Abstract
Background
Gait speed predicts disability, cognitive decline, hospitalization, nursing home admission and mortality. Although gait speed is often measured in clinical practice and research, testing protocols vary widely and their impact on recorded gait speed has yet to be explored.
Objectives
Our purpose is to describe and compare gait speeds obtained from different testing protocols in the same individuals.
Design
Cross-sectional.
Setting
University research setting.
Participants
Subjects were 104 community-dwelling older adults who could ambulate household distances independently (mean age=77.2±6.1).
Measurements
Gait speed was recorded over 4 meters using the protocols: 1) standing start, usual pace over ground, 2) walking start, usual pace over ground with an optokinetic device, 3) walking start, usual pace over ground with a stop watch 4) walking start, usual pace on a computerized walkway, and 5) walking start, fast pace on a computerized walkway. A linear mixed model and pairwise comparisons was used to compare gait speeds within individuals across different protocols.
Results
Mean±SD gait speed for each condition was: standing start, usual pace over ground 0.97±0.23 m/s; walking start, usual pace over ground 1.14±0.25 m/s; walking start, usual pace on walkway 1.01±0.26 m/s; and walking start, fast pace on walkway 1.31±0.34 m/s. On average, the determined gait speed was 0.17 m/s faster during the walking compared to the standing start (p<.001), 0.07 m/s slower on the computerized walkway compared to over ground (p<.001), and 0.25 m/s faster during the fast pace compared to the usual pace walk (p<.001).
Conclusion
Starting protocol (standing vs. walking), testing surface (over ground vs. computerized walkway), and walking pace (usual vs. fast) impact recorded gait speed in older adults. Care should be taken when comparing gait speeds from studies with different testing protocols.
Keywords: gait speed, testing protocol
Gait speed, a significant predictor of important health outcomes,1–3 is frequently used in clinical and epidemiological research studies. However, there is currently no consensus on the ideal testing procedure; therefore testing protocols vary widely, making comparisons difficult.4, 5 Others examining the implications of testing protocol conducted literature reviews comparing protocols between different studies and subjects.4, 5 Because the comparisons were not made between the same subjects, it is difficult to determine if differences in gait speed were due to testing protocol or differences in study sample. Also when evaluating the differences, researchers focused on statistically significant and not clinically significant values (i.e. differences that are clinically noticeable).
Therefore, the purpose of this study is to compare gait speeds obtained from different testing protocols in the same individuals. More specifically, the impact of starting protocol (standing start versus walking start), testing condition (computerized walkway versus over ground walking), measurement method (automatic timing system versus manual stopwatch), and pace (usual pace versus fast pace) on measured gait speed will be investigated taking into account the clinical meaningfulness of the differences in recorded gait speeds.
METHODS
Subjects
The study included 120 older adults, who were recruited from the University of Pittsburgh Pepper Center research registry, a database of older adults who have previously consented to be contacted for studies of this nature. Subjects were community-dwelling adults age 65 and older who could ambulate household distances (50 feet) independently, and who self-reported the ability to tolerate 5 hours of testing. Participants were included in the present analyses if they had completed gait speed measurements for each of the five testing protocols (N =104). The 16 excluded subjects with incomplete data did not differ from the sample on age, gender, self-reported fear of falling, self-reported rating of mobility, or over ground gait speed. The University of Pittsburgh institutional review board approved the study and all participants provided written informed consent.
Measures
Assessments included demographic information, self-report questions of health, mobility, and fear of falling, and measurements of gait speed using five different testing protocols (described below).
Standing Start, Usual Pace, Over Ground (SPPB)
Standing start, usual pace, over ground gait speed was assessed as part of the Short Physical Performance Battery (SPPB).6, 7 Time to complete the 4 meter walk was measured using a stopwatch, and the tester started the time when the participant stepped over the starting line and stopped when the participant had completely crossed the end line.
Walking Start, Usual Pace, Over Ground (Optokinetic)
Gait speed was measured over a 4-meter distance by a motion sensitive light beam system that triggers activation and inactivation of a stopwatch. Participants started walking several steps before the first light beam and continued walking after the final light beam, allowing time to accelerate and decelerate.
Walking Start, Usual Pace, Over Ground (Stopwatch)
During the optokinetic gait speed protocol trials, a tester was simultaneously collecting the participant’s gait speed using a stopwatch. The tester started the stopwatch when the participant crossed the starting line and stopped the time when the participant crossed the final optokinetic sensor.
Walking Start, Usual Pace on a Computerized Walkway
Gait speed was also measured using a computerized walkway, approximately 8 meters in length (GaitMat II). The walkway consists of a central 4-meter segment that collects data, with inactive segments before and after that allow for acceleration and deceleration. The participant walked at their usual pace.
Walking Start, Fast Pace on a Computerized Walkway
This protocol utilized the same setup as described above for walking start, usual pace walking on the computerized walkway. The participant completed additional fast pace trials on the walkway, with the following verbal instructions “walk as quickly as possible, without running or putting yourself at risk for falling”.
Data Analysis
All Analyses were performed using SPSS® version 19.0 (SPSS, Inc., Chicago, IL) and SAS® version 9.3 (SAS Institute, Inc., Cary, North Carolina). We fit a linear mixed model for gait speed with protocol as the main fixed effect of interest, and a subject random effect to account for multiple measurements from each participant. Contrasts were estimated between: 1) standing start, usual pace and walking start, usual pace; 2) walking start, usual pace over ground and walking start, usual pace on the computerized walkway; 3) walking start, usual pace measured by the optokinetic light beam system and walking start, usual pace measures manually with a stop watch; and 4) walking start, usual pace on the computerized walkway and walking start, fast pace on the computerized walkway. Lastly, linear and quadratic regression models were used to create equations to estimate conversions for gait speeds obtained from different testing protocols.
RESULTS
Baseline characteristics of the sample (N = 104) and the average gait speed for each testing protocol are presented in Table 1. The slowest average gait speed was recorded using the standing start, usual pace over ground protocol (0.97 m/s; SD = .23) and the fastest average gait speed was recorded using the fast pace gait speed on the computer walkway (1.31 m/s; SD = .26).
Table 1.
Baseline Characteristics of the Sample
| Characteristic | Study sample (n=104) | Excluded Sample (n=16) |
|---|---|---|
| Age, mean ± SD | 77.1 ± 6.1 | 79.8 ± 4.1 |
| Female, n (%) | 76 (73.1%) | 10 (62.5%) |
| White, n (%) | 91 (87.5%) | 14 (87.5%) |
| ≥High school education, n (%) | 97 (93.2%) | 15 (93.8%) |
| Global mobility fair or poor, n (%) | 25 (24%) | 1 (6.3%) |
| Global health fair or poor, n (%) | 10 (9.6%) | 0 (0%) |
| Afraid of falling, n (%) | 45 (43.3%) | 7 (43.8%) |
| Gait Speed Measures | ||
| Standing start, usual pace over ground (SPPB†), mean ± SD | 0.97 ± .23 | 0.97 ± .25 |
| Walking start, usual pace over ground (optokinetic), mean ± SD | 1.14 ± .25 | -- |
| Walking start, usual pace over ground (stopwatch), mean ± SD | 1.13 ± .25 | -- |
| Walking start, usual pace on computerized walkway, mean ± SD | 1.06 ± .26 | -- |
| Walking start, fast pace on computerized walkway, mean ± SD | 1.31 ± .34 | -- |
SD = standard deviation.
4-meter SPPB
Table 2 displays the mean difference between gait speeds obtained from the various testing protocols. On average, participants walked faster at their usual pace when starting with a walking start compared to a standing start. When comparing the computerized walkway to over ground, participants on average walked slower on the computerized walkway than over ground when instructed to walk at the same pace and starting protocol. These results were similar when measuring usual pace walking over ground with both the optokinetic system and the stopwatch. Walking start, usual pace gait speed over ground was similar when (p>0.05) when measured with the optokinetic system as compared to a stopwatch. Finally, on average, participants walked .25±.16 m/s faster (p<.001) when instructed to walk at their fastest pace than at their usual, comfortable pace.
Table 2.
Mean Differences Between Gait Speeds Obtained from the Different Testing Protocols
| Measures | Estimated Difference | 95% Confidence Interval | p-Value |
|---|---|---|---|
| Walking Start vs. Standing Start | |||
| Optokinetic vs. SPPB† | 0.17 | (0.14, 0.20) | <.0001 |
| Stopwatch vs. SPPB† | 0.16 | (0.14, 0.19) | <.0001 |
| Computerized Walkway vs. SPPB† | 0.10 | (0.07, 0.12) | <.0001 |
| Computerized Walkway vs. Over Ground | |||
| Walkway vs. Optokinetic | −0.07 | (−0.10, −0.04) | <.0001 |
| Walkway vs. Stopwatch | −0.07 | (−0.09, −0.04) | <.0001 |
| Over Ground measures | |||
| Optokinetic vs. Stopwatch | 0.01 | (−0.02–0.03) | 0.7174 |
| Fast Pace vs. Usual Pace | |||
| Fast vs. Usual Computerized Walkway | 0.25 | (0.22–0.27) | <.0001 |
4-meter gait speed from Short Physical Performance Battery (SPPB)
Regression equations to convert gait speed between the various testing protocols are presented in Table 3. The R2 of the equations ranged from 0.74-.83.
Table 3.
Conversion Equations for Gait Speeds Obtained from Different Protocols with Linear and Quadratic Models, and [Proportion of Explained Variability R2]
| Convert to (Y) | ||||||
|---|---|---|---|---|---|---|
| Optokinetic | Stopwatch | SPPB | Walkway Usual | Walkway Fast | ||
| Convert from (X) | Optokinetic | ---- |
Y=0.006+0.990X [94%] Y=−0.025+1.053X −0.030X2 [94%] |
Y=0.074+0.786X [70%] Y=−0.008+0.953X −0.079X2 [70%] |
Y=−0.020+0.954X [83%] Y=−0.022+1.353X −0.190X2 [83%] |
Y=−0.045+1.192X [74%] Y=−0.602+2.323X −0.539X2 [76%] |
| Stopwatch |
Y=0.061+0.951X [94%] Y=−0.229+1.530X −0.272X2 [95%] |
---- |
Y=0.068+0.795X [75%] Y=0.081+0.771X +0.012X2 [75%] |
Y=−0.009+0.949X [85%] Y=−0.185+1.300X −0.165X2 [86%] |
Y=0.014+1.145X [71%] Y=−0.757+2.686X −0.723X2 [74%] |
|
| SPPB |
Y=0.273+0.892X [70%] Y=−0.318+2.237X −0.718X2 [74%] |
Y=0.222+0.939X [75%] Y=−0.106+1.686X −0.399X2 [76%] |
---- |
Y=0.090+1.006X [81%] Y=−0.217+1.706X −0.374X2 [82%] |
Y=0.099+1.251X [72%] Y=−0.792+3.279X −1.082X2 [76%] |
|
| Walkway Usual |
Y=0.213+0.867X [83%] Y=−0.052+1.445X −0.0291X2 [84%] |
Y=0.175+0.898X [85%] Y=0.072+1.122X −0.113X2 [85%] |
Y=0.110+0.806X [81%] Y=0.204+0.601X +0.104X2 [81%] |
---- |
Y=0.042+1.191X [82%] Y=−0.385+2.120X −0.467X2 [83%] |
|
| Walkway Fast |
Y=0.321+0.622X [74%] Y=0.034+1.151X −0.221X2 [76%] |
Y=0.316+0.622X [71%] Y=0.022+1.163X −0.226X2 [73%] |
Y=0.214+0.576X [72%] Y=0.118+0.753X −0.074X2 [72%] |
Y=0.168+0.684X [82%] Y=−0.040+1.066X −0.160X2 [82%] |
---- | |
DISCUSSION
Based on findings of the current study, a small meaningful change in gait speed can be elicited simply by testing an individual’s gait speed over ground instead of on a computerized walkway. Furthermore, a substantial clinically meaningful change can be made by allowing an individual a walking start to accelerate to their true walking pace rather than using a standing start. For practical reasons, there may not be one ideal testing protocol. However, the ease of altering gait speed by clinically meaningful amounts simply by changing the protocol demonstrates the need for consistency in testing protocol and documentation of testing method when noting a gait speed measurement.
For example, gait speed assessment using a walking start may not be feasible in all clinical settings, particularly those that have limited space and would not allow for acceleration/deceleration (i.e., home care, hospital rooms). Using the proper regression equation, standing start gait speed can be readily converted to walking start gait speed, facilitating more accurate comparison of gait speeds obtained with different testing protocols.
Conclusions
While consensus has yet to be reached on the optimal gait speed testing protocol, it is important that both clinicians and researchers are aware of the significant impact of altering aspects of testing protocol on measured gait speed.
Research Findings.
Although gait speed is often measured in clinical practice and research, testing protocols vary widely.
Starting protocol, testing surface, and walking pace all impact recorded gait speed.
Care should be taken when comparing gait speeds from studies with different testing protocols.
Regression equations to convert gait speed between various testing protocols are presented.
Acknowledgments
This work was supported by the Pittsburgh Older Americans Independence Center (NIA P30 AG024827), Beeson Career Development Award (NIA K23 AG026766) and the Pitt Clinical Research Training: Geriatrics/Gerontology (T32 AG021885).
Footnotes
A portion of this work was presented at the American Geriatric Society meeting, May 2012, Seattle, WA.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Reference List
- 1.Abellan Van Kan G, Rolland Y, Andrieu S, et al. Gait speed at usual pace as a predictor of adverse outcomes in community-dwelling older people an international academy on nutrition and aging (IANA) task force. J Nutr Health Aging. 2009;13(10):881–889. doi: 10.1007/s12603-009-0246-z. [DOI] [PubMed] [Google Scholar]
- 2.Studenski S, Perera S, Patel K, et al. Gait speed and survival in older adults. JAMA. 2011;305(1):50–58. doi: 10.1001/jama.2010.1923. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Cesari M, Kritchevsky S, Bauer DC, et al. Prognostic value of usual gait speed in well-functioning older people--results from the Health, Aging and Body Composition Study. J Am Geriatr Soc. 2005;53(10):1675–1680. doi: 10.1111/j.1532-5415.2005.53501.x. [DOI] [PubMed] [Google Scholar]
- 4.Graham JE, Ostir GV, Fisher SR, Ottenbacher KJ. Assessing walking speed in clinical research: a systematic review. J Eval Clin Pract. 2008;14:552–562. doi: 10.1111/j.1365-2753.2007.00917.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Graham JE, Ostir GV, Kuo YF, Fisher SR, Ottenbacher KJ. Relationship between test methodology and mean velocity in timed walk tests: A review. Arch Phys Med Rehabil. 2008;89:865–872. doi: 10.1016/j.apmr.2007.11.029. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Guralnik JM, Simonsick EM, Ferrucci L, et al. A short physical performance battery assessing lower extremity function: Association with self-reported disability and prediction of mortality and nursing home admission. J Gerontol. 1994;49:M85–M94. doi: 10.1093/geronj/49.2.m85. [DOI] [PubMed] [Google Scholar]
- 7.Guralnik JM, Ferrucci L, Simonsick EM, Salive ME, Wallace RB. Lower extremity function in persons over the age of 70 years as a predictor of subsequent disability. New Engl J Med. 1995;332:556–561. doi: 10.1056/NEJM199503023320902. [DOI] [PMC free article] [PubMed] [Google Scholar]