Intrarater and Interrater Reliability of Dynamic Gait Index in Post Stroke Patients With Eye Movement Disorders

Background: The Dynamic Gait Index is a useful scale that has been evaluated for its reliability in patients with vestibular disorder, elderly people and, in chronic stroke population but its reliability has not been evaluated yet in sub-acute stroke patients with eye movement disorders. So the purpose of this study was to evaluate the intra-rater and inter-rater reliability of the Dynamic gait index to measure the dynamic balance, gait and risk of fall in sub-acute stroke population with eye movement disorders. Methods: A total of 30 (18 male, 12 females) stroke patients in the sub-acute phase suffering from eye movement disorders were recruited for this reliability study. Two experienced Physical therapists assessed the intra-rater and inter-rater reliability of the Dynamic gait index in two testing sessions three days apart. In the later session, two raters assessed the patients’ performance simultaneously on the Dynamic gait index. Data analysis was done at 95% condence interval using the intra-class correlation coecient (ICC2, 1). A signicance level was set at P-value <0.05. Results: Mean tests scores by rater 1 in the rst and second assessment were 17.4±2.04 and 18.0±2.05. Mean tests score by rater 2 in the second assessment were 18.6±2.06. Intra-rater (ICC=0.86, CI=0.73-0.93) and inter-rater (ICC=0.91, CI=0.83-0.96) reliability of total DGI scores was found good to excellent while intra-rater (ICC=0.73-0.91) and inter-rater (ICC=0.73-0.93) reliability of individual items of DGI scores were ranged from moderate to good. Item 8 (steps) showed lowest reliability (ICC=0.73). The signicant difference was seen in total and individual scores (P<0.001) of DGI scale in both intra-rater and inter-rater reliability. Conclusions: The Dynamic gait index is found clinically a reliable tool to objectively measure dynamic balance, gait and risk of fall in sub-acute stroke individuals with eye movement disorders. The intra-rater and inter-rater

Conclusions: The Dynamic gait index is found clinically a reliable tool to objectively measure dynamic balance, gait and risk of fall in sub-acute stroke individuals with eye movement disorders. The intra-rater and inter-rater reliability of the total DGI scores was found good to excellent; whereas intra-rater and interrater reliability for individual items of DGI were varied from moderate to good.

Background
Stroke is the second major cause of death and most common cause of disability after ischemic heart disease worldwide (1). It affects around 10.3 million people worldwide every year and estimated to rise by 60-80% in stroke cases by 2050 (2). Its incidence is increasing in Asia and mortality rate due to stroke is also higher in Asia than in America and western Europe (3). In Pakistan the annual expected prevalence of stroke is 250/100,000 population, projecting to 350,000 new cases every year (4).
Visual impairment is one of the most disabling consequences of stoke. Among all visual impairments, eye movement disorders are seen in 68% of the post-stroke population (5). Eye movement disorders that occur as a result of stroke includes, gaze palsies, cranial nerve palsies, strabismus, vergence abnormalities and nystagmus (6). With these eye movement disorders, a patient is unable to respond e ciently to visual inputs from the environment which may result in decreased dynamic stability leading to an increased risk of falls (7). Frequency of falls that occur in the post-stroke population in the initial three to four months is about 34% and the majority of these falls (40-90%) occur during walking (8).Thus, to prevent falls in post-stroke patients with eye movement disorders the evaluation of dynamic balance and gait disorders is of signi cant clinical importance (9).
In clinical practice, various clinical tools have been used to assess dynamic balance and walking performance in stroke patients such as functional independent measure (10), functional classi cation ambulation (11) and timed walking tests (12) but these clinical measures provide limited information to plan treatment and outcome measurements. In addition, it has been reported that dynamic balance of chronic stroke patients during walking cannot be measured with static balance scales (13).
Thus, in order to plan and establish treatment and rehabilitation strategies and to follow up the patients' progress, a reliable and valid outcome measure for dynamic balance and gait related parameters is necessary for stroke patients with eye movement disorders to plan and establish treatment strategies and to record the outcome. Dynamic gait index (DGI), develop by Shumway-Cook and Woollacot (14) is reported to be a clinically useful measure to examine the functional balance and fall risk in a variety of conditions including elderly people (15) and patients with vestibular dysfunction (16) and multiple sclerosis (17). However the reliability of DGI has not been examined yet in subacute stroke patients with eye movement disorders.
Hence, the aim of this study was to assess the intra-rater and interrater reliability of DGI to measure the dynamic balance, gait and risk of fall in sub-acute post-stroke patients suffering from eye movement disorders.

Subjects
A sample of 30 patients suffering from post-stroke eye movement disorders was recruited from the outpatient rehabilitation department of physical therapy at the University Teaching Hospital University of Lahore, Pakistan between December 1, 2019 to May 31 st , 2020. Data were collected after taking approval from the University of Lahore's Institutional Review Board. All methods were performed in accordance with the relevant guidelines and regulations. All the recruited patients signed the written informed consent after taking detailed information about the study procedure. The sample size was estimated based on data from studies done previously in the literature (13). The inclusion criteria were; post-stroke patients of both genders between the age of 19-60 years, diagnosed with an eye movement disorder by a neurophysician following a stroke for the rst time, 3 to 6 months after stroke, ability to walk at least 10m distance with or without an assisted device and able to follow instructions. Patients were excluded from the study if they had a recurrent stroke, any other neurological or orthopedic impairment, organic disorder, and cognitive impairment.(18), (19), (20) Dynamic Gait Index DGI examines the dynamic balance, gait and risk of fall during walking. It is comprised of eight walking items that include; gait on level surfaces, changing speed, gait with horizontal and vertical head turns, pivot turns, step over and around an obstacle and stair climbing. Each individual item is scored on a 0 to 3 scale. The maximum scores that a subject can achieve on DGI are 24. Scores below or equal to 19 speci es a high risk of fall (13), (14).

Procedures
Two experienced physical therapists termed rater 1 and rater 2, independently assessed the patients' DGI based responses. The original version of DGI was used to record patients' scores (14) (see supplementary le). To measure intra-rater reliability, rater 1 assessed the patients twice in two testing sessions three days apart. To measure inter-rater reliability, rater 2 also assessed the patients simultaneously along with rater 1 in second testing session. Both assessors were not permitted to discuss with one another about the tests results and did not have access to prior results. Standard verbal instructions were given to all the patients and each test was performed in a quiet room. All subjects were allowed to use their walking aids in which they feel safe and comfortable. On average each rater took 15 minutes to complete the DGI based assessment for each patient.

Statistics
Descriptive statistics were performed to express patients' demographics and were presented as mean±standard deviation. Intra-rater and inter-rater reliability were calculated using the intra-class correlation coe cient (ICC 2,1) at 95% con dence interval. ICC values<0.05 showed poor reliability, 0.5 to 0.75 indicated moderate, 0.75 to 0.9 indicated good and 0.9 and above indicated excellent reliability (21). P-value was considered signi cant statistically at P<0.05.

Results
Baseline Characteristics:

Discussion
Dynamic balance and gait impairments are the most commonly seen functional limitations in stroke patients with eye movement disorders. DGI is a clinically reliable scale to measure functional balance during walking activities (22).To the author's knowledge, this was the rst research study to assess the intra-rater and interrater reliability of DGI to measure the dynamic balance, gait and risk of fall in subacute stroke population with eye movement disorders. Our ndings provide evidence that DGI is a reliable outcome measure to evaluate the balance and walking functions during gait activities in stroke patients with eye movement disorders.
Our results showed that both the intra-rater and inter-rater reliability of total DGI scores was found good to excellent in this study. Previously, reliability of DGI has been reported moderate to excellent in older people (23), in patients with vestibular dysfunction (24), chronic stroke (25) and Parkinson's patients (26). The reliability of individual items ranged considerably from moderate to excellent in this study, compared to poor to excellent reliability in the studies done previously (23)-(26). Due to difference in study designs, patients' population, and techniques used in statistical analysis, it is di cult to compare results of different studies; though the ndings of previous studies support this view that DGI is clinically a reliable tool when assessment is done based on total scores (27).
In the current study, with regard to the inter-rater reliability, total DGI scores obtained by rater 1 and rater 2 simultaneously in the second assessment were in the range of 14-21and 15-22. It has been described by Shumway-Cook et al. a score equals to19 or below, out of possible 24 speci es a high risk of fall (14). According to this criteria, 40-50% patients in our study were at high risk of fall. Also, mean DGI scores It is more likely that their sample represented a population who were in fair to good condition.
In contrast, another study reported on vestibular dysfunction patients, inter-rater reliability (0.63) using the composite kappa statistics was found as compared to the inter-rater reliability (ICC=0.91) calculated in the present study using intra-class correlation coe cient. There may be different reasons why low reliability was found in vestibular dysfunction patients (29). One reason is that the raters who assessed the patients with vestibular dysfunction on DGI were not experienced and might not have used standardized instructions and were not briefed properly about the scale. Nevertheless, the raters in this study were trained physical therapists and received uniform brie ng before the assessment. So the rater's training in grading the patient's performance on the scale and the patient's behavior would have increased the reliability scores in the present study. Secondly, stroke patients in the present study were having balance problems that have not showed any improvement in performance in three days interval between tests. Whereas patients with vestibular dysfunction exhibited more acute symptoms so even after an hour interval they showed changes in subject's performance leading to low inter-rater reliability.
In the present study, item 8 (steps) showed the lowest DGI scores, while all other items demonstrated good to excellent intra-rater and interrater reliability. These results are similar to the previous study on Parkinson's patients which reported that intra-rater (ICC=.88 1.00) and inter-rater (ICC=.87 -1.00) reliability of all single items demonstrated good to excellent reliability, only item 8 (steps) (ICC=0.73) showed moderate reliability (26). This is because Parkinson's patients had di culty in looking up from a distance during walking and in the present study, post stroke patients with different types of eye movement disorders also had di culty in moving eyes, had rapid eye movements and di culty in looking up towards a task. In both studies, the assessment was performed on day one and with the interval of three days. However, the mean age of the Parkinson's patients was found to be much higher 80.4 years as compared to the mean age of sub-acute stroke patients in the current study i.e, 53.2 years.
In the present study stroke patients with different types of eye movement disorders were observed but due to limited number of patients reported in each type we cannot conclude that which type of eye movement disorder showed more in uence on balance, gait and fall risk. Future studies should be conducted on large sample to determine this difference. However, ndings of this study showed that DGI is found principally a scienti c, reliable, and quick to administer tool to examine the quality of dynamic balance, gait and fall risk during walking activities. This study will help the practitioners to adopt the DGI to plan and establish more scienti c interventions for post-stroke individuals with eye movement disorders.

Limitations Of The Study
In this study, two assessors who assessed the patients' performance were neurologic physical therapists and have previous experience with the DGI scale. It is more likely that reliability of scale may vary if the raters have a different background. Secondly, most of the patients in this study were ambulatory and were capable to walk without assistance but with an assisted device if necessary. Hence, the results of the current study will be generalized to the same population. As with walking aid DGI becomes a 3-point scale. This may result in better reliability of this scale. Thirdly, the study participants who were recruited in the current study were quite younger than the average age of stroke onset. This is because we didn't include more severe patients who could not able to walk or follow instructions. So the ndings of present study are only applicable to the same population.

Strengths
The foremost strength of current study was that, intra-rater and inter-rater reliability of DGI was assessed for the rst time in post-stroke patients with eye movement disorders. No reliability study of DGI has been done before on this population. Secondly, statistical methods used in this study were most appropriate according to the current guidelines and were comparable to the results from other studies. Thirdly, DGI scale takes less than 15 minutes to administer. The scale can be utilized clinically in post-stroke patients with eye movement disorders to evaluate dynamic balance during walking and for training dynamic balance.

Conclusions
Our results suggest that the Dynamic Gait Index is a reliable instrument to objectively assess dynamic balance, gait and risk of fall in sub-acute stroke population with eye movement disorders. The intra-rater and inter-rater reliability of the total DGI scores was found good to excellent; whereas for individual items of DGI it varied from moderate to excellent and thus, dynamic gait index is recommended and can be used in clinical settings by both clinicians and researchers. Further research is needed to investigate its validity, responsiveness and to correlate DGI with other measures to examine the dynamic balance during gait activities and to determine risk of fall in patients with post-stroke eye movement disorders.