Measures of temporal-spatial parameters of gait are commonly used by physical therapists (PTs) when treating children with motor disabilities such as cerebral palsy (CP).1 The measures are used to identify gait deviations and to evaluate the effectiveness of therapeutic interventions. These assessments may occur on the same day (ie, before and after fitting with orthoses) or over periods of time lasting days, weeks, or months. Therefore, when selecting a clinical evaluation tool to measure temporal-spatial gait parameters, an important consideration is test-retest reliability. That is, the stability of the measurement system is critical for attributing change in the measures to the intervention that is being assessed.
PTs have a number of methods for assessing the gait parameters of children with motor disabilities. These include (a) visual observation, (b) temporal evaluations using a stop watch, and (c) temporal-spatial evaluations using either footfall patterns (inkpad, pedographs) or three-dimensional kinematic and kinetic analyses. Visual observation is a common clinical method of assessing gait disorders, yet the lack of agreement between successive observations (ie, reliability) limits the clinical use of this method.2–5 Investigators using a stopwatch for gait analysis have shown moderate to high reliability.6–8 Such gait analysis, however, addresses only temporal aspects of gait and the investigators did not include children with motor disabilities. Measurement of footfall patterns using chalk,9 inkpad, and pedograph methods10–13 has proven reliable, but the amount of time needed to assess these patterns limits their use. More recently, researchers have measured three-dimensional kinematic and kinetic temporal-spatial gait parameters of children with CP using a combination of camera and force-plate systems.14–20 Although the investigators report a high degree of precision when using these three-dimensional systems, these methods are time consuming, expensive, and demand a specialized expertise not found in most clinical settings.
The use of an instrumented walkway system such as the GAITRite® (CIR Systems Inc, Havertown, Pa) to evaluate temporal-spatial gait characteristics is becoming more common in the clinical environment because this system (a) can supply the clinician with quick, objective temporal-spatial gait measurements, (b) is not labor intensive, and (c) demands only a moderate level of computer expertise. High test-retest reliability statistics [ie, intraclass correlation coefficients (ICC), ranging from 0.79 to 0.95] have been reported for the GAITRite walkway system. These studies however involved healthy adults9,21–25 and adults with diabetes.24 Reliability data for temporal-spatial gait parameters on children one to 10 years old without disabilities using the GAITRite system have been reported to be comparable to adult data.26 To date, however, the reliability of gait measurements using the GAITRite or a similar instrumented walkway system for children and adolescents with motor disabilities has not been established.
This article addresses reliability of the instrumented walkway system in evaluating the effectiveness of orthoses on gait patterns, a common and important clinical application of instrumented walkway systems. The effects of orthoses can be evaluated by comparing gait variables between two sessions scheduled on the same day or over an interval of days or weeks. This study addressed same-day comparison. The purpose of this pilot study was to determine same-day test-retest reliability using the GAITRite for children with motor disabilities under two walking conditions: (1) without orthoses or barefoot and (2) while wearing orthoses and shoes.
Our sample of convenience consisted of 19 children (15 girls and four boys). The mean age was 6.8 years (SD = 4.1 years), and the subjects included five with ataxic CP; eight with spastic diplegic CP; three with spastic hemiplegic CP; two with Angelman syndrome, and one with arthrogryposis. The participants were identified from local school districts in a mid-western metropolitan area. Eight of the children were classified at Level I, nine at Level II, and two at Level III according to the Gross Motor Function Classification System (GMFCS).27 All participants were independent ambulators; nine walked without an assistive device, one with quad canes, seven with walkers, and two with gait trainers. All the participants in this study were receiving individual and direct or consultative and indirect physical therapy within the school systems. Inclusion criteria included the ability to maintain a steady-state gait pattern with or without an assistive device for a distance of at least 20 feet. Written consent and verbal assent were obtained from the parents and participants, respectively. Approval for this study was obtained from the University Institutional Review Board.
All participants wore no. 3 or 4 dynamic ankle-foot orthoses (DAFO) (Cascade DAFO™ Inc, Ferndale, Wash) with a contoured footplate made from 2.4-mm-thick polypropylene enclosing the dorsum of the forefoot and ankle and covering the posterior part of the leg to about 2 to 7.5 cm above the malleoli, with straps across the ankle and forefoot. All subjects had been casted for the DAFO using a prefabricated contoured footplate by a PT experienced in casting for this type of orthosis. On the day of testing, to ensure that all children wore an appropriately fitting DAFO, a PT evaluated the orthoses for fit. If the participants’ current orthosis was not fitting appropriately or was damaged so as not to be functional, the PT made a new orthoses, and a new Cascade DAFO was worn to ensure fit and consistency. On the day of testing, six children wore their current Cascade DAFO that had been worn for two months or longer, while 13 wore a new Cascade DAFO.
The GAITRite® Walkway System (GAITRite Gold, CIR Systems, Havertown, PA) is a computer-based instrumental walkway consisting of a portable walkway (ie, a mat) embedded with pressure-activated sensors. The series of six sensor pads are inserted in a grid formation between a layer of vinyl (top cover) and foam rubber (bottom cover). The walkway’s active measurement area is 61 cm wide and 365.8 cm long. Sensors are arranged in a grid pattern (48 × 384) and placed 1.27 cm apart (total of 13,824 sensors) and are activated by mechanical pressure as a patient walks over the walkway. The sampling rate of the system is 80 Hz. Data from the activated sensors are collected by a series of on-board processors and transferred to the computer through a serial port. The system continuously scans the sensors to detect pressures and transfers this information to the computer for calculation of the temporal-spatial gait characteristics. A software program (GAITRite Gold, Version 3.4) was used to calculate temporal- spatial gait parameters (eg, cadence, velocity, step and stride length) for each footfall as well as the overall average of each parameter. These measures were recorded and stored by a DELL compatible computer using GAITRite.
The GAITRite Walkway System software calculations are based on heel-strike data assuming a steady-state gait or consistent cadence (ie, a minimum of four consecutive steps). These fundamental gait events cannot be assumed when evaluating children with motor disabilities, especially children with spastic CP walking barefoot. Therefore, the following method was used to ensure accuracy and objectivity of the gait events.
A Sony Mini DV Digital Handy Cam video camera (1.5 mega pixels, Model # DCR-TRV30 NTSC, Sony Electronics Inc, Park Ridge, NJ) was set up on a tripod six to eight feet from one end of the mat or at a position that would visually capture the child’s walking trial through the whole width and length of the walkway system. The video was used as a permanent record to ensure participant’s identification, type of orthotic, type of assistive device, and walking condition (ie, DAFO or barefoot). In addition, the video was used in conjunction with the GAITRite Walkway System computer software to substantiate the number of footfalls (a minimum of four consecutive steps); that the footfalls fell completely on the electronic walkway (ie, ensure the image on the computer was a complete footprint); and that the computer captured a steady-state gait (ie, no falls, stopping, or readjusting of the assistive device while walking).
The term “footfall” is defined as each separate incidence of the foot hitting the mat (ie, footstep).23,26 The GAITRite Walkway System software analyzes footfall data based on the heel strike reference point. However, the participants in this study usually did not exhibit the typical heel-strike to toe-off pattern of the gait cycle. Therefore, the software option to suspend data analysis was chosen. This permitted the therapist to compare the data collected by the walkway system with the video display to validate that (a) the displayed footfalls were complete footfalls, (b) extraneous data were deleted (ie, assistive device markings) and (c) left and right footfalls were properly identified. This editing process allowed the user to determine that the atypical footfall pattern of the participants was properly identified by the software for further analysis. Once spatial data were substantiated, temporal anomalies were also addressed by comparing spatial footfall data with temporal footfall data to ensure that the data analyzed by the software were part of the steady-state gait pattern.
All participants were tested in a laboratory setting on a university campus. Before data collection, walking on the mat was demonstrated to the participants who then practiced two to four trials until they understood the procedure. Generally, a total of six to 10 walking trials were performed during data collection until three trials for each walking condition that contained at least four consecutive steady-state footfalls were completed. There were two walking conditions, barefoot (ie, without DAFO) and DAFO (ie, shod with DAFO). DAFO walks preceded barefoot walks. To control for fatigue, participants were allowed to rest as needed between trials. No attempt was made to control gait velocity during testing in that encouraging verbal cues were given (eg, “Come on, walk to me”) but no verbal (eg, “walk faster”) or physical prompting was used to initiate or maintain the steady-state gait pattern (ie, no metronome or similar device to control gait was used).
The following six temporal-spatial gait measurements were evaluated: gait velocity (cm/sec), cadence (steps/min), stance time (% gait cycle), stride length (cm), base width (cm; width between first contact of right and first contact of left, base of support), and cycle time (sec).
The stability of the selected temporal-spatial gait parameters was determined by calculating both relative and absolute reliability indices for both conditions: barefoot and with shoes and orthoses.28,29 Relative reliability is the individual variation, both within an individual (the difference in an individual’s scores from trial to trial) and among individuals (consistency in the differences among individuals across trials) on multiple administrations (repeated measures) of the same test.29 Relative reliabilities were determined by calculating two-way mixed-effects ICCs and 95% confidence intervals (95% CIs), with the individual as a random effect and the measurement error as fixed effect. Both single and multiple trial (ie, 3) ICCs were calculated using the Statistical Package for the Social Sciences (v.12.0). All three trials were used in the calculation of the single trial reliability estimate (ICC and 95% CI). Previously reported minimum thresholds of acceptable reliability ranged from 0.75 to 0.8.30,31 Using these as guidelines, the minimum acceptable ICC (95% CI) to demonstrate adequate reliability was set at 0.80 for this study. Absolute reliabilities, defined as the average amount of error (plus or minus) expected in an individual’s observed score,29 were determined by calculating the standard error of measure (SEM) using the following formula:
where SD is the standard deviation of the measures across trials.28 The SEMs were calculated using both the single and multiple trials ICC.
The descriptive characteristics for the participants and the temporal-spatial gait parameters for each of the three walking trials are presented in Tables 1 and 2, respectively. The single and three-trial average relative and absolute reliability estimates are presented in Table 3. For relative reliabilities, the majority (ie, >80%) of ICCs (95% CI) for each condition (ie, barefoot and with shoes and orthoses) for both single and multiple trials met or exceeded the minimum reliability coefficient criteria of 0.80. The ICCs therefore demonstrated considerable stability with multiple administrations of a test (ie, similar scores from trial to trial within and between participants). Of importance was the stability of a majority of the parameters if a single trial was administered. This finding is critical because it lends confidence that an observed score from a single administration of a test is reflective of the participants’ performance. However, several gait parameters failed to achieve the minimal ICC criterion (see above), thereby demonstrating less than acceptable reliability. A single trial to measure stance with either leg (right or left) while barefoot was below the acceptable cutoff (ICC 0.50 and 0.58, respectively), with the 95% CIs ranging from 0.22 to 0.79. The three-trial average also failed to meet the 0.80 criteria for the left leg stance barefoot (ICC = 0.75). The right leg stance barefoot average ICC only achieved the minimal criteria (ICC = 0.80) although the 95% CI indicated the reliability coefficient could be as low as 0.57, indicating the potential for considerable variability across trials. Several other parameters, ie, stride length left-barefoot, base of support left-orthoses, cycle time right-barefoot, and cadence barefoot and orthoses met the ICC criteria of 0.80, yet the lower bounds of the 95% CIs fell below the minimal threshold, suggesting less than acceptable reliabilities for a single measure. Overall, the orthoses condition demonstrated greater single and three-trial average reliability coefficients than the barefoot condition.
In the clinical setting, reproducibility or reliability of same day temporal-spatial gait parameters is critical when determining the effectiveness of therapeutic interventions involving orthoses for children with motor disabilities.32,33 Therefore, the purpose of this pilot study was to determine the same day test-retest reliability using the GAITRite for children with motor disabilities in two walking conditions: (1) barefoot and (2) while wearing orthoses and shoes. Results indicate that with few exceptions, the single and three-trial average relative and absolute reliability estimates the ICCs (95% CI) for each condition (ie, barefoot and shoes with orthoses) met or exceeded the minimum reliability coefficient criteria of 0.80.
Of interest is the higher reliability for gait parameters demonstrated in this study than was recently reported for children of similar age without any disabilities.26 Thorpe et al26 reported ICCs ranging from 0.05 to 0.89 for the same temporal-spatial gait parameters with less than 50% of their ICCs exceeding a minimum reliability coefficient criteria of 0.80 established in the present study. The higher ICCs seen in this study could have resulted from differences in methodologies. First, Thorpe et al26 administered “at least 4” trials to ensure at least four consecutive steady-state footfalls, while six to 10 trials were performed in the present study. Second, testing took place in four different locations (ie, schools) and therefore on different surfaces, while the same surface was used in the present study. Third, gait criteria were verified by observers, whereas the present study was verified by video recordings. Fourth, data collection was performed by four physical therapy students, while a PT performed all data collection and gait analysis in the present study.
Caution needs to be exercised when interpreting the single measure ICC for the temporal-spatial gait parameters presented in Table 3. Specifically, the single measure ICC does not represent a single walking attempt across the GAITRite, but rather a single measure of one “acceptable” trial as outlined in the methods sections (ie, number of footfalls, etc). That is, the single trial ICC represents the reliability of the parameter when one trial was conducted, while the multiple trial ICC reflects the reliability of a given parameter over repeated administrations. In a clinical setting, rarely are multiple trials administered on a diagnostic test; therefore, the single trial ICC corresponds to the expected stability of a parameter with one assessment. This is further extended by the need to conduct several walking trials before obtaining one that meets the requirements for an “acceptable” trial. Thus, although a single measure can be used to assess many of the gait parameters evaluated in this study, the single measure reliability coefficient reflects a single trial only after two to four practice trials followed by one to three walking trials that yield at least four consecutive steady-state footfalls.
A feature of this study that provides additional information was accounting for the confidence interval (95% CI) when evaluating the reliability of measures. Evaluating the confidence interval provides additional information by considering the lower bound of the reliability estimate. Specifically, the minimum acceptable ICC was 0.80, yet the 95% CI of several of the ICC estimates had a lower bound than our established minimum criteria. Therefore, although many of our ICCs exceeded our 0.80 criteria, the lower bound of several of the 95% CI suggest a cautious approach when interpreting their reliability.
Limitations of this study include the small number of participants with varied ages and diagnosis and at varied levels of the GMFCS. However, given these limitations, the results of this study represent an initial attempt to evaluate the reliability of measuring temporal-spatial gait parameters of children with motor disabilities using an instrumented walkway system.
1. Ôunpuu S, Davis RB, DeLuca PA. Joint kinetics: methods, interpretation and treatment decision-making in children with cerebral palsy and myelomeningocele. Gait Posture.
2. Goodkin R, Diller L. Reliability among physical therapists in diagnosis and treatment of gait deviations in hemiplegics. Percept Mot Skills.
3. Krebs DE, Edelstein JE, Fishman S. Reliability of observational kinematic gait analysis. Phys Ther.
4. Eastlack ME, Arvidson J, Snyder-Mackler L, et al. Interrater reliability of videotaped observational gait-analysis assessments. Phys Ther.
5. Coutts F. Gait analysis in the therapeutic environment. Man Ther.
6. Youndas JW. Atwood. Measurement of temporal aspects of gait obtained with a multimemory stopwatch. J Orthop Sports Phys Ther.
7. Morris S, Morris M, Iansek R. Reliability of measurements obtained with the timed ‘Up & Go’ test in people with Parkinson disease. Phys Ther.
8. Wall JC, Scarbrough J. Use of a multimemory stopwatch to measure temporal gait parameters. J Orthop Sports Phys Ther.
9. McDonough AL, Batavia M, Chen S, et al. The concurrent validity of the Functional Ambulation Profile and the GAITRite® system. Phys Ther.
10. Adams MA, Chandler LS, Schuhmann K. Gait changes in children with cerebral palsy following a neurodevelopmental treatment course. Ped Phys Ther.
11. Gaudet G, Goodman R, Landry M, Russell G, Wall JC. Measurement of step length and step width: a comparison of videotape and direct measurements. Physiother Can.
12. Heitmann DK, Gossman MR, Shaddeau SA, et al. Balance performance and step width in noninstitutionalized, elderly, female fallers and nonfallers. Phy Ther.
13. Clarkson B. Absorbent paper method for recording foot placement during gait: suggestions from the field. Phys Ther.
14. Buckon CE, Thomas SS, Jakobson-Huston S, et al. Comparison of three ankle-foot orthoses configurations for children with spastic diplegia. Develop Med Child Neurol.
15. Radtka SA, Skinner SR, Dixon DM, et al. A comparison of gait with solid, dynamic, and no ankle-foot orthoses in children with spastic cerebral palsy. Phy Ther.
16. Abel MF, Juhl GA, Vaughan CL, et al. Gait Assessment of fixed ankle-foot orthoses in children with spastic diplegia. Arch Phy Med Rehabil.
17. Rethlefsen S, Kay R, Dennis S, et al. The effects of fixed and articulated ankle-foot orthoses on gait patterns in subjects with cerebral palsy. J Pediatr Ortho.
18. Smiley SJ, Jacobsen FS, Mielke C, et al. A comparison of the effects of solid, articulated, and posterior leaf-spring ankle-foot orthoses and shoes alone on gait and energy expenditure in children with spastic diplegic cerebral palsy. Orthopedics.
19. White H, Jenkins J, Neace WP, et al. Clinically prescribed orthoses demonstrate an increase in velocity of gait in children with cerebral palsy: a retrospective study. Develop Med Child Neuro.
20. Romkes J, Rrunner R. Comparison of a dynamic and a hinged ankle-foot orthoses by gait analysis in patients with hemiplegic cerebral palsy. Gait Posture.
21. Cutlip RG, Mancinelli C, Huber F, et al. Evaluation of an instrument walkway for measurement of the kinematic parameters of gait. Gait Posture.
22. Bilney B, Morris M, Webster K. Concurrent related validity of the GAITRite® walkway system for quantification of the spatial and temporal parameters of gait. Gait Posture.
23. Webster KE, Wittwer JE, Feller JA. Validity of the GAITRite® walkway system for the measurement of averaged and individual step parameters in gait. Gait Posture.
24. Menz HB, Latt MD, Tiedemann A, et al. Reliability of the GAITRite walkway system for the quantification of temporospatial parameters of gait in young and older people. Gait & Posture
25. van Uden CJT, Besser MP. Test-retest reliability of temporal and spatial gait characteristics measured with an instrumented walkway system (GAITRite®). BMC Musculoskelet Disord.
26. Thorpe DE, Dusing SC, Moore CG. Repeatability of temporalspatial gait measures in children using the GAITRite electronic walkway. Arch Phys Med Rehabil.
27. Palisano R, Rosenbaum P, Walter S, et al. Gross motor function classification system. Dev Med Child Neurol.
28. Domholdt E. Measurement theory. In: Rehabilitation Research: Principles and Applications
. 3rd ed. St. Louis, MO: Elsevier Saunders; 2005.
29. Safrit MJ, Wood TM. Measurement Concepts in Physical Education and Exercise Science.
Champaign, IL: Human Kinetics; 1989.
30. Baumgartner TA, Jackson AS, Mahar MT, et al. Measurement for Evaluation in Physical Education and Exercise Science.
7th ed. New York, NY: McGraw-Hill; 2003:94–95.
31. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull.
32. Whittle MW. Gait Analysis.
3rd ed. Edinburgh, UK: Elsevier Science; 2003.
33. Sutherland DH, Olshen RA, Biden EN, et al. The Development of Mature Walking.
London, UK: Mac Keith Press; 1988:59–61.