Journal of Pediatric Rehabilitation Medicine: An Interdisciplinary Approach 7 (2014) 233–240 DOI 10.3233/PRM-140292 IOS Press
233
Measuring reliability and validity of the ActiGraph GT3X accelerometer for children with cerebral palsy: A feasibility study Margaret E. O’Neila,∗ , Maria A. Fragala-Pinkhamb, Jeffrey L. Formanc and Stewart G. Trostd a
Department of Physical Therapy and Rehabilitation Sciences, College of Nursing and Health Professions, Drexel University, Philadelphia, PA, USA b Franciscan Hospital for Children, Research Center for Children with Special Health Care Needs, Brighton, MA, USA c Department of Pediatric Rehabilitation Medicine, Franciscan Hospital for Children, Brighton, MA, USA d Queensland Institute of Technology, Institute of Health and Biomedical Innovation, St Lucia, QLD, Australia
Accepted 3 March 2014
Abstract. PURPOSE: The purposes of this study were to: 1) establish inter-instrument reliability between left and right hip accelerometer placement; 2) examine procedural reliability of a walking protocol used to measure physical activity (PA); and 3) confirm concurrent validity of accelerometers in measuring PA intensity as compared to the gold standard of oxygen consumption measured by indirect calorimetry. METHODS: Eight children (mean age: 11.9; SD: 3.2, 75% male) with CP (GMFCS levels I – III) wore ActiGraph GT3X accelerometers on each hip and the Cosmed K4b2 portable indirect calorimeter during two measurement sessions in which they performed the six minute walk test (6MWT) at three self-selected speeds (comfortable/slow, brisk, fast). Oxygen consumption (VO2) and accelerometer step and activity count data were recorded. RESULTS: Inter-instrument reliability of ActiGraph GT3X accelerometers placed on left and right hips was excellent (ICC = 0.96–0.99, CI95 : 0.81–0.99). Reproducibility of the protocol was good/excellent (ICC = 0.75–0.95, CI95 : 0.75–0.98). Concurrent validity of accelerometer count data and VO2 was fair/good (rho = 0.67, p < 0.001). The correlation between step count and VO2 was not significant (rho = 0.29, p = 0.2). CONCLUSION: This preliminary research suggests that ActiGraph GT3X accelerometers are reliable and valid devices to monitor PA during walking in children with CP and may be appropriate in rehabilitation research and clinical practice. ActiGraph GTX3 step counts were not valid for this sample and further research is warranted. Keywords: Physical activity measurement, accelerometry, cerebral palsy
1. Introduction and purpose Physical activity (PA) is identified as a leading health indicator in Healthy People 2020 and is important for children’s growth and development [1,2]. PA is ∗ Corresponding author: Margaret E. O’Neil, Drexel University, Mail Stop 7502, 1601 Cherry Street, Phila, PA 19102, USA. Tel.: +1 267 359 5546; Fax: +1 267 359 5576; E-mail: [email protected].
defined as “any bodily movement produced by skeletal muscles that results in energy expenditure (above resting)” [3,4]. It is a complex behavior with several dimensions (frequency, duration, intensity and type) and domains (leisure, occupational, chores and activities of daily living) [2]. The recommendation for school-aged children and youth participation in daily PA is defined as a minimum of 60 minutes of moderate to vigorous PA in-
c 2014 – IOS Press and the authors. All rights reserved 1874-5393/14/$27.50
234
M.E. O’Neil et al. / Measuring reliability and validity of the ActiGraph GT3X accelerometer for children with cerebral palsy
cluding muscle and bone strengthening activities [5]. Among children with typical development moderate to vigorous PA is associated with healthy weight, cardiorespiratory endurance, bone health and psychological health [6]. Despite known health benefits, the majority of children in the United States (58%) do not achieve 60 minutes of daily moderate to vigorous PA [7]. Evidence from cross-sectional studies suggests that children with cerebral palsy (CP) participate in significantly less PA compared to their peers with typical development [8,9]. Factors that contribute to decreased PA in children with CP span the International Classification of Function Model (ICF) dimensions of body structure and function, activity and participation associated with this diagnosis [9–11]. Specifically, children with CP often have decreased strength, motor coordination, balance, muscular endurance, and limitations in functional mobility and motor skills [9]. Children with CP also have decreased aerobic fitness (VO2 peak) which may contribute to their limited daily PA [11]. Children with CP are at high risk for secondary health conditions such as osteoporosis, chronic pain, fatigue, and obesity, which can influence participation in PA [9, 12]. Limited access to active recreation opportunities and decreased motivation and interest also influence level of PA participation in children with CP [13–15]. Pediatric rehabilitation services include health promotion, PA and exercise interventions for children with CP to promote active, healthy lifestyles. Providers involved in delivery of these services may include physical therapists, occupational therapists, exercise scientists and physicians. Although evidence supports the effectiveness of short-term aerobic exercise interventions lasting 2–6 months [11,16], no studies have been published on long-term PA interventions (> 6 months) aimed at improving health and physical function in children with CP. Researchers and clinicians acknowledge the importance of PA intensity as a critical component of interventions for children with CP. Furthermore additional research on PA intensity dosing is needed to guide clinical practice [9]. Therefore reliable and valid measures of PA frequency, intensity, and duration are important to examine PA participation in activity-based rehabilitation programs. PA measures include self-report questionnaires, pedometers, and accelerometers [2]. Quantitative fieldbased measures of “real world” PA are especially important when measuring habitual PA [8,17–22]. Selfreport measures have been validated for youth with typical development, however they often have inher-
ent response bias and may be difficult for children to complete, especially those younger than 10 years of age [2]. These measures have not been validated for children with CP. Child self-report and parent proxy measures of PA and participation specifically for children with disabilities describe PA patterns and behaviors but do not provide accurate or precise measures of PA intensity or frequency [19,22]. Pedometers and step-counters are useful and reliable measures of the frequency and duration of daily walking [2,4,8], but do not provide information on PA intensity [8,22]. Further, some pedometers may be susceptible to missing data due to design limitations [4]. Accelerometers have been used to measure PA frequency, duration and intensity in adults and children with typical development for several decades [2,4] and recently have been used among children with CP [8, 17,21,23]. Clanchy et al. [21] established the validity of ActiGraph uniaxial accelerometers in their ability to discriminate PA intensity in ambulatory children with CP during a walking protocol in which children walked at different self-selected speeds. However, the study did not examine inter-instrument reliability (left versus right hip placement). Inter-instrument reliability has been established for uniaxial and triaxial ActiGraph accelerometers among children with typical development [6], but has not been established among children with CP. It is important to examine monitor placement and establish interinstrument reliability to determine best placement of ActiGraph accelerometers for accurate PA data collection. Monitor placement may be critical in children with movement asymmetry and different distributions of CP (i.e., hemiplegia, diplegia, quadriplegia). Further, it is important to determine if a walking protocol for children with CP consisting of three different walking speeds is reproducible across measurement sessions to ensure protocol stability and reduce potential measurement error. The 6-minute walk test (6MWT) is reproducible in children with CP when walking at a fast pace [24,25], however, reproducibility of the 6MWT protocol with three self-selected walking speeds used in the current study has not been established. In real world settings, children with CP use a variety of walking speeds; therefore, it is important to determine if ActiGraph GT3X accelerometers can accurately measure PA intensity for different walking speeds. The purposes of this study were to: 1) examine interinstrument reliability between left and right hip ActiGraph GT3X accelerometer placement; 2) establish re-
M.E. O’Neil et al. / Measuring reliability and validity of the ActiGraph GT3X accelerometer for children with cerebral palsy
producibility of a walking protocol to measure PA; and 3) confirm concurrent validity of ActiGraph GT3X accelerometers in measuring PA intensity when compared to oxygen consumption (VO2) measured by indirect calorimetry as established in a previous study [23].
2. Methods 2.1. Participants A sample of eight children with CP participated in this study. Participants were 6–14 years old (mean = 11.9, SD = 3.2) and most (n = 6, 75%) were boys. Four children were classified at Gross Motor Function Classification System [26] (GMFCS) level I, one at GMFCS level II and three at GMFCS level III. Four children presented with a motor distribution of diplegia, three with hemiplegia and one with quadriplegia. Six children had healthy weight (BMI at 5th – 85th percentile) and two were obese (BMI 95th percentile) [27] (Table 1). All participants met the study’s inclusion criteria which were: 1) diagnosis of CP, 2) able to walk independently with or without an assistive device, 3) between the ages of 6–18 years, 4) no exercise restrictions or uncontrolled seizures, and 5) no lower extremity Botox injections less than three months prior to the study or orthopedic interventions less than six months prior to the study since these procedures initially may alter ambulation skills. Three of the eight participants used assistive devices (crutches or walker) and five used lower extremity braces. Three children were on seizure medications; two children had lower extremity Botox injections six or more months prior to testing and orthopedic surgery a year or more prior to testing. At the time of the study five children received physical and/or occupational therapy services in school and/or outpatient settings; three children received speech and language pathology services and special education services. 2.2. Procedure Institutional Review Board approval was obtained from Franciscan Hospital for Children (FHC) and Drexel University. Children were recruited from the outpatient clinics at FHC and from professional colleagues and parent groups in the area. Flyers were posted in clinic areas and were disseminated via email to colleagues and parents.
235
Parents and children provided informed consent or assent. Children participated in two measurement sessions at least one week apart. Each data collection session was approximately 1.5–2 hours. In the first data collection session parents completed a demographic questionnaire and medical information questionnaire about their children. One of the researchers (MFP) classified children on the GMFCS levels. The clinical and movement diagnoses were obtained from the medical record or by parent report. Height and weight were recorded for each child using a stadiometer (Shorrboard 420, QuickMedical, Issaquah, WA) and a digital scale (Tanita UM26, Arlington Heights, IL). Tricep and calf skin folds were measured using Lange calipers (Beta Technology, Santa Cruz, CA) and a standard protocol [28]. The Cosmed K4b2 (Rome, Italy) portable metabolic unit was used to measure VO2 during a 5 minute sitting rest and during three 6-minute walk tests. Children also wore a Polar heart rate (HR) monitor (Polar T-34, Lake Success, NY) which recorded HR during the activities. Children were fitted with a face mask which was secured to the head using a nylon head harness. The face mask was connected to the Cosmed K4b2 portable metabolic unit and the unit and battery were placed into a chest harness worn by the child. The children also wore a pair of ActiGraph GT3X accelerometers (ActiGraph Corporation, Pensacola, FL) which were attached at each side of the child’s waist using an elastic belt. Placement for each accelerometer was at the midaxilla line at the superior edge of the iliac crest. At the beginning of each data collection session, all equipment was calibrated or initialized according to manufacturer guidelines. Children sat quietly in a chair with back support for 5 minutes. Then each child participated in a walking protocol which consisted of three 6WMT trials each at a different self-selected speed (comfortable/slow; brisk; and fast) [21]. For the 6MWT trials, children were given the same verbal instruction to help pace their walking. Instructions for the three trials were: 1) Comfortable /Slow: “Walk at a comfortable slow pace like when you are at the mall or walking in your neighborhood or at school but you are not in a hurry”; 2) Brisk: “This time we want you to walk a little faster so please walk at a faster pace like when you are hurrying to get to class after the bell has rung” and 3) Fast: “This time we want you to walk as fast as you possibly can without falling or running.” [21] The trials were conducted in a hospital outpatient setting. A physical therapist (MFP) walked with each child to
236
M.E. O’Neil et al. / Measuring reliability and validity of the ActiGraph GT3X accelerometer for children with cerebral palsy Table 1 Child demographics, anthropometric & health data
ID
Gender
Age (yrs)
1 2 3 4 5 6 7 8
M M M M F M F M
12.2 13.6 14.00 15.92 6.42 8.75 10.17 14.25
∗ Body
Clinical diagnosis Diplegia Hemiplegia Hemiplegia Hemiplegia Diplegia Quadriplegia Diplegia Diplegia
Movement diagnosis Spasticity Spasticity Spasticity Spasticity Dystonia Dystonia Spasticity Spasticity
GMFCS level I I I I II III III III
Height (cm) 154.7 165.5 156 170.5 109.4 116.7 126.4 164.5
Weight (kg) 59.1 45.1 44.1 66.1 18.2 21.6 30.8 72.6
BMI % ile 95.3 11.9 33.4 75.6 47.6 45.1 79.1 95.7
% Body Fat∗ 65.3 15.9 10.9 23.6 17.3 7.62 34.38 51.7
RMR∗∗ kcal/24 hrs 1688.1 1475.4 1446.2 1823.5 857.0 990.1 1098.4 1920.9
fat estimate (Slaughter equation for tricep and calf skinfold measures); ∗∗ RMR = resting metabolic rate estimates (Schofield Equations).
encourage a consistent pace throughout the entire trial and for safety. Distance walked was recorded for each trial and walking speed was calculated. Children rested between each 6MWT trial until HR returned to resting level and when they reported having had sufficient rest. During the second data collection session, children repeated the 5 minute sitting rest and the walking protocol described above. 2.3. Measures The Cosmed K4b2 is a portable indirect calorimeter that is a reliable and valid measure of oxygen consumption (VO2) and energy expenditure (CO2 production). Indirect calorimetry is considered a criterion measure of PA intensity for children with CP [19]. Data collected by the Cosmed K4b2 were breath-bybreath relative VO2 (ml/kg/min) and CO2 production. Metabolic equivalents and resting metabolic rates were estimated from these data [29,30]. The ActiGraph GT3X accelerometer was used to measure activity and step counts. ActiGraph GT3X is a triaxial accelerometer designed to detect three dimensional accelerations from 0.05–2.00 g magnitude with a frequency response of 0.25–2.50 Hz. The ActiGraph accelerometer filters and processes the acceleration data over a specific time interval (epoch). For this study, the epoch interval was set at 1 second to synchronize count data to the VO2 and HR data measured by the Cosmed K4b2 unit and the Polar HR monitor [2, 4]. Activity and step counts were “summed” during each epoch and stored by the monitor with high counts reflecting high PA intensity and frequency. 2.4. Data reduction and analysis Body mass index (BMI) was calculated from height and weight data and weight status was determined from Center for Disease Control and Prevention (CDC)
growth charts [27]. Body fat estimates were calculated using skinfold data and the appropriate age and gender specific Slaughter equations [29]. Estimates of resting metabolic rate were calculated using the Schofield equations [30]. Descriptive statistics (means and standard deviations) were calculated for walking distance, walking speed, HR, VO2, activity and step counts and metabolic equivalents (METs) generated during the 5 minute sitting rest and the three 6MWTs. Descriptive data are reported from session 2 since both sessions had similar outcomes. Mean values for the VO2 and activity and step count data were calculated for the time interval of 3.0– 4.5 minutes for the sitting rest period and 3.5–5.5 minutes for each walking trial. Data from minutes 3.5–5.5 were used in the analysis to ensure that “steady state” had been achieved and to avoid using data from the final 30 seconds in which children may have slowed their pace in anticipation of the end of the trial. For each participant, steady state was confirmed by inspection of HR and VO2 values [21,31]. METs (estimates of PA intensity) were calculated (mean VO2/resting metabolic rate) [29]. Activity counts were classified into an intensity category based on the cut-points using the Evenson youth prediction equation. Although there are no youth prediction equations for children and youth with CP, Clanchy et al. determined that the Evenson equation was sufficient for interpreting activity counts for children with CP who were ambulatory [21,32]. Intra-class correlation coefficients (ICCs) were calculated for activity and step count data for each test session to examine inter-instrument reliability (left to right hip monitor placement). ICCs also were used to examine reproducibility of the 6MWT walking protocol between test sessions. Spearman correlations were calculated to determine if ActiGraph GT3X step and activity counts are valid measures of PA intensity as compared to VO2.
M.E. O’Neil et al. / Measuring reliability and validity of the ActiGraph GT3X accelerometer for children with cerebral palsy
237
Table 2 Descriptive data for the three 6MWT trials (from Session 2) 6MWT trials Distance (meters) Speed (m/min) Heart rate (bpm) VO2 (ml/kg/min) Activity Counts Mean (SD) Mean (SD) Mean (SD) Mean (SD) (Counts/min) (Left and Right) Mean (SD) Rest n/a n/a 78.67(8.82) 4.77 (1.10) n/a
Step Counts METs (Counts/min) Mean (SD) (Left and Right) Mean (SD) n/a n/a
Comfortable/ 339.67 (70.39) Slow
56.61(11.73)
122.11(28.28)
18.05 (8.36)
L: 2521.5 (953.6) R: 2520.8 (900.9)
L: 93.2 (31.2) R: 92.2 (31.4)
Brisk
400.38(113.80)
66.18(19.18)
144.72 (27.27)
21.08 (6.53)
L: 3753.2 (1604.8) L: 102.9 (33.7) R: 3761.9 (1633.1) R: 102.8 (33.1)
4.56 (1.79)
Fast
384.38(217.03)
64.07(36.18)
164.02(20.76)
25.53 (8.31)
L: 4542.8 (2016.1) L: 108 (37.1) R: 4439.8 (1907.1) R: 106.8 (38.6)
5.61 (2.81)
3.88 (1.92)
3. Results
3.4. Validity
3.1. Descriptive data
ActiGraph GT3X right hip count data were used to evaluate concurrent validity. Fair to good validity (rho = 0.67, p < 0.001) was established between ActiGraph activity count data and VO2. The correlation between ActiGraph GT3X step count and VO2 was not significant (rho = 0.29, p = 0.2).
ActiGraph GT3X activity and step counts (counts/ min) increased on both sides (left and right hips) when participants increased self-selected speed in the three 6MWT trials. Mean distance walked and speed increased during the 6MWT trials between comfortable/slow to brisk trials but not between brisk and fast trials. Also, mean distance and speed increased between comfortable/slow and fast trials. HR (bpm) was, on average, 78.67 bpm (SD = 8.82) during sitting rest and increased to a mean of 164.02 (SD = 20.76) for the fast speed 6MWT trial. Relative oxygen consumption (VO2 ml/kg/min) ranged from a mean of 4.77 (SD = 1.10) at rest to a mean of 25.53 (SD = 8.31) during the fast speed 6MWT trial. METS increased from a mean of 3.88 (SD = 1.92) during the comfortable/slow speed 6MWT trial to a mean of 5.61 (SD = 2.81) during the fast speed 6MWT trial (Table 2). 3.2. Inter-instrument reliability Results indicate that ActiGraph GT3X accelerometers had excellent inter-instrument reliability for both activity and step count data across the 6MWT trials at the three self-selected speeds. The range for the ICCs for all three 6MWT trials was 0.98 (CI95 : 0.92–0.99) to 0.98 (CI95 : 0.87–0.99) for activity counts and 0.99 (CI95 : 0.98–1.00) to 0.99 for step counts (CI95 : 0.99– 1.00) (Table 3). 3.3. Walking protocol reproducibility Good test-retest reproducibility was demonstrated between the two testing sessions. The ICC range for comfortable/slow to fast walking trials activity counts was 0.75 (CI95 : 0.02–0.95) to 0.95 (CI95 : 0.81–0.99) and for step counts was 0.88 (CI95 : 0.49–0.98) to 0.97 (CI95 : 0.88–0.99) (Table 4).
4. Discussion Findings from this study suggest that the ActiGraph GT3X accelerometer provides a reliable and valid measure of PA intensity during walking for children with CP. The ActiGraph data were reliable for bilateral hip placement suggesting that children with CP can follow the same protocol as children with typical development and wear ActiGraph GT3X accelerometers on the right side regardless of GMFCS level and motor distribution (diplegia, hemiplegia, quadriplegia). However since our sample size was small, replication of this study using a larger sample size of children with unilateral and bilateral motor involvement is recommended. ActiGraph GT3X activity counts were significantly and positively correlated with VO2 for the walking trials; however step counts were not. These findings indicate that for this device activity counts may be a better outcome measure than step counts when the goal is to increase PA intensity during walking for children with CP. When comparing the brisk and fast walking trials, participants exercised at a higher intensity during the fast walk as indicated by increased HR and VO2 but only took a few more steps and had a slight decrease in distance walked in 6 minutes. Activity counts; however, did increase across all three walking trials. Increased accessory trunk and upper extremity motions were observed during brisk and fast walking trials especially for participants classified at GMFCS levels II–
238
M.E. O’Neil et al. / Measuring reliability and validity of the ActiGraph GT3X accelerometer for children with cerebral palsy Table 3 Test Session: Inter-instrument Reliability (left to right monitor placement) 6MWT trials Comfortable/Slow Walk : L vs. R Brisk Walk: L vs. R Fast Walk: L vs. R
Activity counts ICC 95% CI 0.98 0.92–0.99 0.99 0.98–1.00 0.98 0.87–0.99
ICC 0.99 0.96 0.99
Step counts 95% CI 0.98–1.00 0.81–0.99 0.99–1.00
L = Left hip; R = Right hip. Table 4 Walking protocol reproducibility 6MWT trials Comfortable/Slow Walk: L1 vs. L2 Comfortable/Slow Walk: R1 vs. R2 Brisk Walk: L1 vs. L2 Brisk Walk: R1 vs. R2 Fast Walk: L1 vs. L2 Fast Walk: R1 vs. R2
Activity counts ICC 95% CI 0.75 0.02–0.95 0.86 0.48–0.97 0.88 0.52–0.98 0.95 0.59–0.99 0.89 0.47–0.98 0.95 0.81–0.99
ICC 0.88 0.96 0.88 0.98 0.88 0.97
Step counts 95% CI 0.49–0.98 0.82–0.99 0.47–0.98 0.92–0.99 0.48–0.98 0.88–0.99
L1 = Left hip Session 1; L2 = Left hip Session 2; R1 = Right hip Session 1; R2 = Right hip Session 2.
III which may be recorded as increased activity counts. These accessary motions seemed to increase especially at the end of the fast walking trial for children who were fatigued. Child performance on the 6MWT trials was reproducible between the two testing sessions, suggesting that this is a feasible and stable protocol for future PA measurement studies. It may also be a good test protocol to examine the effectiveness of activity-based interventions that are designed to increase PA walking intensity. In this study, children increased distance and speed between the comfortable/slow and brisk 6MWT trials. The distance covered in the brisk 6MWT trial was similar to the mean reported in one study (340.8 meters) [24] and less than the mean of 449.1 meters reported in another study of youth with CP [25]. The difference in the second study [25] could be that children were older (mean age = 13.6; SD = 1.6 years) and fewer were classified at GMFCS III. In the current study, on average children walked a shorter distance and at a slower speed during the fast 6MWT trial compared to the brisk trial. On inspection of the data, it was observed that 6 of the 8 children increased speed from their brisk to fast walk. One child had the same speed and two children with more motor involvement (GMFCS Level II and III) decreased in their speed from brisk to fast. This decreased distance and speed may be due to fatigue and de-conditioning because the fast speed 6MWT trial was the final trial in the protocol. The two children who demonstrated a decrease in speed were inactive and deconditioned according to parent report. Children were given a rest period to
allow them to re-establish baseline HR between trials. However, in future protocols additional rest time may be required for recovery of muscular endurance for some children in order to maintain a consistent fast pace for 6 minutes. To further illustrate this point, differences in speeds within a walking trial were evaluated. On average during the fast 6MWT trial, children walked faster at the one minute mark (mean = 83.22 meters/minute, SD = 24.79) compared to the three (65.87 meters/minute, SD = 38.02) and six minute marks (64.07 meters/minute, SD = 36.18). The mean speed at the one minute mark for the fast walk trial was faster than the speed during the brisk 6MWT trial (mean = 66.18 meters/min, SD = 19.18). These findings reflect the difficulty children had with maintaining consistent intensity during the fast 6MWT. Findings on the fast 6MWT trial suggest that further study is needed to determine if a longer rest period in between tests or a shorter walk test would be more appropriate for testing PA intensity in children with CP. The full 6MWT protocol used in the current study (three trials with self-selected speeds) might be a good outcome when the goal is to measure an increase in walking endurance; however further research is needed to confirm this recommendation. During the three 6MWT trials, the participants exhibited a steady increase in HR, VO2 and METs. Using a maximal heart rate (MHR) of 194 bpm [33], participants were, on average, at 63% MHR during the comfortable/slow 6MWT trial; 74.6% MHR during the brisk 6MWT trial; and 84.5% MHR during the fast 6MWT trial. These MHR percentages are in agreement
M.E. O’Neil et al. / Measuring reliability and validity of the ActiGraph GT3X accelerometer for children with cerebral palsy
with MET values (Table 2) and reflect the PA intensity range (light, moderate and vigorous) during the three 6MWT trials, respectively. The increased values in the children’s physiologic data (VO2 and HR) are in line with the increased activity count data recorded during each progressively faster 6MWT trial. However, the activity counts for children in this study are higher for the comfortable/slow and brisk 6MWT trials than reported previously for adolescents with CP [21]. Children in the current study had higher activity counts for the comfortable/slow trial because, on average, they self-selected a faster speed for this trial than children in the previous study. In contrast, they walked slightly slower for the brisk and fast walk trials than the previous study but had higher activity counts. This may be due to the fact that they were younger and shorter with more severe CP (i.e., a larger proportion of children in the current study are at GMFCS level III). Also, a few children exhibited increased non-purposeful movement with unstable hip and trunk movement and decreased balance. These child characteristics may contribute to increased activity counts. Overall, children in the current study did have higher activity counts and slightly higher step counts in the fast 6MWT trial compared to the brisk 6MWT trial despite slightly decreased distance and speed during the fast 6MWT trial. Again, these findings may be due to changes in walking patterns such as decreased step length while increasing number of steps for at least part of the trial or related to changes in speed at the end of the trial secondary to fatigue. Although this study is limited by the small sample size and inability to examine PA dimensions across GMFCS levels and clinical diagnoses, it does provide evidence to guide further research in this area. Future directions for establishing ActiGraph GT3X as a valid and reliable measure of PA for children and youth with CP should include a larger sample size with equal distribution among GMFCS levels I–III, motor distribution(hemiplegia, diplegia), and movement diagnoses (spasticity, dystonia). This will assist in determining if ActiGraph GT3X accelerometers are valid for children and youth with different gait patterns. A larger sample size will provide data to determine if youth prediction equations specific for children with CP are needed to estimate PA intensity from activity counts. Although these findings suggest that the ActiGraph GTX3 is valid for measuring PA intensity for walking, it does not provide information on other types of PA during daily life. In future research, it would be useful to expand the activity protocol of three different walking speeds to include other PA tasks in which children participate in their daily lives. Participation in daily
239
tasks, including daily chores and active recreation, are often goals of therapy so it is important to examine the reliability and validity of accelerometry in measuring PA intensity across a variety of activities specifically for children with CP. Additional research is needed to compare activity and step counts to determine if both can be used to determine PA intensity for children with CP in natural environments. Another consideration in PA outcome measurement is that the ActiGraph GT3X accelerometer is just one type of device used to measure PA in youth. There are several other types of accelerometers that have been validated and are currently being used to record PA parameters in children with typical development. Information from this study on ActiGraph GT3X accelerometers cannot be generalized to other devices. Each type of accelerometer should be validated to determine research and clinical utility of the device for youth with CP. 5. Conclusion This study contributes to the literature on quantitative measures of PA in children with CP by providing preliminary information on the inter-rater reliability of ActiGraph GT3X activity and step counts and the reproducibility of a walking protocol for children with CP. It adds to previous research on the ability of the ActiGraph GT3X activity counts to estimate PA intensity for walking. It also suggests that further research is needed on the use of step counts to estimate PA intensity in children with CP. Information on PA frequency and intensity is important to rehabilitation professionals for examining effectiveness of activity-based interventions. Conflict of interest The authors have no conflicts of interest to declare. Acknowledgement This study was partially supported by an APTA, Section on Pediatrics Research Grant. References [1]
U.S. Department of Health and Human Services. Healthy People 2020 2012 [updated January 26, 2012; cited February 2, 2012]. Available from: www.healthypeople.gov.
240 [2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
M.E. O’Neil et al. / Measuring reliability and validity of the ActiGraph GT3X accelerometer for children with cerebral palsy Trost S. Measurement of physical activity in children and adolescents. American Journal of Lifestyle Medicine. 2007; 1(4): 299-314. Caspersen C, Powell K, Christenson G. Physical activity, exercise, and physical fitness: Definitions and distinctions for health-related research. 1985 March-April Report No.: Contract No. 2. Trost S. Objective measurement of physical activity in youth: current issues, future directions. Exercise and Sport Sciences Review. 2001; 29(1): 32-6. US Department of Health and Human Services. Physical Activity Guidelines for Americans 2008 [updated October 16, 2008; cited January 16, 2014]. Available from: http://www. health.gov/Paguidelines/guidelines/Chapter3.aspx. Trost S, McIver K, Pate R. Conducting accelerometer-based activity assessments in field-based research. Medicine & Science in Sports & Exercise. 2005; 37(Suppl 11): S531-S43. Troiano R, Berrigan D, Dodd K, Masse L, Tilert T, McDowell M. Physical activity in the United States measured by accelerometer. Medicine & Science in Sports & Exercise. 2008; 40(1): 181-8. Bjornson K, Belza B, Kartin D, Logsdon R, McLaughlin J. Ambulatory physical activity performance in youth with cerebral palsy and youth who are developing typically. Physical Therapy. 2007; 87(3): 248-57. Fowler E, Kolobe TH, Damiano D, Thorpe D, Morgan D, Brunstrum J, et al. Promotion of physical fitness and prevention of secondary conditions for children with cerebral palsy: Section on Pediatrics Research Summit Proceedings. Physical Therapy. 2007; 87(11): 1495-510. King G, Law M, King S, Rosenbaum P, Kertoy M, Young N. A conceptual model of the factors affecting the recreation and leisure participation of children with disabilities. Physical & Occupational Therapy in Pediatrics. 2003; 23(1): 63-90. Rogers A, Furler B, Brinks S, Darrah J. A systematic review of the effectiveness of aerobic exercise interventions for children with cerebral palsy: An AACPDM evidence report. Developmental Medicine & Child Neurology. 2008; 50(11): 808-14. Rogozinski B, Davids J, Davis R, Christopher L, Anderson J, Jameson G, et al. Prevalence of obesity in ambulatory children with cerebral palsy. The Journal of Bone and Joint Surgery. 2007; 89(11): 2421-6. Rimmer J, Rowland J. Physical activity for youth with disabilities: a critical need in an underserved population. Developmental Neurorehabilitation. 2008; 11(2): 141-8. Law M, Petrenchik T, King G, Hurley P. Perceived environmental barriers to recreational, community, and school participation for children and youth with physical disabilities. Archives of Physical Medicine and Rehabilitation. 2007; 88(12): 1636-42. Verschuren O, Wiart L, Hermans D, Ketelaar M. Identification of facilitators and barriers to physical activity in children and adolescents with cerebral palsy. Journal of Pediatrics. 2012; 161(3): 488-94. Verschuren O, Ketelaar M, Gorter J, Helders P, Uiterwaal C, Takken T. Exercise training programs in children and adolescents with cerebral palsy. Archives of Pediatrics and Adolescent Medicine. 2007; 161(11): 1075-81. Capio C, Sit C, Abernethy B. Physical activity measurement using MTI (actigraph) among children with cerebral palsy. Archives of Physical Medicine and Rehabilitation. 2010; 91(8): 1283-90.
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
Balemans A, Fragala-Pinkham M, Lennon N, Thorpe D, Boyd R, O’Neil M, et al. Systematic review of the clinimetric properties of laboratory and field-based aerobic and anaerobic fitness measures in children with cerebral palsy. Archives of Physical Medicine and Rehabilitation. 2013; 94: 287-301. Clanchy K, Tweedy S, Boyd R. Measurement of habitual physical activity performance in adolescents with cerebral palsy: A systematic review. Developmental Medicine & Child Neurology. 2011; 53(6): 499-505. Capio C, Sit C, Abernethy B, Rotor E. Physical activity measurement instruments for children with cerebral palsy: A systematic review. Developmental Medicine & Child Neurology. 2010; 52(10): 908-16. Clanchy K, Tweedy S, Boyd R, Trost S. Validity of accelerometry in ambulatory children and adolescents with cerebral palsy. European Journal of Applied Physiology. 2011; 111(12): 2951-9. van Wely L, Becher J, Balemans A, Dallmeijer A. Ambulatory activity of children with cerebral palsy: which characteristics are important? Developmental Medicine & Child Neurology. 2012; 54(5): 436-42. Gorter J, Noorduyn S, Obeid J, Timmons B. Accelerometry: a feasible method to quantify physical activity in ambulatory and nonambulatory adolescents with cerebral palsy. International Journal of Pediatrics. 2012; doi: 10.1155/2012/329284 (Epub 2012): Jun 26. Thompson P, Beath T, Bell J, Jacobson G, Phair T, NM S, et al. Test-retest reliability of the 10-meter fast walk test and 6-minute walk test in ambulatory school-aged children with cerebral palsy. Developmental Medicine & Child Neurology. 2008; 50(5): 370-6. Maher C, Williams M, Olds T. The six minute walk test for children with cerebral palsy. International Journal of Rehabilitation Research. 2008; 31(2): 185-8. Palisano R, Rosenbaum P, Bartlett D, Livingston M. Content validity of the expanded and revised Gross Motor Function Classification System. Developmental Medicine and Child Neurology. 2008; 50(10): 744-50. National Center for Health Statistics. CDC growth charts: United States: Centers for Disease Control and Prevention, [cited May 30,2012]. Zemel B, Riley E, Stallings V. Evaluation of methodology for nutritional assessment in children: anthropometry, body composition, and energy expenditure. Annual Review of Nutrition. 1997; 17: 211-35. Slaughter M, Lohman T, Boileau R, Horswill C, Stillman R, Van Loan M, et al. Skinfold equations for estimation of body fatness in children and youth. Human Biology. 1988; 60(5): 709-23. Schofield W. Predicting basal metabolic rate, new standards and review of previous work. Human Nutrition and Clinical Nutrition. 1985; 39: 5-41. Corry I, Duffy C, Cosgrave A, Graham H. Measurement of oxygen consumption in disabled children by the Cosmed K2 portable telemetry system. Developmental Medicine & Child Neurology. 1996; 38(7): 585-93. Evenson K, Catellier D, Gill K, Ondrak K, McMurray R. Calibration of two objective measures of physical activity for children. Journal for Sports Sciences. 2008; 26(14): 1557-65. Verschuren O, Maltais D, Takken T. The 220-age equation does not predict maximum heart rate in children and adolescents. Developmental Medicine and Child Neurology. 2011; 53(9): 861-4.
Copyright of Journal of Pediatric Rehabilitation Medicine is the property of IOS Press and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
233
Measuring reliability and validity of the ActiGraph GT3X accelerometer for children with cerebral palsy: A feasibility study Margaret E. O’Neila,∗ , Maria A. Fragala-Pinkhamb, Jeffrey L. Formanc and Stewart G. Trostd a
Department of Physical Therapy and Rehabilitation Sciences, College of Nursing and Health Professions, Drexel University, Philadelphia, PA, USA b Franciscan Hospital for Children, Research Center for Children with Special Health Care Needs, Brighton, MA, USA c Department of Pediatric Rehabilitation Medicine, Franciscan Hospital for Children, Brighton, MA, USA d Queensland Institute of Technology, Institute of Health and Biomedical Innovation, St Lucia, QLD, Australia
Accepted 3 March 2014
Abstract. PURPOSE: The purposes of this study were to: 1) establish inter-instrument reliability between left and right hip accelerometer placement; 2) examine procedural reliability of a walking protocol used to measure physical activity (PA); and 3) confirm concurrent validity of accelerometers in measuring PA intensity as compared to the gold standard of oxygen consumption measured by indirect calorimetry. METHODS: Eight children (mean age: 11.9; SD: 3.2, 75% male) with CP (GMFCS levels I – III) wore ActiGraph GT3X accelerometers on each hip and the Cosmed K4b2 portable indirect calorimeter during two measurement sessions in which they performed the six minute walk test (6MWT) at three self-selected speeds (comfortable/slow, brisk, fast). Oxygen consumption (VO2) and accelerometer step and activity count data were recorded. RESULTS: Inter-instrument reliability of ActiGraph GT3X accelerometers placed on left and right hips was excellent (ICC = 0.96–0.99, CI95 : 0.81–0.99). Reproducibility of the protocol was good/excellent (ICC = 0.75–0.95, CI95 : 0.75–0.98). Concurrent validity of accelerometer count data and VO2 was fair/good (rho = 0.67, p < 0.001). The correlation between step count and VO2 was not significant (rho = 0.29, p = 0.2). CONCLUSION: This preliminary research suggests that ActiGraph GT3X accelerometers are reliable and valid devices to monitor PA during walking in children with CP and may be appropriate in rehabilitation research and clinical practice. ActiGraph GTX3 step counts were not valid for this sample and further research is warranted. Keywords: Physical activity measurement, accelerometry, cerebral palsy
1. Introduction and purpose Physical activity (PA) is identified as a leading health indicator in Healthy People 2020 and is important for children’s growth and development [1,2]. PA is ∗ Corresponding author: Margaret E. O’Neil, Drexel University, Mail Stop 7502, 1601 Cherry Street, Phila, PA 19102, USA. Tel.: +1 267 359 5546; Fax: +1 267 359 5576; E-mail: [email protected].
defined as “any bodily movement produced by skeletal muscles that results in energy expenditure (above resting)” [3,4]. It is a complex behavior with several dimensions (frequency, duration, intensity and type) and domains (leisure, occupational, chores and activities of daily living) [2]. The recommendation for school-aged children and youth participation in daily PA is defined as a minimum of 60 minutes of moderate to vigorous PA in-
c 2014 – IOS Press and the authors. All rights reserved 1874-5393/14/$27.50
234
M.E. O’Neil et al. / Measuring reliability and validity of the ActiGraph GT3X accelerometer for children with cerebral palsy
cluding muscle and bone strengthening activities [5]. Among children with typical development moderate to vigorous PA is associated with healthy weight, cardiorespiratory endurance, bone health and psychological health [6]. Despite known health benefits, the majority of children in the United States (58%) do not achieve 60 minutes of daily moderate to vigorous PA [7]. Evidence from cross-sectional studies suggests that children with cerebral palsy (CP) participate in significantly less PA compared to their peers with typical development [8,9]. Factors that contribute to decreased PA in children with CP span the International Classification of Function Model (ICF) dimensions of body structure and function, activity and participation associated with this diagnosis [9–11]. Specifically, children with CP often have decreased strength, motor coordination, balance, muscular endurance, and limitations in functional mobility and motor skills [9]. Children with CP also have decreased aerobic fitness (VO2 peak) which may contribute to their limited daily PA [11]. Children with CP are at high risk for secondary health conditions such as osteoporosis, chronic pain, fatigue, and obesity, which can influence participation in PA [9, 12]. Limited access to active recreation opportunities and decreased motivation and interest also influence level of PA participation in children with CP [13–15]. Pediatric rehabilitation services include health promotion, PA and exercise interventions for children with CP to promote active, healthy lifestyles. Providers involved in delivery of these services may include physical therapists, occupational therapists, exercise scientists and physicians. Although evidence supports the effectiveness of short-term aerobic exercise interventions lasting 2–6 months [11,16], no studies have been published on long-term PA interventions (> 6 months) aimed at improving health and physical function in children with CP. Researchers and clinicians acknowledge the importance of PA intensity as a critical component of interventions for children with CP. Furthermore additional research on PA intensity dosing is needed to guide clinical practice [9]. Therefore reliable and valid measures of PA frequency, intensity, and duration are important to examine PA participation in activity-based rehabilitation programs. PA measures include self-report questionnaires, pedometers, and accelerometers [2]. Quantitative fieldbased measures of “real world” PA are especially important when measuring habitual PA [8,17–22]. Selfreport measures have been validated for youth with typical development, however they often have inher-
ent response bias and may be difficult for children to complete, especially those younger than 10 years of age [2]. These measures have not been validated for children with CP. Child self-report and parent proxy measures of PA and participation specifically for children with disabilities describe PA patterns and behaviors but do not provide accurate or precise measures of PA intensity or frequency [19,22]. Pedometers and step-counters are useful and reliable measures of the frequency and duration of daily walking [2,4,8], but do not provide information on PA intensity [8,22]. Further, some pedometers may be susceptible to missing data due to design limitations [4]. Accelerometers have been used to measure PA frequency, duration and intensity in adults and children with typical development for several decades [2,4] and recently have been used among children with CP [8, 17,21,23]. Clanchy et al. [21] established the validity of ActiGraph uniaxial accelerometers in their ability to discriminate PA intensity in ambulatory children with CP during a walking protocol in which children walked at different self-selected speeds. However, the study did not examine inter-instrument reliability (left versus right hip placement). Inter-instrument reliability has been established for uniaxial and triaxial ActiGraph accelerometers among children with typical development [6], but has not been established among children with CP. It is important to examine monitor placement and establish interinstrument reliability to determine best placement of ActiGraph accelerometers for accurate PA data collection. Monitor placement may be critical in children with movement asymmetry and different distributions of CP (i.e., hemiplegia, diplegia, quadriplegia). Further, it is important to determine if a walking protocol for children with CP consisting of three different walking speeds is reproducible across measurement sessions to ensure protocol stability and reduce potential measurement error. The 6-minute walk test (6MWT) is reproducible in children with CP when walking at a fast pace [24,25], however, reproducibility of the 6MWT protocol with three self-selected walking speeds used in the current study has not been established. In real world settings, children with CP use a variety of walking speeds; therefore, it is important to determine if ActiGraph GT3X accelerometers can accurately measure PA intensity for different walking speeds. The purposes of this study were to: 1) examine interinstrument reliability between left and right hip ActiGraph GT3X accelerometer placement; 2) establish re-
M.E. O’Neil et al. / Measuring reliability and validity of the ActiGraph GT3X accelerometer for children with cerebral palsy
producibility of a walking protocol to measure PA; and 3) confirm concurrent validity of ActiGraph GT3X accelerometers in measuring PA intensity when compared to oxygen consumption (VO2) measured by indirect calorimetry as established in a previous study [23].
2. Methods 2.1. Participants A sample of eight children with CP participated in this study. Participants were 6–14 years old (mean = 11.9, SD = 3.2) and most (n = 6, 75%) were boys. Four children were classified at Gross Motor Function Classification System [26] (GMFCS) level I, one at GMFCS level II and three at GMFCS level III. Four children presented with a motor distribution of diplegia, three with hemiplegia and one with quadriplegia. Six children had healthy weight (BMI at 5th – 85th percentile) and two were obese (BMI 95th percentile) [27] (Table 1). All participants met the study’s inclusion criteria which were: 1) diagnosis of CP, 2) able to walk independently with or without an assistive device, 3) between the ages of 6–18 years, 4) no exercise restrictions or uncontrolled seizures, and 5) no lower extremity Botox injections less than three months prior to the study or orthopedic interventions less than six months prior to the study since these procedures initially may alter ambulation skills. Three of the eight participants used assistive devices (crutches or walker) and five used lower extremity braces. Three children were on seizure medications; two children had lower extremity Botox injections six or more months prior to testing and orthopedic surgery a year or more prior to testing. At the time of the study five children received physical and/or occupational therapy services in school and/or outpatient settings; three children received speech and language pathology services and special education services. 2.2. Procedure Institutional Review Board approval was obtained from Franciscan Hospital for Children (FHC) and Drexel University. Children were recruited from the outpatient clinics at FHC and from professional colleagues and parent groups in the area. Flyers were posted in clinic areas and were disseminated via email to colleagues and parents.
235
Parents and children provided informed consent or assent. Children participated in two measurement sessions at least one week apart. Each data collection session was approximately 1.5–2 hours. In the first data collection session parents completed a demographic questionnaire and medical information questionnaire about their children. One of the researchers (MFP) classified children on the GMFCS levels. The clinical and movement diagnoses were obtained from the medical record or by parent report. Height and weight were recorded for each child using a stadiometer (Shorrboard 420, QuickMedical, Issaquah, WA) and a digital scale (Tanita UM26, Arlington Heights, IL). Tricep and calf skin folds were measured using Lange calipers (Beta Technology, Santa Cruz, CA) and a standard protocol [28]. The Cosmed K4b2 (Rome, Italy) portable metabolic unit was used to measure VO2 during a 5 minute sitting rest and during three 6-minute walk tests. Children also wore a Polar heart rate (HR) monitor (Polar T-34, Lake Success, NY) which recorded HR during the activities. Children were fitted with a face mask which was secured to the head using a nylon head harness. The face mask was connected to the Cosmed K4b2 portable metabolic unit and the unit and battery were placed into a chest harness worn by the child. The children also wore a pair of ActiGraph GT3X accelerometers (ActiGraph Corporation, Pensacola, FL) which were attached at each side of the child’s waist using an elastic belt. Placement for each accelerometer was at the midaxilla line at the superior edge of the iliac crest. At the beginning of each data collection session, all equipment was calibrated or initialized according to manufacturer guidelines. Children sat quietly in a chair with back support for 5 minutes. Then each child participated in a walking protocol which consisted of three 6WMT trials each at a different self-selected speed (comfortable/slow; brisk; and fast) [21]. For the 6MWT trials, children were given the same verbal instruction to help pace their walking. Instructions for the three trials were: 1) Comfortable /Slow: “Walk at a comfortable slow pace like when you are at the mall or walking in your neighborhood or at school but you are not in a hurry”; 2) Brisk: “This time we want you to walk a little faster so please walk at a faster pace like when you are hurrying to get to class after the bell has rung” and 3) Fast: “This time we want you to walk as fast as you possibly can without falling or running.” [21] The trials were conducted in a hospital outpatient setting. A physical therapist (MFP) walked with each child to
236
M.E. O’Neil et al. / Measuring reliability and validity of the ActiGraph GT3X accelerometer for children with cerebral palsy Table 1 Child demographics, anthropometric & health data
ID
Gender
Age (yrs)
1 2 3 4 5 6 7 8
M M M M F M F M
12.2 13.6 14.00 15.92 6.42 8.75 10.17 14.25
∗ Body
Clinical diagnosis Diplegia Hemiplegia Hemiplegia Hemiplegia Diplegia Quadriplegia Diplegia Diplegia
Movement diagnosis Spasticity Spasticity Spasticity Spasticity Dystonia Dystonia Spasticity Spasticity
GMFCS level I I I I II III III III
Height (cm) 154.7 165.5 156 170.5 109.4 116.7 126.4 164.5
Weight (kg) 59.1 45.1 44.1 66.1 18.2 21.6 30.8 72.6
BMI % ile 95.3 11.9 33.4 75.6 47.6 45.1 79.1 95.7
% Body Fat∗ 65.3 15.9 10.9 23.6 17.3 7.62 34.38 51.7
RMR∗∗ kcal/24 hrs 1688.1 1475.4 1446.2 1823.5 857.0 990.1 1098.4 1920.9
fat estimate (Slaughter equation for tricep and calf skinfold measures); ∗∗ RMR = resting metabolic rate estimates (Schofield Equations).
encourage a consistent pace throughout the entire trial and for safety. Distance walked was recorded for each trial and walking speed was calculated. Children rested between each 6MWT trial until HR returned to resting level and when they reported having had sufficient rest. During the second data collection session, children repeated the 5 minute sitting rest and the walking protocol described above. 2.3. Measures The Cosmed K4b2 is a portable indirect calorimeter that is a reliable and valid measure of oxygen consumption (VO2) and energy expenditure (CO2 production). Indirect calorimetry is considered a criterion measure of PA intensity for children with CP [19]. Data collected by the Cosmed K4b2 were breath-bybreath relative VO2 (ml/kg/min) and CO2 production. Metabolic equivalents and resting metabolic rates were estimated from these data [29,30]. The ActiGraph GT3X accelerometer was used to measure activity and step counts. ActiGraph GT3X is a triaxial accelerometer designed to detect three dimensional accelerations from 0.05–2.00 g magnitude with a frequency response of 0.25–2.50 Hz. The ActiGraph accelerometer filters and processes the acceleration data over a specific time interval (epoch). For this study, the epoch interval was set at 1 second to synchronize count data to the VO2 and HR data measured by the Cosmed K4b2 unit and the Polar HR monitor [2, 4]. Activity and step counts were “summed” during each epoch and stored by the monitor with high counts reflecting high PA intensity and frequency. 2.4. Data reduction and analysis Body mass index (BMI) was calculated from height and weight data and weight status was determined from Center for Disease Control and Prevention (CDC)
growth charts [27]. Body fat estimates were calculated using skinfold data and the appropriate age and gender specific Slaughter equations [29]. Estimates of resting metabolic rate were calculated using the Schofield equations [30]. Descriptive statistics (means and standard deviations) were calculated for walking distance, walking speed, HR, VO2, activity and step counts and metabolic equivalents (METs) generated during the 5 minute sitting rest and the three 6MWTs. Descriptive data are reported from session 2 since both sessions had similar outcomes. Mean values for the VO2 and activity and step count data were calculated for the time interval of 3.0– 4.5 minutes for the sitting rest period and 3.5–5.5 minutes for each walking trial. Data from minutes 3.5–5.5 were used in the analysis to ensure that “steady state” had been achieved and to avoid using data from the final 30 seconds in which children may have slowed their pace in anticipation of the end of the trial. For each participant, steady state was confirmed by inspection of HR and VO2 values [21,31]. METs (estimates of PA intensity) were calculated (mean VO2/resting metabolic rate) [29]. Activity counts were classified into an intensity category based on the cut-points using the Evenson youth prediction equation. Although there are no youth prediction equations for children and youth with CP, Clanchy et al. determined that the Evenson equation was sufficient for interpreting activity counts for children with CP who were ambulatory [21,32]. Intra-class correlation coefficients (ICCs) were calculated for activity and step count data for each test session to examine inter-instrument reliability (left to right hip monitor placement). ICCs also were used to examine reproducibility of the 6MWT walking protocol between test sessions. Spearman correlations were calculated to determine if ActiGraph GT3X step and activity counts are valid measures of PA intensity as compared to VO2.
M.E. O’Neil et al. / Measuring reliability and validity of the ActiGraph GT3X accelerometer for children with cerebral palsy
237
Table 2 Descriptive data for the three 6MWT trials (from Session 2) 6MWT trials Distance (meters) Speed (m/min) Heart rate (bpm) VO2 (ml/kg/min) Activity Counts Mean (SD) Mean (SD) Mean (SD) Mean (SD) (Counts/min) (Left and Right) Mean (SD) Rest n/a n/a 78.67(8.82) 4.77 (1.10) n/a
Step Counts METs (Counts/min) Mean (SD) (Left and Right) Mean (SD) n/a n/a
Comfortable/ 339.67 (70.39) Slow
56.61(11.73)
122.11(28.28)
18.05 (8.36)
L: 2521.5 (953.6) R: 2520.8 (900.9)
L: 93.2 (31.2) R: 92.2 (31.4)
Brisk
400.38(113.80)
66.18(19.18)
144.72 (27.27)
21.08 (6.53)
L: 3753.2 (1604.8) L: 102.9 (33.7) R: 3761.9 (1633.1) R: 102.8 (33.1)
4.56 (1.79)
Fast
384.38(217.03)
64.07(36.18)
164.02(20.76)
25.53 (8.31)
L: 4542.8 (2016.1) L: 108 (37.1) R: 4439.8 (1907.1) R: 106.8 (38.6)
5.61 (2.81)
3.88 (1.92)
3. Results
3.4. Validity
3.1. Descriptive data
ActiGraph GT3X right hip count data were used to evaluate concurrent validity. Fair to good validity (rho = 0.67, p < 0.001) was established between ActiGraph activity count data and VO2. The correlation between ActiGraph GT3X step count and VO2 was not significant (rho = 0.29, p = 0.2).
ActiGraph GT3X activity and step counts (counts/ min) increased on both sides (left and right hips) when participants increased self-selected speed in the three 6MWT trials. Mean distance walked and speed increased during the 6MWT trials between comfortable/slow to brisk trials but not between brisk and fast trials. Also, mean distance and speed increased between comfortable/slow and fast trials. HR (bpm) was, on average, 78.67 bpm (SD = 8.82) during sitting rest and increased to a mean of 164.02 (SD = 20.76) for the fast speed 6MWT trial. Relative oxygen consumption (VO2 ml/kg/min) ranged from a mean of 4.77 (SD = 1.10) at rest to a mean of 25.53 (SD = 8.31) during the fast speed 6MWT trial. METS increased from a mean of 3.88 (SD = 1.92) during the comfortable/slow speed 6MWT trial to a mean of 5.61 (SD = 2.81) during the fast speed 6MWT trial (Table 2). 3.2. Inter-instrument reliability Results indicate that ActiGraph GT3X accelerometers had excellent inter-instrument reliability for both activity and step count data across the 6MWT trials at the three self-selected speeds. The range for the ICCs for all three 6MWT trials was 0.98 (CI95 : 0.92–0.99) to 0.98 (CI95 : 0.87–0.99) for activity counts and 0.99 (CI95 : 0.98–1.00) to 0.99 for step counts (CI95 : 0.99– 1.00) (Table 3). 3.3. Walking protocol reproducibility Good test-retest reproducibility was demonstrated between the two testing sessions. The ICC range for comfortable/slow to fast walking trials activity counts was 0.75 (CI95 : 0.02–0.95) to 0.95 (CI95 : 0.81–0.99) and for step counts was 0.88 (CI95 : 0.49–0.98) to 0.97 (CI95 : 0.88–0.99) (Table 4).
4. Discussion Findings from this study suggest that the ActiGraph GT3X accelerometer provides a reliable and valid measure of PA intensity during walking for children with CP. The ActiGraph data were reliable for bilateral hip placement suggesting that children with CP can follow the same protocol as children with typical development and wear ActiGraph GT3X accelerometers on the right side regardless of GMFCS level and motor distribution (diplegia, hemiplegia, quadriplegia). However since our sample size was small, replication of this study using a larger sample size of children with unilateral and bilateral motor involvement is recommended. ActiGraph GT3X activity counts were significantly and positively correlated with VO2 for the walking trials; however step counts were not. These findings indicate that for this device activity counts may be a better outcome measure than step counts when the goal is to increase PA intensity during walking for children with CP. When comparing the brisk and fast walking trials, participants exercised at a higher intensity during the fast walk as indicated by increased HR and VO2 but only took a few more steps and had a slight decrease in distance walked in 6 minutes. Activity counts; however, did increase across all three walking trials. Increased accessory trunk and upper extremity motions were observed during brisk and fast walking trials especially for participants classified at GMFCS levels II–
238
M.E. O’Neil et al. / Measuring reliability and validity of the ActiGraph GT3X accelerometer for children with cerebral palsy Table 3 Test Session: Inter-instrument Reliability (left to right monitor placement) 6MWT trials Comfortable/Slow Walk : L vs. R Brisk Walk: L vs. R Fast Walk: L vs. R
Activity counts ICC 95% CI 0.98 0.92–0.99 0.99 0.98–1.00 0.98 0.87–0.99
ICC 0.99 0.96 0.99
Step counts 95% CI 0.98–1.00 0.81–0.99 0.99–1.00
L = Left hip; R = Right hip. Table 4 Walking protocol reproducibility 6MWT trials Comfortable/Slow Walk: L1 vs. L2 Comfortable/Slow Walk: R1 vs. R2 Brisk Walk: L1 vs. L2 Brisk Walk: R1 vs. R2 Fast Walk: L1 vs. L2 Fast Walk: R1 vs. R2
Activity counts ICC 95% CI 0.75 0.02–0.95 0.86 0.48–0.97 0.88 0.52–0.98 0.95 0.59–0.99 0.89 0.47–0.98 0.95 0.81–0.99
ICC 0.88 0.96 0.88 0.98 0.88 0.97
Step counts 95% CI 0.49–0.98 0.82–0.99 0.47–0.98 0.92–0.99 0.48–0.98 0.88–0.99
L1 = Left hip Session 1; L2 = Left hip Session 2; R1 = Right hip Session 1; R2 = Right hip Session 2.
III which may be recorded as increased activity counts. These accessary motions seemed to increase especially at the end of the fast walking trial for children who were fatigued. Child performance on the 6MWT trials was reproducible between the two testing sessions, suggesting that this is a feasible and stable protocol for future PA measurement studies. It may also be a good test protocol to examine the effectiveness of activity-based interventions that are designed to increase PA walking intensity. In this study, children increased distance and speed between the comfortable/slow and brisk 6MWT trials. The distance covered in the brisk 6MWT trial was similar to the mean reported in one study (340.8 meters) [24] and less than the mean of 449.1 meters reported in another study of youth with CP [25]. The difference in the second study [25] could be that children were older (mean age = 13.6; SD = 1.6 years) and fewer were classified at GMFCS III. In the current study, on average children walked a shorter distance and at a slower speed during the fast 6MWT trial compared to the brisk trial. On inspection of the data, it was observed that 6 of the 8 children increased speed from their brisk to fast walk. One child had the same speed and two children with more motor involvement (GMFCS Level II and III) decreased in their speed from brisk to fast. This decreased distance and speed may be due to fatigue and de-conditioning because the fast speed 6MWT trial was the final trial in the protocol. The two children who demonstrated a decrease in speed were inactive and deconditioned according to parent report. Children were given a rest period to
allow them to re-establish baseline HR between trials. However, in future protocols additional rest time may be required for recovery of muscular endurance for some children in order to maintain a consistent fast pace for 6 minutes. To further illustrate this point, differences in speeds within a walking trial were evaluated. On average during the fast 6MWT trial, children walked faster at the one minute mark (mean = 83.22 meters/minute, SD = 24.79) compared to the three (65.87 meters/minute, SD = 38.02) and six minute marks (64.07 meters/minute, SD = 36.18). The mean speed at the one minute mark for the fast walk trial was faster than the speed during the brisk 6MWT trial (mean = 66.18 meters/min, SD = 19.18). These findings reflect the difficulty children had with maintaining consistent intensity during the fast 6MWT. Findings on the fast 6MWT trial suggest that further study is needed to determine if a longer rest period in between tests or a shorter walk test would be more appropriate for testing PA intensity in children with CP. The full 6MWT protocol used in the current study (three trials with self-selected speeds) might be a good outcome when the goal is to measure an increase in walking endurance; however further research is needed to confirm this recommendation. During the three 6MWT trials, the participants exhibited a steady increase in HR, VO2 and METs. Using a maximal heart rate (MHR) of 194 bpm [33], participants were, on average, at 63% MHR during the comfortable/slow 6MWT trial; 74.6% MHR during the brisk 6MWT trial; and 84.5% MHR during the fast 6MWT trial. These MHR percentages are in agreement
M.E. O’Neil et al. / Measuring reliability and validity of the ActiGraph GT3X accelerometer for children with cerebral palsy
with MET values (Table 2) and reflect the PA intensity range (light, moderate and vigorous) during the three 6MWT trials, respectively. The increased values in the children’s physiologic data (VO2 and HR) are in line with the increased activity count data recorded during each progressively faster 6MWT trial. However, the activity counts for children in this study are higher for the comfortable/slow and brisk 6MWT trials than reported previously for adolescents with CP [21]. Children in the current study had higher activity counts for the comfortable/slow trial because, on average, they self-selected a faster speed for this trial than children in the previous study. In contrast, they walked slightly slower for the brisk and fast walk trials than the previous study but had higher activity counts. This may be due to the fact that they were younger and shorter with more severe CP (i.e., a larger proportion of children in the current study are at GMFCS level III). Also, a few children exhibited increased non-purposeful movement with unstable hip and trunk movement and decreased balance. These child characteristics may contribute to increased activity counts. Overall, children in the current study did have higher activity counts and slightly higher step counts in the fast 6MWT trial compared to the brisk 6MWT trial despite slightly decreased distance and speed during the fast 6MWT trial. Again, these findings may be due to changes in walking patterns such as decreased step length while increasing number of steps for at least part of the trial or related to changes in speed at the end of the trial secondary to fatigue. Although this study is limited by the small sample size and inability to examine PA dimensions across GMFCS levels and clinical diagnoses, it does provide evidence to guide further research in this area. Future directions for establishing ActiGraph GT3X as a valid and reliable measure of PA for children and youth with CP should include a larger sample size with equal distribution among GMFCS levels I–III, motor distribution(hemiplegia, diplegia), and movement diagnoses (spasticity, dystonia). This will assist in determining if ActiGraph GT3X accelerometers are valid for children and youth with different gait patterns. A larger sample size will provide data to determine if youth prediction equations specific for children with CP are needed to estimate PA intensity from activity counts. Although these findings suggest that the ActiGraph GTX3 is valid for measuring PA intensity for walking, it does not provide information on other types of PA during daily life. In future research, it would be useful to expand the activity protocol of three different walking speeds to include other PA tasks in which children participate in their daily lives. Participation in daily
239
tasks, including daily chores and active recreation, are often goals of therapy so it is important to examine the reliability and validity of accelerometry in measuring PA intensity across a variety of activities specifically for children with CP. Additional research is needed to compare activity and step counts to determine if both can be used to determine PA intensity for children with CP in natural environments. Another consideration in PA outcome measurement is that the ActiGraph GT3X accelerometer is just one type of device used to measure PA in youth. There are several other types of accelerometers that have been validated and are currently being used to record PA parameters in children with typical development. Information from this study on ActiGraph GT3X accelerometers cannot be generalized to other devices. Each type of accelerometer should be validated to determine research and clinical utility of the device for youth with CP. 5. Conclusion This study contributes to the literature on quantitative measures of PA in children with CP by providing preliminary information on the inter-rater reliability of ActiGraph GT3X activity and step counts and the reproducibility of a walking protocol for children with CP. It adds to previous research on the ability of the ActiGraph GT3X activity counts to estimate PA intensity for walking. It also suggests that further research is needed on the use of step counts to estimate PA intensity in children with CP. Information on PA frequency and intensity is important to rehabilitation professionals for examining effectiveness of activity-based interventions. Conflict of interest The authors have no conflicts of interest to declare. Acknowledgement This study was partially supported by an APTA, Section on Pediatrics Research Grant. References [1]
U.S. Department of Health and Human Services. Healthy People 2020 2012 [updated January 26, 2012; cited February 2, 2012]. Available from: www.healthypeople.gov.
240 [2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
M.E. O’Neil et al. / Measuring reliability and validity of the ActiGraph GT3X accelerometer for children with cerebral palsy Trost S. Measurement of physical activity in children and adolescents. American Journal of Lifestyle Medicine. 2007; 1(4): 299-314. Caspersen C, Powell K, Christenson G. Physical activity, exercise, and physical fitness: Definitions and distinctions for health-related research. 1985 March-April Report No.: Contract No. 2. Trost S. Objective measurement of physical activity in youth: current issues, future directions. Exercise and Sport Sciences Review. 2001; 29(1): 32-6. US Department of Health and Human Services. Physical Activity Guidelines for Americans 2008 [updated October 16, 2008; cited January 16, 2014]. Available from: http://www. health.gov/Paguidelines/guidelines/Chapter3.aspx. Trost S, McIver K, Pate R. Conducting accelerometer-based activity assessments in field-based research. Medicine & Science in Sports & Exercise. 2005; 37(Suppl 11): S531-S43. Troiano R, Berrigan D, Dodd K, Masse L, Tilert T, McDowell M. Physical activity in the United States measured by accelerometer. Medicine & Science in Sports & Exercise. 2008; 40(1): 181-8. Bjornson K, Belza B, Kartin D, Logsdon R, McLaughlin J. Ambulatory physical activity performance in youth with cerebral palsy and youth who are developing typically. Physical Therapy. 2007; 87(3): 248-57. Fowler E, Kolobe TH, Damiano D, Thorpe D, Morgan D, Brunstrum J, et al. Promotion of physical fitness and prevention of secondary conditions for children with cerebral palsy: Section on Pediatrics Research Summit Proceedings. Physical Therapy. 2007; 87(11): 1495-510. King G, Law M, King S, Rosenbaum P, Kertoy M, Young N. A conceptual model of the factors affecting the recreation and leisure participation of children with disabilities. Physical & Occupational Therapy in Pediatrics. 2003; 23(1): 63-90. Rogers A, Furler B, Brinks S, Darrah J. A systematic review of the effectiveness of aerobic exercise interventions for children with cerebral palsy: An AACPDM evidence report. Developmental Medicine & Child Neurology. 2008; 50(11): 808-14. Rogozinski B, Davids J, Davis R, Christopher L, Anderson J, Jameson G, et al. Prevalence of obesity in ambulatory children with cerebral palsy. The Journal of Bone and Joint Surgery. 2007; 89(11): 2421-6. Rimmer J, Rowland J. Physical activity for youth with disabilities: a critical need in an underserved population. Developmental Neurorehabilitation. 2008; 11(2): 141-8. Law M, Petrenchik T, King G, Hurley P. Perceived environmental barriers to recreational, community, and school participation for children and youth with physical disabilities. Archives of Physical Medicine and Rehabilitation. 2007; 88(12): 1636-42. Verschuren O, Wiart L, Hermans D, Ketelaar M. Identification of facilitators and barriers to physical activity in children and adolescents with cerebral palsy. Journal of Pediatrics. 2012; 161(3): 488-94. Verschuren O, Ketelaar M, Gorter J, Helders P, Uiterwaal C, Takken T. Exercise training programs in children and adolescents with cerebral palsy. Archives of Pediatrics and Adolescent Medicine. 2007; 161(11): 1075-81. Capio C, Sit C, Abernethy B. Physical activity measurement using MTI (actigraph) among children with cerebral palsy. Archives of Physical Medicine and Rehabilitation. 2010; 91(8): 1283-90.
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
Balemans A, Fragala-Pinkham M, Lennon N, Thorpe D, Boyd R, O’Neil M, et al. Systematic review of the clinimetric properties of laboratory and field-based aerobic and anaerobic fitness measures in children with cerebral palsy. Archives of Physical Medicine and Rehabilitation. 2013; 94: 287-301. Clanchy K, Tweedy S, Boyd R. Measurement of habitual physical activity performance in adolescents with cerebral palsy: A systematic review. Developmental Medicine & Child Neurology. 2011; 53(6): 499-505. Capio C, Sit C, Abernethy B, Rotor E. Physical activity measurement instruments for children with cerebral palsy: A systematic review. Developmental Medicine & Child Neurology. 2010; 52(10): 908-16. Clanchy K, Tweedy S, Boyd R, Trost S. Validity of accelerometry in ambulatory children and adolescents with cerebral palsy. European Journal of Applied Physiology. 2011; 111(12): 2951-9. van Wely L, Becher J, Balemans A, Dallmeijer A. Ambulatory activity of children with cerebral palsy: which characteristics are important? Developmental Medicine & Child Neurology. 2012; 54(5): 436-42. Gorter J, Noorduyn S, Obeid J, Timmons B. Accelerometry: a feasible method to quantify physical activity in ambulatory and nonambulatory adolescents with cerebral palsy. International Journal of Pediatrics. 2012; doi: 10.1155/2012/329284 (Epub 2012): Jun 26. Thompson P, Beath T, Bell J, Jacobson G, Phair T, NM S, et al. Test-retest reliability of the 10-meter fast walk test and 6-minute walk test in ambulatory school-aged children with cerebral palsy. Developmental Medicine & Child Neurology. 2008; 50(5): 370-6. Maher C, Williams M, Olds T. The six minute walk test for children with cerebral palsy. International Journal of Rehabilitation Research. 2008; 31(2): 185-8. Palisano R, Rosenbaum P, Bartlett D, Livingston M. Content validity of the expanded and revised Gross Motor Function Classification System. Developmental Medicine and Child Neurology. 2008; 50(10): 744-50. National Center for Health Statistics. CDC growth charts: United States: Centers for Disease Control and Prevention, [cited May 30,2012]. Zemel B, Riley E, Stallings V. Evaluation of methodology for nutritional assessment in children: anthropometry, body composition, and energy expenditure. Annual Review of Nutrition. 1997; 17: 211-35. Slaughter M, Lohman T, Boileau R, Horswill C, Stillman R, Van Loan M, et al. Skinfold equations for estimation of body fatness in children and youth. Human Biology. 1988; 60(5): 709-23. Schofield W. Predicting basal metabolic rate, new standards and review of previous work. Human Nutrition and Clinical Nutrition. 1985; 39: 5-41. Corry I, Duffy C, Cosgrave A, Graham H. Measurement of oxygen consumption in disabled children by the Cosmed K2 portable telemetry system. Developmental Medicine & Child Neurology. 1996; 38(7): 585-93. Evenson K, Catellier D, Gill K, Ondrak K, McMurray R. Calibration of two objective measures of physical activity for children. Journal for Sports Sciences. 2008; 26(14): 1557-65. Verschuren O, Maltais D, Takken T. The 220-age equation does not predict maximum heart rate in children and adolescents. Developmental Medicine and Child Neurology. 2011; 53(9): 861-4.
Copyright of Journal of Pediatric Rehabilitation Medicine is the property of IOS Press and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
