Gait variability in older adults: A structured review of testing protocol and clinimetric properties

Abstract

Gait variability (stride-to-stride fluctuations) is used increasingly as a marker for gait performance and future mobility status, cognitive status, and falls. This structured review explicitly examined literature that reported on the reliability, validity and responsiveness of gait variability in older adults. We searched Medline, Embase, Web of Science, Scopus, CINAHL, PEDRO, Biomechanics, SportDiscus and PsycInfo databases. Two independent reviewers undertook data extraction, with adjudication by a third reviewer in cases of disagreement. Twenty-two full papers were screened and 10 met the predefined inclusion criteria, involving 1036 participants who were mainly community dwelling older adults in their 8th decade. A wide range of gait variability parameters, testing protocols and calculations of gait variability were reported. Reliability estimates varied, but were mostly fair to moderate. Concurrent validity was established for stance time variability and change estimates were reported for stance time and swing time. Standard of reporting was generally poor, with insufficient detail provided for aspects of measurement and testing protocols. Further research is required to standardise testing procedures and establish reliability, responsiveness and validity for confident use of gait variability as a robust measure.

Introduction

Recent research suggests that gait variability (stride-to-stride fluctuations) may provide a more discriminative measure of gait performance than routine spatio-temporal measures such as average gait speed or step time. Gait variability was originally considered to represent noise of instrumentation or physiological noise. More recently it has been proposed to reflect the underlying neural control of gait with demonstrated sensitivity to pathological and ageing processes [1]. It discerns between older adults with and without mobility impairment and cognitive impairment [2], [3]; identifies subtle pathological gait changes prior to more readily observed spatio-temporal changes [4]; and predicts falls [5] and future cognitive decline [6]. Gait variability has also been reported to increase when the demands of walking increase, for example under dual and multiple task conditions, emphasising the contribution of cognitive control to locomotion [7]. Different parameters are used to describe gait variability. For example, stride time variability reflects the ability to generate consistent rhythmical step cycles, whilst step width and double limb support (DLS) (time spent with both feet in contact with the floor), reflects postural control mechanisms during gait [8].

There is no agreed standardised protocol for measuring gait variability, as evidenced by inconsistencies in data capture techniques, distance walked, instrumentation and analytical algorithms. Earlier reports suggest that measurement accuracy improves with longer distances [1], but this has not been verified. The situation is further confused by use of different and interchangeable terminology. For example ‘relative variability’ or coefficient of variation (CV or CoV), (SD/mean × 100%), is a statistical measure of the dispersion of group data points in a data series relative to the mean, often used to evaluate reliability (stability) of measures including gait outcomes (e.g. [9]) and gait variability itself [10]. However, the concept of gait variability relevant to this review is concerned with the variability inherent in an individual's stride-to-stride fluctuations. Calculation of gait variability (CV) is derived from the total number of steps or strides for each individual, for each gait parameter of interest.

Unlike routine measures, the clinimetric properties of gait variability have not been well established. Discriminative validity 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 (e.g. [2], [3], [4], [5], [6], [7], [11], [12], [13]), although these properties have not been formally examined. More recently, gait variability has been selected as a primary outcome for a randomised controlled trial for falls [14], although responsiveness has not been firmly established.

Lack of standardisation of measurement protocol and knowledge of clinimetric properties limits the interpretation of gait variability from evaluative, diagnostic and prognostic studies. This review has three aims: (1) evaluate studies that explicitly test clinimetric properties of gait variability in older adults; (2) report on the testing protocols used and justification for their use; and (3) make recommendations concerning the utility of gait variability as a measure.