- Email: [email protected]
Physical activity in pediatric onset multiple sclerosis: Validating a questionnaire for clinical practice and research
Multiple Sclerosis and Related Disorders 10 (2016) 26–29
Contents lists available at ScienceDirect
Multiple Sclerosis and Related Disorders journal homepage: www.elsevier.com/locate/msard
Physical activity in pediatric onset multiple sclerosis: Validating a questionnaire for clinical practice and research Dominique Kinnett-Hopkins a, Stephanie A. Grover b, E. Ann Yeh b,c, Robert W. Motl a,d,n a
Department of Kinesiology and Community Health, University of Illinois at Urbana-Champaign, USA Division of Neurology, Department of Neurosciences and Mental Health, Hospital for Sick Children, Toronto, ON, Canada c Department of Pediatrics, The Faculty of Medicine, University of Toronto, Toronto, ON, Canada d Department of Physical Therapy, University of Alabama at Birmingham, USA b
art ic l e i nf o
a b s t r a c t
Article history: Received 8 June 2016 Received in revised form 17 August 2016 Accepted 19 August 2016
Background: Knowledge regarding physical activity (PA) and its benefits in pediatric onset multiple sclerosis (POMS) is growing and suggests high levels of inactivity. The utility of a validated screening tool for clinical settings is unknown. This study evaluated the Godin Leisure-Time Exercise Questionnaire (GLTEQ) as a measure of PA in POMS. Methods: POMS patients (n¼27) and healthy controls (n¼45) wore an accelerometer over a 7-day period and then completed the GLTEQ. Results: The GLTEQ captured expected group differences in PA for vigorous, moderate, and moderate-tovigorous physical activity (MVPA), confirmed by accelerometry. There was a large, positive correlation between GLTEQ and accelerometry scores for vigorous PA in POMS (r ¼0.736, p¼0.001), and a nearly significant and moderate, positive correlation between MVPA scores (r¼ 0.319, p¼ .053). Conclusion: We provide evidence that supports the validity of GLTEQ scores as measures of vigorous and MVPA in POMS. Researchers and clinicians might adopt this scale for measuring PA. & 2016 Elsevier B.V. All rights reserved.
Keywords: Multiple sclerosis Physical activity Pediatric Measurement Validity
1. Introduction Pediatric onset multiple sclerosis (POMS) constitutes approximately 3–5% of all cases of multiple sclerosis (MS) (Chitnis et al., 2009). POMS is characterized by higher disease burden on magnetic resonance imaging (MRI) (Yeh et al., 2009; Chabas and Pelletier, 2009) and slower time to irreversible motor deficit compared with adult onset MS (Renoux et al., 2007; Simone et al., 2002). Youth with POMS experience cognitive dysfunction (Amato et al., 2008), frequent and severe relapses, high rates of fatigue and depression, and reduced quality of life (QOL) (Simone et al., 2002; Amato et al., 2008; Parrish et al., 2013; Goretti et al., 2012). There is evidence that physical activity participation in adults with MS aids in the management of the manifestations of this disease (Motl et al., 2015) and thus can be promoted through healthcare providers (Vollmer et al., 2012). There is a need for similar evidence in POMS, and this necessitates a validated assessment of physical activity levels for researchers as well as clinicians interested in screening whether or not POMS patients are active enough for n Correspondence to: 1705 University Blvd, SHPB 336, Birmingham, AL 35233, USA. E-mail address: [email protected] (R.W. Motl).
http://dx.doi.org/10.1016/j.msard.2016.08.010 2211-0348/& 2016 Elsevier B.V. All rights reserved.
health benefits. Increasing knowledge suggests that certain lifestyle attributes, including physical activity, may affect outcomes in POMS (Yeh et al., 2015). This interest in lifestyle is based, in part, on observations that high body mass index (BMI) was associated with POMS (Gianfrancesco et al., 2014; Langer-Gould et al., 2013) and strenuous physical activity participation, as measured by the Godin Leisure-Time Exercise Questionnaire (GLTEQ), was associated with outcomes such as lower T2 lesion volumes and relapse rates in POMS (Grover et al., 2015) as well as the evidence of the benefits of physical activity participation in adults with MS (Motl et al., 2015). Nevertheless, much scientific inquiry still needs to be undertaken regarding physical activity and POMS. The continued study of physical activity in POMS has been highlighted in a recent review (Yeh et al., 2015) and requires measures with evidence for the validity of score inferences (Motl et al., 2015). Indeed, the scientific study of physical activity among adults with MS was significantly accelerated based on evidence establishing the construct validity of scores from self-report outcomes such as the GLTEQ (Godin and Shephard, 1985; Motl et al., 2006). Such evidence of validity is necessary in POMS because its manifestations (e.g., cognitive impairment, fatigue, and depression) could influence the validity of self-report measures of physical activity, and
D. Kinnett-Hopkins et al. / Multiple Sclerosis and Related Disorders 10 (2016) 26–29
one cannot simply assume that the evidence for adults with MS (Gosney et al., 2007) or children from the general population (Koo and Rohan, 1999) support applications in POMS. The validity of the GLTEQ in POMS can be established by comparison with objective accelerometry using a two-part approach, namely the known-groups and construct validity paradigms (Motl and Sandroff, 2010). This two-part approach requires (a) comparing mean scores from the GLTEQ and accelerometry for capturing expected differences of vigorous, moderate and MVPA between POMS and healthy controls (HCs) and (b) examining and comparing the magnitude of the correlation between scores from the GLTEQ with accelerometry in combined samples of POMS and HC and POMS separately. Accordingly, evidence of score validity (i.e., the degree to which scores from a measurement can be interpreted as a measure of the intended construct) (Messick, 1995) can be established based on the expectation of capturing (a) lower levels of vigorous, moderate and MVPA in POMS than HC with the GLTEQ, which are confirmed with accelerometry, considering the presence and the degree of disease manifestation such as cognitive impairments, fatigue and depression (Parrish et al., 2013; Goretti et al., 2012) and (b) moderate or strong correlations between scores from the GLTEQ with accelerometer outcomes in the combined samples of POMS and HCs that are confirmed in the POMS sample alone. This study examined the validity of scores from the GLTEQ in POMS patients. We validated the scores using an objective measure of physical activity (i.e., accelerometry) in combination with a nomological net (i.e. hypothesized pattern of differences between groups and correlations among scores) (Cronbach and Meehl, 1955; Landy, 1986). We opted for comparison with accelerometry, as this is a validated and commonly used outcome of physical activity in pediatric populations (Crouter et al., 2013; Puyau et al., 2004) and provides objective metrics for comparison with selfreported behavior. Of note, we have focused on validating scores from the GLTEQ so that this simple and practical measure is amenable for use by clinicians to determine POMS physical activity risk and for inclusion in large, population-based studies of PA in POMS and applications wherein researchers and clinicians do not have resources or experience to include accelerometry.
27
Table 1 Demographic characteristics of POMS and HC. Variable
POMS (N ¼ 27)
HCs (N¼ 45)
p-Value
Statistic
Sex (% female (n)) Age (years) Race (% white (n)) MS Duration (years) EDSS
18/66.7 15.73(3.2) 12/44.4 2.03(2) 1.5 (0.5)
30/66.7 14.76(3.8) 33/73.3
1.000 0.030 0.027
X2 ¼0.000 z ¼ 0.051 X2 ¼6.009
Note. Values are reported as median (IQR) and were analyzed using the MannWhitney U Test (age) or chi-square test (sex and race). POMS ¼ pediatric-onset multiple sclerosis; HCs¼ healthy controls; EDSS ¼Expanded Disability Status Scale.
that measure the frequency (0–7) of strenuous (e.g. jogging), moderate (e.g. fast walking) and mild (e.g. easy walking) exercise for sessions more than 15 min during one's free time in the preceding week (Godin, 2011). The weekly frequencies of strenuous, moderate, and mild activities are multiplied by 9, 5 and 3 metabolic equivalents (METs), respectively, and then summed into a measure of total leisure activity (0–119) in MET/min per week. Importantly, the score of 9 for vigorous/strenuous reflects one, 15minute bout per week, whereas the score of 9 for light/mild reflects three, 15-minute bouts per week. The health contribution score (i.e., equivalent of time spent in MVPA) is calculated from the frequency of only strenuous and moderate activities. The frequencies for strenuous and moderate activities are multiplied by 9 and 5 METs, respectively, and then summed into a health contribution score (0–98) that reflects MET/min per week. The scores are then classified into three categories: active (substantial benefits; 24 or more MET/min per week), moderately active (some benefits; 14–23 MET/min per week), and insufficiently active (less substantial or low benefits; 13 or fewer MET/min per week) (Godin, 2011). 2.3. Accelerometer
The procedures for this study were approved by the Hospital for Sick Children Research Ethics Board, and all participants and legal guardians provided written informed consent and/or assent. We recruited a sample of 72 subjects for this study (POMS¼ 27; HC ¼45). The POMS patients were recruited from the Pediatric MS and Neuroinflammatory Clinic at The Hospital for Sick Children, Toronto, Canada. HCs subjects were recruited via word of mouth and flyers around the hospital campus. POMS participants were included if they were (a) 8–18 years old, (b) had Expanded Disability Status Scale (EDSS) scores o4.0, (c) had not received steroids or experienced a relapse in the last 30 days, and (d) if their MS was considered stable. Participants were excluded if there was a presence of non-specific white matter abnormalities and metabolic or infectious etiologies for white matter abnormalities. HCs were included if they (a) were 8–18 years of age and (b) had no previous neurological problems. The demographic and clinical characteristics of the participants are provided in Table 1.
We used the ActiGraph model 7164, accelerometer (ActiGraph Corporation, Pensacola, Florida) as an objective measure of physical activity based on monitoring over a seven-day period under free-living conditions. The device is a small (2.0 1.6 0.6 in.) and lightweight (1.5 ounces) accelerometer worn in a pouch on a snug elastic belt around the waist on the non-dominant hip. The single, vertical axis piezoelectric bender element within the accelerometer generates an electrical signal proportional to the force acting on it that is converted into activity counts over a pre-specified period of time (epoch) and then stored in random access memory within the device. The epoch in this study was set to be 1 min. The data from the accelerometers were processed using ActiLife software. The data from each participant's accelerometer were processed into 2 separate Microsoft Excel (Microsoft, Redmond, Washington) files representing wear time and time spent participating in differing levels of physical activity indicated by the number of counts per minute (CPM) recorded during each epoch: sedentary (i.e., r799 CPM), light (i.e., 800-3199 CPM), moderate (i.e., 3200-8199 CPM), and vigorous (i.e., Z 8200 CPM) (Puyau et al., 2002). Accelerometer wear time data were checked against participant recorded wear times from the log sheet, and only valid days (Z10 h of wear time without periods of continuous zeros exceeding 60 min indicative of compliance) were included in the analysis. The outcomes for the analyses were average daily minutes for total wear time and minutes of light, moderate and vigorous physical activity per day across valid days of data.
2.2. Instruments
2.4. Procedures
2. Methods 2.1. Participants
GLTEQ. The GLTEQ is comprised of three open-ended questions
Participants underwent a neurologic exam by a neurologist
28
D. Kinnett-Hopkins et al. / Multiple Sclerosis and Related Disorders 10 (2016) 26–29
specializing in MS for generation of an Expanded Disability Status Scale (EDSS) score. We familiarized the participants with the objective monitors and asked them to wear the device during waking hours, except while showering, bathing, and swimming, for seven consecutive days. Compliance was verified by comparing the times of use and non-use recorded in the daily logs with the minute-byminute activity counts automatically recorded by the accelerometer. Participants were instructed to complete GLTEQ at the end of the seven-day period and return the study materials in a postage-paid, preaddressed padded Xpresspost envelope through Canada Post. 2.5. Data analysis All analyses were performed using SPSS version 22.0 (SPSS Inc, Chicago, IL). Descriptive data are presented as median (IQR). The descriptive statistics for the measures were provided per physical activity intensity level, such that (a) light activity as measured by the accelerometer was compared with mild activity as measured by the GLTEQ; (b) moderate activity as measured by the accelerometer was compared with moderate activity from the GLTEQ; and (c) vigorous activity as measured by the accelerometer was compared with strenuous activity from the GLTEQ. Total physical activity levels and MVPA were measured by the accelerometer and GLTEQ. We compared differences in physical activity scores between HCs and POMS using independent samples median MannWhitney U Tests and corresponding estimates of effect sizes as Cliff's dominance (d) (Puyau et al., 2002). Cliff's d ranges from 1 to 1 and indicates the difference in probability that an observation in one sample exceeds an observation in the other sample (e.g., POMS vs. HCs). The relationships among scores representing light, moderate, vigorous, and MVPA across both measures were examined using Spearman rho rank-order correlation in POMS and the overall sample. Light activity as measured by the accelerometer was correlated with mild activity as measured by the GLTEQ. Moderate activity as measured by the accelerometer was correlated with moderate activity from the GLTEQ. Vigorous activity as measured by the accelerometer was correlated with strenuous activity as measured by the GLTEQ. MVPA as measured by the accelerometer was correlated with the health contribution score as measured by the GLTEQ. We further used Cohen's guidelines (Cohen, 1977) of 0.1, 0.3, and 0.5 for judging the magnitude of correlations as small, moderate, and large, respectively in the sample.
3. Results Descriptive data are presented for the samples in Table 1. The POMS sample was 66.7% female, median age of 15.73 (IQR ¼3.2) years, 44.4% white, median MS duration of 2.03 (IQR ¼2) years, median EDSS of 1.5 (IQR ¼ 1). The HCs sample was 66.7% female, median age of 14.76 (IQR ¼3.8) years, and 73.3% white. There was a significant difference in age and race between the two groups as measured by the Mann-Whitney U test (p¼ 0.030 & p¼ 0.027 respectively).
Table 2 Descriptive and inferential statistics along with effect size for the measures in POMS (n¼ 27) and HC (n ¼45). Activity type
Measure POMS
HCs
p-Value Cliff's d
Median IQR
Median IQR
0.14
1.83
0.80
Accel GLTEQ
9.00
27.00
31.50
36.00 0.009
0.364
Moderate
Accel GLTEQ
9.00 10.00
15.31 15.00
15.00 25.00
17.10 0.025 35.00 0.031
0.318 0.302
Light/mild
Accel GLTEQ
91.00 9.00
48.29 88.14 21.00 15.00
47.07 15.75
0.930 0.103
0.012 0.222
Total
Accel GLTEQ
106.33 36.00
60.11 41.00
78.55 0.534 36.50 0.003
0.088 0.416
MVPA/HC
Accel GLTEQ
9.50 30.00
16.00 15.50 46.00 52.00
24.56 0.013 40.75 0.005
0.351 0.393
109.00 65.00
5.04
0.001
0.477
Vigorous/ strenuous
Note. POMS¼ Pediatric Onset Multiple Sclerosis; HCs¼Healthy Controls; SD ¼Standard Deviation; MVPA ¼ Moderate-to-Vigorous Physical Activity; HC¼Health Contribution Score Accel¼Accelerometry; GLTEQ¼ Godin Leisure-Time Exercise Questionnaire. Values for accelerometry are reported as total minutes per day. Values for the GLTEQ are reported as metabolic equivalent (MET) minutes per week. The differences between variables were analyzed using the independent samples median Mann-Whitney U Tests.
differences are consistent with those detected by accelerometry, with the exception of total physical activity. There were no differences between POMS and HC for light activity by accelerometry or the GLTEQ. Cliff's d indicated that we captured comparable magnitudes of difference between accelerometry and the GLTEQ for all levels except total physical activity. 3.2. Construct validity analysis The correlations among scores from the two measures of physical activity are presented below for the overall sample and then the POMS sub-sample. 3.2.1. Overall sample There was a large, positive correlation between accelerometer and GLTEQ scores for vigorous physical activity (r ¼0.479, p¼ 0.001) and a moderate, positive correlation between accelerometer and GLTEQ scores for MVPA (r ¼0.450, p ¼0.001). There was a small, positive correlation between accelerometer and GLTEQ scores for moderate physical activity (r ¼0.281, p ¼0.008). There was no statistically significant correlation between accelerometer and GLTEQ scores for light activity (r ¼0.073, p ¼0.270 respectively). 3.2.2. POMS sub-sample There was a large, positive correlation between accelerometer and GLTEQ scores for vigorous (r ¼0.736, p ¼0.001) and a nearly significant and moderate, positive correlation for MVPA (r ¼0.319, p¼ 0.053). There were no statistically significant correlations between accelerometer and GLTEQ scores for moderate or light activity (r ¼0.151, p ¼0.226 & r ¼ 0.056, p¼ 0.391 respectively).
3.1. Known-groups analysis Table 2 provides the descriptive data for each activity intensity, total physical activity and MVPA from the measures of physical activity. The accelerometry metrics captured the expected significant differences in physical activity levels between POMS and HC for vigorous, moderate, and MVPA. The GLTEQ captured similar statistically significant differences in strenuous, moderate, and MVPA as well as total physical activity, and the magnitude of
4. Discussion and conclusion This study highlights the utility of a straightforward, pencil and paper questionnaire for measuring MVPA in children and adolescents with MS. We have provided novel evidence of score validity for the GLTEQ as a measure of physical activity in POMS. Based on a two-part approach in combination with a nomological net, we
D. Kinnett-Hopkins et al. / Multiple Sclerosis and Related Disorders 10 (2016) 26–29
have provided evidence of (a) expected group differences in vigorous, moderate, and MVPA by way of the GLTEQ between POMS and HCs (i.e., POMS engage in less MVPA than HCs) that were confirmed by accelerometry; (b) a strong, positive correlation between scores from accelerometry and GLTEQ for vigorous activity in POMS; and (c) a nearly statistically significant, but moderate, positive correlation between scores from accelerometry and GLTEQ for MVPA in POMS. This provides evidence that scores from the GLTEQ measures of vigorous and perhaps MVPA can be validly interpreted among individuals with POMS. The GLTEQ has evidence of score validity for measuring total physical activity. Accelerometry did not capture the total physical activity difference and this is likely due to the more inclusive nature of the self-report measure. Accelerometry primarily captures ambulatory physical activity, whereas the GLTEQ is not restricted by ambulation and may better capture different types of physical activity that contribute into total physical activity levels. Such information is important as one considers conducting large scale investigations, investigations with limited resources, or considering tools that can be used in a clinical setting to enter discussions with children and families regarding physical activity levels. Regarding clinical research, the ideal measurement approach would include accelerometry and the GLTEQ, as it can deliver a more complete picture of physical activity in POMS. On the other hand, the GLTEQ provides a practical and relatively quick means for clinicians and other health care providers to screen for physical activity based this straightforward, three-tiered scoring system (active, moderately active, and insufficiently active), thus highlighting its utility in clinical practice. Limitations of this study include (1) data collection in a tertiary care center, possibly leading to selection bias of the population; and (2) significant differences between HC and POMS subjects in age and race. We feel, however, that as this center serves as a regional pediatric and MS center, recruitment likely captured most POMS patients in the region. Future research should include age and race matched HCs. Our results provide evidence that scores from the GLTEQ can be validly interpreted as a measure of physical activity among POMS and can be adopted in clinical practice and by researchers to determine the physical activity profile of POMS patients. Given the high rate of inactivity of children with MS, and known benefits of adequate MVPA in this population, future studies are needed to determine how this scale can be integrated effectively into clinical practice to improve patient outcomes.
Declaration of conflicting interests DKH and SG have no conflicts of interest to disclose. EAY receives funding from the National MS Society, the Canadian Institutes of Health Research, the Dairy Farmers of Ontario, SickKids Foundation, SickKids Innovation Fund, CMSMS/PHAC, the Canadian Multiple Sclerosis Scientific Research Foundation and the MS Society of Canada. RWM has received speaker honoraria from EMD Serono and funding from Biogen Idec, Acorda Therapeutics and Sun Health Technologies.
Funding sources National MS Society pilot study Grant (Grant #: PP2256) and the Mario Batalli Foundation. Author contributions DKH and RWM conceived and wrote the first draft of the
29
manuscript. SG and EAY contributed to data collection. DKH and RWM analyzed the data. All authors contributed in reviewing data and editing of the manuscript.
Acknowledgements We are grateful to the parents, guardians and children at the Hospital for Sick Children for being part of the study.
References Amato, M.P., Goretti, B., Ghezzi, A., et al., 2008. Cognitive and psychosocial features of childhood and juvenile MS. Neurology 70 (20), 1891–1897. Chabas, D., Pelletier, D., 2009. Sorting through the pediatric MS spectrum with brain MRI. Nat. Rev. Neurol. 5 (4), 186–188. Chitnis, T., Glanz, B., Jaffin, S., Healy, B., 2009. Demographics of pediatric-onset multiple sclerosis in an MS center population from the Northeastern United States. Mult. Scler. 15 (5), 627–631. Cohen, J., 1977. Statistical Power Analysis for the Behavioral Sciences. Cronbach, L.J., Meehl, P.E., 1955. Construct validity in psychological tests. Psychol. Bull. 52 (4), 281–302. Crouter, S.E., Horton, M., Bassett Jr., D.R., 2013. Validity of ActiGraph child-specific equations during various physical activities. Med Sci. Sports Exerc. 45 (7), 1403–1409. Gianfrancesco, M.A., Acuna, B., Shen, L., et al., 2014. Obesity during childhood and adolescence increases susceptibility to multiple sclerosis after accounting for established genetic and environmental risk factors. Obes. Res Clin. Pract. 8 (5), e435–447. Godin, G., 2011. The Godin-Shephard leisure-time physical activity questionnaire. Health Fit. J. Can. 4 (1), 18–22. Godin, G., Shephard, R.J., 1985. A simple method to assess exercise behavior in the community. Can. J. Appl. Sport Sci. 10 (3), 141–146. Goretti, B., Portaccio, E., Ghezzi, A., et al., 2012. Fatigue and its relationships with cognitive functioning and depression in paediatric multiple sclerosis. Mult. Scler. 18 (3), 329–334. Gosney, J.L., Scott, J.A., Snook, E.M., Motl, R.W., 2007. Physical activity and multiple sclerosis: validity of self-report and objective measures. Fam. Community Health 30 (2), 144–150. Grover, S.A., Aubert-Broche, B., Fetco, D., et al., 2015. Lower physical activity is associated with higher disease burden in pediatric multiple sclerosis. Neurology 85 (19), 1663–1669. Koo, M.M., Rohan, T.E., 1999. Comparison of four habitual physical activity questionnaires in girls aged 7–15 yr. Med. Sci. Sports Exerc. 31 (3), 421–427. Landy, F.J., 1986. Stamp collecting versus science - validation as hypothesis-testing. Am. Psychol. 41 (11), 1183–1192. Langer-Gould, A., Brara, S.M., Beaber, B.E., Koebnick, C., 2013. Childhood obesity and risk of pediatric multiple sclerosis and clinically isolated syndrome. Neurology 80 (6), 548–552. Motl, R.W., Sandroff, B.M., 2010. Objective monitoring of physical activity behavior in multiple sclerosis. Phys. Ther. Rev. 15.3, 204–211. Motl, R.W., McAuley, E., Snook, E.M., Scott, J.A., 2006. Validity of physical activity measures in ambulatory individuals with multiple sclerosis. Disabil. Rehabil. 28 (18), 1151–1156. Motl, R.W., Learmonth, Y.C., Pilutti, L.A., Gappmaier, E., Coote, S., 2015. Top 10 research questions related to physical activity and multiple sclerosis. Res. Q. Exerc. Sport 86 (2), 117–129. Parrish, J.B., Weinstock-Guttman, B., Smerbeck, A., Benedict, R.H., Yeh, E.A., 2013. Fatigue and depression in children with demyelinating disorders. J. Child Neurol. 28 (6), 713–718. Puyau, M.R., Adolph, A.L., Vohra, F.A., Butte, N.F., 2002. Validation and calibration of physical activity monitors in children. Obes. Res. 10 (3), 150–157. Puyau, M.R., Adolph, A.L., Vohra, F.A., Zakeri, I., Butte, N.F., 2004. Prediction of activity energy expenditure using accelerometers in children. Med. Sci. Sports Exerc. 36 (9), 1625–1631. Renoux, C., Vukusic, S., Mikaeloff, Y., et al., 2007. Natural history of multiple sclerosis with childhood onset. N. Engl. J. Med. 356 (25), 2603–2613. Simone, I.L., Carrara, D., Tortorella, C., et al., 2002. Course and prognosis in earlyonset MS: comparison with adult-onset forms. Neurology 59 (12), 1922–1928. Messick, S., 1995. Validity of psychological assessment: Validity of inferences from persons' responses and performances as scientific inquiry into score meaning. Am Psychol 50 (9), 741–749. Vollmer, T.L., Benedict, R., Bennett, S., et al., 2012. Exercise as prescriptive therapy in multiple sclerosis: a consensus conference white paper. Int. J. MS Care. Yeh, E.A., Chitnis, T., Krupp, L., et al., 2009. Pediatric multiple sclerosis. Nat. Rev. Neurol. 5 (11), 621–631. Yeh, E.A., Kinnett-Hopkins, D., Grover, S.A., Motl, R.W., 2015. Physical activity and pediatric multiple sclerosis: Developing a research agenda. Mult. Scler. 21 (13), 1618–1625.
Contents lists available at ScienceDirect
Multiple Sclerosis and Related Disorders journal homepage: www.elsevier.com/locate/msard
Physical activity in pediatric onset multiple sclerosis: Validating a questionnaire for clinical practice and research Dominique Kinnett-Hopkins a, Stephanie A. Grover b, E. Ann Yeh b,c, Robert W. Motl a,d,n a
Department of Kinesiology and Community Health, University of Illinois at Urbana-Champaign, USA Division of Neurology, Department of Neurosciences and Mental Health, Hospital for Sick Children, Toronto, ON, Canada c Department of Pediatrics, The Faculty of Medicine, University of Toronto, Toronto, ON, Canada d Department of Physical Therapy, University of Alabama at Birmingham, USA b
art ic l e i nf o
a b s t r a c t
Article history: Received 8 June 2016 Received in revised form 17 August 2016 Accepted 19 August 2016
Background: Knowledge regarding physical activity (PA) and its benefits in pediatric onset multiple sclerosis (POMS) is growing and suggests high levels of inactivity. The utility of a validated screening tool for clinical settings is unknown. This study evaluated the Godin Leisure-Time Exercise Questionnaire (GLTEQ) as a measure of PA in POMS. Methods: POMS patients (n¼27) and healthy controls (n¼45) wore an accelerometer over a 7-day period and then completed the GLTEQ. Results: The GLTEQ captured expected group differences in PA for vigorous, moderate, and moderate-tovigorous physical activity (MVPA), confirmed by accelerometry. There was a large, positive correlation between GLTEQ and accelerometry scores for vigorous PA in POMS (r ¼0.736, p¼0.001), and a nearly significant and moderate, positive correlation between MVPA scores (r¼ 0.319, p¼ .053). Conclusion: We provide evidence that supports the validity of GLTEQ scores as measures of vigorous and MVPA in POMS. Researchers and clinicians might adopt this scale for measuring PA. & 2016 Elsevier B.V. All rights reserved.
Keywords: Multiple sclerosis Physical activity Pediatric Measurement Validity
1. Introduction Pediatric onset multiple sclerosis (POMS) constitutes approximately 3–5% of all cases of multiple sclerosis (MS) (Chitnis et al., 2009). POMS is characterized by higher disease burden on magnetic resonance imaging (MRI) (Yeh et al., 2009; Chabas and Pelletier, 2009) and slower time to irreversible motor deficit compared with adult onset MS (Renoux et al., 2007; Simone et al., 2002). Youth with POMS experience cognitive dysfunction (Amato et al., 2008), frequent and severe relapses, high rates of fatigue and depression, and reduced quality of life (QOL) (Simone et al., 2002; Amato et al., 2008; Parrish et al., 2013; Goretti et al., 2012). There is evidence that physical activity participation in adults with MS aids in the management of the manifestations of this disease (Motl et al., 2015) and thus can be promoted through healthcare providers (Vollmer et al., 2012). There is a need for similar evidence in POMS, and this necessitates a validated assessment of physical activity levels for researchers as well as clinicians interested in screening whether or not POMS patients are active enough for n Correspondence to: 1705 University Blvd, SHPB 336, Birmingham, AL 35233, USA. E-mail address: [email protected] (R.W. Motl).
http://dx.doi.org/10.1016/j.msard.2016.08.010 2211-0348/& 2016 Elsevier B.V. All rights reserved.
health benefits. Increasing knowledge suggests that certain lifestyle attributes, including physical activity, may affect outcomes in POMS (Yeh et al., 2015). This interest in lifestyle is based, in part, on observations that high body mass index (BMI) was associated with POMS (Gianfrancesco et al., 2014; Langer-Gould et al., 2013) and strenuous physical activity participation, as measured by the Godin Leisure-Time Exercise Questionnaire (GLTEQ), was associated with outcomes such as lower T2 lesion volumes and relapse rates in POMS (Grover et al., 2015) as well as the evidence of the benefits of physical activity participation in adults with MS (Motl et al., 2015). Nevertheless, much scientific inquiry still needs to be undertaken regarding physical activity and POMS. The continued study of physical activity in POMS has been highlighted in a recent review (Yeh et al., 2015) and requires measures with evidence for the validity of score inferences (Motl et al., 2015). Indeed, the scientific study of physical activity among adults with MS was significantly accelerated based on evidence establishing the construct validity of scores from self-report outcomes such as the GLTEQ (Godin and Shephard, 1985; Motl et al., 2006). Such evidence of validity is necessary in POMS because its manifestations (e.g., cognitive impairment, fatigue, and depression) could influence the validity of self-report measures of physical activity, and
D. Kinnett-Hopkins et al. / Multiple Sclerosis and Related Disorders 10 (2016) 26–29
one cannot simply assume that the evidence for adults with MS (Gosney et al., 2007) or children from the general population (Koo and Rohan, 1999) support applications in POMS. The validity of the GLTEQ in POMS can be established by comparison with objective accelerometry using a two-part approach, namely the known-groups and construct validity paradigms (Motl and Sandroff, 2010). This two-part approach requires (a) comparing mean scores from the GLTEQ and accelerometry for capturing expected differences of vigorous, moderate and MVPA between POMS and healthy controls (HCs) and (b) examining and comparing the magnitude of the correlation between scores from the GLTEQ with accelerometry in combined samples of POMS and HC and POMS separately. Accordingly, evidence of score validity (i.e., the degree to which scores from a measurement can be interpreted as a measure of the intended construct) (Messick, 1995) can be established based on the expectation of capturing (a) lower levels of vigorous, moderate and MVPA in POMS than HC with the GLTEQ, which are confirmed with accelerometry, considering the presence and the degree of disease manifestation such as cognitive impairments, fatigue and depression (Parrish et al., 2013; Goretti et al., 2012) and (b) moderate or strong correlations between scores from the GLTEQ with accelerometer outcomes in the combined samples of POMS and HCs that are confirmed in the POMS sample alone. This study examined the validity of scores from the GLTEQ in POMS patients. We validated the scores using an objective measure of physical activity (i.e., accelerometry) in combination with a nomological net (i.e. hypothesized pattern of differences between groups and correlations among scores) (Cronbach and Meehl, 1955; Landy, 1986). We opted for comparison with accelerometry, as this is a validated and commonly used outcome of physical activity in pediatric populations (Crouter et al., 2013; Puyau et al., 2004) and provides objective metrics for comparison with selfreported behavior. Of note, we have focused on validating scores from the GLTEQ so that this simple and practical measure is amenable for use by clinicians to determine POMS physical activity risk and for inclusion in large, population-based studies of PA in POMS and applications wherein researchers and clinicians do not have resources or experience to include accelerometry.
27
Table 1 Demographic characteristics of POMS and HC. Variable
POMS (N ¼ 27)
HCs (N¼ 45)
p-Value
Statistic
Sex (% female (n)) Age (years) Race (% white (n)) MS Duration (years) EDSS
18/66.7 15.73(3.2) 12/44.4 2.03(2) 1.5 (0.5)
30/66.7 14.76(3.8) 33/73.3
1.000 0.030 0.027
X2 ¼0.000 z ¼ 0.051 X2 ¼6.009
Note. Values are reported as median (IQR) and were analyzed using the MannWhitney U Test (age) or chi-square test (sex and race). POMS ¼ pediatric-onset multiple sclerosis; HCs¼ healthy controls; EDSS ¼Expanded Disability Status Scale.
that measure the frequency (0–7) of strenuous (e.g. jogging), moderate (e.g. fast walking) and mild (e.g. easy walking) exercise for sessions more than 15 min during one's free time in the preceding week (Godin, 2011). The weekly frequencies of strenuous, moderate, and mild activities are multiplied by 9, 5 and 3 metabolic equivalents (METs), respectively, and then summed into a measure of total leisure activity (0–119) in MET/min per week. Importantly, the score of 9 for vigorous/strenuous reflects one, 15minute bout per week, whereas the score of 9 for light/mild reflects three, 15-minute bouts per week. The health contribution score (i.e., equivalent of time spent in MVPA) is calculated from the frequency of only strenuous and moderate activities. The frequencies for strenuous and moderate activities are multiplied by 9 and 5 METs, respectively, and then summed into a health contribution score (0–98) that reflects MET/min per week. The scores are then classified into three categories: active (substantial benefits; 24 or more MET/min per week), moderately active (some benefits; 14–23 MET/min per week), and insufficiently active (less substantial or low benefits; 13 or fewer MET/min per week) (Godin, 2011). 2.3. Accelerometer
The procedures for this study were approved by the Hospital for Sick Children Research Ethics Board, and all participants and legal guardians provided written informed consent and/or assent. We recruited a sample of 72 subjects for this study (POMS¼ 27; HC ¼45). The POMS patients were recruited from the Pediatric MS and Neuroinflammatory Clinic at The Hospital for Sick Children, Toronto, Canada. HCs subjects were recruited via word of mouth and flyers around the hospital campus. POMS participants were included if they were (a) 8–18 years old, (b) had Expanded Disability Status Scale (EDSS) scores o4.0, (c) had not received steroids or experienced a relapse in the last 30 days, and (d) if their MS was considered stable. Participants were excluded if there was a presence of non-specific white matter abnormalities and metabolic or infectious etiologies for white matter abnormalities. HCs were included if they (a) were 8–18 years of age and (b) had no previous neurological problems. The demographic and clinical characteristics of the participants are provided in Table 1.
We used the ActiGraph model 7164, accelerometer (ActiGraph Corporation, Pensacola, Florida) as an objective measure of physical activity based on monitoring over a seven-day period under free-living conditions. The device is a small (2.0 1.6 0.6 in.) and lightweight (1.5 ounces) accelerometer worn in a pouch on a snug elastic belt around the waist on the non-dominant hip. The single, vertical axis piezoelectric bender element within the accelerometer generates an electrical signal proportional to the force acting on it that is converted into activity counts over a pre-specified period of time (epoch) and then stored in random access memory within the device. The epoch in this study was set to be 1 min. The data from the accelerometers were processed using ActiLife software. The data from each participant's accelerometer were processed into 2 separate Microsoft Excel (Microsoft, Redmond, Washington) files representing wear time and time spent participating in differing levels of physical activity indicated by the number of counts per minute (CPM) recorded during each epoch: sedentary (i.e., r799 CPM), light (i.e., 800-3199 CPM), moderate (i.e., 3200-8199 CPM), and vigorous (i.e., Z 8200 CPM) (Puyau et al., 2002). Accelerometer wear time data were checked against participant recorded wear times from the log sheet, and only valid days (Z10 h of wear time without periods of continuous zeros exceeding 60 min indicative of compliance) were included in the analysis. The outcomes for the analyses were average daily minutes for total wear time and minutes of light, moderate and vigorous physical activity per day across valid days of data.
2.2. Instruments
2.4. Procedures
2. Methods 2.1. Participants
GLTEQ. The GLTEQ is comprised of three open-ended questions
Participants underwent a neurologic exam by a neurologist
28
D. Kinnett-Hopkins et al. / Multiple Sclerosis and Related Disorders 10 (2016) 26–29
specializing in MS for generation of an Expanded Disability Status Scale (EDSS) score. We familiarized the participants with the objective monitors and asked them to wear the device during waking hours, except while showering, bathing, and swimming, for seven consecutive days. Compliance was verified by comparing the times of use and non-use recorded in the daily logs with the minute-byminute activity counts automatically recorded by the accelerometer. Participants were instructed to complete GLTEQ at the end of the seven-day period and return the study materials in a postage-paid, preaddressed padded Xpresspost envelope through Canada Post. 2.5. Data analysis All analyses were performed using SPSS version 22.0 (SPSS Inc, Chicago, IL). Descriptive data are presented as median (IQR). The descriptive statistics for the measures were provided per physical activity intensity level, such that (a) light activity as measured by the accelerometer was compared with mild activity as measured by the GLTEQ; (b) moderate activity as measured by the accelerometer was compared with moderate activity from the GLTEQ; and (c) vigorous activity as measured by the accelerometer was compared with strenuous activity from the GLTEQ. Total physical activity levels and MVPA were measured by the accelerometer and GLTEQ. We compared differences in physical activity scores between HCs and POMS using independent samples median MannWhitney U Tests and corresponding estimates of effect sizes as Cliff's dominance (d) (Puyau et al., 2002). Cliff's d ranges from 1 to 1 and indicates the difference in probability that an observation in one sample exceeds an observation in the other sample (e.g., POMS vs. HCs). The relationships among scores representing light, moderate, vigorous, and MVPA across both measures were examined using Spearman rho rank-order correlation in POMS and the overall sample. Light activity as measured by the accelerometer was correlated with mild activity as measured by the GLTEQ. Moderate activity as measured by the accelerometer was correlated with moderate activity from the GLTEQ. Vigorous activity as measured by the accelerometer was correlated with strenuous activity as measured by the GLTEQ. MVPA as measured by the accelerometer was correlated with the health contribution score as measured by the GLTEQ. We further used Cohen's guidelines (Cohen, 1977) of 0.1, 0.3, and 0.5 for judging the magnitude of correlations as small, moderate, and large, respectively in the sample.
3. Results Descriptive data are presented for the samples in Table 1. The POMS sample was 66.7% female, median age of 15.73 (IQR ¼3.2) years, 44.4% white, median MS duration of 2.03 (IQR ¼2) years, median EDSS of 1.5 (IQR ¼ 1). The HCs sample was 66.7% female, median age of 14.76 (IQR ¼3.8) years, and 73.3% white. There was a significant difference in age and race between the two groups as measured by the Mann-Whitney U test (p¼ 0.030 & p¼ 0.027 respectively).
Table 2 Descriptive and inferential statistics along with effect size for the measures in POMS (n¼ 27) and HC (n ¼45). Activity type
Measure POMS
HCs
p-Value Cliff's d
Median IQR
Median IQR
0.14
1.83
0.80
Accel GLTEQ
9.00
27.00
31.50
36.00 0.009
0.364
Moderate
Accel GLTEQ
9.00 10.00
15.31 15.00
15.00 25.00
17.10 0.025 35.00 0.031
0.318 0.302
Light/mild
Accel GLTEQ
91.00 9.00
48.29 88.14 21.00 15.00
47.07 15.75
0.930 0.103
0.012 0.222
Total
Accel GLTEQ
106.33 36.00
60.11 41.00
78.55 0.534 36.50 0.003
0.088 0.416
MVPA/HC
Accel GLTEQ
9.50 30.00
16.00 15.50 46.00 52.00
24.56 0.013 40.75 0.005
0.351 0.393
109.00 65.00
5.04
0.001
0.477
Vigorous/ strenuous
Note. POMS¼ Pediatric Onset Multiple Sclerosis; HCs¼Healthy Controls; SD ¼Standard Deviation; MVPA ¼ Moderate-to-Vigorous Physical Activity; HC¼Health Contribution Score Accel¼Accelerometry; GLTEQ¼ Godin Leisure-Time Exercise Questionnaire. Values for accelerometry are reported as total minutes per day. Values for the GLTEQ are reported as metabolic equivalent (MET) minutes per week. The differences between variables were analyzed using the independent samples median Mann-Whitney U Tests.
differences are consistent with those detected by accelerometry, with the exception of total physical activity. There were no differences between POMS and HC for light activity by accelerometry or the GLTEQ. Cliff's d indicated that we captured comparable magnitudes of difference between accelerometry and the GLTEQ for all levels except total physical activity. 3.2. Construct validity analysis The correlations among scores from the two measures of physical activity are presented below for the overall sample and then the POMS sub-sample. 3.2.1. Overall sample There was a large, positive correlation between accelerometer and GLTEQ scores for vigorous physical activity (r ¼0.479, p¼ 0.001) and a moderate, positive correlation between accelerometer and GLTEQ scores for MVPA (r ¼0.450, p ¼0.001). There was a small, positive correlation between accelerometer and GLTEQ scores for moderate physical activity (r ¼0.281, p ¼0.008). There was no statistically significant correlation between accelerometer and GLTEQ scores for light activity (r ¼0.073, p ¼0.270 respectively). 3.2.2. POMS sub-sample There was a large, positive correlation between accelerometer and GLTEQ scores for vigorous (r ¼0.736, p ¼0.001) and a nearly significant and moderate, positive correlation for MVPA (r ¼0.319, p¼ 0.053). There were no statistically significant correlations between accelerometer and GLTEQ scores for moderate or light activity (r ¼0.151, p ¼0.226 & r ¼ 0.056, p¼ 0.391 respectively).
3.1. Known-groups analysis Table 2 provides the descriptive data for each activity intensity, total physical activity and MVPA from the measures of physical activity. The accelerometry metrics captured the expected significant differences in physical activity levels between POMS and HC for vigorous, moderate, and MVPA. The GLTEQ captured similar statistically significant differences in strenuous, moderate, and MVPA as well as total physical activity, and the magnitude of
4. Discussion and conclusion This study highlights the utility of a straightforward, pencil and paper questionnaire for measuring MVPA in children and adolescents with MS. We have provided novel evidence of score validity for the GLTEQ as a measure of physical activity in POMS. Based on a two-part approach in combination with a nomological net, we
D. Kinnett-Hopkins et al. / Multiple Sclerosis and Related Disorders 10 (2016) 26–29
have provided evidence of (a) expected group differences in vigorous, moderate, and MVPA by way of the GLTEQ between POMS and HCs (i.e., POMS engage in less MVPA than HCs) that were confirmed by accelerometry; (b) a strong, positive correlation between scores from accelerometry and GLTEQ for vigorous activity in POMS; and (c) a nearly statistically significant, but moderate, positive correlation between scores from accelerometry and GLTEQ for MVPA in POMS. This provides evidence that scores from the GLTEQ measures of vigorous and perhaps MVPA can be validly interpreted among individuals with POMS. The GLTEQ has evidence of score validity for measuring total physical activity. Accelerometry did not capture the total physical activity difference and this is likely due to the more inclusive nature of the self-report measure. Accelerometry primarily captures ambulatory physical activity, whereas the GLTEQ is not restricted by ambulation and may better capture different types of physical activity that contribute into total physical activity levels. Such information is important as one considers conducting large scale investigations, investigations with limited resources, or considering tools that can be used in a clinical setting to enter discussions with children and families regarding physical activity levels. Regarding clinical research, the ideal measurement approach would include accelerometry and the GLTEQ, as it can deliver a more complete picture of physical activity in POMS. On the other hand, the GLTEQ provides a practical and relatively quick means for clinicians and other health care providers to screen for physical activity based this straightforward, three-tiered scoring system (active, moderately active, and insufficiently active), thus highlighting its utility in clinical practice. Limitations of this study include (1) data collection in a tertiary care center, possibly leading to selection bias of the population; and (2) significant differences between HC and POMS subjects in age and race. We feel, however, that as this center serves as a regional pediatric and MS center, recruitment likely captured most POMS patients in the region. Future research should include age and race matched HCs. Our results provide evidence that scores from the GLTEQ can be validly interpreted as a measure of physical activity among POMS and can be adopted in clinical practice and by researchers to determine the physical activity profile of POMS patients. Given the high rate of inactivity of children with MS, and known benefits of adequate MVPA in this population, future studies are needed to determine how this scale can be integrated effectively into clinical practice to improve patient outcomes.
Declaration of conflicting interests DKH and SG have no conflicts of interest to disclose. EAY receives funding from the National MS Society, the Canadian Institutes of Health Research, the Dairy Farmers of Ontario, SickKids Foundation, SickKids Innovation Fund, CMSMS/PHAC, the Canadian Multiple Sclerosis Scientific Research Foundation and the MS Society of Canada. RWM has received speaker honoraria from EMD Serono and funding from Biogen Idec, Acorda Therapeutics and Sun Health Technologies.
Funding sources National MS Society pilot study Grant (Grant #: PP2256) and the Mario Batalli Foundation. Author contributions DKH and RWM conceived and wrote the first draft of the
29
manuscript. SG and EAY contributed to data collection. DKH and RWM analyzed the data. All authors contributed in reviewing data and editing of the manuscript.
Acknowledgements We are grateful to the parents, guardians and children at the Hospital for Sick Children for being part of the study.
References Amato, M.P., Goretti, B., Ghezzi, A., et al., 2008. Cognitive and psychosocial features of childhood and juvenile MS. Neurology 70 (20), 1891–1897. Chabas, D., Pelletier, D., 2009. Sorting through the pediatric MS spectrum with brain MRI. Nat. Rev. Neurol. 5 (4), 186–188. Chitnis, T., Glanz, B., Jaffin, S., Healy, B., 2009. Demographics of pediatric-onset multiple sclerosis in an MS center population from the Northeastern United States. Mult. Scler. 15 (5), 627–631. Cohen, J., 1977. Statistical Power Analysis for the Behavioral Sciences. Cronbach, L.J., Meehl, P.E., 1955. Construct validity in psychological tests. Psychol. Bull. 52 (4), 281–302. Crouter, S.E., Horton, M., Bassett Jr., D.R., 2013. Validity of ActiGraph child-specific equations during various physical activities. Med Sci. Sports Exerc. 45 (7), 1403–1409. Gianfrancesco, M.A., Acuna, B., Shen, L., et al., 2014. Obesity during childhood and adolescence increases susceptibility to multiple sclerosis after accounting for established genetic and environmental risk factors. Obes. Res Clin. Pract. 8 (5), e435–447. Godin, G., 2011. The Godin-Shephard leisure-time physical activity questionnaire. Health Fit. J. Can. 4 (1), 18–22. Godin, G., Shephard, R.J., 1985. A simple method to assess exercise behavior in the community. Can. J. Appl. Sport Sci. 10 (3), 141–146. Goretti, B., Portaccio, E., Ghezzi, A., et al., 2012. Fatigue and its relationships with cognitive functioning and depression in paediatric multiple sclerosis. Mult. Scler. 18 (3), 329–334. Gosney, J.L., Scott, J.A., Snook, E.M., Motl, R.W., 2007. Physical activity and multiple sclerosis: validity of self-report and objective measures. Fam. Community Health 30 (2), 144–150. Grover, S.A., Aubert-Broche, B., Fetco, D., et al., 2015. Lower physical activity is associated with higher disease burden in pediatric multiple sclerosis. Neurology 85 (19), 1663–1669. Koo, M.M., Rohan, T.E., 1999. Comparison of four habitual physical activity questionnaires in girls aged 7–15 yr. Med. Sci. Sports Exerc. 31 (3), 421–427. Landy, F.J., 1986. Stamp collecting versus science - validation as hypothesis-testing. Am. Psychol. 41 (11), 1183–1192. Langer-Gould, A., Brara, S.M., Beaber, B.E., Koebnick, C., 2013. Childhood obesity and risk of pediatric multiple sclerosis and clinically isolated syndrome. Neurology 80 (6), 548–552. Motl, R.W., Sandroff, B.M., 2010. Objective monitoring of physical activity behavior in multiple sclerosis. Phys. Ther. Rev. 15.3, 204–211. Motl, R.W., McAuley, E., Snook, E.M., Scott, J.A., 2006. Validity of physical activity measures in ambulatory individuals with multiple sclerosis. Disabil. Rehabil. 28 (18), 1151–1156. Motl, R.W., Learmonth, Y.C., Pilutti, L.A., Gappmaier, E., Coote, S., 2015. Top 10 research questions related to physical activity and multiple sclerosis. Res. Q. Exerc. Sport 86 (2), 117–129. Parrish, J.B., Weinstock-Guttman, B., Smerbeck, A., Benedict, R.H., Yeh, E.A., 2013. Fatigue and depression in children with demyelinating disorders. J. Child Neurol. 28 (6), 713–718. Puyau, M.R., Adolph, A.L., Vohra, F.A., Butte, N.F., 2002. Validation and calibration of physical activity monitors in children. Obes. Res. 10 (3), 150–157. Puyau, M.R., Adolph, A.L., Vohra, F.A., Zakeri, I., Butte, N.F., 2004. Prediction of activity energy expenditure using accelerometers in children. Med. Sci. Sports Exerc. 36 (9), 1625–1631. Renoux, C., Vukusic, S., Mikaeloff, Y., et al., 2007. Natural history of multiple sclerosis with childhood onset. N. Engl. J. Med. 356 (25), 2603–2613. Simone, I.L., Carrara, D., Tortorella, C., et al., 2002. Course and prognosis in earlyonset MS: comparison with adult-onset forms. Neurology 59 (12), 1922–1928. Messick, S., 1995. Validity of psychological assessment: Validity of inferences from persons' responses and performances as scientific inquiry into score meaning. Am Psychol 50 (9), 741–749. Vollmer, T.L., Benedict, R., Bennett, S., et al., 2012. Exercise as prescriptive therapy in multiple sclerosis: a consensus conference white paper. Int. J. MS Care. Yeh, E.A., Chitnis, T., Krupp, L., et al., 2009. Pediatric multiple sclerosis. Nat. Rev. Neurol. 5 (11), 621–631. Yeh, E.A., Kinnett-Hopkins, D., Grover, S.A., Motl, R.W., 2015. Physical activity and pediatric multiple sclerosis: Developing a research agenda. Mult. Scler. 21 (13), 1618–1625.