15.1) (4.7) (5.4) (16.4) (10.8)
Values are presented as mean ± SD. a For differences in men and women. *p < 0.05 for differences between HF and LF; **p < 0.005 for differences between HF and LF; ***p < 0.0005 for differences HF and LF.
Self-reported and objectively measured PA
5
Table II. HF vs. LF for GPAQ data
Domain Vigorous occupational activity Moderate occupational activity MVPA occupational Travelling activity Vigorous leisure time activity Moderate leisure time activity MVPA leisure time activity Total vigorous activity (all domains) Total moderate activity (all domains) Total MVPA (all domains)
All valid (n = 113) 6.90 11.68 18.58 11.72 42.61 50.71 93.32 184.25 241.33 425.58
(49.26) (48.68) (90.39) (23.79) (34.04) (45.35) (63.51) (237.35) (328.76) (500.77)
High fit (HF; n = 71) 10.99 16.06 27.04 15.35 52.11 57.11 109.22 247.75 304.86 552.61
(61.95) (57.33) (111.46) (27.27) (33.54) (46.79) (63.13) (271.38) (382.76) (578.69)
Low fit (LF; n = 42) 0.00 4.29 4.29 5.60 26.55 39.88 66.43 76.90 133.93 210.83
(0.00) (27.77) (27.77) (14.66) (28.70) (41.08) (55.08) (96.75) (162.22) (195.20)
Diff between HF and LF group (p) 0.254 0.216 0.197 0.035* 0.00007** 0.0504 0.0004** 0.0001** 0.007* 0.0003**
Downloaded by [University of Sussex Library] at 07:36 25 March 2015
HF, higher fitness; LF, lower fitness; MVPA, moderate-to-vigorous physical activity. Values are represented as a 7-day average in minutes; mean ± SD. *p < 0.05 for differences HF and LF groups; **p < 0.0005 for differences HF and LF groups.
on ACSM criteria; ACSM, 2010) were placed in the LF group (n = 42). Seventy-one participants (62.8%) were categorised as HF, with 69.2% (n = 36) of the men and 57.4% (n = 35) of the women falling in the HF category, respectively. Therefore, by design, mean V_ O2 max for the HF and LF group were 43.9 ml/kg/min and 30.8 ml/kg/min, respectively (p < 0.001).
Self-reported PA results: GPAQ Table II illustrates all differences between the HF and LF groups for self-reported PA. Seven participants referred to themselves as currently ‘inactive’, and recalled no moderate or vigorous activity in the leisure domain. Average time spent in self-reported vigorous activity in the leisure domain for the total group per day was 43 ± 34 minutes, 95% confidence interval (CI) = 36; 49 minutes, median = 45 minutes. The HF group reported significantly more minutes of leisure time vigorous PA (52 ± 34 minutes) per day than the LF group (27 ± 29 minutes; p < 0.001; t = −4.13). Similarly, total time spent in vigorous activity, which includes all domains for a 7-day week, was significantly different between groups, with the HF group reporting more minutes (248 ± 271) than the LF group (77 ± 97); p = 0.0001; t = −3.93). Average time spent in self-reported moderate intensity activity per day for the total sample was 51 ± 45 minutes, 95% CI = 42; 59 minutes, median = 45 minutes. The HF group reported significantly more total moderate intensity PA, measured for all domains for a 7-day week, than the LF group (p < 0.001). There were no significant differences between the men and women for any self-reported moderate or vigorous intensity activity.
Objective measures of PA Participants wore the accelerometers for an average of 6.1 ± 1.5 days. All participants (n = 113) had 4 or more valid days of wear. In cases where participants wore the accelerometer for more than 7 days, only the first 7 days of data were used in the statistical analyses. There were no significant differences between men and women for objectively measured vigorous intensity or moderate intensity PA. However, there was a significant difference in objectively measured moderate intensity PA between the HF and LF groups (p = 0.003; t = −2.96) (Figure 1). The difference between the HF and LF groups for objectively measured vigorous intensity activity approached significance (p = 0.08; t = −1.71). The
Figure 1. Differences in objectively measured PA between the HF and LF groups.
6
S. A. Tomaz et al.
HF group did significantly more objectively measured MVPA (measured as total PA) than the LF group (p = 0.001; t = −3.31; shown in Figure 1).
Downloaded by [University of Sussex Library] at 07:36 25 March 2015
Bland–Altman plots Figure 2A represents the difference between selfreported and objectively measured moderate intensity PA. Only two participants accurately reported moderate intensity activity (δ = 0). Just over a third of the total sample (34.5%, n = 39) over-reported moderate PA, with more over-reporters being in the HF group (n = 25). Figure 2B represents the difference in reported vigorous PA values from the GPAQ and objectively measured vigorous minutes. Few participants (14.2%, n = 16) reported vigorous intensity activity accurately. However, these participants did not do any vigorous intensity activity according to the accelerometer, and because they had reported no
vigorous activity in the GPAQ, their report was accurate. There were 82 participants (72.6% of the total sample) who over-reported vigorous PA. Figure 2B represents a homoscedastic distribution. As the duration of reported vigorous activity increased, so did the discrepancy between reported and measured vigorous activity increase. Objective vs. self-reported PA results Spearman’s correlations showed that the selfreported moderate PA correlated significantly with objectively measured moderate PA (p < 0.05, r = 0.20). Correlations for vigorous PA were significant (p < 0.05, r = 0.31). Delta minutes were not significantly different between the HF and LF groups for moderate intensity PA (37 minutes vs. −19 minutes; p = 0.304; t = 1.03). The LF group over-reported vigorous activity to a significantly lesser extent than the HF group (−63 minutes vs. −211 minutes; p = 0.001; t = 3.32). Similarly, delta for MVPA was significantly different between the HF and LF groups (−26 minutes vs. −230 minutes; p = 0.012; t = 2.54). The mean ICC value between 2 and 7 days for this study was calculated as r = 0.28. The Spearman– Brown Prophecy formula was used to determine the amount of days required to achieve an ICC of 0.8 (being a ‘desirable ICC’) (Baranowski & de Moor, 2000). The Spearman–Brown Prophecy formula suggests that we needed 10 days of monitoring of MVPA data. This is unusually high because 4 days of MVPA monitoring is described as sufficient for the analysis of MVPA data (Matthews et al., 2002). Furthermore, it suggests that the day-to-day variability is quite high; hence, the need for more days to get a better idea of activity. Discussion
Figure 2. Bland–Altman plot showing the difference the HF and LF groups for (A) bouted moderate intensity and (B) vigorous intensity PA.
The first important finding of this research study is that the level of agreement between self-reported activity and objectively measured vigorous activity was weaker in the HF group. This suggests that the HF group over-estimates their time spent in vigorous intensity PA, when compared with accelerometry. Alternatively, another explanation might be that the HF group misclassify moderate activity to be of a more vigorous nature. Our findings are different to previous research which has shown that self-reported PA measured using the IPAQ is well correlated with objectively measured vigorous intensity PA (Hagstromer et al., 2006). With respect to moderate PA, the same study by Hagströmer et al. supported our findings although they used the IPAQ, which was found to correlate poorly with moderate intensity PA
Downloaded by [University of Sussex Library] at 07:36 25 March 2015
Self-reported and objectively measured PA (Hagstromer et al., 2006). However, there is limited data describing the extent to which fitness may impact on difference in self-reported and objectively measured activity. Irwin et al. concluded that PA may influence estimation of energy expenditure. Their protocol used doubly labelled water to measure energy expenditure and a 7-day PA recall, and lacked an objective measure of cardiorespiratory fitness (V_ O2 max ) (Irwin, Ainsworth, & Conway, 2001b). These researchers’ findings also suggested that more active individuals were more likely to under-report activity, specifically vigorous PA. Thus, these findings are contrary to those observed in our study. These differences may be due to the fact that doubly labelled water is considered the criterion method for energy expenditure, and therefore will yield different results for activity than accelerometry. Furthermore, the Irwin study collected data from adult males (Irwin et al., 2001b), whereas our sample included women. Another reason for the HF groups’ over-reporting could relate to the manner in which the data were analysed. This is evident when we compare our findings to that of a recent study done in the USA NHANES cohort. Our sample self-reported ±60.79 min/day of MVPA, which is similar to that reported in the NHANES study (54.8 ± 1.9 min/day) (Atienza et al., 2011). However, the objectively measured MVPA in our study was far higher than what was observed in the NHANES sample (±38.74 vs. 6.7 min/day) (Atienza et al., 2011). This may be due to the fact that USA NHANES researchers used the Troiano’s cut points, whereas we used the Matthews cut points. The Matthews cut points for moderate PA is lower than that of Troiano, 760–5998 counts per minute vs. 2020 and 5998 counts per minute, respectively. Consequently, our data analysis would have resulted in more PA being classified as moderate intensity. There are many methods and equations used to analyse and making accelerometer data meaningful, as well as many different sets of cut points. Currently, there are no specific equations or data reduction methods used for different ActiGraph models. Therefore, it may be necessary to use validated cut points for the newer models. Equations may need to be more attentive to the new technology found in the newer generation ActiGraph models. Another possible explanation for the over-estimation of vigorous intensity PA may be that the HF group participate in more PA than LF individuals, creating more room for recall error. Field notes and general comments from participants suggested that attending gym, jogging and doing activities such as aerobics were classified as vigorous. However, it is possible that the accelerometer may not capture these activities as vigorous intensity. Many of our participants reported engaging in weight and/or
7
strength training, but because this mode of training does not commonly involve continuous ambulation, an accelerometer attached to one’s hip would not be sensitive to weight training activities such as a bench press or a bicep curl (Matthews, 2005). Another limitation of using accelerometers includes the inability to wear the accelerometer under water (Martinez-Gomez et al., 2010); however, this limitation may have impacted on only a small number of participants in this study. Limitations and strengths One of the limitations of this research study was that the participants comprised of individuals who were completing the health and fitness assessment for their medical insurance company. This could have resulted in some bias as people who are of a HF level or more physically active being more likely to volunteer compared with a LF individual. In an attempt to address the limitations of the small sample size as well as large difference in numbers in the HF and LF groups, we co-varied for confounding variables in the statistical analysis of the data. To the knowledge of the researchers, this is the first research study to look at the impact of cardiovascular fitness on the difference between selfreported and objective measures of PA. Summary Cardiovascular fitness appears to play a role in the accuracy of self-reported total, moderate and vigorous intensity PA levels. The findings of this research study did not support our original hypothesis that the HF group would have better agreement than the LF group. This may be due to HF individuals doing more PA and therefore having more room for reporting error. Additional research is warranted to validate and measure the extent to which fitness can influence self-reported activity in a larger group of people which includes a range of fitness levels. Acknowledgements The funding for this project was received from an internal fund of author T.-L. Kolbe-Alexander, from the University of Cape Town. We appreciate the assistance with recruitment from the Sports Science Institute of South Africa and Lander and Pursad Biokineticists. A special thank you to Mike Lambert for his assistance with the statistical methods. References ACSM. (2010). ACSM’s Guidelines for Exercise Testing and Prescription (8th ed.). Philadelphia: Lippincott Williams & Wilkins.
Downloaded by [University of Sussex Library] at 07:36 25 March 2015
8
S. A. Tomaz et al.
Armstrong, T., & Bull, F. C. (2006). Development of the World Health Organization global physical activity questionnaire. Journal of Public Health, 14(2), 66–70. doi:10.1007/s10389006-0024-x Atienza, A. A., Moser, R. P., Perna, F., Dodd, K., BallardBarbash, R., Troiano, R. P., & Berrigan, D. (2011). Self-reported and objectively measured activity related to biomarkers using NHANES. Medicine & Science in Sports & Exercise, 43, 815–821. doi:10.1249/MSS.0b013e3181fdfc32 Baranowski, T., & de Moor, C. (2000). How many days was that? Intra-individual variability and physical activity assessment. Research Quarterly for Exercise & Sport, 71(2 Suppl.), S74–S78. Bull, F. C., Maslin, T. S., & Armstrong, T. (2009). Global physical activity questionnaire (GPAQ): Nine country reliability and validity study. Journal of Physical Activity and Health, 6, 790–804. Craig, C. L., Marshall, A. L., Sjostrom, M., Bauman, A. E., Booth, M. L., Ainsworth, B. E., … Oja, P. (2003). International physical activity questionnaire: 12-country reliability and validity. Medicine & Science in Sports & Exercise, 35, 1381–1395. doi:10.1249/01.MSS.0000078924.61453.FB Crouter, S. E., Dellavalle, D. M., Haas, J. D., Frongillo, E. A., & Bassett, D. R. (2013). Validity of actiGraph 2-regression model, Matthews cut-points, and NHANES cut-points for assessing free-living physical activity. Journal of Physical Activity and Health, 10, 504–514. Downs, A., Van Hoomissen, H. J., Lafrenz, A., & Julka, D. L. (2013). Accelerometer-measured versus self-reported physical activity in college students: Implications for research and practice. Journal of American College Health. Dugas, L. R., Lambert, E. V., Plattner, T., Gay, J., Lambert, M. I., & Noakes, T. D. (2006). A new step test-derived fitness index may be used to predict maximal functional capacity in healthy persons. Medicine & Science in Sports & Exercise, 38 (5), S510. Dugas, L. R., van der Merwe, L., Odendaal, H., Noakes, T. D., & Lambert, E. V. (2005). A novel energy expenditure prediction equation for intermittent physical activity. Medicine & Science in Sports & Exercise, 37(12), 2154–2161. doi:10.1249/01.mss. 0000177742.12931.50 Durnin, J., & Womersley, J. (1974). Body fat assessed from total body density and its estimation from skinfold thickness: Measurements on 481 men and women aged from 16 to 72 Years. British Journal of Nutrition, 32(1), 77–97. Fillipas, S., Cicuttini, F., Holland, A. E., & Cherry, C. L. (2010). The international physical activity questionnaire overestimates moderate and vigorous physical activity in HIV-infected individuals compared with accelerometry. Journal of the Association of Nurses in AIDS Care, 21(2), 173–181. doi:10.1016/j. jana.2009.11.003 Fogelholm, M., Malmberg, J., Suni, J., Santtila, M., Kyrolainen, H., Mantysaari, M., & Oja, P. (2006). International physical activity questionnaire: Validity against fitness. Medicine & Science in Sports & Exercise, 38, 753–760. doi:10.1249/01. mss.0000194075.16960.20 Gerrard, P. (2012). Accuracy of self-reported physical activity as an indicator of cardiovascular fitness depends on education level. Archives of Physical Medicine and Rehabilitation, 93, 1872– 1874. doi:10.1016/j.apmr.2012.04.007 Guthold, R., Louazani, S. A., Riley, L. M., Cowan, M. J., Bovet, P., Damasceno, A., … Armstrong, T. P. (2011). Physical activity in 22 African countries: Results from the World Health Organization STEPwise approach to chronic disease risk factor surveillance. American Journal of Preventive Medicine, 41(1), 52–60. doi:10.1016/j.amepre.2011.03.008 Hagströmer, M., Oja, P., & Sjöström, M. (2006). The international physical activity questionnaire (IPAQ): A study of
concurrent and construct validity. Public Health Nutrition, 9, 755–762. doi:10.1079/PHN2005898 Irwin, M. L., Ainsworth, B. E., & Conway, J. M. (2001). Estimation of energy expenditure from physical activity measures: Determinants of accuracy. Obesity Research, 9, 517–525. doi:10.1038/oby.2001.68 Keytel, L., Goedecke, J., Noakes, T., Hiiloskorpi, H., Laukkanen, R., van der Merwe, L., & Lambert, E. V. (2005). Prediction of energy expenditure from heart rate monitoring during submaximal exercise. Journal of Sports Sciences, 23, 289–297. doi:10.1080/02640410470001730089 Le Masurier, G. C., & Tudor-Locke, C. (2003). Comparison of pedometer and accelerometer accuracy under controlled conditions. Medicine & Science in Sports & Exercise, 35, 867–871. doi:10.1249/01.MSS.0000064996.63632.10 Lyden, K., Kozey, S. L., Staudenmeyer, J. W., & Freedson, P. S. (2011). A comprehensive evaluation of commonly used accelerometer energy expenditure and MET prediction equations. European Journal of Applied Physiology, 111(2), 187–201. doi:10.1007/s00421-010-1639-8 Martinez-Gomez, D., Puertollano, A., Warnberg, J., Calabro. M. A., Welk, G. J., Sjostrom, M., … Marcos, A. (2010). Comparison of the ActiGraph accelerometer and Bouchard diary to estimate energy expenditure in Spanish adolescents. Nutricion Hospitalaria, 24, 701–710. Masse, L. C., Fuemmeler, B. F., Anderson, C. B., Matthews, C. E., Trost, S. G., Catellier, D. J., & Treuth, M. (2005). Accelerometer data reduction: A comparison of four reduction algorithms on select outcome variables. Medicine & Science in Sports & Exercise, 37(Suppl.), S544–S554. doi:10.1249/01.mss. 0000185674.09066.8a Matthews, C. E. (2005). Calibration of accelerometer output for adults. Medicine & Science in Sports & Exercise, 37(Suppl.), S512–S522. doi:10.1249/01.mss.0000185659.11982.3d Matthews, C. E., Ainsworth, B. E., Thompson, R. W., & Bassett, D. R., Jr. (2002). Sources of variance in daily physical activity levels as measured by an accelerometer. Medicine & Science in Sports & Exercise, 34, 1376–1381. doi:10.1097/00005768200208000-00021 McClain, J. J., Sisson, S. B., & Tudor-Locke, C. (2007). Actigraph accelerometer interinstrument reliability during free-living in adults. Medicine & Science in Sports & Exercise, 39, 1509–1514. doi:10.1249/mss.0b013e3180dc9954 Pillay, J. D., Kolbe-Alexander, T. L., van Mechelen, W., & Lambert, E. V. (2012). Steps that count – the association between the number and intensity of steps accumulated and fitness and health measures. Journal of Physical Activity and Health, 11, 10–17. Plasqui, G., & Westerterp, K. R. (2006). Accelerometry and heart rate as a measure of physical fitness: Cross-validation. Medicine & Science in Sports & Exercise, 38, 1510–1514. doi:10.1249/01. mss.0000228942.55152.84 Plasqui, G., & Westerterp, K. R. (2007). Physical activity assessment with accelerometers: An evaluation against doubly labeled water. Obesity, 15, 2371–2379. doi:10.1038/oby. 2007.281 Prince, S. A., Adamo, K. B., Hamel, M., Hardt, J., Gorber, S. C., & Tremblay, M. (2008). A comparison of direct versus self-report measures for assessing physical activity in adults: A systematic review. The International Journal of Behavioural Nutrition and Physical Activity, 5(1), 56. doi:10.1186/1479-5868-5-56 Santos-Lozano, A., Marín, P. J., Torres-Luque, G., Ruiz, J. R., Lucía, A., & Garatachea, N. (2012). Technical variability of the GT3X accelerometer. Medical Engineering & Physics, 34, 787– 790. doi:10.1016/j.medengphy.2012.02.005 Santos-Lozano, A., Torres-Luque, G., Marín, P., Ruiz, J., Lucia, A., & Garatachea, N. (2012). Intermonitor variability of GT3X
Downloaded by [University of Sussex Library] at 07:36 25 March 2015
Self-reported and objectively measured PA accelerometer. International Journal of Sports Medicine, 33, 994– 999. doi:10.1055/s-0032-1312580 Schuna, J. M., Jr., Johnson, W. D., & Tudor-Locke, C. (2013). Adult self-reported and objectively monitored physical activity and sedentary behavior: NHANES 2005–2006. International Journal of Behavior and Nutrition Physical Activity, 10, 126. Tehard, B., Saris, W. H. M., Astrup, A., Martinez, J. A., Taylor, M. A., Barbe, P., Richterova, B., Guy-Grand, B., Sørensen, T. I., & Oppert, J. M. (2005). Comparison of two physical activity questionnaires in obese subjects: The NUGENOB study. Medicine & Science in Sports & Exercise, 37, 1535–1541. doi:10.1249/01.mss.0000177464.68521.3b Timperio, A., Salmon, J., & Crawford, D. (2003). Validity and reliability of a physical activity recall instrument among overweight and non-overweight men and women. Journal of Science and Medicine in Sport, 6, 477–491. doi:10.1016/S1440-2440 (03)80273-6 Troiano, R. P., Berrigan, D., Dodd, K. W., Mâsse, L. C., Tilert, T., & McDowell, M. (2008). Physical activity in the United States measured by accelerometer. Medicine & Science in Sports & Exercise, 40(1), 181–188. doi:10.1249/mss.0b013e 31815a51b3 Trost, S. G., McIver, K. L., & Pate, R. R. (2005). Conducting accelerometer-based activity assessments in field-based research.
9
Medicine & Science in Sports & Exercise, 37(Suppl.), S531–S543. doi:10.1249/01.mss.0000185657.86065.98 Tucker, J. M., Welk, G. J., & Beyler, N. K. (2011). Physical activity in U.S.: Adults compliance with the physical activity guidelines for Americans. American Journal of Preventive Medicine, 40, 454–461. doi:10.1016/j.amepre.2010.12.016 Vanhelst, J., Mikulovic, J., Bui-Xuan, G., Dieu, O., Blondeau, T., Fardy, P., & Béghin, L. (2012). Comparison of two ActiGraph accelerometer generations in the assessment of physical activity in free living conditions. BMC Research Notes, 5(1), 187. doi:10.1249/MSS.0b013e3181c29e90 Ward, D. S., Evenson, K. R., Vaughn, A., Rodgers, A. B., & Troiano, R. P. (2005). Accelerometer use in physical activity: Best practices and research recommendations. Medicine & Science in Sports & Exercise, 37(Suppl.), S582–S588. doi:10.1249/01.mss.0000185292.71933.91 WHO. (2010a). Global recommendations on physical activity for health. Retrieved from http://whqlibdoc.who.int/publications/ 2010/9789241599979_eng.pdf WHO (2010b). Global status report on noncommunicable diseases. Retrieved from http://www.who.int/nmh/publications/ncd_report_full_en.pdf
