Secondary Logo

Journal Logo

Clinical Gait Measures for Ambulatory Children with Cerebral Palsy: A Review

Gilbertson, Torey J. PT, PhD; Bjornson, Kristie F. PT, PhD; McDonald, Cody CPO; Hafner, Brian J. PhD

Journal of Prosthetics and Orthotics: January 2016 - Volume 28 - Issue 1 - p 2–12
doi: 10.1097/JPO.0000000000000080

Purpose Valid and reliable measures of walking ability are needed to document the effectiveness of orthotic and therapeutic interventions for ambulatory children with cerebral palsy (CP). Selection of measures suited to evaluating children with CP can be informed by evidence of a measure's clinical utility and psychometric performance in the population of interest. The purpose of this review was to identify clinical measures that have been used to evaluate gait of children with CP, to review measures' evidence of psychometric testing, and to determine which measures are most appropriate for use with children with CP in a clinical setting.

Materials and Methods PubMed, Web of Science, and PsycINFO databases were searched for measures with evidence of psychometric testing in children with CP.

Results Eleven measures suitable for clinical administration were identified across 22 articles. Three measures had sufficient psychometric evidence to support clinical use in children with CP.

Conclusions The 6-minute walk test, 10-meter walk test, and the Edinburgh Gait Scale are recommended for directly measuring walking capacity, speed, and rating walking form, respectively. Results of this review suggest that, although a range of measures exist for evaluating pediatric gait, additional work is needed to assess and document measures' psychometric properties in children with CP.

TOREY J. GILBERTSON, PT, PhD; KRISTIE F. BJORNSON, PT, PhD; CODY MCDONALD, CPO; and BRIAN J. HAFNER, PhD, are affiliated with the Department of Rehabilitation Medicine, University of Washington Seattle, Washington.

KRISTIE BJORNSON, PT, PhD, is affiliated with the Department of Pediatrics, University of Washington; and the Seattle Children's Research Institute, Seattle, Washington.

Disclosure: The authors declare no conflict of interest.

Correspondence to: Torey J. Gilbertson, PT, PhD, Department of Rehabilitation Medicine, University of Washington, 1959 NE Pacific St., Box 356490, Seattle, WA 98195; email:

Cerebral palsy (CP) is a group of permanent movement and posture disorders attributed to nonprogressive damage to the fetal or infant brain that affects three to four of every 1000 school-aged children in the United States.1,2 Although persons with CP may present with a variety of motor impairments, CP is typically categorized by the dominant movement disorder (e.g., spasticity, ataxia, dystonia, or athetosis), distribution of affected body parts (e.g., one, two, three, or four limbs), and/or topography of impairment (e.g., bilateral lower limbs, arm and leg on one side).1 Spasticity is common in children with CP and often impairs basic movements such as walking.3,4

Walking is a fundamental motor skill that enables participation in numerous daily activities such as play and interaction with siblings or peers at home, school, and within the community.5 As such, interventions intended to positively affect walking ability are key modalities for members of an interdisciplinary rehabilitation team (e.g., orthotists and therapists).6 Ankle-foot orthoses (AFOs), for example, have the potential to improve walking ability and gait efficiency in ambulatory children with CP.7 However, AFOs also may impair aspects of walking ability as well as transitional gross motor skills (e.g., floor-to-stand) by restricting ankle joint motion. The ability to accurately measure children's walking ability before, during, and after intervention is therefore essential to effective evaluation, planning, and treatment.

Use of standardized outcome instruments is critical for accurate assessment and documentation of a variety of treatment outcomes, including gait, both within and across treatment sessions.8–10 Standardized instruments selected to assess patient outcomes need to have sufficient evidence of psychometric testing (e.g., reliability, validity, and responsiveness), and evidence of the use of measures in specific populations or diagnostic groups is essential.10 Furthermore, information gained from measures administered in a clinic is often generalized to environments a child may encounter (e.g., home, school, and community) outside of the clinic; however, evidence of this relationship is yet to be documented in children with CP. However, there is literature that suggests clinical walking tests can predict community-based walking activity in persons poststroke.11 Thus, measures used with children with CP require review to determine if any clinical measures possess characteristics from which ambulation in community or daily life can be predicted, as ambulation beyond the clinic should be the goal of orthotic and therapeutic intervention.

Tests and standardized measures that are often used to assess gait in children with CP include measures of body segment motions (e.g., joint kinematics) or other spatiotemporal outcomes (e.g., stride length, velocity, and cadence). Many measures are intended for administration in a laboratory environment. Fewer are suited to clinical administration where space, equipment, and personnel are generally limited. For most clinicians, selection of measures will be restricted to those that can be administered in clinical settings (e.g., examination rooms, hallways, or training areas). Thus, practical aspects of the measure, such as required space, necessary equipment, administration time, and patient time, will likely be important to consider in the selection process.

Instruments used to evaluate pediatric patients must have sufficient evidence of psychometric testing (e.g., reliability, validity, and responsiveness) and use in specific populations or diagnostic groups (e.g., children with CP).10 The intended application of the measure will often dictate the psychometric properties of importance. For example, if an instrument is to be used to monitor a patient's progress over time, evidence of test-retest reliability will be of key importance. Similarly, if the measure is intended to be used by different members of a rehabilitation team (i.e., both the therapist and orthotist), evidence of interrater reliability will be important. Thus, to optimally evaluate longitudinal changes in gait function, measures are desired that are both suited to clinical settings and possess sound psychometric properties in the population of interest.

Recent reviews of instruments suited to evaluating adults and children with neurological disorders underscore the need for information about outcome measures' psychometric quality and clinical utility.12,13 Recent reviews have examined psychometric properties of observational12 and performance-based14 measures suited to evaluating gait in children with a variety of neurological disorders. Another recent review by Seale13 identified instruments suited to measuring balance in pediatric and adult AFO users. However, to our knowledge, a review of instruments' validity, reliability, and responsiveness specific to children with CP does not exist.

The purpose of this review was therefore to identify and appraise, based on evidence of reliability and validity with this specific patient population, clinical instruments that may be appropriate for measuring walking ability in children with CP. The primary aims were to 1) inform clinicians about measures suited to evaluating walking capacity and/or joint movement characteristics in clinical practice and 2) discuss situations where use of these measures may (or may not) be considered best practice for evaluating the effect of interventions (e.g., orthoses) to enhance walking in children with CP.

Back to Top | Article Outline


PubMed, Web of Science, and PsycINFO scientific databases were searched to identify publications that described outcome measures suitable for assessing walking ability and specifically examined at least one aspect of reliability or validity in children with CP, as more relevant clinical outcomes such as minimal detectable difference (MDD) or minimal detectable change (MDC), is derived from reliability. These databases were chosen because they index journals that have historically published research evidence related to children with CP. Search terms related to the population and activity of interest, as well as evidence of psychometric testing, were used to locate candidate publications. The population of interest was identified using the phrase “cerebral palsy” and the acronym “CP.” To identify publications that described the activity of interest, the root terms “gait,” “walk*,” and “ambulat*” were selected. Lastly, to identify publications that presented new evidence of psychometric testing, the root terms “psychometric,” “valid*,” and “reliab*” were chosen. Phrases and key words were combined to yield a single search strategy that was used across the selected databases: (“cerebral palsyor CP) and (gait or walk* or ambulat*) and (valid* or reliab* or psychometric*). The authors elected to confine the search to English literature published between January 1, 2003 and the original search date of December 31, 2013 so as to identify outcome measures used in the most recent decade. A priori inclusion and exclusion criteria (Table 1) were defined to identify relevant publications.

Table 1

Table 1

Titles and abstracts were screened to identify candidate publications and eliminate those that clearly did not meet the described inclusion and exclusion criteria. The full text of each remaining publication was then reviewed using the specified selection criteria to identify articles that were ultimately included in this review. As part of the evaluation, pearling (inspection) of the references was conducted and appropriate articles were included. A separate search of each measure identified also was conducted for potential inclusion and facilitating discussion of the measure's clinical utility based on the type of instrument. For example, a search using “Edinburgh and gait” was conducted in each database once the Edinburgh Gait Scale (EGS) was identified and included in this review. Identified measures were grouped into one of two categories: “direct measurement” or “observed rating,” based on the type of instrument. Direct measurements included instruments designed to assess distance walked over a set amount of time (e.g., 6 minutes) or time required to walk a set distance (e.g., 10 m). Observed rating measures included clinical tools developed to describe joint or body segment position at various phases of gait via visual observation.

Elements extracted from each publication included name of the described outcome measure(s), number of participants and/or raters, participants' ages (range and mean), percentage of males, Gross Motor Function Classification System (GMFCS)15,16 level of participants, distribution of involvement, other participant diagnoses, test-retest reliability, interrater reliability, intrarater reliability, and evidence of validity. Relative strength of each measure's psychometric properties was determined using rating scales defined by Portney and Watkins10 (e.g., Intraclass Correlation Coefficient [ICC] values above 0.75 are indicative of good reliability). The research design (e.g., randomized controlled trial, case study) and statistical method (e.g., ICC, percentage of agreement) were also identified during appraisal. Recommendations for clinical use of measures in each category were made by the review authors based on the methodological aspects of the study (i.e., number of children with CP involved in the reviewed studies, the research designs, the quality of the reported validity and reliability testing, and the discussion of the measures by the study investigators).

Back to Top | Article Outline


The search yielded a total of 22 articles and 11 different measures (Figure 1). Identified direct measurements included the 1-minute walk test (1MWT), the 10-m walk test (10MWT) or 10-m fast walk test (10MFWT), the 6-minute walk test (6MWT), the 10-minute walk test (10MWT), and the timed up and go (TUG). Observed rating measures identified included the EGS, observational gait analysis (OGA), Observational Gait Scale (OGS), Physician Rating Scale (PRS), Salford Gait Tool (SF-GT), and Visual Gait Assessment Scale (VGAS). Brief descriptions of each measure, including scoring instructions, were created (Table 2). A summary of published evidence, including populations studied, reported psychometric properties, and notes on clinical applications was similarly developed to facilitate critical review of the measures included in this review (Table 3).

Figure 1

Figure 1

Table 2

Table 2

Table 3

Table 3

Back to Top | Article Outline


McDowell et al.17 compared the 1MWT to Gross Motor Function Measure (GMFM) scores in 34 children. The GMFM is an evaluative measure with five dimensions: lying and rolling, sitting, crawling and kneeling, standing and walking, and running and jumping.18,19 Distance on the 1MWT was significantly correlated with GMFM scores (Pearson r = 0.92, P < 0.001).17 This evidence supports convergent validity. The investigators reported significant differences between 1MWT distances and GMFCS levels (P < 0.001) with differences between those in levels II and III (P = 0.003) and levels III and IV (P = 0.008), but not significant between levels I and II (P = 0.42). This indicates good evidence of discriminant validity for the 1MWT. In addition, McDowell et al.20 reported good test-retest reliability (ICC = 0.94) with the 1MWT over a 1-week period in 17 children.

The 1MWT was reported by Chong et al.21 to have good evidence of convergent validity through strong correlations between both the GMFCS (Pearson r = −0.75) and ABILOCO-kids questionnaire of locomotion ability (Pearson r = 0.70) using 60 children. Chong et al. also reported good concurrent validity via a strong correlations to the 6MWT (Pearson r = 0.81). In the three studies mentioned previously, children wore their typical orthoses during walking, although the number of children who wore these was not reported.

Kerr et al.22 examined 46 children (22 who wore solid or dynamic AFOs) and reported a correlation between the 1MWT distances and oxygen consumption (r2 = 0.477; P < 0.001). Through evidence of strong concurrent validity, the 1MWT is not valid for assessing energy consumption.

Back to Top | Article Outline


Test-rest reliability of the 10MFWT was examined by Thompson et al.23 on 31 children, who found very strong correlation of ICC (2,1) = 0.81 at a mean of 10.6 days (range, 7–31 days). The test-retest reliability was inversely correlated with function (0.59 for GMFCS level I, 0.70 for level II, and 0.78 for level III). Typically used assistive devices and orthoses were used during testing and were consistent between trials.

Back to Top | Article Outline


The reliability and validity of the 6MWT was also assessed by Thompson et al.23 A 1 to 2 week test-retest ICC (2,1) of 0.98 was found for the same 31 children as reported with the 10MFWT. Differences in 6MWT distances among GMFCS levels (I–III) using a paired t-test with correction for multiple comparisons (P < 0.015) were reported. This demonstrates good discriminant validity. As in the 10MWT, Thompson et al. allowed children to wear their typical orthoses, but did not report number of children who used orthoses in their study.

Maher et al.,24 in a study of 41 older children (eight who used AFOs of various types and three who used in-shoe orthoses), reported an ICC of 0.98 for 30-minute test-retest reliability. Similarly, Nsenga Leunkeu et al.25 studied 1-week test-retest reliability and reported an ICC of 0.87 on 12 older children. They also studied 24 older children to compare 6MWT performance to energy expenditure. VO2 peak was significantly correlated with 6MWT (Pearson r = 0.95, P < 0.001), supporting convergent validity. Children in the Nsenga Leunkeu et al.25 study were not allowed to use orthoses or assistive devices to “avoid augmenting the energy cost of exercise.” Good convergent validity between the 6MWT and both GMFCS levels (Pearson r = −0.75) and ABILOCO-kids questionnaire (Pearson r = 0.70) was also reported by Chong et al.,21 who allowed children in their study to wear their typical orthoses.

Back to Top | Article Outline


In a study of 7-day, test-retest reliability in 29 children with either CP or meningomyelocele, Pirpiris et al.26 reported an ICC of 0.91 (95% confidence interval, 0.77–0.99). The investigators also noted a significant difference (P < 0.01) between 10MWT walking speed as compared with a multi-trial average speed of 10-m walks evaluated using three-dimensional gait analysis (3DGA) (Pearson r = 0.65), suggesting poor criterion validity. Because of the different distances walked by children in these tests, the investigators hypothesized that the 10MWT may be a better indicator of community walking speed than the 10-m walking test. In this study, children who were obligatory users of orthoses were allowed to wear them for instrumented gait analysis; otherwise, children were barefoot. Comparisons were only made if orthoses or assistive devices were consistent between trials.

Back to Top | Article Outline


The TUG was designed to assess mobility and dynamic balance in various populations.27 It measures the time required for an individual to rise from a seated position, walk 3 m, turn around, and return to a seated position. Williams et al.28 examined the TUG in a study with 41 children, 33 of whom had CP. Williams et al. conducted two initial trials (time 1) and two more trials after 10 to 20 minutes (time 2) and documented immediate test-retest reliabilities (ICC [1,1] = 0.98 within both time 1 and 2) and same-day retest between time 1 and 2 (ICC [1,3] = 0.99).28 Williams et al. found significant differences (P = 0.01) in TUG times between levels (I–III) using a Kruskal-Wallis analysis of variance (ANOVA), which suggests evidence of discriminant validity. Dhote et al.29 also studied test-retest reliability of the TUG in 30 children. They performed both 30-minute and 1-week retests and reported ICCs of 0.99 for each. In both of the aforementioned studies, children were allowed to wear orthoses if typically worn.

Back to Top | Article Outline


Edinburgh Gait Scale

The EGS (also known as the Edinburgh Visual Gait Analysis Interval Test [GAIT]) was developed as an OGA tool. It is composed of 17 ratings of joint position (i.e., 0 points for normal, 1 point for moderate flexion or extension, and 2 points for marked deviation into flexion or extension) for each lower limb evaluated on six anatomical sites: trunk, pelvis, hip, knee, ankle, and foot for a theoretical total of 34 points.30 The parameters are analyzed in both sagittal and frontal planes through videotape observation. Bella et al.31 compared the EGS to an OGS32 and the VGAS33 in eight children. There was high interrater agreement (κ = 0.47–1.00) among ratings made using all three scales (i.e., OGS, VGAS, and EGS). Most intrarater combinations had moderate to excellent weighted κ index frequencies (κ = 0.20–0.81). The reliability of the EGS using 10 videos of children was studied by Viehweger et al.34 They found moderate to very strong intrarater reliability (ICC = 0.59–0.96) and weak to strong interrater agreement (21.4%–67.9%) among eight raters with varying degrees of clinical and gait observation experience. Better agreement was found between items in the stance phase than those in the swing phase. Greater correlations with increasing clinical and gait analysis experience suggested that those with more experience often yielded more consistent scores.34 In both the Bella et al. and Viehweger et al. studies, children walked barefoot for analysis.

Hillman et al.35 compared the EGS to the Gillette Gait Index (GGI) and the Gillette Functional Assessment Questionnaire (FAQ) in 58 children and adults walking barefoot, 25 of whom were analyzed a second time wearing their typical orthoses. The FAQ is a parent-reported measure, so was not included as part of this review. The correlation between the EGS and GGI was very strong (Spearman r = 0.83–0.89), documenting criterion validity as the GGI was developed from the criterion standard 3DGA. The correlation between the EGS and the FAQ was lower (Spearman r = −0.52). The observed inverse relationship was expected as higher FAQ scores indicate higher functional walking and higher EGS scores show higher deviation from “normal,” suggestive of evidence of convergent validity.

In a comparison in five children by Read et al.,30 the EGS and 3DGA had varying levels of agreement between the different joints (47%–83% agreement; mean, 64%). Intrarater reliability showed least significant difference (LSD) differences between 2.63 and 4.01 (mean, 3.20). Interrater agreement among five clinicians ranged from 55% to 96% (mean, 70%). Read et al.30 did not specify whether their data were collected in barefoot or in orthoses, although use of 3DGA suggests barefoot.

Back to Top | Article Outline

Observational Gait Analysis

The OGA consists of 10 items evaluating general positioning of the pelvis, hip, knees, and ankle at various times during the gait cycle and in all three planes of motion.36 Kawamura et al.36 assessed 10 periods in the gait cycle using OGA in 50 older children who walked without use of AFOs. Four different raters evaluated position (joint angle) of the hip, knee, and ankle as primary interest. Very weak to strong agreement (κ = 0.01–0.65) was reported between OGA and 3DGA. Joint and period of gait influenced interrater agreement, which was weak to very strong (κ = 0.25–0.88). Between OGA and 3DGA, the highest agreement (κ = 0.21–0.65) was reported for the knee joint angle at three different periods of gait (initial contact, terminal stance, and initial swing). Criterion validity for OGA seems to be dependent upon rater and joint.

Back to Top | Article Outline

Physician Rating Scale

The PRS was initially developed for evaluation of gait in children post botulinum toxin A injection.37 The PRS emphasizes position of the foot and knee in stance with six categorical ratings. Wren et al.38 modified the PRS to include hip rating and greater specify the rating of knee and ankle positions in a study of 30 children and young adults. Ratings of joint positions from four raters were compared with positions simultaneously collected with 3DGA. Differences between visual rating and 3DGA were found to be significant for all joints (P < 0.0001) indicating poor criterion validity. The investigators also assessed interrater and intrarater reliability among four raters using live observations and full-speed and slow-motion video. Interrater agreement varied across joints with the highest agreement on crouch (κ = 0.71) and hip (κ = 0.40) using live observation. Slow-motion video agreement was highest for foot contact (κ = 0.74) and dorsiflexion (DF) (κ = 0.52). Agreement for the knee (κ = 0.65) was highest using full speed video. Intrarater agreement also varied (κw= 0.50–0.78 for foot contact, 0.71–0.80 for crouch, 0.26–0.44 for hip flexion, 0.60–0.86 for knee flexion, and 0.39–0.61 for DF). Although not specified, it is assumed based on the scoring criteria, children did not wear AFOs for this study.

Back to Top | Article Outline

Observational Gait Scale

The OGS is a modification of the PRS which includes eight sections evaluated in frontal and sagittal planes (22 points per leg)39 and places a higher emphasis on the foot and knee during stance than the PRS.37 Mackey et al.37 studied the reliability of the OGS in 20 children walking barefoot and found moderate intrarater agreement (κw = 0.30–0.86) of scores across joints was slightly higher than interrater agreement (mean κ = 0.69 vs. 0.62).37 Good criterion validity was documented by moderate to high agreement between OGS and 3DGA (κw = 0.38–0.98).

Borel et al.40 used the OGS in 12 children, assessing 20 gait videos. Traditional visual rating of OGS was compared with ratings using Dartfish41 analysis software. Interrater reliabilities with visual rating (κ = 0.16–0.80) and using Dartfish (κ = 0.35–0.85) were reported. Thus, the analysis software was found not only to be more efficient in terms of time during assessment, but also more reliable. Similar to the Read et al.30 study, Borel et al. did not specify that their participants were barefoot, but it is assumed based on description.

Back to Top | Article Outline

Visual Gait Assessment Scale

The VGAS is also adapted from the PRS.42 It consists of seven parameters (hip and knee position in terminal stance and mid-swing, initial foot contact, foot contact in stance, and timing of heel rise). Two raters measured barefoot walking ability in 31 children, using the VGAS and 3DGA.33 Agreement between measurement methods ranged from κ = −0.11 to 0.51, depending on the parameter, making criterion validity very weak to moderate. Intrarater agreement (κ = −0.04 to 0.86) and interrater agreement (κ = 0.44–0.89) were also reported. Dickens and Smith33 noted the ratings at the hip were the least reliable. The VGAS and EGS were compared by Bella et al.31 in eight children during barefoot walking. Intrarater reliability among three raters ranged from κw = 0.54 to 1.00. In another study of four children in bare feet and 10 raters of differing experience, Brown et al.42 found Bland-Altman coefficients of intrarater reliability of 4.06 for experienced raters and 5.94 for inexperienced raters, suggesting the VGAS was more reliable with experienced observers. They also determined that more experienced raters generally had higher intrarater (50.0%–87.6% vs. 41.7%–83.3%) and interrater (45.8%–100.0% vs. 38.3%–91.7%) agreements on ratings. Agreements were higher for ratings at the ankle and foot than those at the knee or hip.

Back to Top | Article Outline

Salford Gait Tool

The SF-GT is a sagittal plane observational measure used to evaluate hip, knee, and ankle positioning at six specific events during the gait cycle (initial contact, end double support, mid-stance, start double support, toe-off, and mid-swing). Toro et al.43,44 compared the SF-GT to 3DGA and found LSDs between SF-GT and 3DGA varied from 1.41 to 27.14 degrees. This suggests moderate to low evidence of criterion validity. Those investigators also examined interrater and intrarater reliability of SF-GT ratings with 13 children and 23 raters evaluating videos of walking. The interrater reliability was high in nearly all instances with reported agreement at or over 63% in nearly every joint and position rating with means of 77% at the hip, 81% at the knee, and 75% at the ankle (though lower in children with severe crouch [56%]). Mean intrarater agreement was reported as 72% for the hip, 78% for the knee, and 73% for the ankle (overall agreement between 51% and 90%).44 Neither of these studies specified barefoot walking conditions, but this should be assumed based on the ankle scoring criteria.

Back to Top | Article Outline


Selection of measures to evaluate the effectiveness of an intervention is essential for orthotists and therapists to assess and document changes in patient function. The goal of this review was to identify and discuss those measures most suited to evaluating walking performance and quality of movement in children with CP in a clinical setting. Review and analysis of clinical utility and psychometric performance of measures in children with CP revealed that the 6MWT and EGS are well suited to measurement of children with CP. However, other measures may be appropriate for use in select situations. Recommendations for use of each measure are discussed later.

Back to Top | Article Outline


Of tools classified as direct measurements, the 6MWT bears the most substantial documented evidence of reliability and validity in the measurement of walking capacity of children with CP compared with other measures and is therefore recommended for most children with CP. Three studies, with sample sizes of 24 to 41 and similar levels of test-retest reliability, suggest the 6MWT as a highly reliable test for longitudinal assessments.23–25 The Thompson et al.23 study allows for generalization to other clinicians based on ICC model used (2,1). The high correlation with VO2 also suggest that the 6MWT may be an effective measure of energy expenditure25 and could therefore also be used to measure general physical functioning. Lastly, there is also evidence of discriminant validity between the 6MWT and GMFCS level.21 Although the 6MWT has better psychometric properties than the 1MWT, children who are not able to walk for 6 minutes due to physical or cognitive functioning may be better suited for evaluation using the 1MWT.

The 1MWT test, while clinically friendly, should be used with caution as test-retest reliability was assessed in only one study and an unreported ICC model was used.20 Chong et al.21 found a high correlation between the 1MWT and the 6MWT in children with CP, suggesting that the 1MWT may be as useful for assessing walking capacity, but requires less time to administer. However, further research is needed to substantiate this claim. The correlation between the 1MWT and O2 consumption was low, so may not be a sound proxy for energy expenditure.22 Studies showed positive relationships to the 1MWT with GMFM scores17 and GMFCS level.21

Based on this review, the 10MFWT (or 10MWT) is a psychometrically appropriate measure for evaluation of speed of ambulation over short distances (e.g., within a child's house) and had good test-retest reliability with the ICC model used (2,1) suggesting generalizability to other clinicians.23 This test was also strongly related to physical functioning based on the GMFCS level, especially for those at GMFCS level III. Using ANOVA instead of the paired t-test, as deemed more appropriate by these authors, may have yielded different results and altered the recommendation for use. Although the 10MFWT shows promise for use in children with CP, caution should be taken with its use until further evidence is available as psychometric evidence was derived from only one study.

The published 10MWT literature suggests evidence of good test-retest reliability, using an ICC model that supports generalizability.26 However, the 10MWT may not be appropriate for children functioning at a lower physical level (e.g., GMFCS III) due to the difficulties of sustaining walking for 10 minutes.26 Based on the current evidence for children, the 10MWT is not recommended for clinical use at this time as it requires more time to administer than the 6MWT and does not seem to possess superior psychometric properties.

The TUG has been studied frequently in children with CP, although only two studies are more recent. These studies included large sample sizes and employed ICC models (i.e., 1,1 and 1,3) suited to generalizability of the results.29 The TUG showed promising evidence of test-retest reliability with assessments occurring on the same day28 and a week apart.29 The TUG is recommended as a longitudinal assessment tool due to test-retest reliability over both short and longer periods. Because the TUG requires a child to stand up, sit down, and turn 180 degrees, it is not solely a measure of ambulation. However, it may be a useful and reliable tool to assess general mobility.

The majority of the studies of direct measurement allowed children to wear AFOs if typically worn. The lone exception was the Nsenga Leunkeu et al. study which required barefoot walking due to the energy expenditure. Based on data by Brehm et al.,45 children walk with greater efficiency while in AFOs. Thus, the publications reviewed here are most likely reflective of the children's best performance. None of these publications reported MDD.

Back to Top | Article Outline


Of the observed rating measures reviewed, the EGS seems to be the most appropriate tool for visual kinematic assessment of gait in a clinical setting. This is in direct agreement with the findings by Rathinam et al.12 Psychometric testing of the EGS showed good intrarater reliability,30,31,34 the strongest interrater reliability of the measures reviewed,30,31,34 and good concurrent and criterion validity.30,35 All studies that examined the EGS had moderate to strong intrarater reliability. However, these studies included relatively small sample sizes (i.e., n = 4(30), 8(27), and 10(29)) and thus may limit generalizability of the reliability reported. Interrater reliability of the EGS varied and was generally lower than intrarater reliability. However, evidence of EGS interrater reliability was still stronger than the other observed rating measures reviewed. Both Read et al.30 and Viehweger et al.34 reported lowest agreement with knee position at terminal swing and with peak extension during stance, respectively, with other correlations lower at the knee than other joints. This may be due to vagueness in the instructions and the definition of extension when a contracture is present, so better definitions and training for knee ratings may improve reliability.

The EGS had the highest correlations reported for concurrent and criterion validity, suggesting it may most closely measure the underlying construct of gait. There did not seem to be any discrepancies between agreements in swing phase versus stance phase ratings, which makes this tool appropriate for evaluation of both phases of gait. Hillman et al.35 studied children in both barefoot walking and while wearing orthoses, but did not report separate correlations for these two conditions. Thus it is unknown whether reliability is altered with children wearing orthoses versus walking in bare feet.

Bella et al.31 reported that the EGS was the most user-friendly tool (i.e., easiest to understand and use) of the three they evaluated (EGS, VGAS, and Brazilian OGS). The EGS was also the only measure for which MDD was reported where Read et al.30 showed that a change in three units is likely indicative of true change. Lastly, the EGS has a greater number of joint position ratings than the other observational gait assessments found in this study, thus providing more information for the clinician. This greater number of ratings extends the time required to score the test to approximately 25 minutes.46 Unlike some of the other measures, the EGS also used both sagittal and coronal plane video, which may be of interest to clinicians who want to assess coronal plane aspects of gait (i.e., hind foot valgus, foot rotation, and lateral trunk shift). Through the use of both planes of video and the recommended use of rotation blocks, movement in the transverse plane is also assessed. After training, the high intrarater agreement indicates the EGS could be useful for longitudinal assessment of children with CP.

The OGA was reviewed in only one study, though was suggested that it had slightly better interrater reliability than the EGS. It was less correlated to the criterion standard of 3DGA and may therefore not as well reflect the measurement of gait kinematics.36 Because the OGA was only reported in one study over the period of this review, these authors feel its use is not well justified for clinical applications when compared with the EGS.

Studies examining the psychometric properties of the OGS reported mixed results.37,40 Although the OGS was highly correlated to the criterion standard of 3DGA, the intrarater and interrater reliabilities were quite mixed, depending on the rated joint. Implications for use of this instrument include, for example, that it may not reliably measure base of support or hind-foot position,37 which could be important with interventions that target hip abduction or evaluate use of foot orthoses. This measure was reported to have increased efficiency and reliability when used with companion analysis software (e.g., Dartfish).40 Thus, it is recommended to use analysis software for improved reliability when using the OGS clinically. Time for administration of this measure was not reported.

The SF-GT shows promise as a measure for children with CP, but the single study and reliability statistic (i.e., percent agreement) used weakens the interpretation of repeatability. The study included a fairly small n (i.e., 13 videos), but a high number of raters (i.e., 23 therapists).44 This number of raters may make the interrater reliability agreements more generalizable despite the small sample of source videos. The wide variation between SF-GT scores and 3DGA may suggest the SF-GT is slightly less robust of an assessment than the EGS. However, the SF-GT may be more useful than the EGS if sagittal-plane kinematics is of interest (e.g., hip or knee joint position beyond mid-stance or mid-swing phases of gait).

Psychometric testing of the VGAS across multiple studies revealed lower interrater and intrarater reliability than the EGS.31,33,42 Similarly, the evidence of validity was less convincing than that identified for the EGS. The VGAS was the only measure to have been used exclusively on children with hemiplegia with both the Dickens and Smith33 and Brown et al.42 studies. Bella et al.31 stated the VGAS was less user-friendly than the EGS. Therefore, the VGAS may not be clinically appropriate as a longitudinal measure of gait kinematics in children with CP.

Choosing the correct assessment tool is directly related to the goals established and the focus of rehabilitation strategy (e.g., therapeutic exercise or orthoses). Time- or distance-based measures, specifically the 6MWT, are recommended for assessing walking capacity or speed, although they are unable to provide information on movement quality. If gait quality is a goal and focus of intervention, then use of observed rating measures, and specifically the EGS, is recommended. The EGS, for example, could be used to assess heel position during initial contact and document changes associated with orthoses or therapeutic intervention intended to improve ankle DF. While choice of an appropriate measure is ultimately at the discretion of the orthotist or therapist, the results of this review can be used to guide their selection.

The aforementioned suggestions for use of gait measures in children with CP were derived from careful analysis and review of published evidence within this review. However, the authors acknowledge that several limitations exist. First, this review was not a systematic review and, although rigor of the included publication was analyzed, publications were not scored using an established methodological appraisal tool. Extraction of data from the publications was performed by only one author (T.G.), which could bias the evidence selected for review. The evidence search also included only three databases and additional evidence may have been identified if other databases (e.g., CINAHL) were used. Lastly, sensitivity to change was rarely reported; therefore, recommendations were not based on this psychometric property.

Back to Top | Article Outline


Clinicians who provide care to ambulatory children with CP must often perform gait evaluations to develop, plan, and evaluate orthotic and/or therapeutic interventions. Numerous measures exist to objectively assess patients' walking ability. However, busy clinicians may be challenged to thoroughly review instruments and assess their applicability for children with CP. In this review, the authors identified instruments that may be used to measure walking ability, reviewed available evidence of reliability and validity, and provided suggestions for clinical instruments that may be most appropriate for measuring various aspects of walking in children with CP.

Based on this review, the 10MFWT seems to be a sound measure to assess peak gait speed in children with CP. Similarly, the 6MWT seems to be a useful direct measurement to assess walking ability over longer distances. The 6MWT is recommended over the 1MWT and 10MWT, primarily due to stronger available evidence of validity and reliability. At this time, the 1MWT may only be recommended for children with CP who may not be able to complete the 6MWT. Similarly, the 10MFWT and the 6MWT are recommended over the TUG when the primary goal is to measure walking ability, as the TUG requires transition activities.

The EGS seems to be the most psychometrically sound measure for clinical evaluation of gait quality in children with CP. It has been reported as the most user-friendly observed rating measure and has the highest reliability of the measures reviewed. The EGS closely correlates with the criterion standard of 3DGA and integrates both sagittal and coronal plane observations. Regardless of the observed rating measure used, the use of video recording is advised over real-time measurement. Supplementary assessment with companion analysis software is also recommended for enhanced reliability of the evaluation. Evidence suggests that reliability of any observational rating measure may be higher when performed by a consistent, experienced rater.

Of the reviewed gait measures, few have substantial documented psychometric evidence among children with CP. It is vital that measures intended for longitudinal use are capable of accurately reflecting true changes in patients' walking abilities over time. The reliability of any measure only reflects the population studied and may only be confidently used in children closely resembling those studied in terms of age, topographical distribution, movement disorder, and GMFCS level. Thus, further psychometric testing is needed to establish the acceptable evidence of reliability and validity of clinical gait measures for children with CP.

Back to Top | Article Outline


1. Rosenbaum P, Paneth N, Leviton A, et al. A report: the definition and classification of cerebral palsy April 2006. Dev Med Child Neurol Suppl 2007; 109: 8–14.
2. Aisen ML, Kerkovich D, Mast J, et al. Cerebral palsy: clinical care and neurological rehabilitation. Lancet Neurol 2011; 10( 9): 844–852.
3. Cans C, Guillem P, Baille F, et al. Surveillance of cerebral palsy in Europe: a collaboration of cerebral palsy surveys and registers. Dev Med Child Neurol 2000; 42( 12): 816–824.
4. Bar-On L, Molenaers G, Aertbeliën E, et al. The relation between spasticity and muscle behavior during the swing phase of gait in children with cerebral palsy. Res Dev Disabil 2014; 35( 12): 3354–3364.
5. Bult MK, Verschuren O, Jongmans MJ, et al. What influences participation in leisure activities of children and youth with physical disabilities? A systematic review. Res Dev Disabil 2011; 32( 5): 1521–1529.
6. Sorsdahl AB, Moe-Nilssen R, Strand LI. Test-retest reliability of spatial and temporal gait parameters in children with cerebral palsy as measured by an electronic walkway. Gait Posture 2008; 27( 1): 43–50.
7. Morris C, Bowers R, Ross K, et al. Orthotic management of cerebral palsy: recommendations from a consensus conference. NeuroRehabilitation 2011; 28( 1): 37–46.
8. Oeffinger D, Bagley A, Rogers S, et al. Outcome tools used for ambulatory children with cerebral palsy: responsiveness and minimum clinically important differences. Dev Med Child Neurol 2008; 50( 12): 918–925.
9. Finch E, Brooks D, Stratford PW, Mayo NE. Physical Rehabilitation Outcome Measures Second. Hamilton, Ontario, Canada: Lippincott Williams & Wilkins; 2002.
10. Portney LG, Watkins MP. Foundations of Clinical Research: Applications to Practice. Upper Saddle River, NJ: Pearson/Prentice Hall; 2009.
11. Mudge S, Stott NS. Timed walking tests correlate with daily step activity in persons with stroke. Arch Phys Med Rehabil 2009; 90( 2): 296–301.
12. Rathinam C, Bateman A, Peirson J, Skinner J. Observational gait assessment tools in paediatrics—a systematic review. Gait Posture 2014; 40( 2): 279–285.
13. Seale J. Valid and Reliable Instruments for the Clinical Assessment of the Effect of an Ankle-Foot Orthoses on Balance. J Prosthet Orthot 2010; 22( 10): P38–P45.
14. Ammann-Reiffer C, Bastiaenen CH, de Bie RA, van Hedel HJ. Measurement properties of gait-related outcomes in youth with neuromuscular diagnoses: a systematic review. Phys Ther 2014; 94( 8): 1067–1082.
15. Palisano R, Rosenbaum P, Walter S, et al. Development and reliability of a system to classify gross motor function in children with cerebral palsy. Dev Med Child Neurol 1997; 39( 4): 214–223.
16. Palisano RJ, Rosenbaum P, Bartlett D, Livingston MH. Content validity of the expanded and revised Gross Motor Function Classification System. Dev Med Child Neurol 2008; 50( 10): 744–750.
17. McDowell BC, Kerr C, Parkes J, Cosgrove A. Validity of a 1 minute walk test for children with cerebral palsy. Dev Med Child Neurol 2005; 47( 11): 744–748.
18. Russell DJ, Rosenbaum PL, Cadman DT, et al. The gross motor function measure: a means to evaluate the effects of physical therapy. Dev Med Child Neurol 1989; 31( 3): 341–352.
19. Russell DJ, Avery LM, Rosenbaum PL, et al. Improved scaling of the gross motor function measure for children with cerebral palsy: evidence of reliability and validity. Phys Ther 2000; 80( 9): 873–885.
20. McDowell BC, Humphreys L, Kerr C, Stevenson M. Test-retest reliability of a 1-min walk test in children with bilateral spastic cerebral palsy (BSCP). Gait Posture 2009; 29( 2): 267–269.
21. Chong J, Mackey AH, Broadbent E, Stott NS. Relationship between walk tests and parental reports of walking abilities in children with cerebral palsy. Arch Phys Med Rehabil 2011; 92( 2): 265–270.
22. Kerr C, McDowell BC, Cosgrove A. Oxygen cost versus a 1-minute walk test in a population of children with bilateral spastic cerebral palsy. J Pediatr Orthop 2007; 27( 3): 283–287.
23. Thompson P, Beath T, Bell J, et al. Test-retest reliability of the 10-metre fast walk test and 6-minute walk test in ambulatory school-aged children with cerebral palsy. Dev Med Child Neurol 2008; 50( 5): 370–376.
24. Maher CA, Williams MT, Olds TS. The six-minute walk test for children with cerebral palsy. Int J Rehabil Res 2008; 31( 2): 185–188.
25. Nsenga Leunkeu A, Shephard RJ, Ahmaidi S. Six-minute walk test in children with cerebral palsy gross motor function classification system levels I and II: reproducibility, validity, and training effects. Arch Phys Med Rehabil 2012; 93( 12): 2333–2339.
26. Pirpiris M, Wilkinson AJ, Rodda J, et al. Walking speed in children and young adults with neuromuscular disease: comparison between two assessment methods. J Pediatr Orthop 2003; 23( 3): 302–307.
27. Podsiadlo D, Richardson S. The timed “Up & Go”: a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc 1991; 39( 2): 142–148.
28. Williams EN, Carroll SG, Reddihough DS, et al. Investigation of the timed ‘up & go’ test in children. Dev Med Child Neurol 2005; 47( 8): 518–524.
29. Dhote SN, Khatri PA, Ganvir SS. Reliability of “Modified timed up and go” test in children with cerebral palsy. J Pediatr Neurosci 2012; 7( 2): 96–100.
30. Read HS, Hazlewood ME, Hillman SJ, et al. Edinburgh visual gait score for use in cerebral palsy. J Pediatr Orthop 2003; 23( 3): 296–301.
31. Bella GP, Rodrigues NB, Valenciano PJ, et al. Correlation among the visual gait assessment scale, Edinburgh visual gait scale and observational gait scale in children with spastic diplegic cerebral palsy. Rev Bras Fisioter 2012; 16( 2): 134–140.
32. Araujo PA, Kirkwood RN, Figueiredo EM. Validity and intra- and inter-rater reliability of the Observational Gait Scale for children with spastic cerebral palsy. Rev Bras Fisioter 2009; 13( 3): 267–273.
33. Dickens WE, Smith MF. Validation of a visual gait assessment scale for children with hemiplegic cerebral palsy. Gait Posture 2006; 23( 1): 78–82.
34. Viehweger E, Zürcher Pfund L, Hélix M, et al. Influence of clinical and gait analysis experience on reliability of observational gait analysis (Edinburgh Gait Score Reliability). Ann Phys Rehabil Med 2010; 53( 9): 535–546.
35. Hillman SJ, Hazlewood ME, Schwartz MH, et al. Correlation of the Edinburgh Gait Score with the Gillette Gait Index, the Gillette Functional Assessment Questionnaire, and dimensionless speed. J Pediatr Orthop 2007; 27( 1): 7–11.
36. Kawamura CM, de Morais Filho MC, Barreto MM, et al. Comparison between visual and three-dimensional gait analysis in patients with spastic diplegic cerebral palsy. Gait Posture 2007; 25( 1): 18–24.
37. 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( 1): 4–11.
38. Wren TA, Rethlefsen SA, Healy BS, et al. Reliability and validity of visual assessments of gait using a modified physician rating scale for crouch and foot contact. J Pediatr Orthop 2005; 25( 5): 646–650.
39. Boyd RN, Graham HK. Objective measurement of clinical findings in the use of botulinum toxin type A for the management of children with cerebral palsy. Eur J Neurol 1999; 6: S23–S35.
40. Borel S, Schneider P, Newman CJ. Video analysis software increases the interrater reliability of video gait assessments in children with cerebral palsy. Gait Posture 2011; 33( 4): 727–729.
41. Dartfish. Dartfish. Fribourg: Dartfish.
42. Brown CR, Hillman SJ, Richardson AM, et al. Reliability and validity of the Visual Gait Assessment Scale for children with hemiplegic cerebral palsy when used by experienced and inexperienced observers. Gait Posture 2008; 27( 4): 648–652.
43. 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( 3): 321–327.
44. 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( 3): 328–332.
45. Brehm MA, Harlaar J, Schwartz M. Effect of ankle-foot orthoses on walking efficiency and gait in children with cerebral palsy. J Rehabil Med 2008; 40( 7): 529–534.
46. Maathuis KG, van der Schans CP, van Iperen A, et al. Gait in children with cerebral palsy: observer reliability of Physician Rating Scale and Edinburgh Visual Gait Analysis Interval Testing scale. J Pediatr Orthop 2005; 25( 3): 268–272.

cerebral palsy; children; gait; outcome measures; psychometric properties; evaluation

© 2016 by the American Academy of Orthotists and Prosthetists.