Post-stroke Visual Gait Measure for Developing Countries: A Reliability and Validity Study
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.266273
Source of Support: None, Conflict of Interest: None
Keywords: Gait assessment, rehabilitation, reliability, stroke, validity
Gait deviation is a common manifestation of post-stroke hemiparesis. The deviation occurs because of the asymmetry caused by the half-side body paresis. In addition to the paresis, abnormal muscle activation, synergy, increased tone, imbalance between agonist and antagonist muscles, and contracture also contribute to the hemiparetic gait. Foot drop, knee hyperextension, and pelvis hiking are some of the commonly observed disturbances in stroke patients.,, As and when patients start walking, they usually adopt a stereotypical abnormal gait pattern. Most spatiotemporal variables such as stance time and step length on the affected leg and walking speed are altered. The pathomechanics leads to inefficient, unsafe, or nonfunctional walking. Furthermore, post-stroke patients have low or no perception for the actual gait asymmetries. Most subjects do not achieve functional and community ambulation. The disability further affects the quality of life. Post-stroke gait rehabilitation primarily depends upon determining the specific deviations. Thus, an objective investigation is essential to plan a successful intervention program.
Various methods are available in clinical and research practice to identify and quantify the impairment. A complex phenomenon such as gait, being a complex dynamic phenomenon, can be analyzed by sophisticated kinematic and kinetic analysis. However, gait analysis are too expensive to use in the regular practice, especially in developing countries.
Visual gait observation, a clinical and cost-effective technique, is very useful in discerning deviations during walking among post-stroke hemiparetic subjects.,, Although clinical observation of the gross deviations can be performed by the naked eye, the detection and quantification of all the possible impairments require a structured approach. There are few observation-based tools available to assess gait related impairments for neurological disorders including stroke. The Rivermead Visual Gait Assessment (RVGA), the Gait Assessment and Intervention Tool, the adapted New York Medical School Orthotic Gait Analysis work sheet, the Hemiplegic Gait Analysis Form, and the Wisconsin Gait Scale are some of the identified measures for observational gait evaluation in stroke. RVGA assesses kinematic aspect of the gait deviations and is considered to be the clinically useful measure. The scale has items for the upper limb, trunk, pelvis, hip, knee, and ankle to discern typical alterations exhibited by a hemiparetic subject during walking. The head-to-toe order of items allows an assessor to be at ease during visualizing the impairment. The items are separately categorized for stance and swing phases of the gait. The RVGA also has a provision to evaluate opposite aspects of deviation; for instance, knee excessive flexion or extension as well as ankle in excess plantar flexion or dorsi flexion. The extremes allow the measurement possible deviations that might be adapted by the different stroke patients. The vision of examiner is the medium to recognize and compute the gait-deviations. Thus, underlying knowledge and expertise in gait-biomechanics play a decisive role for a skillful rater.
The developer of RVGA has already established its initial reliability and validity. No further studies have been conducted to confirm and expand the findings. The available information on psycho-clinocometric properties of the measure is not adequate. Further, the gait assessment by RVGA may be enhanced using videos. However, the reliability and validity of the same has not been established yet. The objective of the study was to establish the reliability and validity of the video-based RVGA in post-stroke subjects.
Subjects were recruited from a lower extremity trial conducted in a rehabilitation institute. The present study was approved by the ethics committee of Pandit Deendayal Upadhyaya National Institute for Persons with Physical Disabilities, New Delhi, India. The participants were enrolled in the study if they met the following inclusion criteria: (i) unilateral stroke of >6 months, (ii) hemiparesis of either right or left side, (iii) age between 30 to 60 years, (iv) normal ambulation prior to stroke, and (v) able to walk for at least 10 m without any physical assistance, foot orthosis, and walking-device. The subjects were excluded if they exhibited the following features: (i) any neurological or orthopedic disorder affecting locomotion and balance, (ii) severe cognitive and perceptual deficit, and (iii) cardiovascular instability.
The present study was a prospective observational design.
The Rivermead visual gait assessment
The RVGA, a visual-based gait measure used to discern gait deviations in neurologically disorders, especially stroke. It consists of 20 items (2 arms, 11 stance phase, and 7 swing phase) which assess the deviations on the affected side. Each item is scored on 4-point scale ranging from 0 to 3 (0 = normal, 1 = mild deviation, 2 = moderate deviation, and 3 = severe deviation). Only item 2 is scored up to moderate alteration; thus, the total score varies from 0 (normal gait) to 59 (severely affected gait). Visual observations are made on the impaired lower limb while the patient walks at a normal place with orthotic devices as applicable. However, in the present study, participants having ability to walk without any such appliances for at least 10 m were recruited. This allowed for the true presentation of gait deviation as well and increased visibility of the ankle-foot for observation were clearly visible for the observation.
In the preliminary study, the developer of the RVGA determined it to be it a reliable and valid tool for the assessment of gait deviation in a history of a neurological problems, such as stroke or multiple sclerosis patients.
Validity related measures
Fugl–Meyer assessment (lower extremity)
The Fugl–Meyer assessment (FMA) is the most commonly used motor measure to quantify recovery of post-stroke hemiparetic patients. The FMA comprises subsection exclusively for the lower extremity (FMA-LE). In general, FMA-LE has six items not varying from reflex activity, pure and combined synergistic movements, and voluntary performance. The item further consists of 17 sub-items in total. The motor performance is scored on a three-point ordinal scale (0, no performance to 2, faultless performance) with a maximum score of 34 indicating good motor recovery and minimum score of 0 signifying poor recovery.
The FMA has been found to be a reliable (0.95 to 0.99) and valid (0.71 to 0.76) tool., Specifically, FMA-LE exhibited adequate reliability and validity to assess motor recovery of the lower limb post-stroke.
The 10-meter walk test (10MWT) was used to estimate the gait speed, through analysis of the time taken to cover the 10 M distance., The participants were to walk at their comfortable and maximal speed for a distance of 10 m. The time was not considered for the first and the last 1 m to allow for acceleration and deceleration, respectively. The mean time of three readings was calculated for the performance time. The gait speeds [comfortable (Speed C) and maximal (Speed M)] in meter/second (m/s) were estimated using the relation = 8/the mean time of performance. The 10MWT exhibited considerable reliability (0.95 to 0.99) and validity (0.60 to 0.94) in patients with stroke.,,
Time up and go test
The Time up and go test (TUG) is a time measure of sit-to-stand and 3-m walk, to and fro. The test is demonstrated to have excellent reliability (0.95) and considerable validity (0.62 to 0.96) in stroke patients.,
Berg Balance Scale
The Berg Balance Scale (BBS) assesses balance and fall risks among stroke subjects. Although the scale was developed for elderly, it is commonly used and found to be reliable and valid in post-stroke patients. This 14-item performance-based measure is scored on a five-point ordinal scale range from 0 (poor balance) to 4 (good balance). BBS showed good to excellent reliability (0.72 to 0.95) and validity (0.82 to 0.94).,,
Prior to the trial intervention, all the participants were made to walk at comfortable speed on a 15-m obstacle-free pathway. A reference line was drawn on the floor to provide a guiding direction during walking. The lower dress of the subjects was folded up to above knee to have a clear view of the lower leg and joints. A bicolor VELCRO strap was also tied around the waist (along the line joining both anterior and posterior superior iliac spines) to view the deviation related to pelvis level. The participants were not allowed to wear any foot wear during the test. The walking was video-taped from anterior aspect, posterior aspect, affected side, and less-affected side. Four separate rounds of the walking were conducted to take the videos from each view. The videotaping was conducted by staff not part of the study at the research site at the research-site. After coding the tapes, the staff provided them to the four different raters in random order. Two raters had more than 20 years of experience in neurorehabilitation, and the others were novice in the field. A common discussion on each item of RVGA was held to clarify the observation and scoring as per the guidelines of the scale developers. Prior to the study, three trial videos were scored by the raters to acquaint with the method of assessment.
Each rater scored the coded video on the RVGA data collection sheets twice, one at the baseline and other after one month to eliminate any recollection of the initial assessment. The assessors were not informed about the paretic side of the subjects to reduce any bias of negligible to mild deviations. During each scoring session, the rater played the coded video clip of a subject using VLC Media Player version 2.2.1. The playing speed of clip was slowed to half. The reduced speed of video allowed the assessor to perceive the deviations more accurately for each body part during every phase of the gait cycle. Only one clip was scored by each rater in a day to avoid any possible error because of stereotypical observations. Further, each rater scored the measure separately on an individual computer system.
10MWT, TUG, and BBS were conducted in a separate session as per the standard guidelines of each measure.
The assessments were considered from an ongoing trial in which as conventional management, the affected lower extremity received motor therapy according to neurophysiological approaches such as Brunnstorm and Bobath., The intervention comprised of reflexive, synergistic, out-of-synergy, active, weight bearing, and anti-deformity movements. The experimental protocol consisted of task-based mirror therapy. Object based movements such as ball rolling, rocking-on-board, wiping, pedaling, shifting were performed by the less-affected side in sitting/long sitting position in front of a mirror frame while hiding the affected limb. The movement of the less-affected lower limb was reflected and perceived as the affected extremity.
Reliability and validity analysis of the data in addition definition of the descriptive and inferential statistics was conducted using IBM SPSS, Version 22.0. The reliability for the global RVGA score and validity was analyzed using the Pearson correlation coefficient (r). First assessments of all the raters were considered for validity estimation. A P value ≤0.05 was considered as significant. The coefficients 0.50 to 0.69, 0.70 to 0.89, and >0.09 were considered as moderately, highly, and very highly correlated, respectively. Furthermore, level of agreement between and within raters for individual RVGA item was calculated by the Kappa measure of agreement (κ). The κ value of <0.2, 0.21 to 0.4, 0.41 to 0.6, 0.61 to 0.8, and > 0.8 were considered as poor, fair, moderate, substantial, and perfect agreement, respectively.
All enrolled subjects signed the informed consent form and completed the assessment protocol. Out of 40 subjects, there were 34 males and 6 female, the average age of the participants was 52 years. The mean post-stroke duration was 19 months, with 30 subjects having ischemic stroke and 22 patients exhibiting right side paresis. The detail demographic characteristics of the participants are given in [Table 1].
The analysis for total RVGA scores between the experienced raters exhibited a very high correlation (r = 0.940 P < 0.001). Similar strength of relation was also found for the novice raters (r = 0.941, P < 0.001). The individual item analysis demonstrated κ value ranging from 0.213 (P = 0.013) to 0.903 (P < 0.001) for the experienced raters. The individual item analysis for the novice therapist varied from 0.278 (P = 0.006) to 0.774 (P < 0.001). However, three and four items of RVGA showed no agreement for the experienced and novice therapists, respectively [Table 2]a and [Table 2]c.
The correlation between the first and the second assessment was found to be very high for all the raters, r ranging from 0.951 to 0.958 (P < 0.001). The κ for individual RVGA item varied from 0.222 (P = 0.018) to 0.776 (P < 0.001) and 0.252 (P = 0.05) to 0.695 (P < 0.001) for both the experienced raters, respectively. For the novice raters, the agreement level was found to be 0.393 (P < 0.001) to 0.788 (P < 0.001) and 0.212 (P = 0.001) to 0.624 (P < 0.001) [Table 2]b and [Table 2]c. One item of RVGA each for an experienced rater and a novice rater did not show any agreement.
The total RVGA score exhibited significant positive relation with BBS for all the raters (r = 0.386 to 0.477, P = 0.014 to 0.002). However, RVGA was not found to be correlated with FMA-LE, 10MWT, or TUG (P > 0.05) [Table 3].
Gait impairment among post-stroke survivors interferes with the functional recovery and increases the disability level. Management requires critical assessment of the deficits. Various tools of various merit and demerit available ranging from the observational to sophisticated analysis. Because the gait is a complex and swift phenomenon, observational scales are usually considered as less reliable as compared to the machine-based measures. Gait-analysis are costly, complex, and time-taking tools. Because of the inclusion of smart phones in daily life, obtaining video of walking is quite easy and economical for clinicians. RVGA is an old gait measure for neurologically impaired subjects, especially stroke, developed at a time when videography was more difficult to capture effortless. Although the items of RVGA are most appropriate for post-stroke gait evaluation, the chances of missing a deviation are common during the eye-observation method. Thus, the present study was conducted to validate the assessment process of RVGA using gait videos.
The findings suggest that the RVGA is a reliable and valid way of assessing gait in patients with stroke. These results confirmed the preliminary psychometric reports conducted on a small sample by the developers of the scale. The present study is the first investigation to utilize the items of RVGA for quantifying deviations by videography. The scale is useful in evaluating array of gait-deviations exhibited by post-stroke hemiparetic subjects. The slow-motion assessment of video-tape from four different views (anterior, posterior, affected side, and less-affected side) allowed raters to decisively discern the deviations without causing discomfort to the patient.
Observational gait measures are commonly used in clinical practice. In addition to RVGA, other measures are also recommended. Video-based assessment using scales such as Wisconsin Gait Scale and the Gait Abnormality Rating Scale have also been conducted. Both the measures were found to be reliable and valid using video-protocol., In addition, there are a range of gait related measures used in post-stroke studies; however, most of them are nonvideo-based.,, The tools are described in [Table 4]. All they should be considered based on their advanatges and disadvangates and then used. The video-based validation of RVGA was a need. The present study authenticates the use of RVGA adjunct by videos in stroke-gait assessment.
The inter-rater reliability of RVGA was found to be good by its developer. However, the reliability was observed up to very high in the present study. This may be because of the use of gait videos. This validates the use of videos in the observational gait assessment. The experience of rater was not found to be an important factor for using RVGA. Few swing-phase items did not show any agreement between the raters, two (trunk flexion and pelvis backward rotation) of them are common between the experience and novice raters. This could be because of the dynamic and complex aspect of the swing phase which was difficult to be discerned by the raters. Both the group of therapists demonstrated almost same level of intra-rater reliability. All the raters exhibited acceptable agreement for almost all the items. The majority of the items had fair to substantial agreement followed by negligible perfect and poor agreement.
The relation between RVGA and BBS is in concordance with that of the developer studies. However, no association of RVGA was observed with FMA-LE, 10MWT, or TUG. This could be attributed to the constructs of the measures. The FMA-LE assesses the motor recovery components in the static positions such as lying, sitting, and standing. Further, it does not consider the trunk status and less-affected limb. In contrast, the RVGA examines the deviations involving the trunk, pelvis, lower limbs, and upper limbs in various dynamic aspects. In addition, the post-stroke gait deviation is a maladaptive learned behavior that may not be influenced by the recovery of individual FMA-LE item. The construct of 10MWT and TUG is not stroke-specific. Post-stroke subjects maintain the speed of walking for functional achievements despite the deviations. Furthermore, TUG has exhibited no relation with the kinetic and kinematic variables in stroke.
In view of the cost and time factor for gait analyses, video-based RVGA may have a wider application in developing countries. In the present study, the videos were captured using a smart-phone. Structured procedures to record the clips need to be developed in the future. The RVGA exhibited no relation with the speed measures of gait. This indicates that the post-stroke subject could maintain functional walking in spite of the gait deviations. Although recognition of deviations was undemanding, the raters occasionally experienced difficulty in separating the mild and moderate deviations. The RVGA can only be applied to independent ambulatory post-stroke subjects. Future studies are recommended to establish the validity of RVGA using gait analyzer. The other type of validity is also needed to be determined. Further, there is a need to develop a standard training manual for the RVGA using videos to have uniformity.
The RVGA is a reliable and valid tool to assess the gait-related impairment in post-stroke hemiparetic patients. It is a simple and economical method to assess the stroke-gait, a complex phenomenon in clinical and research purpose.
Financial support and sponsorship
Pandit Deendayal Upadhyaya National Institute for Persons with Physical Disabilities, New Delhi, India.
Conflicts of interest
There are no conflicts of interest.
[Table 1], [Table 2], [Table 3], [Table 4]