Reliability and Validity of the Shaw Gait Assessment Tool for Temporospatial Gait Assessment in People With Hemiparesis
Presented to the American Physical Therapy Association, June 10–13, 2009, Baltimore, MD.
,
,
Abstract
Reid S, Held JM, Lawrence S. Reliability and validity of the Shaw gait assessment tool for temporospatial gait assessment in people with hemiparesis.
Objective
To assess the intra-/interrater reliability and the validity of the free web-based Shaw Gait Assessment Tool (with visual and numerical output) for assessing speed, cadence, step length, and limb advance time in people with hemiplegic gait.
Design
Intra-/interrater reliability and concurrent validity with 2 raters using the Shaw Gait Assessment Tool and 1 rater using a multimemory stopwatch.
Setting
Busy outpatient rehabilitation gym at a tertiary care medical center.
Participants
Convenience sample of adults with hemiplegic gait after cerebrovascular accident or traumatic brain injury.
Interventions
Not applicable.
Main Outcome Measures
Intraclass correlation coefficients (ICCs) and Pearson product-moment correlation coefficients.
Results
ICCs for intrarater reliability ranged from 0.94 (95% CI, 0.88–0.97) to 0.98 (95% CI, 0.96–0.99), (P<.001), and for interrater reliability from 0.95 (95% CI, 0.88–0.98) to 0.99 (95% CI, 0.99–0.99), (P<.001). The Shaw Gait Assessment Tool correlated with the stopwatch for all measured gait parameters with Pearson product-moment correlation coefficients (range, r=0.95 to r= 0.99, P<.001).
Conclusions
The Shaw Gait Assessment Tool is a free, easy-to-use tool that gives reliable and valid results for 4 temporospatial parameters of hemiplegic gait.
GAIT IMPAIRMENTS AND deviations resulting from brain injury are common and often lead to disability.1, 2, 3 The ability to walk has been rated as one of the most important goals of rehabilitation by patients with stroke.4 Gait assessment and treatment are integral components of the rehabilitation process from both the clinicians' and patients' viewpoints. PTs are well trained to evaluate and treat gait deficits; however, they need reliable, valid, and efficient methods of measuring gait.
Gait has been examined in numerous ways. On one end of the spectrum, there is the very basic observational gait analysis,5 which is probably the most commonly used clinical tool, simply because there is no equipment required. On the other end of the spectrum, there are motion and gait analysis laboratories, which can include video monitoring, electromyogram, and ground reaction force assessments.5 Observational gait analysis does not give quantifiable temporospatial information, and with the exception of push-off, has been shown to be only moderately reliable.6, 7 Motion analysis laboratories provide detailed information on many aspects of gait, but they tend to be located in university settings or specialized research hospitals and are therefore not available to most clinicians.
Over the years, a variety of tools have been developed for clinical use. The stopwatch is an affordable tool for measuring gait speed and, if the stopwatch has multimemory capability, cadence, average stride length.8 For any temporospatial aspect other than speed, post hoc calculations are required, which are time consuming, not feasible in most busy clinical environments, and do not provide immediate visual feedback for the patient. Researchers have used video cameras with a grid,9, 10 in conjunction with a personal computer,11 or ink footprints and a stopwatch12, 13 to record gait and measure various temporospatial parameters. These systems, although affordable, also require time-consuming post hoc analysis. Tools like the Stride Analyzer or the GAITRite System remain too costly for many clinics or hospitals.
Based on a recent survey, 1826 physiotherapists in the United Kingdom indicated that management of abnormal gait constituted a major aspect of physiotherapy practice, yet there was no systematic use of standardized tools.14 Over 50% of the therapists in the survey had not received formal training in gait assessment. Of the therapists who responded, the 5 most frequently stated features of an ideal gait assessment tool were that it be: (1) easy to use, (2) quick to use, (3) reliable, (4) clear and easy to understand, and (5) valid. Operationally, the tool is easy to use and quickly translated. It takes 10 to 20 minutes to administer the assessment and obtain results.14
To meet the need, a PT clinician developed the Shaw Gait Assessment Toola in 2001. It is a free, web-based temporospatial gait assessment tool, which is accessible by anyone with Internet access, and includes simple to follow directions and a tutorial for new users. Users can choose either English or metric measurements and can also specify the length of the testing walk. The data are immediately available in a graphical format for the clinician to review with the patient and are norm referenced for speed, cadence, and step length (fig 1). To our knowledge, no other freely available gait assessment tool gives such information instantaneously. This tool has been tested on a healthy adult population and found to have acceptable reliability and validity for the temporospatial measures of speed, cadence, step length, and limb advance time (an indirect measure of symmetry).15 The tool has not been tested on people with gait impairments.
Fig 1
Sample printout from the Shaw Gait Assessment Tool. The dark section represents the patient's performance as compared to the paler age and sex-based norms.
The purposes of this study were to assess the intra- and interrater reliability and the concurrent validity of the Shaw Gait Assessment Tool for speed, cadence, step length, and right/left limb advance time in people with hemiplegic gait. Because this was a clinical setting, we wanted to use commonly accepted clinical measures, and the multimemory stopwatch offered us the ability to test for concurrent validity. Based on the work of Youdas et al,8 the stopwatch is reliable and valid for speed, cadence, step length, and limb advance time in people with impairments, while Handa et al16 have found the stopwatch to be reliable and valid for these measures in a healthy population.
Our hypotheses were that the Shaw Gait Assessment Tool would demonstrate good inter- and intrarater reliability and good concurrent validity with the stopwatch in people with hemiparesis.
Methods
Participants
Participants were recruited from the inpatient and outpatient populations of our rehabilitation center. Inclusion criteria were as follows: men or women aged 18 years or older, acute or chronic hemiparesis acquired in adulthood, stable medical condition, FIM score of 4 or higher, and the ability to walk 6m for 5 times without becoming distracted to the point of stopping their forward progress. Participants needed to understand the information in the informed consent and Health Insurance Portability and Accountability Act forms and to be able to follow study instructions. People were excluded if they had a medical condition that might limit their participation in the study, for example, recent total joint replacement or unstable cardiac condition.
To evaluate reliability, we determined that we would need 26 subjects using the ICC2,1 analysis of variance, a large effect size (F=0.40), power equal to 80%, and alpha equal to 0.05.17 To evaluate concurrent validity, for a correlation of r equal to 0.9 (or r2=0.8) with 95% power for significance, we needed 8 subjects. For an r equals 0.95 (r2=0.9), we needed 9 subjects.18 The study was approved by the Committee on Human Research at the University of Vermont.
Rater Training
The 2 raters using the Shaw Gait Assessment Tool and laptops took the short online tutorial. With healthy adults as subjects, the 3 raters then met for several practice sessions and resolved any questions or differences in execution, much as any PTs using a new tool might do.
Testing Protocol
Testing took place in a busy outpatient gym during normal treatment hours, but did not interfere with participants' ongoing therapy. Black electrical tape was used to mark off a 10-m walkway. Measurements were taken for a 6-m distance, with 2m before and after the start/finish lines to allow participants to start and stop the walk with a gradual increase/decrease in speed. (fig 2). One practice walk was followed by 2 test walks “at a comfortable pace” (separated by 2min). After a 15-minute rest, 2 more test walks were performed separated by a 2 minute rest, for a total of 4 test walks. Four walks were chosen to provide more data for reliability testing, while still remaining within 1 test session and causing less inconvenience to the subjects. Participants wore a gait belt for safety and an aide provided guarding from the nonhemiplegic side. Because guarding by the aide could have obstructed a good view of foot contact, each walk was in the same direction with the aide on the side of the participant farthest from the testers. (People were wheeled back to the original start line to minimize fatigue.)
Fig 2
Schematic representation of the position of the walkway and the 3 testers. Please note, the walkway was simply a series of tape marks on the gym floor.
We anticipated that people might lose their balance during the walk. Loss of balance was defined as a disruption of the person's specific gait pattern, staggering, or need for assistance to continue with the walk. If loss of balance occurred, data from that walk were not saved. Based on the therapists' clinical judgment and the willingness of the participant, 1 replacement walk could be performed as long as there was at least a 2-minute rest after the previous walk. If a replacement walk was not possible, only data from the walks without loss of balance were used for analysis.
Infrequently, a therapist realized that she had missed counting a step or had pressed the right (or left) arrow key twice consecutively. If this happened during testing in a clinical situation, the test would be repeated, so we felt that therapist fault was another valid reason for asking the participant to perform an additional walk. Because only 1 additional walk was permitted, if a therapist made an error during more than 1 test walk, only data from the remaining therapists were used in analysis for that test walk. Of the 312 test walks (26 participants × 4 test walks × 3 testers), we lacked data for 13 test walks.
Data were collected by 2 therapists using the Shaw Gait Assessment Tool on laptop computers and by 1 therapist holding a multimemory stopwatch. The 2 Shaw Gait Assessment Tool testers sat approximately 1.8m (6ft) apart on each side of the midpoint of the 6-m walkway. The person using the multimemory stopwatch stood between and slightly behind the Shaw Gait Assessment Tool testers (see fig 2).
As study participants crossed the start line, testers using the laptops pressed the right arrow key if the right foot crossed first and the left arrow key if the left foot crossed first. Testers continued to hit right and left arrow keys coinciding with right and left foot strikes thus counting the number of steps the subject took to cover the 6m. When subjects crossed the finish line, the testers hit the s (for stop) key. Based on the step count (c), the distance covered (d), and the time elapsed (t), the program then calculated the speed, cadence, step length, and limb advance time. The computer performed calculations for each walking trial: (1) gait speed: S=d/t, (2) cadence or step frequency: C=c/t, (3) average step length: L=d/c, and (4) limb advance time average: Tb=[t/c]both, Tl=[t/c]left, Tr=[t/c]right.
Because gait can be asymmetric even in people without pathology, but definitely in people with hemiplegic gait, the Shaw Gait Assessment Tool derives a measure of symmetry based on the limb advance time (a term seldom used by clinicians). Any systematic variability in the time interval separating 1 footfall from the next is quantified by averaging these intervals for all right and left steps to yield the right (and left) limb advance time in seconds.
The person holding the stopwatch used the lap/split feature of the stopwatch to record consecutive footfalls. If a participant crossed the start line with his/her left foot first, the stopwatch tester bent her left elbow to 90° forming an L (for left). She kept her arm down by her side if the right foot crossed the start line first. During the breaks between test walks, the stopwatch tester wrote down all the stopwatch data using the recall button, and indicated whether the test began with the right or left foot. Later that day, the stopwatch tester manually entered the data (elapsed time for each step and whether test started with right or left foot) into a program that replicated the Shaw Gait Assessment Tool calculations. The program (provided by the Shaw Gait Assessment Tool developer) was used to derive step length, cadence, speed, and limb advance time.
Data Analysis
All analyses were conducted using SPSS 15.0 for Windowsb An alpha level of 0.05 was set as acceptable. ICCs with 95% confidence intervals were calculated to evaluate intra- and interrater reliability for all variables: speed, cadence, step length, limb advance time left, and limb advance time right. Intrarater reliability for each of the 3 raters (laptop 1, laptop 2, and stopwatch) was determined using a 2-way mixed model and absolute agreement type. Interrater reliability (laptop 1, laptop 2) was determined using a 2-way random effects model and absolute agreement type. Absolute agreement was chosen because it measures whether raters assign the same absolute score. An ICC greater than 0.9 was set as acceptable.
The Pearson product-moment correlation coefficient was used to determine the concurrent validity of the Shaw Gait Assessment Tool compared to the multimemory stopwatch for gait speed, cadence, step length, limb advance time left, and limb advance time right. Separate correlation coefficients were obtained by comparing each laptop's data with the stopwatch result.
Results
Sixteen men and 10 women participated in the study. Their ages ranged from 19 to 85 years (mean ± SD, 58±19y). Length of time since onset of brain injury ranged from 0.5 to 66 months (mean ± SD, 11.7±16.96mo). Fifteen people had right-sided hemiplegia, while 11 had left-sided involvement. Sixteen participants used some type of lower-extremity bracing, while 10 did not; 19 people used a cane, 2 used a rolling walker, and 5 used no assistive device. Table 1 presents the means, SDs, and coefficients of variation for the gait parameters of speed, time, average step length, and cadence for all 3 raters.
| Raters | Time (s) | Speed (m/s) | Step Length (cm) | Cadence (steps/min) |
|---|---|---|---|---|
| Laptop 1 (n=97) | ||||
| Mean ± SD | 10.36±4.66 | 1.19±0.46 | 46.80±12.21 | 83.97±20.52 |
| CV (%) | 44.98 | 38.63 | 26.09 | 24.44 |
| Laptop 2 (n=99) | ||||
| Mean ± SD | 10.40±4.72 | 1.18±0.47 | 46.63±12.60 | 84.28±20.69 |
| CV (%) | 45.38 | 39.64 | 27.02 | 24.56 |
| Stopwatch (n=100) | ||||
| Mean ± SD | 10.16±4.58 | 1.20±0.46 | 47.45±11.85 | 84.71±19.86 |
| CV (%) | 45.08 | 38.11 | 24.97 | 23.44 |
Abbreviation: CV, coefficient of variation.
Table 2 shows the ICCs for intrarater reliability for both laptops and the stopwatch for all gait parameters. The intrarater reliability for both methods was excellent for all gait parameters measured, with the ICCs (model 2) for the laptops ranging from 0.94 to 0.98, and for the stopwatch ranging from 0.96 to 0.98 (all P<.001).
P<.001.
Table 3 shows the range of ICCs for interrater reliability for all trials and parameters for the laptop users. ICCs (model 2) ranged from 0.95 to 0.99 (P<.001), demonstrating excellent interrater reliability.
P<.001
Pearson product-moment correlations for each laptop with the stopwatch are shown in table 4 for all gait parameters. All correlations were 0.95 or higher, with P<.001, demonstrating that the Shaw Gait Assessment Tool is valid compared to the multimemory stopwatch
Discussion
In evaluating reliability and validity of the Shaw Gait Assessment Tool, we wanted to ensure that whatever results we obtained would represent its use in a typical clinical setting. To that end, the study was carried out in a busy outpatient rehabilitation gym with many distractions (for both participants and raters), and not in a quiet gait lab or testing room. The greater than 0.94 ICCs for all variables demonstrate excellent intra- and interrater reliability in a real-life setting. As stated earlier, we deliberately used the ICC because it assesses not only correlation, but agreement and generalizability as well.17
Strengthening the generalizability of the Shaw Gait Assessment Tool is the fact that the 2 Shaw Gait Assessment Tool raters trained themselves using the online tutorial, as any new user would need to do. New users of this tool will not have access to the developer, so it is encouraging to know that with the online tutorial and a few practice sessions, raters can expect excellent inter- and intrarater reliability.
Also adding to the generalizability of the study was the heterogeneity of our study population. Length of time since onset of hemiplegia ranged from 0.5 to 66 months; the ages spanned from 18 to 85 years; gait speeds were as slow as .42m/s and as fast as 2.8m/s, while average step length varied from 17.14cm to 66.67cm. Table 1 summarizes the temporospatial data for speed, time, average step length, and cadence for all 3 raters. Based on the small sample size, it would not be prudent to use these data to calculate a minimal detectable change, however.
The data in table 1 are provided so that clinicians can compare their patients to our study population. Limb advance time is not included in the table because the Shaw Gait Assessment Tool does not differentiate involved from noninvolved side, only right from left. Because there were 16 people with right-sided hemiplegia and 11 with left-sided hemiplegia, longer-limb advance times would be expected on different sides. These aggregate data therefore essentially negate the measure of asymmetry compared to looking at 1 person's results.
It may be helpful to elaborate on the parameter step length. As explained in the Methods section, average step length is calculated by dividing the distance covered (walkway length) by the step count. Because a subject's first or last footfall may be very close or up to a step length away from the start or finish line, the actual distance covered may not be identical to the length of the walkway. The Shaw Gait Assessment Tool uses the walkway length for the average step length calculation; therefore, one might question the validity of the Shaw Gait Assessment Tool calculated step length. The Elite motion analysis system uses an actual distance traveled to calculate step length. In the study with healthy adults where the Shaw Gait Assessment Tool was compared to the Elite, Pearson product-moment correlations ranged from 0.87 to 0.92.15 The issue of distance of the initial/last footfalls from the start/finish lines is the same whether a person is a healthy adult or has hemiparetic gait. Despite the potential differences when using actual distance walked versus walkway length, Shaw Gait Assessment Tool average step length calculations appear to be valid.15
The Shaw Gait Assessment Tool does not have the ability to perform separate averages for right and left, so it calculates an overall step length average. If a person's gait is very asymmetric, this average step length may not be very informative when viewed in isolation. Considering the right- and left-limb advance times will give a clinician an idea of the symmetry or asymmetry of the gait pattern. The goal would be to have right- and left-limb advance times be as close to one another as possible. However, we would not want patients to improve their symmetry at the expense of their step length. Two-step lengths equal 1 stride; therefore, noting improvements in symmetry of limb advance time without a decrease in step/stride length may help clinicians chart patients' progress.
It could be argued that having the laptop and stopwatch users stationed fairly close to one another may have resulted in inadequate blinding. The laptops users were seated about 1.8m (6ft) apart, with the stopwatch user slightly behind, but in between them. As mentioned above, the testing took place in a busy, noisy outpatient rehabilitation gym. Occasionally, people would inadvertently walk across the walkway during testing. Testers had to remain very focused on the actual walk, while attempting to ignore the distractions. The testers needed to look straight ahead to be able to observe the subjects walk and could not have seen other testers' keystrokes. In addition, both the arrow keys and the stopwatch buttons produced some level of noise when pressed, so in effect, 3 people's keystrokes would have needed to synchronize to skew the results—possible, but we feel unlikely.
We chose the stopwatch as our measure of concurrent validity because it is so widely used in clinical practice due to its availability and affordability. Pearson product-moment correlations greater than 0.95 for all variables reveal a strong correlation between the stopwatch and the Shaw Gait Assessment Tool, leading to the conclusion that the Shaw Gait Assessment Tool has strong concurrent validity with the stopwatch. The Shaw Gait Assessment Tool has also proven reliable and valid with healthy adults when using the Elite motion analysis system as the criterion standard.15
One might question the necessity of using the Shaw Gait Assessment Tool when a stopwatch is so readily available and inexpensive. As stated previously, speed, cadence, and step length each require a posttest calculation. Limb advance time would require a much more time consuming calculation. None of the results calculated with the stopwatch would be available for visual reinforcement and many people respond well to visual learning and information.
Computers are becoming an integral part of many clinics and hospital rehabilitation facilities, which are also offering Wi-Fi Internet access for patients and guests; this is certainly the case in the United States. In fact, many facilities have already switched to an electronic medical record, with additional facilities changing over almost daily, with ready access to computers for clinical staff, and Wi-Fi availability ensuring easy access to the Internet. After completing a test, users have the option of copying the numerical (not graphic) results into an electronic health record or printing the test results, an option which will preserve the graphs, so there should be no issues of patient confidentiality.
Assessment of gait speed has become more prevalent and important over the past few years. Gait speed is a powerful indicator of function and prognosis after stroke. It can be stratified into clinically meaningful functional ambulation classes and is a strong measure of physical decline.19
An improvement in gait speed that puts a person in a higher class of ambulation (eg, household vs community) results in better function and quality of life.19 The Shaw Gait Assessment Tool provides more data than just speed; however, and information such as step length, cadence, and left-right symmetry (as measured by right- and left-limb advance times) can be used to aid in the choice of an ankle foot orthosis, with some devices producing greater improvements than others.20
Study Limitations
We did not happen to enroll anyone with foot drag into our study. Because the Shaw Gait Assessment Tool calculations are based on foot contact, this tool would not be useful in people with foot drag and is a limitation of the tool.
Conclusions
The results of this study provide evidence to support strong inter- and intrarater reliability and concurrent validity of the Shaw Gait Assessment Tool in people with hemiplegic gait. Clinicians can now have access to a tool that is efficient, affordable, and easy to use, and which will give them information other than gait speed in a timely manner. Future research with this tool could include populations such as people with amputation, people with other neurologic diagnoses, or people after orthopedic surgery.
Suppliers
aFree web tool. Available from: http://www.theralink.com/apps/sga/.
bSPSS Corp, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606.
Acknowledgements
We would like to thank Tony Shaw, PT, for developing the Shaw Gait Assessment Tool and continuing to improve it based on clinical needs; Karen Stoneman, PTA, for participating in data collection as one of the raters; and Alan Howard, MS, for ongoing statistical help. Support in the form of time for working on the project and data collection was provided by Rehabilitation Therapies at Fletcher Allen Health Care.
References
1Hillier, S.L., Sharpe, M.H., and Metzer, J. Outcomes 5 years post-traumatic brain injury with further reference to neurophysical impairment and disability. Brain Inj. 1997; 11: 661–675
2Ochi, F., Esquenazi, A., Hirai, B., and Talaty, M. Temporal-spatial feature of gait after traumatic brain injury. J Head Trauma Rehabil. 1999; 14: 105–115
3Bohannon, R.W., Horton, M.G., and Wikholm, J.B. Importance of four variables of walking of patients with stroke. Int J Rehabil Res. 1991; 14: 246–250
4Mumma, C.M. Perceived losses following stroke. Rehabil Nurs. 1986; 11: 19–24
5Whittle, M.W. Methods of gait analysis. in: M.W. Whittle (Ed.) Gait analysis: an introduction. 4th ed. Butterworth Heinemann, Burlington; 2007: 15597
6Krebs, D.E., Edelstein, J.E., and Fishman, S. Reliability of observational kinematic gait analysis. Phys Ther.1985; 65: 1027–1033
7McGinley, J.L., Goldie, P.A., Greenwood, K.M., and Olney, S.J. Accuracy and reliability of observational gait analysis data: judgments of push-off in gait after stroke. Phys Ther. 2003; 83: 146–160
8Youdas, J.W., Atwood, A.L., Harris-Love, M.O., Stiller, T.L., Egan, K.S., and Therneau, T.M. Measurements of temporal aspects of gait obtained with a multimemory stopwatch in persons with gait impairments. J Orthop Sports Phys Ther. 2000; 30: 279–286
9Wall, J.C., Devlin, J., Khirchof, R., and Lackey, B. Measurement of step widths and step lengths: a comparison of measurements made directly from a grid with those from a video recording. J Orthop Sports Phys Ther. 2000;30: 410–417
10Wall, J.C. and Crosbie, J. Temporal gait analysis using slow motion video and a personal computer.Physiotherapy. 1997; 83: 109–115
11Gaudet, G., Goodman, R., Landry, M., Russell, G., and Wall, J.C. Measurement of step length and step width: a comparison of videotape and direct measurements. Physiother Can. 1990; 42: 12–15
12Holden, M.K., Gill, K.M., Magliozzi, M.R., Nathan, J., and Piehl-Baker, L. Clinical gait assessment in the neurologically impaired (Reliability and meaningfulness) . Phys Ther. 1984; 64: 35–40
13Riley, M., Goodman, M., and Fritz, V.U. An objective measure of gait using ink footprints. South African Journal of Physical Therapy. 1999; 55: 8–11
14Toro, B., Nester, C.J., and Farren, P.C. The status of gait assessment among physiotherapists in the United Kingdom. Arch Phys Med Rehabil. 2003; 84: 1878–1884
15Reid, S., Shaw, A., and Haugh, L.D. Validity and reliability of an internet-based temporal gait assessment tool with healthy adults: a pilot study. Arch Phys Med Rehabil. 2005; 86: 1014–1018
16Handa, T., Sahara, R., Yoshikazi, K. et al. Examination of reliability and validity of walking speed, cadence, stride length – comparison with stopwatch and three-dimension motion analyzer. J Phys Ther Sci. 2007; 19: 213–222
17Portney, L.G. and Watkins, M.P. Foundations of clinical research. Applications to practice. 3rd ed. Pearson Prentice Hall, Upper Saddle River; 2009
18QueryAdvisorAvailable at. http://www.statsol.ie/nquery/nquery.htm. (Accessed June 22, 2005)
19Schmidt, A., Duncan, P.W., Studenski, S. et al. Improvements in speed-based gait classifications are meaningful. Stroke. 2007; 38: 2096–2100
20Gök, H., Küçükdeveci, A., Altinkaynak, H., Yavuzer, G., and Ergin, S. Effects of ankle-foot orthoses on hemiparetic gait. Clin Rehabil. 2003; 17: 137–139
Nenhum comentário:
Postar um comentário