Elsevier

Gait & Posture

Volume 42, Issue 2, July 2015, Pages 133-137
Gait & Posture

Gait Deviation Index, Gait Profile Score and Gait Variable Score in children with spastic cerebral palsy: Intra-rater reliability and agreement across two repeated sessions

https://doi.org/10.1016/j.gaitpost.2015.04.019Get rights and content

Highlights

  • Excellent reliability and acceptable agreement of the GDI and GPS was found.

  • Large variation in reliability and agreement of the GVS was found.

  • The clinimetric properties of the GDI and GPS justify their use as measures in research and clinical practice.

Abstract

The Gait Deviation Index (GDI) and Gait Profile Score (GPS) are the most used summary measures of gait in children with cerebral palsy (CP). However, the reliability and agreement of these indices have not been investigated, limiting their clinimetric quality for research and clinical practice. The aim of this study was to investigate the intra-rater reliability and agreement of summary measures of gait (GDI; GPS; and the Gait Variable Score (GVS) derived from the GPS).

The intra-rater reliability and agreement were investigated across two repeated sessions in 18 children aged 5–12 years diagnosed with spastic CP. No systematic bias was observed between the sessions and no heteroscedasticity was observed in Bland–Altman plots. For the GDI and GPS, excellent reliability with intraclass correlation coefficient (ICC) values of 0.8–0.9 was found, while the GVS was found to have fair to good reliability with ICCs of 0.4–0.7. The agreement for the GDI and the logarithmically transformed GPS, in terms of the standard error of measurement as a percentage of the grand mean (SEM%) varied from 4.1 to 6.7%, whilst the smallest detectable change in percent (SDC%) ranged from 11.3 to 18.5%. For the logarithmically transformed GVS, we found a fair to large variation in SEM% from 7 to 29% and in SDC% from 18 to 81%.

The GDI and GPS demonstrated excellent reliability and acceptable agreement proving that they can both be used in research and clinical practice. However, the observed large variability for some of the GVS requires cautious consideration when selecting outcome measures.

Introduction

Three-dimensional instrumented gait analysis (3DGA) has become an important examination in children with cerebral palsy (CP) in both research and clinical practice [1]. Healthcare professionals such as physiotherapists and biomedical engineers perform 3DGA. The examination provides a large amount of complex and interdependent data, which have led to the development of indices that can describe the quality of the gait pattern in a single score [2]. The summary measures most commonly used are the Gait Deviation Index (GDI) [3] and the Gait Profile Score (GPS) [4], which both provide a single score of the quality of the patient's kinematics during gait.

The GDI is based on the calculation of the distance between the patient's data and the average from the reference dataset on 15 gait features of gait kinematics of the pelvis, hip, knee and ankle [3]. The GPS is obtained from the same gait kinematics as the GDI and is calculated on all gait features representing the root mean square difference between the patient's data and the average from the reference dataset [4].

The Gait Variable Score (GVS), which consists of nine gait variables for each side of the body, can be derived from the GPS score. An overall GVS for the pelvis is used and by convention the left side value is taken [4].

The GDI and GPS are different ways of scaling the same underlying construct and therefore there is little point in using both outcome measures [4]. There is debate about the use of the GDI and GPS in clinical practice and research. At present, there are pros and cons for both indices and choosing one over the other is often based on personal preference [2].

Despite being frequently used in research, the clinimetric properties of the GDI and GPS have only partly been described in the literature. Studies of children with CP have shown that the GDI and GPS demonstrate satisfactory concurrent validity when compared with gold standard measures of gait and gross motor function [3], [4], [5], [6]. They have also shown responsiveness to surgical lengthening of the gastrocnemius muscle [7], [8], and the GDI has been reported as a reliable measure within a single session [9]. However, intra-tester reliability and agreement across two separate sessions have, to our knowledge, only been investigated for the GDI in typically developing children, demonstrating limits of agreement of ±10 points and a non-significant difference between the two sessions [9]. The variability between two separate sessions can be described as ‘intrinsic variation’, which reflect biological variation within individuals under investigation and ‘extrinsic errors’, such as inconsistent marker placement, anthropometric measures, data sampling and processing [1].

In addition, only a few studies have investigated the reliability and agreement of 3DGA on children with CP. A systematic review found three studies reporting reliability and agreement of 3DGA [1]. Two of the papers used a recently criticized method, coefficient of multiple correlation [10], to investigate reliability in gait and reported values between sessions above 0.7 [11], [12]. The third paper reported intraclass correlation coefficient (ICC) above 0.6 [13]. The above mentioned three studies compare differences in the joint movement between sessions, in contrast to the present study, where variations from reference dataset between sessions are compared.

To be clinically and scientifically useful, measurements must be both valid and reliable [1], [14]. Thus, to provide complete clinimetric properties of the GDI, GPS and GVS, there is a need to investigate their reliability and agreement to inform a decision about which indices to use in future research investigating the quality of gait kinematics. Therefore, the aim of this study was to investigate the intra-rater reliability and agreement of the GDI, GPS and GVS in children with CP across two repeated sessions.

Section snippets

Methods

A test–retest trial was conducted to evaluate the intra-rater reliability and agreement across two repeated sessions of the GDI, GPS and GVS in children diagnosed with CP. Ethics approval was obtained from the Committee for Medical Research Ethics in the Region of Southern Denmark (S-20120162) and the Danish Data Protection Agency (12/25588). The current study conforms to the format described in the article ‘Guidelines for Reporting Reliability and Agreement Studies’ (GRRAS) [14].

Results

Three teams of two assessors conducted the data collection. Two assessors had a background as physiotherapists and one was a biomedical engineer. The assessors had 6–18 months of experience at our gait laboratory, and they all underwent a training period before the sampling began. In total, 18 children volunteered to participate. The characteristics of the participants are presented in Table 1. The two 3DGA sessions were separated by 1–9 days (mean 4.4 days, SD 2.9). The data include the scores

Discussion

Despite the wide use of summary measures of gait (GDI, GPS and GVS) in research and clinical practice for, their reliability and agreement have not previously been demonstrated.

Given that the GDI and GPS are different ways of scaling the same underlying construct, it might seem redundant to investigate both of the indices [2]. However, despite the similarity in the underlying approach, different mathematical methodologies are used and consequently, the reliability of each kinematic variable

Conclusion

Excellent reliability and acceptable agreement and no systematic bias between test sessions were found for the Gait Deviation Index (GDI) and Gait Profile Score (GPS), the GDI and GPS can be used to document changes in deviations from normal gait in children with spastic cerebral palsy. Furthermore, our study showed large variability in both reliability and agreement for the Gait Variable Score, which might be important information in the interpretation of gait analysis in clinical practice and

Conflict of interest statement

All the authors of this manuscript declare that they have no conflicts of interest related to the current study.

Acknowledgements

We thank all volunteers in the project; physiotherapist Lotte Slot Jensen for helping with data collection and Suzanne Capell for English language proofreading of the manuscript. The authors would also like to acknowledge the support of the University of Southern Denmark, Odense; Odense University Hospital Research grants (2012), the Region of Southern Denmark Research grants and Ph.D. grants (2012), the Physiotherapy Praxis Foundation, Ludvig and Sara Elsass Foundation and The Danish

Cited by (0)

View full text