- Email: [email protected]
Ambulant adults with spastic cerebral palsy: The validity of lower limb joint angle measurements from sagittal video recordings
Gait & Posture 35 (2012) 186–191
Contents lists available at SciVerse ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Ambulant adults with spastic cerebral palsy: The validity of lower limb joint angle measurements from sagittal video recordings Kerstin L. Larsen a,*, Grethe Maanum a,b, Kathrine F. Frøslie a,c, Reidun Jahnsen a,d a
Sunnaas Rehabilitation Hospital, Nesoddtangen, Norway Faculty of Medicine, University of Oslo, Norway c Department of Biostatistics, University of Oslo, Norway d Oslo University Hospital, Department for Child Neurology, Rikshospitalet, Oslo, Norway b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 18 February 2011 Received in revised form 30 August 2011 Accepted 4 September 2011
Background: In the development of a clinical program for ambulant adults with cerebral palsy (CP), we investigated the validity of joint angles measured from sagittal video recordings and explored if movements in the transversal plane identified with three-dimensional gait analysis (3DGA) affected the validity of sagittal video joint angle measurements. Methods: Ten observers, and 10 persons with spastic CP (19–63 years), Gross Motor Function Classification System I–II, participated in the study. Concurrent criterion validity between video joint angle measurements and 3DGA was assessed by Bland–Altman plots with mean differences and 95% limits of agreement (LoA). Pearson’s correlation coefficients (r) and scatter plots were used supplementary. Transversal kinematics 2 SD from our reference band were defined as increased movement in the transversal plane. Results: The overall mean differences in degrees between joint angles measured by 3DGA and video recordings (38, 58 and 78 for the hip, knee and ankle respectively) and corresponding LoA (188, 108 and 158 for the hip, knee and ankle, respectively) demonstrated substantial discrepancies between the two methods. The correlations ranged from low (r = 0.39) to moderate (r = 0.68). Discrepancy between the two measurements was seen both among persons with and without the presence of deviating transversal kinematics. Conclusion: Quantifying lower limb joint angles from sagittal video recordings in ambulant adults with spastic CP demonstrated low validity, and should be conducted with caution. This gives implications for selecting evaluation method of gait. ß 2011 Elsevier B.V. All rights reserved.
Keywords: Gait Observational Validity Cerebral palsy Rehabilitation
1. Introduction Cerebral palsy (CP) is a disorder of motor control and posture resulting from a non-progressive lesion of the developing brain [1]. As gait impairment is a common functional limitation in persons with CP [2], several observational gait analysis tools have been developed [3–9]. Young adults with CP frequently experience declined walking ability [10]. As they may benefit from rehabilitation programs to restore function and/or prevent further deterioration, it is essential to have valid clinical tools that describe their gait systematically, and are cost-effective and easy to use.
* Corresponding author at: Sunnaas Rehabilitation Hospital, Clinic for NeuroRehabilitation, 1450 Nesoddtangen, Norway. Tel.: +47 66 96 90 00; fax: +47 66 91 25 76. E-mail addresses: [email protected], [email protected] (K.L. Larsen). 0966-6362/$ – see front matter ß 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2011.09.004
The validity of a measurement tool describes whether it is appropriate and useful and is divided into face-, construct-, content- and criterion validity [11]. Criterion validity describes systematic equivalence between two similar tools. In gait analysis this is best achieved by comparing the actual gait assessment tool to the criterion standard of gait, which is regarded to be threedimensional gait analysis (3DGA) [2,12]. 3DGA is the most advanced and specific gait analysis tool used in the specialised multidisciplinary follow-up program of ambulant children with CP [13]. However, it includes expensive technology and time consuming analyses, resulting in restricted access. Therefore alternative video-based gait assessment tools (VGA) have been developed. Most of them originate from the need for outcome measurements in clinical practice and in research and are based on categorisation of different gait strategies into ordinal scales from experts in the field of gait impairment in children with spastic CP [3–8]. Several studies, investigating the effects of botulinum toxin A (BoNT-A) on gait in children with spastic CP, have used a VGA as outcome [3,4,14,15]. Ubhi et al. reported both
K.L. Larsen et al. / Gait & Posture 35 (2012) 186–191
reliability, clinically relevant change, as well as the limitations of VGA compared with 3DGA [4]. However, other BoNT-A studies do not discuss the limitations of assessing gait with VGA [3,12,14,15]. In contrast to the categorical approach in most VGA-tools, the Salford Gait Tool (SF-GT) [9] is based on visually estimation of exact joint angles in the hip, knee and ankle at six selected events during the gait cycle, and then categorisation of them. The validity and reliability or SF-GT in children with CP has been established by comparing SF-GT with sagittal 3DGA kinematics [9,16]. Another study using a quantitative approach [17], compares measurements of six selected joint angles on a video screen using software program with simultaneous 3DGA sagittal kinematics. To our knowledge, no study has investigated the validity of measuring joint angles from a video screen in adults with CP. Due to the given exact joint angles, which creates transparency with regard to validity the SF-GT scoring sheet was chosen in this study [9]. Further, the literature on two-dimensional VGA tools is limited with regards to presenting simultaneous data on movements in the sagittal and transverse plane during gait. Thus, in the development of a clinical program for ambulant adults with CP and decreased walking ability, we investigated concurrent criterion validity by comparing joint angles measured from sagittal video recordings and sagittal plane kinematics from 3DGA. We also explored if movement deviations in the transverse plane identified with 3DGA affected the validity of sagittal video joint angle measurements. 2. Methods 2.1. Ethics The study was approved by the Data protection supervisor and the Regional Committee for Research Ethics. All the participants and observers gave written consent. 2.2. Participants A subset of 10 participants was randomly drawn from a larger study population of well functioning adults with spastic CP recruited through advertisements in newspapers and on websites [18]. The participants were five women and five men, age range 19–63 years, five with unilateral- and five with bilateral CP. They all
187
walked independently, four being classified at level I and six at level II of the Gross Motor Function Classification System (GMFCS) [19,20]. Ten observers out of about 40 staff members were recruited through information on e-mail, and nine physiotherapists and one certified prosthetic/orthotist volunteered. Their clinical experience ranged from one up to 40 years. Five observers had experience with 3DGA. Before the assessments the observers attended a lecture on gait, gait parameters, and gait assessment methods, in addition to instruction in the use of SF-GT and goniometry measurement on a screen. One of the authors (RJ) participated as an observer. For the purpose of this study, the observers assessed the video recordings individually once. 2.3. Procedure 3DGA data were collected in a Motion Analysis Laboratory with six infrared cameras (MX13, 100 Hz, VICON Motion System, Oxford, England), two force plates (AMTI OR 6-7, Advanced Mechanical Technology Inc, Watertown, USA), and two digital video cameras (25 Hz). Fifteen reflective markers (14 mm diameter) were attached to the pelvis and lower limbs according to the Plug-In-Gait model (VICON). The gait data were collected by two of the authors (GM and KLL), both having expert knowledge of the underlying biomechanical model. The sagittal video recordings of the 10 participants were copied to CDs. The observers played the CD using PCs with 17 thumbs flat screens. They measured joint angles of the hip, knee, and ankle from the right limb in absolute degrees at six events during a gait cycle; initial contact (IC), end double support (EDS), mid stance (MST), start double support (SDS), toe off (TO) and, mid swing (MSW) according to a modified SF-GT sheet (Appendix A). It was possible to play the video recordings back and forth one frame at a time, and to ‘‘freeze’’ the picture when performing the measurements. A 30 cm long goniometer with each degree marked was used. The observers worked independently with no collaboration with each other. The 3DGA data were processed in Workstation (VICON) by the first author. The joint angles, as described in SF-GT, were extracted to Excel from Workstation by the use of Pipeline External Communication Server (PECS, VICON, Oxford, England), except for the joint angles in MST and MSW that were transferred manually. Further data management was performed in SPSS version 15.0. The transverse plane kinematics for pelvis, hip and foot-progression were exported to Polygon for graphical curves together with the Motion Analysis Laboratory’s own reference material of healthy adults (n = 48) [21]. 2.4. Analysis Concurrent criterion validity was assessed by comparing the joint angles from the same gait cycle measured by SF-GT and sagittal plane kinematics from 3DGA. Agreement between actual measurements in degrees was calculated using Bland– Altman plots with mean differences and 95% limits of agreement (LoA), i.e. 2 standard deviations of the differences (2 SDdiff) [22]. Scatter plots with Pearson’s r were used supplementary, and the magnitudes of the r were evaluated as little
Fig. 1. Kinematic data from 3DGA for all joints and events showing the variability in gait pattern for all participants. The measurements were done at events defined by the SFGT tool, and interpolation lines were added to the chart.
K.L. Larsen et al. / Gait & Posture 35 (2012) 186–191
188
Fig. 2. Data from SF-GT for all joints and events defined in the SF-GT tool, showing the variability in assessment of 10 participants, each evaluated by 10 different observers.
(0–0.25), low (0.26–0.49), moderate (0.50–0.69), high (0.70–0.89) and very high (0.90–1.0) [11]. To explore if increased movements in transverse plane could explain differences between measurements in 3DGA and SF-GT, the transverse plane 3DGA curves for pelvis, hip and foot-progression were visually compared with scatter plots of IC knee joint angle from the SF-GT measurements and the sagittal plane 3DGA.
3. Results The 10 participants had varying gait strategies as demonstrated by 3DGA kinematics (Fig. 1). There was one missing data from the SF-GT measurements (ankle in MSW), resulting in 1799 joint angles. Totally 180 measurements were sampled from the 3DGA. The SF-GT measurements from the 10 observers are summarised in Fig. 2. For each of the 18 different joint angles displayed in Figs. 1 and 2, 10 Bland–Altman plots and 10 scatter plots were made, and mean differences, LoA and Pearson’s correlation coefficients were calculated, one for each of the observers’ concordance with 3DGA. A summary of these numbers are shown in Table 1. The best agreement between SF-GT and 3DGA was found for the knee joint, with mean differences for the different events ranging from 08 to 178 and LoA in the range 8–158. Clinically relevant examples of the ranges of individual observer’s agreements are mean difference
(LoA) of 18 (108) and 138 (118) for knee IC, and 28 (68) and 118 (118) for knee MST. The individual results revealed no differences between observers from the gait-laboratory and observers from the clinic. The overall mean correlations for the hip, knee and ankle were 0.39, 0.68, and 0.39 respectively (Table 1). Fig. 3 shows the visually estimated joint angles of IC for the knee for all the 10 participants against the simultaneous 3DGA joint angles, as well as transverse plane curves for one gait cycle. Discrepancy between the 3DGA joint angles and the SFGT measurements was seen both with and without deviations from our reference band in transverse kinematics. There were no differences between the observer groups (i.e. gait laboratory and clinic) (Fig. 3). 4. Discussion The purpose of this study was to investigate the validity of measuring lower limb joint angles on a video-screen in 10 ambulant adults with spastic CP. Concurrent criterion validity was evaluated by comparing joint angles registered with SF-GT and sagittal plane kinematics from 3DGA, which is considered to be the criterion standard of gait analysis. We also explored if increased
Table 1 Descriptive statistics for the agreement and correlation between 3DGA and SF-GT measurements. Each cell in the table is based on 10 sets of paired observations; one for each rater and 3DGA. The mean differences and 95% limits of agreement (LoA) (2 SDdiff) are the means based on the 10 Bland–Altman plots constructed from the paired observations sets. The r-mean are the mean of the corresponding correlation coefficients. Knee
Hip
IC EDS MST SDS TO MSW Overall mean
Ankle
Mean difference
LoA
r [range]
Mean difference
LoA
r [range]
3 5 4 6 11 3
24 20 14 13 16 19
0.26 0.43 0.52 0.63 0.28 0.24
3 0 5 5 17 5
10 9 9 8 15 12
0.84 0.39 0.65 0.77 0.74 0.69
3
18
0.39
5
10
0.68
[ 0.07 to 0.65] [0.10 to 0.80] [0.36 to 0.82] [0.39 to 0.80] [ 0.11 to 0.63] [ 0.12 to 0.46]
[0.69 [0.03 [0.41 [0.19 [0.20 [0.51
Mean difference to to to to to to
0.92] 0.77] 0.76] 0.90] 0.87] 0.84]
LoA
r [range]
1 6 7 14 6 6
16 12 12 13 18 17
0.52 0.19 0.42 0.64 0.27 0.31
7
15
0.39
IC: initial contact, EDS: end double support, MST: mid stance, SDS: start double support, TO: toe off, and MSW: mid swing.
[0.14 to 0.82] [ 0.18 to 0.50] [0.12 to 0.65] [0.36 to 0.88] [0.00 to 0.57] [ 0.03 to 0.50]
K.L. Larsen et al. / Gait & Posture 35 (2012) 186–191
189
Fig. 3. Scatter plots of IC knee joint angle from the SF-GT measurements and the sagittal plane 3DGA, and the transversal plane 3DGA curves for pelvis, hip and foot-progression. Since some of the observers have identical measures, there are not eleven data points for knee angle at IC for each participant. There are no missing data for the knee.
movements in the transverse plane affected the validity of joint angles measured on a video recording. Analysis of mean differences and corresponding LoA demonstrated substantial discrepancies and lack of agreement between the two methods. We did not find increasing discrepancy between video recordings and 3DGA when increase movements in transverse plane were present. Our study demonstrated that the measured knee joint angles achieved the highest agreement with the concurrent 3DGA results. Toro et al. [9], using the least significant difference (LSD) statistics, reported a fairly high mean difference between SF-GT and 3DGA for all joints (168) and a wide range (2–638). Even if 80–81% of the mean measurements were within the range of 168; this indicates uncertainty about the validity. A recent study [17] investigated the agreement between joint angles measured with 3DGA and computer-based video screen measurement (VSM). Despite the use of a VSM, a wide range between the measurements was found. This study [17], using one investigator, demonstrated highest agreement for peak knee flexion in swing (mean difference (LoA) = 1.48 (118)), and the lowest agreement for peak ankle dorsiflexion in swing (mean difference (LoA) = 128 (14.68)). These results are similar to our study, where 10 investigators evaluated the joint angles by using a plastic goniometer on a screen. We expected that the use of a goniometer would strengthen the focus on the joints being measured [23]. However, the similar results, despite different measuring methods, support the findings of poor validity of measuring joint angles from video recordings, indicating that issues other than the measurement methods are important limitations when estimating joint angles on a video-screen.
The observers had different levels of experience with gait analysis and the use of goniometry. Since the same gait cycles were compared, the reflex markers used in 3DGA were also visible on the video recordings. Consequently, the observers with expert knowledge of 3DGA potentially could have an advantage when estimating joint angles. Despite this, the overall results revealed no differences between observers from the gait-laboratory and observers from the clinic. Grunt [17] concludes that VSM should be used with caution and that rotational deficit could cause problems in the interpretation. However, Grunt did not present data to support this, and to our knowledge no previous studies on VGA tools have explored this issue. Our study failed to demonstrate that increased movements in the transverse plane kinematics resulted in a higher discrepancy in the joint angles measurements from video recordings. Though, this may be due to the limited sample. As numerous interventions are directed towards improving gait strategies for persons with CP, it is appealing to use joint angles as outcome, when you are restricted to use video instead of 3DGA. With the introduction of a VSM, this becomes even more attractive. However, the use of a goniometer, even if computer-based, has its well known limitations [24]. Thus, visual analysis, by ‘‘naked eye’’ or video, will always be subjective, having qualitative strengths and quantitative limitations. When estimating joint angles in degrees from a video-screen, difficulties arise because the limb may not be viewed from the correct angle [12,13,25]. When a person demonstrates a pathological gait strategy in the transverse plane, this will give a false image when estimating joint angles by the use of a sagittal plane video recording.
190
K.L. Larsen et al. / Gait & Posture 35 (2012) 186–191
Several studies have reported the reliability of VGA to be satisfactory [4–8,16]. However, reliability does not guarantee validity, and responsiveness includes them both [11]. Neither Scholtes et al. [14], who used joint angles measured from video recordings and the Edinburgh Visual Gait Score (EVGS) [5], nor Kim et al. [15], who used the Physician’s Rating Scale [3], discuss the limitations of using VGA with regard to what they consider as a clinically important improvement. On the other hand a recent publication, which used VGA to evaluate the effects of selective dorsal rhizotomy on gait in persons with CP, addresses the limitations of measuring gait in only two dimensions [26]. After exploring the issue of validity in measuring complex movements, such as gait in persons with CP in only two dimensions, we suggest that VGAs are considered as subjective qualitative measures. The EVGS, considered the leading VGA tool, is assessing joint motion in sagittal and coronal plane using an ordinal scale, representing overall gait deviation in CP [5,13]. The qualitative strength of this instrument is shown in the studies by Grunt et al. and Hillman et al. [26,27]. Thus, VGAs have its strengths in respect of giving an overall descriptive picture of gait [5,13], as well as being easy to use [13,26]. The correlation between EVGS and Gillette Gait Index reflects the qualitative strength of EVGS in giving a score for overall gait strategy using an ordinal scale [27]. To improve accuracy of VGA techniques ordinal scales are recommended, developed out of needed detail level and known common gait strategies for the population of interest. For clinical as well as research purposes, we suggest that quantitative continuous data obtained from measures of basic gait parameters, such as cadence, stride length and speed replace the ongoing practice on measuring joint angles on a two dimensional screen. Continuous gait parameters could be achieved by simple methods, such as using a stopwatch with a marked walkway and powder on the feet [12], or more technical instruments, such as portable walkways, for example GAITRite [28]. In addition functional tests, such as test of functional mobility (the Timed Up and Go test) [29] and functional walking capacity (6-minute walk test) [30], have recently been suggested as appropriate outcomes reflecting activities in daily life, for adults with spastic CP, GMFCS levels I–II [18]. The small sample size of the present work is the major limitation. However, our conclusions are based on means and SDs of observed differences. Such numbers are adequately estimated in small sample sizes, provided that the data are normally distributed. Another limitation in our study was that cameras with 25 frames per second were used. This resulted in blurry pictures, especially for the ankle in swing, due to high speed of the movement of this body segment. The observers therefore might have chosen different frames for the gait events measured so contributed to the discrepancy of measurements. Higher picture frequency would probably reduce this problem. A disagreement with the 3DGA does not necessarily mean that the VGA is the only inaccurate measurement method. The validity of 3DGA can be affected by imprecise marker placement, imprecise anthropometric measurements, soft tissue movements and shortcomings of the Plug-In-Gait marker protocol with regard to the joint angles [25]. The staff members collecting the 3DGA data were specially trained and a standardised protocol was used. Thus, we assume precise marker placement and anthropometric measurements, resulting in the same quality of data as from comparable gait laboratories [31]. However, both soft tissue movements and the shortcomings of the foot model remains. 5. Conclusion This study demonstrates the shortcomings of quantifying joint angles from a video recording, giving implications when selecting evaluation methods of gait. Since VGA is more available
than 3DGA in both clinical programs and research settings, we suggest that a qualitative approach is preferable for the videoanalysis. Other measurements, such as spatiotemporal parameters and functional tests on mobility, could generate continuous quantitative data. Acknowledgements This project has been financed by EXTRA funds from the Norwegian Foundation for Health and Rehabilitation, the East Regional Health Administration and Sunnaas Rehabilitation Hospital. We are grateful to all the adults with CP and observers who participated, and we acknowledge the contribution of the Norwegian CP Association. Conflict of interest statement The authors have no conflict of interest to disclose.
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.gaitpost.2011.09.004. References [1] Bax M, Goldstein M, Rosenbaum P, Leviton A, Paneth N, Dan B, et al. Proposed definition and classification of cerebral palsy. April 2005. Dev Med Child Neurol 2005;47:571–6. [2] Gage JR, Schwartz MH, Koop SE, Novacheck TF. Identification and treatment of gait problems in cerebral palsy. Clinics in Developmental Medicine Nos 180181, 2nd edn, London: Mac Keith Press; 2009. [3] Koman LA, Mooney III JF, Smith BP, Goodman A, Mulvaney T. Management of spasticity in cerebral palsy with botulinum—a toxin: report of a preliminary, randomized, double-blind trial. J Pediatr Orthop 1994;14:299–303. [4] Ubhi T, Bhakta BB, Ives HL, Allgar V, Roussounis SH. Randomised double blind placebo controlled trial of the effect of botulinum toxin on walking in cerebral palsy. Arch Dis Child 2000;83:481–7. [5] Read HS, Hazlewood ME, Hillman SJ, Prescott RJ, Robb JE. Edinburgh visual gait score for use in cerebral palsy. J Pediatr Orthop 2003;23:296–301. [6] Mackey AH, Lobb GL, Walt SE, Stott NS. Reliability and validity of the Observational Gait Scale in children with spastic diplegia. Dev Med Child Neurol 2003;45:4–11. [7] Wren TA, Rethlefsen SA, Healy BS, Do KP, Dennis SW, Kay RM. Reliability and validity of visual assessments of gait using a modified physician rating scale for crouch and foot contact. J Pediatr Orthop 2005;25:646–50. [8] Dickens WE, Smith MF. Validation of a visual gait assessment scale for children with hemiplegic cerebral palsy. Gait Posture 2006;23:78–82. [9] Toro B, Nester CJ, Farren PC. The development and validity of the Salford Gait Tool: an observation-based clinical gait assessment tool. Arch Phys Med Rehabil 2007;88:321–7. [10] Opheim A, Jahnsen R, Olsson E, Stanghelle JK. Walking function, pain, and fatigue in adults with cerebral palsy: a 7-year follow-up study. Dev Med Child Neurol 2009;51:381–8. [11] Domholdt E. Rehabilitation research: principles and applications. St. Louis, MS: Elsevier Saunders; 2005. [12] Rose SA, Ounpuu S, DeLuca PA. Strategies for the assessment of pediatric gait in the clinical setting. Phys Ther 1991;71:961–80. [13] Harvey A, Gorter JW. Video gait analysis for ambulatory children with cerebral palsy: why, when, where and how! Gait Posture 2010. [14] Scholtes VA, Dallmeijer AJ, Knol DL, Speth LA, Maathuis CG, Jongerius PH, et al. Effect of multilevel botulinum toxin a and comprehensive rehabilitation on gait in cerebral palsy. Pediatr Neurol 2007;36:30–9. [15] Kim K, Shin HI, Kwon BS, Kim SJ, Jung IY, Bang MS. Neuronox versus Botox for spastic equinus gait in children with cerebral palsy: a randomized, doubleblinded, controlled multicentre clinical trial. Dev Med Child Neurol 2011;53: 239–44. [16] Toro B, Nester CJ, Farren PC. Inter- and intraobserver repeatability of the Salford Gait Tool: an observation-based clinical gait assessment tool. Arch Phys Med Rehabil 2007;88:328–32. [17] Grunt S, van Kampen PJ, van der Krogt MM, Brehm MA, Doorenbosch CA, Becher JG. Reproducibility and validity of video screen measurements of gait in children with spastic cerebral palsy. Gait Posture 2010;31:489–94. [18] Maanum G, Jahnsen R, Stanghelle JK, Sandvik L, Keller A. Effects of botulinum toxin A in ambulant adults with spastic cerebral palsy: a randomized doubleblind placebo controlled-trial. J Rehabil Med 2011;43:338–47.
K.L. Larsen et al. / Gait & Posture 35 (2012) 186–191 [19] Palisano R, Rosenbaum P, Walter S, Russell D, Wood E, Galuppi B. Development and reliability of a system to classify gross motor function in children with cerebral palsy. Dev Med Child Neurol 1997;39:214–23. [20] Jahnsen R, Aamodt G, Rosenbaum P. Gross Motor Function Classification System used in adults with cerebral palsy; agreement of self-reported versus professional rating. Dev Med Child Neurol 2006;48:734–8. [21] Roislien J, Skare O, Gustavsen M, Broch NL, Rennie L, Opheim A. Simultaneous estimation of effects of gender, age and walking speed on kinematic gait data. Gait Posture 2009;30:441–5. [22] Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10. [23] van de Pol RJ, van TE, Lucas C. Inter-rater reliability for measurement of passive physiological range of motion of upper extremity joints is better if instruments are used: a systematic review. J Physiother 2010;56:7–17. [24] McDowell BC, Hewitt V, Nurse A, Weston T, Baker R. The variability of goniometric measurements in ambulatory children with spastic cerebral palsy. Gait Posture 2000;12:114–21.
191
[25] Perry J, Burnfield JM. Gait analysis: normal and pathological function, 2nd edn., Thorofare, N.J: SLACK; 2010. [26] Grunt S, Henneman WJ, Bakker MJ, Harlaar J, van der Ouwerkerk WJ, van SP, et al. Effect of selective dorsal rhizotomy on gait in children with bilateral spastic paresis: kinematic and EMG-pattern changes. Neuropediatrics 2010;41:209–16. [27] Hillman SJ, Hazelwood ME, Schwartz MH, van der Linden ML, Robb JE. Correlation of the Edinburgh Gait Score with the Gillette Gait Index, the Gillette Functional Assessment Questionnaire, and Dimensionless Speed. J Pediatr Orthop 2007; 27:7–11. [28] Webster KE, Wittwer JE, Feller JA. Validity of the GAITRite walkway system for the measurement of averaged and individual step parameters of gait. Gait Posture 2005;22:317–21. [29] Podsiadlo D, Richardson S. The timed ‘‘Up & Go’’: a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc 1991;39:142–8. [30] ATS statement: guidelines for the six-minute walk test. Am J Respir Crit Care Med 2002;166:111–7. [31] McGinley JL, Baker R, Wolfe R, Morris ME. The reliability of three-dimensional kinematic gait measurements: a systematic review. Gait Posture 2009;29:360–9.
Contents lists available at SciVerse ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Ambulant adults with spastic cerebral palsy: The validity of lower limb joint angle measurements from sagittal video recordings Kerstin L. Larsen a,*, Grethe Maanum a,b, Kathrine F. Frøslie a,c, Reidun Jahnsen a,d a
Sunnaas Rehabilitation Hospital, Nesoddtangen, Norway Faculty of Medicine, University of Oslo, Norway c Department of Biostatistics, University of Oslo, Norway d Oslo University Hospital, Department for Child Neurology, Rikshospitalet, Oslo, Norway b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 18 February 2011 Received in revised form 30 August 2011 Accepted 4 September 2011
Background: In the development of a clinical program for ambulant adults with cerebral palsy (CP), we investigated the validity of joint angles measured from sagittal video recordings and explored if movements in the transversal plane identified with three-dimensional gait analysis (3DGA) affected the validity of sagittal video joint angle measurements. Methods: Ten observers, and 10 persons with spastic CP (19–63 years), Gross Motor Function Classification System I–II, participated in the study. Concurrent criterion validity between video joint angle measurements and 3DGA was assessed by Bland–Altman plots with mean differences and 95% limits of agreement (LoA). Pearson’s correlation coefficients (r) and scatter plots were used supplementary. Transversal kinematics 2 SD from our reference band were defined as increased movement in the transversal plane. Results: The overall mean differences in degrees between joint angles measured by 3DGA and video recordings (38, 58 and 78 for the hip, knee and ankle respectively) and corresponding LoA (188, 108 and 158 for the hip, knee and ankle, respectively) demonstrated substantial discrepancies between the two methods. The correlations ranged from low (r = 0.39) to moderate (r = 0.68). Discrepancy between the two measurements was seen both among persons with and without the presence of deviating transversal kinematics. Conclusion: Quantifying lower limb joint angles from sagittal video recordings in ambulant adults with spastic CP demonstrated low validity, and should be conducted with caution. This gives implications for selecting evaluation method of gait. ß 2011 Elsevier B.V. All rights reserved.
Keywords: Gait Observational Validity Cerebral palsy Rehabilitation
1. Introduction Cerebral palsy (CP) is a disorder of motor control and posture resulting from a non-progressive lesion of the developing brain [1]. As gait impairment is a common functional limitation in persons with CP [2], several observational gait analysis tools have been developed [3–9]. Young adults with CP frequently experience declined walking ability [10]. As they may benefit from rehabilitation programs to restore function and/or prevent further deterioration, it is essential to have valid clinical tools that describe their gait systematically, and are cost-effective and easy to use.
* Corresponding author at: Sunnaas Rehabilitation Hospital, Clinic for NeuroRehabilitation, 1450 Nesoddtangen, Norway. Tel.: +47 66 96 90 00; fax: +47 66 91 25 76. E-mail addresses: [email protected], [email protected] (K.L. Larsen). 0966-6362/$ – see front matter ß 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2011.09.004
The validity of a measurement tool describes whether it is appropriate and useful and is divided into face-, construct-, content- and criterion validity [11]. Criterion validity describes systematic equivalence between two similar tools. In gait analysis this is best achieved by comparing the actual gait assessment tool to the criterion standard of gait, which is regarded to be threedimensional gait analysis (3DGA) [2,12]. 3DGA is the most advanced and specific gait analysis tool used in the specialised multidisciplinary follow-up program of ambulant children with CP [13]. However, it includes expensive technology and time consuming analyses, resulting in restricted access. Therefore alternative video-based gait assessment tools (VGA) have been developed. Most of them originate from the need for outcome measurements in clinical practice and in research and are based on categorisation of different gait strategies into ordinal scales from experts in the field of gait impairment in children with spastic CP [3–8]. Several studies, investigating the effects of botulinum toxin A (BoNT-A) on gait in children with spastic CP, have used a VGA as outcome [3,4,14,15]. Ubhi et al. reported both
K.L. Larsen et al. / Gait & Posture 35 (2012) 186–191
reliability, clinically relevant change, as well as the limitations of VGA compared with 3DGA [4]. However, other BoNT-A studies do not discuss the limitations of assessing gait with VGA [3,12,14,15]. In contrast to the categorical approach in most VGA-tools, the Salford Gait Tool (SF-GT) [9] is based on visually estimation of exact joint angles in the hip, knee and ankle at six selected events during the gait cycle, and then categorisation of them. The validity and reliability or SF-GT in children with CP has been established by comparing SF-GT with sagittal 3DGA kinematics [9,16]. Another study using a quantitative approach [17], compares measurements of six selected joint angles on a video screen using software program with simultaneous 3DGA sagittal kinematics. To our knowledge, no study has investigated the validity of measuring joint angles from a video screen in adults with CP. Due to the given exact joint angles, which creates transparency with regard to validity the SF-GT scoring sheet was chosen in this study [9]. Further, the literature on two-dimensional VGA tools is limited with regards to presenting simultaneous data on movements in the sagittal and transverse plane during gait. Thus, in the development of a clinical program for ambulant adults with CP and decreased walking ability, we investigated concurrent criterion validity by comparing joint angles measured from sagittal video recordings and sagittal plane kinematics from 3DGA. We also explored if movement deviations in the transverse plane identified with 3DGA affected the validity of sagittal video joint angle measurements. 2. Methods 2.1. Ethics The study was approved by the Data protection supervisor and the Regional Committee for Research Ethics. All the participants and observers gave written consent. 2.2. Participants A subset of 10 participants was randomly drawn from a larger study population of well functioning adults with spastic CP recruited through advertisements in newspapers and on websites [18]. The participants were five women and five men, age range 19–63 years, five with unilateral- and five with bilateral CP. They all
187
walked independently, four being classified at level I and six at level II of the Gross Motor Function Classification System (GMFCS) [19,20]. Ten observers out of about 40 staff members were recruited through information on e-mail, and nine physiotherapists and one certified prosthetic/orthotist volunteered. Their clinical experience ranged from one up to 40 years. Five observers had experience with 3DGA. Before the assessments the observers attended a lecture on gait, gait parameters, and gait assessment methods, in addition to instruction in the use of SF-GT and goniometry measurement on a screen. One of the authors (RJ) participated as an observer. For the purpose of this study, the observers assessed the video recordings individually once. 2.3. Procedure 3DGA data were collected in a Motion Analysis Laboratory with six infrared cameras (MX13, 100 Hz, VICON Motion System, Oxford, England), two force plates (AMTI OR 6-7, Advanced Mechanical Technology Inc, Watertown, USA), and two digital video cameras (25 Hz). Fifteen reflective markers (14 mm diameter) were attached to the pelvis and lower limbs according to the Plug-In-Gait model (VICON). The gait data were collected by two of the authors (GM and KLL), both having expert knowledge of the underlying biomechanical model. The sagittal video recordings of the 10 participants were copied to CDs. The observers played the CD using PCs with 17 thumbs flat screens. They measured joint angles of the hip, knee, and ankle from the right limb in absolute degrees at six events during a gait cycle; initial contact (IC), end double support (EDS), mid stance (MST), start double support (SDS), toe off (TO) and, mid swing (MSW) according to a modified SF-GT sheet (Appendix A). It was possible to play the video recordings back and forth one frame at a time, and to ‘‘freeze’’ the picture when performing the measurements. A 30 cm long goniometer with each degree marked was used. The observers worked independently with no collaboration with each other. The 3DGA data were processed in Workstation (VICON) by the first author. The joint angles, as described in SF-GT, were extracted to Excel from Workstation by the use of Pipeline External Communication Server (PECS, VICON, Oxford, England), except for the joint angles in MST and MSW that were transferred manually. Further data management was performed in SPSS version 15.0. The transverse plane kinematics for pelvis, hip and foot-progression were exported to Polygon for graphical curves together with the Motion Analysis Laboratory’s own reference material of healthy adults (n = 48) [21]. 2.4. Analysis Concurrent criterion validity was assessed by comparing the joint angles from the same gait cycle measured by SF-GT and sagittal plane kinematics from 3DGA. Agreement between actual measurements in degrees was calculated using Bland– Altman plots with mean differences and 95% limits of agreement (LoA), i.e. 2 standard deviations of the differences (2 SDdiff) [22]. Scatter plots with Pearson’s r were used supplementary, and the magnitudes of the r were evaluated as little
Fig. 1. Kinematic data from 3DGA for all joints and events showing the variability in gait pattern for all participants. The measurements were done at events defined by the SFGT tool, and interpolation lines were added to the chart.
K.L. Larsen et al. / Gait & Posture 35 (2012) 186–191
188
Fig. 2. Data from SF-GT for all joints and events defined in the SF-GT tool, showing the variability in assessment of 10 participants, each evaluated by 10 different observers.
(0–0.25), low (0.26–0.49), moderate (0.50–0.69), high (0.70–0.89) and very high (0.90–1.0) [11]. To explore if increased movements in transverse plane could explain differences between measurements in 3DGA and SF-GT, the transverse plane 3DGA curves for pelvis, hip and foot-progression were visually compared with scatter plots of IC knee joint angle from the SF-GT measurements and the sagittal plane 3DGA.
3. Results The 10 participants had varying gait strategies as demonstrated by 3DGA kinematics (Fig. 1). There was one missing data from the SF-GT measurements (ankle in MSW), resulting in 1799 joint angles. Totally 180 measurements were sampled from the 3DGA. The SF-GT measurements from the 10 observers are summarised in Fig. 2. For each of the 18 different joint angles displayed in Figs. 1 and 2, 10 Bland–Altman plots and 10 scatter plots were made, and mean differences, LoA and Pearson’s correlation coefficients were calculated, one for each of the observers’ concordance with 3DGA. A summary of these numbers are shown in Table 1. The best agreement between SF-GT and 3DGA was found for the knee joint, with mean differences for the different events ranging from 08 to 178 and LoA in the range 8–158. Clinically relevant examples of the ranges of individual observer’s agreements are mean difference
(LoA) of 18 (108) and 138 (118) for knee IC, and 28 (68) and 118 (118) for knee MST. The individual results revealed no differences between observers from the gait-laboratory and observers from the clinic. The overall mean correlations for the hip, knee and ankle were 0.39, 0.68, and 0.39 respectively (Table 1). Fig. 3 shows the visually estimated joint angles of IC for the knee for all the 10 participants against the simultaneous 3DGA joint angles, as well as transverse plane curves for one gait cycle. Discrepancy between the 3DGA joint angles and the SFGT measurements was seen both with and without deviations from our reference band in transverse kinematics. There were no differences between the observer groups (i.e. gait laboratory and clinic) (Fig. 3). 4. Discussion The purpose of this study was to investigate the validity of measuring lower limb joint angles on a video-screen in 10 ambulant adults with spastic CP. Concurrent criterion validity was evaluated by comparing joint angles registered with SF-GT and sagittal plane kinematics from 3DGA, which is considered to be the criterion standard of gait analysis. We also explored if increased
Table 1 Descriptive statistics for the agreement and correlation between 3DGA and SF-GT measurements. Each cell in the table is based on 10 sets of paired observations; one for each rater and 3DGA. The mean differences and 95% limits of agreement (LoA) (2 SDdiff) are the means based on the 10 Bland–Altman plots constructed from the paired observations sets. The r-mean are the mean of the corresponding correlation coefficients. Knee
Hip
IC EDS MST SDS TO MSW Overall mean
Ankle
Mean difference
LoA
r [range]
Mean difference
LoA
r [range]
3 5 4 6 11 3
24 20 14 13 16 19
0.26 0.43 0.52 0.63 0.28 0.24
3 0 5 5 17 5
10 9 9 8 15 12
0.84 0.39 0.65 0.77 0.74 0.69
3
18
0.39
5
10
0.68
[ 0.07 to 0.65] [0.10 to 0.80] [0.36 to 0.82] [0.39 to 0.80] [ 0.11 to 0.63] [ 0.12 to 0.46]
[0.69 [0.03 [0.41 [0.19 [0.20 [0.51
Mean difference to to to to to to
0.92] 0.77] 0.76] 0.90] 0.87] 0.84]
LoA
r [range]
1 6 7 14 6 6
16 12 12 13 18 17
0.52 0.19 0.42 0.64 0.27 0.31
7
15
0.39
IC: initial contact, EDS: end double support, MST: mid stance, SDS: start double support, TO: toe off, and MSW: mid swing.
[0.14 to 0.82] [ 0.18 to 0.50] [0.12 to 0.65] [0.36 to 0.88] [0.00 to 0.57] [ 0.03 to 0.50]
K.L. Larsen et al. / Gait & Posture 35 (2012) 186–191
189
Fig. 3. Scatter plots of IC knee joint angle from the SF-GT measurements and the sagittal plane 3DGA, and the transversal plane 3DGA curves for pelvis, hip and foot-progression. Since some of the observers have identical measures, there are not eleven data points for knee angle at IC for each participant. There are no missing data for the knee.
movements in the transverse plane affected the validity of joint angles measured on a video recording. Analysis of mean differences and corresponding LoA demonstrated substantial discrepancies and lack of agreement between the two methods. We did not find increasing discrepancy between video recordings and 3DGA when increase movements in transverse plane were present. Our study demonstrated that the measured knee joint angles achieved the highest agreement with the concurrent 3DGA results. Toro et al. [9], using the least significant difference (LSD) statistics, reported a fairly high mean difference between SF-GT and 3DGA for all joints (168) and a wide range (2–638). Even if 80–81% of the mean measurements were within the range of 168; this indicates uncertainty about the validity. A recent study [17] investigated the agreement between joint angles measured with 3DGA and computer-based video screen measurement (VSM). Despite the use of a VSM, a wide range between the measurements was found. This study [17], using one investigator, demonstrated highest agreement for peak knee flexion in swing (mean difference (LoA) = 1.48 (118)), and the lowest agreement for peak ankle dorsiflexion in swing (mean difference (LoA) = 128 (14.68)). These results are similar to our study, where 10 investigators evaluated the joint angles by using a plastic goniometer on a screen. We expected that the use of a goniometer would strengthen the focus on the joints being measured [23]. However, the similar results, despite different measuring methods, support the findings of poor validity of measuring joint angles from video recordings, indicating that issues other than the measurement methods are important limitations when estimating joint angles on a video-screen.
The observers had different levels of experience with gait analysis and the use of goniometry. Since the same gait cycles were compared, the reflex markers used in 3DGA were also visible on the video recordings. Consequently, the observers with expert knowledge of 3DGA potentially could have an advantage when estimating joint angles. Despite this, the overall results revealed no differences between observers from the gait-laboratory and observers from the clinic. Grunt [17] concludes that VSM should be used with caution and that rotational deficit could cause problems in the interpretation. However, Grunt did not present data to support this, and to our knowledge no previous studies on VGA tools have explored this issue. Our study failed to demonstrate that increased movements in the transverse plane kinematics resulted in a higher discrepancy in the joint angles measurements from video recordings. Though, this may be due to the limited sample. As numerous interventions are directed towards improving gait strategies for persons with CP, it is appealing to use joint angles as outcome, when you are restricted to use video instead of 3DGA. With the introduction of a VSM, this becomes even more attractive. However, the use of a goniometer, even if computer-based, has its well known limitations [24]. Thus, visual analysis, by ‘‘naked eye’’ or video, will always be subjective, having qualitative strengths and quantitative limitations. When estimating joint angles in degrees from a video-screen, difficulties arise because the limb may not be viewed from the correct angle [12,13,25]. When a person demonstrates a pathological gait strategy in the transverse plane, this will give a false image when estimating joint angles by the use of a sagittal plane video recording.
190
K.L. Larsen et al. / Gait & Posture 35 (2012) 186–191
Several studies have reported the reliability of VGA to be satisfactory [4–8,16]. However, reliability does not guarantee validity, and responsiveness includes them both [11]. Neither Scholtes et al. [14], who used joint angles measured from video recordings and the Edinburgh Visual Gait Score (EVGS) [5], nor Kim et al. [15], who used the Physician’s Rating Scale [3], discuss the limitations of using VGA with regard to what they consider as a clinically important improvement. On the other hand a recent publication, which used VGA to evaluate the effects of selective dorsal rhizotomy on gait in persons with CP, addresses the limitations of measuring gait in only two dimensions [26]. After exploring the issue of validity in measuring complex movements, such as gait in persons with CP in only two dimensions, we suggest that VGAs are considered as subjective qualitative measures. The EVGS, considered the leading VGA tool, is assessing joint motion in sagittal and coronal plane using an ordinal scale, representing overall gait deviation in CP [5,13]. The qualitative strength of this instrument is shown in the studies by Grunt et al. and Hillman et al. [26,27]. Thus, VGAs have its strengths in respect of giving an overall descriptive picture of gait [5,13], as well as being easy to use [13,26]. The correlation between EVGS and Gillette Gait Index reflects the qualitative strength of EVGS in giving a score for overall gait strategy using an ordinal scale [27]. To improve accuracy of VGA techniques ordinal scales are recommended, developed out of needed detail level and known common gait strategies for the population of interest. For clinical as well as research purposes, we suggest that quantitative continuous data obtained from measures of basic gait parameters, such as cadence, stride length and speed replace the ongoing practice on measuring joint angles on a two dimensional screen. Continuous gait parameters could be achieved by simple methods, such as using a stopwatch with a marked walkway and powder on the feet [12], or more technical instruments, such as portable walkways, for example GAITRite [28]. In addition functional tests, such as test of functional mobility (the Timed Up and Go test) [29] and functional walking capacity (6-minute walk test) [30], have recently been suggested as appropriate outcomes reflecting activities in daily life, for adults with spastic CP, GMFCS levels I–II [18]. The small sample size of the present work is the major limitation. However, our conclusions are based on means and SDs of observed differences. Such numbers are adequately estimated in small sample sizes, provided that the data are normally distributed. Another limitation in our study was that cameras with 25 frames per second were used. This resulted in blurry pictures, especially for the ankle in swing, due to high speed of the movement of this body segment. The observers therefore might have chosen different frames for the gait events measured so contributed to the discrepancy of measurements. Higher picture frequency would probably reduce this problem. A disagreement with the 3DGA does not necessarily mean that the VGA is the only inaccurate measurement method. The validity of 3DGA can be affected by imprecise marker placement, imprecise anthropometric measurements, soft tissue movements and shortcomings of the Plug-In-Gait marker protocol with regard to the joint angles [25]. The staff members collecting the 3DGA data were specially trained and a standardised protocol was used. Thus, we assume precise marker placement and anthropometric measurements, resulting in the same quality of data as from comparable gait laboratories [31]. However, both soft tissue movements and the shortcomings of the foot model remains. 5. Conclusion This study demonstrates the shortcomings of quantifying joint angles from a video recording, giving implications when selecting evaluation methods of gait. Since VGA is more available
than 3DGA in both clinical programs and research settings, we suggest that a qualitative approach is preferable for the videoanalysis. Other measurements, such as spatiotemporal parameters and functional tests on mobility, could generate continuous quantitative data. Acknowledgements This project has been financed by EXTRA funds from the Norwegian Foundation for Health and Rehabilitation, the East Regional Health Administration and Sunnaas Rehabilitation Hospital. We are grateful to all the adults with CP and observers who participated, and we acknowledge the contribution of the Norwegian CP Association. Conflict of interest statement The authors have no conflict of interest to disclose.
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.gaitpost.2011.09.004. References [1] Bax M, Goldstein M, Rosenbaum P, Leviton A, Paneth N, Dan B, et al. Proposed definition and classification of cerebral palsy. April 2005. Dev Med Child Neurol 2005;47:571–6. [2] Gage JR, Schwartz MH, Koop SE, Novacheck TF. Identification and treatment of gait problems in cerebral palsy. Clinics in Developmental Medicine Nos 180181, 2nd edn, London: Mac Keith Press; 2009. [3] Koman LA, Mooney III JF, Smith BP, Goodman A, Mulvaney T. Management of spasticity in cerebral palsy with botulinum—a toxin: report of a preliminary, randomized, double-blind trial. J Pediatr Orthop 1994;14:299–303. [4] Ubhi T, Bhakta BB, Ives HL, Allgar V, Roussounis SH. Randomised double blind placebo controlled trial of the effect of botulinum toxin on walking in cerebral palsy. Arch Dis Child 2000;83:481–7. [5] Read HS, Hazlewood ME, Hillman SJ, Prescott RJ, Robb JE. Edinburgh visual gait score for use in cerebral palsy. J Pediatr Orthop 2003;23:296–301. [6] Mackey AH, Lobb GL, Walt SE, Stott NS. Reliability and validity of the Observational Gait Scale in children with spastic diplegia. Dev Med Child Neurol 2003;45:4–11. [7] Wren TA, Rethlefsen SA, Healy BS, Do KP, Dennis SW, Kay RM. Reliability and validity of visual assessments of gait using a modified physician rating scale for crouch and foot contact. J Pediatr Orthop 2005;25:646–50. [8] Dickens WE, Smith MF. Validation of a visual gait assessment scale for children with hemiplegic cerebral palsy. Gait Posture 2006;23:78–82. [9] Toro B, Nester CJ, Farren PC. The development and validity of the Salford Gait Tool: an observation-based clinical gait assessment tool. Arch Phys Med Rehabil 2007;88:321–7. [10] Opheim A, Jahnsen R, Olsson E, Stanghelle JK. Walking function, pain, and fatigue in adults with cerebral palsy: a 7-year follow-up study. Dev Med Child Neurol 2009;51:381–8. [11] Domholdt E. Rehabilitation research: principles and applications. St. Louis, MS: Elsevier Saunders; 2005. [12] Rose SA, Ounpuu S, DeLuca PA. Strategies for the assessment of pediatric gait in the clinical setting. Phys Ther 1991;71:961–80. [13] Harvey A, Gorter JW. Video gait analysis for ambulatory children with cerebral palsy: why, when, where and how! Gait Posture 2010. [14] Scholtes VA, Dallmeijer AJ, Knol DL, Speth LA, Maathuis CG, Jongerius PH, et al. Effect of multilevel botulinum toxin a and comprehensive rehabilitation on gait in cerebral palsy. Pediatr Neurol 2007;36:30–9. [15] Kim K, Shin HI, Kwon BS, Kim SJ, Jung IY, Bang MS. Neuronox versus Botox for spastic equinus gait in children with cerebral palsy: a randomized, doubleblinded, controlled multicentre clinical trial. Dev Med Child Neurol 2011;53: 239–44. [16] Toro B, Nester CJ, Farren PC. Inter- and intraobserver repeatability of the Salford Gait Tool: an observation-based clinical gait assessment tool. Arch Phys Med Rehabil 2007;88:328–32. [17] Grunt S, van Kampen PJ, van der Krogt MM, Brehm MA, Doorenbosch CA, Becher JG. Reproducibility and validity of video screen measurements of gait in children with spastic cerebral palsy. Gait Posture 2010;31:489–94. [18] Maanum G, Jahnsen R, Stanghelle JK, Sandvik L, Keller A. Effects of botulinum toxin A in ambulant adults with spastic cerebral palsy: a randomized doubleblind placebo controlled-trial. J Rehabil Med 2011;43:338–47.
K.L. Larsen et al. / Gait & Posture 35 (2012) 186–191 [19] Palisano R, Rosenbaum P, Walter S, Russell D, Wood E, Galuppi B. Development and reliability of a system to classify gross motor function in children with cerebral palsy. Dev Med Child Neurol 1997;39:214–23. [20] Jahnsen R, Aamodt G, Rosenbaum P. Gross Motor Function Classification System used in adults with cerebral palsy; agreement of self-reported versus professional rating. Dev Med Child Neurol 2006;48:734–8. [21] Roislien J, Skare O, Gustavsen M, Broch NL, Rennie L, Opheim A. Simultaneous estimation of effects of gender, age and walking speed on kinematic gait data. Gait Posture 2009;30:441–5. [22] Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10. [23] van de Pol RJ, van TE, Lucas C. Inter-rater reliability for measurement of passive physiological range of motion of upper extremity joints is better if instruments are used: a systematic review. J Physiother 2010;56:7–17. [24] McDowell BC, Hewitt V, Nurse A, Weston T, Baker R. The variability of goniometric measurements in ambulatory children with spastic cerebral palsy. Gait Posture 2000;12:114–21.
191
[25] Perry J, Burnfield JM. Gait analysis: normal and pathological function, 2nd edn., Thorofare, N.J: SLACK; 2010. [26] Grunt S, Henneman WJ, Bakker MJ, Harlaar J, van der Ouwerkerk WJ, van SP, et al. Effect of selective dorsal rhizotomy on gait in children with bilateral spastic paresis: kinematic and EMG-pattern changes. Neuropediatrics 2010;41:209–16. [27] Hillman SJ, Hazelwood ME, Schwartz MH, van der Linden ML, Robb JE. Correlation of the Edinburgh Gait Score with the Gillette Gait Index, the Gillette Functional Assessment Questionnaire, and Dimensionless Speed. J Pediatr Orthop 2007; 27:7–11. [28] Webster KE, Wittwer JE, Feller JA. Validity of the GAITRite walkway system for the measurement of averaged and individual step parameters of gait. Gait Posture 2005;22:317–21. [29] Podsiadlo D, Richardson S. The timed ‘‘Up & Go’’: a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc 1991;39:142–8. [30] ATS statement: guidelines for the six-minute walk test. Am J Respir Crit Care Med 2002;166:111–7. [31] McGinley JL, Baker R, Wolfe R, Morris ME. The reliability of three-dimensional kinematic gait measurements: a systematic review. Gait Posture 2009;29:360–9.