The minimal clinically important difference for the Gait Profile Score
Highlights
► A minimally clinically important difference is derived for the Gait Profile Score. ► Based on variability of typically developing children. ► Based on difference between Functional Assessment Questionnaire levels. ► MCID is 1.6°.
Introduction
There is an increasing emphasis in clinical research into establishing whether outcomes are clinically meaningful as well as statistically significant. The concept of minimal clinical important difference [MCID [1]] is widely accepted but there is little consensus on precise definitions [2], [3]. The range of methods has been divided into anchor based methods, which compare a new measure to other measures of clinical evidence, and distribution based methods which are based on its statistical or psychometric properties [3]. Both have limitations, anchor based methods, in validating one measure on the basis of another, have the potential to be “circular in logic and fraught with potential bias” [2]. Distribution based methods “cannot address the question of clinical importance, which is central to the concept of MCID” [2].
None of the existing measures of gait quality [4], [5], [6], [7] have had an MCID defined. Oeffinger et al. [8] reported a six year multi-centre study to define MCID for range of measures of walking function but did not include any gait quality indices. The Gait Profile Score [GPS [4], [5]] represents the root mean square (RMS) difference between a particular gait trial and averaged data from people with no gait pathology. It has an advantage over the other indices as it is comprised of a number of gait variable scores (GVSs) representing an equivalent RMS difference for different kinematic variables. These can be displayed as a bar chart known as the Movement Analysis Profile (MAP). The aim of this paper is thus to define an MCID for the GPS.
The cross-sectional approach is an anchor based method that compares groups that are different in terms of a clinically relevant disease related criterion [3]. Within clinical gait analysis two such criterion measures of functional mobility are widely used. The Functional Assessment Questionnaire [FAQ [9]] is a 10 point scale (6–10 describe functional walkers with 10 being most able) which was designed specifically to be used as an outcome measure. The Gross Motor Function Classification System [GMFCS [10], [11]] is designed for children with cerebral palsy (CP) with five levels (functional walkers are classified within levels I–III, with I being the most able). Both scales were derived by clinicians to define groups of children whose physical function is clinically different and both were developed independently of any consideration of instrumented gait analysis (IGA). The GMFCS was derived using a Delphi process involving 48 clinicians from a range of backgrounds [10]. It is more difficult to establish the construct validity of the FAQ from the original publication [9] but this has now been cited over 100 times reflecting widespread acceptance within the clinical research community (and thus high face validity). They thus appear ideal candidates for defining MCID for a single gait index defined from IGA. The FAQ is not restricted to any particular age group or disease condition and the larger number of five categories for functional walkers allow for greater discrimination across the wide spectrum of walking abilities than the three levels of the GMFCS that represent functional walking ability. For this reason it will be used as the primary measure for determining MCID of the GPS. The GMFCS is probably more widely accepted and has been more rigorously validated for children with CP, it is a classification system rather than an outcome measure but the different levels still represent differences in physical function. Analysis of differences in GPS between GMFCS levels will therefore be presented as a secondary analysis to provide further justification for the MCID determined from FAQ data.
Whilst the focus of this study will be on this cross-sectional approach the results will be discussed in the context of alternate methods. Sample variation [3] is a distribution based approach defining MCID in terms of a measure's variability within a particular sample population. Selection of an appropriate reference population is dependent on how clinicians typically use the data within clinical reasoning and practice. Almost all clinical gait analysis compares data from individuals against control data from a reference population without gait pathology [12]. The standard deviation of this population thus seems an obvious parameter against which to consider any proposed MCID.
Another approach [13] is to define a threshold for MCID as an appropriate percentage of a score of disease severity. A consortium working with pain suggested a generic 30% reduction from baseline as appropriate [13]. Whilst this approach appears sensible, subjective semi-qualitative pain scores are quite different from objective quantitative measures such as the GPS and a 30% reduction may be too conservative. Schwartz et al. [14] suggested a threshold of 10% improvement for a range of measures of gait pathology. Extending this approach to the GPS would have to be modified as the healthy population has a non-zero GPS.
Section snippets
Methods
The GPS [5] is based upon a number of gait variable scores (GVS) each of which is the root mean square difference between a specific time normalized gait variable and the mean data from some reference population calculated across the gait cycle. Thus if xi,t is the value of gait variable i calculated at a specific point in the gait cycle t, and is the mean value of that variable at the same point in the gait cycle for the reference population then the ith gait variable score is given
Results
Three hundred and eighty two children were recorded as having a FAQ level which range from 6 to 10 and 268 of those with cerebral palsy as having a GMFCS level ranging from I to III. Table 1 records the numbers in each of these groups, their median and inter quartile range GPS (also plotted in Fig. 1). The gradient of the line of regression (Fig. 1) is taken as a measure of the average difference of GPS median score between adjacent levels and is 1.6° (s.e. 0.2°) for FAQ and 2.9° (s.e. 0.4°)
Discussion
The linearity of the correlation between GPS median values and both FAQ and GMFCS levels is quite remarkable given that definitions of both levels have been developed independently of each other and prior to the conception of the GPS. (Even more remarkably inclusion, of the TD group as an extension of the classification system has little effect on the r2 values.) This provides a strong rationale for taking the gradient of the regression line with FAQ score as indicative of an MCID for GPS
Acknowledgments
Adrienne Fosang, Tandy Hastings-Ison and Adrienne Harvey were responsible for capturing the data on which this study is based. Oren Tirosh and Sarah Beynon were responsible for the original data collation.
Conflict of interest statement
None.
References (16)
- et al.
Measurement of health status. Ascertaining the minimal clinically important difference
Control Clin Trials
(1989) - et al.
Defining clinically meaningful change in health-related quality of life
J Clin Epidemiol
(2003) - et al.
Correlations of the Gait Profile Score and the Movement Analysis Profile relative to clinical judgments
Gait Posture
(2010) - et al.
The gait profile score and movement analysis profile
Gait Posture
(2009) - et al.
The gait deviation index: a new comprehensive index of gait pathology
Gait Posture
(2008) - et al.
An index for quantifying deviations from normal gait
Gait Posture
(2000) - et al.
Core outcome measures for chronic pain clinical trials: IMMPACT recommendations
Pain
(2005) - et al.
GaitaBase: web-based repository system for gait analysis
Comput Biol Med
(2010)
Cited by (132)
Botulinum neurotoxin type A responders among children with spastic cerebral palsy: Pattern-specific effects
2024, European Journal of Paediatric NeurologyGait velocity control using projection mapping for children with spastic diplegia cerebral palsy
2023, Clinical BiomechanicsNormative dataset selection affects gait profile scores of children with cerebral palsy
2023, Gait and Posture