Elsevier

Gait & Posture

Volume 37, Issue 4, April 2013, Pages 580-585
Gait & Posture

Is gait variability reliable in older adults and Parkinson's disease? Towards an optimal testing protocol

https://doi.org/10.1016/j.gaitpost.2012.09.025Get rights and content

Abstract

Background

Despite the widespread use of gait variability in research and clinical studies, testing protocols designed to optimise its reliability have not been established. This study evaluates the impact of testing protocol and pathology on the reliability of gait variability.

Objective

To (i) estimate the reliability of gait variability during continuous and intermittent walking protocols in older adults and people with Parkinson's disease (PD), (ii) determine optimal number of steps for acceptable levels of reliability of gait variability and (iii) provide sample size estimates for use in clinical trials.

Methods

Gait variability was measured twice, one week apart, in 27 older adults and 25 PD participants. Participants walked at their preferred pace during: (i) a continuous 2 min walk and (ii) 3 intermittent walks over a 12 m walkway. Gait variability was calculated as the within-person standard deviation for step velocity, length and width, and step, stance and swing duration.

Results

Reliability of gait variability ranged from poor to excellent (intra class correlations .041–.860; relative limits of agreement 34–89%). Gait variability was more reliable during continuous walks. Control and PD participants demonstrated similar reliability. Increasing the number of steps improved reliability, with most improvement seen across the first 30 steps.

Conclusions

In this study, we identified testing protocols that improve the reliability of measuring gait variability. We recommend using a continuous walking protocol and to collect no fewer than 30 steps. Early PD does not appear to impact negatively on the reliability of gait variability.

Highlights

► Reliability of gait variability ranged from poor to excellent across variables. ► Parkinson's disease did not influence reliability of gait variability. ► Continuous walking protocols were more reliable than intermittent protocols. ► Over 30 steps are recommended to calculate acceptable levels of gait variability.

Introduction

Gait variability (stride-to-stride fluctuations) may discriminate important clinical features better than routine spatio-temporal measures such as mean gait speed or step time [1]. Gait variability is sensitive to pathology and ageing [2], [3]; discriminates between older adults with and without mobility and cognitive impairment [2], [4]; identifies subtle pathological gait impairments preceding more readily observed spatio-temporal changes [5]; and predicts falls [6] and future cognitive decline [7]. More recently, gait variability has also been reported as a primary outcome for a randomised controlled trial for falls [8].

Despite the widespread use of gait variability in research and clinical studies, its clinimetric properties are not clearly established. Whilst discriminative and predictive validity may be inferred from a significant body of cross sectional and longitudinal studies concerned with gait and mobility outcomes for older adults [3], reliability has not been firmly established. A recent study has estimated meaningful change of gait variability scores based on two annual follow up analyses in 241 community dwelling older adults [9]. However, stability of measures in the shorter term and in pathology is unclear. Lack of standardised testing protocol and knowledge of clinimetric properties limits the interpretation of gait variability from evaluative, diagnostic and prognostic studies.

A current systematic review of gait variability in older adults highlighted a number of critical issues such as: inconsistency in measurement protocol; uncertainty about the number of steps required for optimal measurement; effect of pathology on reliability estimates; and the wide range of variability outcomes described with no consensus for robust estimates [3]. Gait is less variable when walking continuously compared to performing multiple short intermittent walks [10] highlighting the influence of testing protocol. Gait variability is also not a single entity, but is derived from a range of spatio-temporal characteristics which may themselves differ in reliability. In healthy older adults for example, reliability ranged from poor to excellent, with intra-class correlations (ICC) ranging from .11 to .98 depending on the variables reported [3]. Unsurprisingly, pathology influences reliability estimates which are less reliable for older adults with dementia [11] and impaired balance [12]. Earlier reports also suggest that measurement accuracy improves with longer distances which allow a greater number of steps or strides to be included [13] however, this has not been verified in people with PD.

The purpose of this study was therefore to address the reported inconsistencies by carrying out a comprehensive evaluation of the reliability of gait variability in order to guide the use of variability as an outcome in research and clinical studies. To achieve this, we aimed to: (1) evaluate the effect of testing protocol (continuous walking compared to intermittent walks repeated over a short distance); (2) evaluate the effect of pathology (control and PD participants) on reliability; (3) identify the optimal number of steps for acceptable levels of reliability; and (4) calculate sample size estimates to inform clinical trials.

Section snippets

Participants

Participants were excluded if they had any neurological (other than PD), orthopaedic or cardiothoracic conditions that may have markedly affected their walking or safety during the testing sessions. In addition, PD participants had to be diagnosed with idiopathic PD according to the UK Parkinson's Disease Brain Bank criteria and were excluded if they presented with significant memory impairment (Mini Mental State Exam (MMSE)  24 [14]), dementia with Lewy bodies, drug induced parkinsonism,

Results

Twenty-seven controls and 25 people with mild-to-moderate PD were recruited into this study (Table 1). The control group were slightly older than the PD sample and tended to have a higher proportion of women, but were well matched for BMI. People with PD also tended to walk more slowly than the control participants however this was not statistically significant.

Prior to addressing the primary aims of this study, we tested whether gait variability was more reliable when it was estimated from

Discussion

This is the first study to carry out a comprehensive evaluation of the reliability of gait variability in older adults and people with PD. We have shown that variability of steps when a participant walks continuously provides a more reliable estimate of gait variability than intermittent walking trials, however, the impact of pathology was not a critical factor for reliability. Reliability differs depending upon the variable of interest and improves as the number of steps included in estimates

Acknowledgements

The research was funded by the National Institute for Health Research (NIHR) Newcastle Biomedical Research Centre based at Newcastle upon Tyne Hospitals NHS Foundation Trust and Newcastle University. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health. We would also like to recognise the support of Dadirayi Mhiripiri and Julia Greenland during data collection and processing.
Conflicts of interest

The authors have no conflicts of

References (21)

There are more references available in the full text version of this article.

Cited by (192)

View all citing articles on Scopus
View full text