Abstract
BACKGROUND:
Visual vertigo (VV), triggered by environmental or dynamic visual stimuli and repetitive visual patterns, can affect daily life activities. The Visual Vertigo Analogue Scale (VVAS) is a valid and reliable self-administered questionnaire to assess VV, which has been culturally adapted to the Argentine population but has not been validated.
OBJECTIVE:
To validate the Argentine version of VVAS (VVAS-A) by confirming its psychometric properties in patients with vestibular disorders.
METHODS:
Vestibular patients (n = 82) completed the VVAS-A and the Dizziness Handicap Inventory Argentine version (DHI-A) during their initial visit and one week later. The VVAS-A's internal consistency, test retest reliability, ceiling and floor effects, and construct validity were determined. Test-retest data (n = 71) was used to calculate reliability using the intraclass correlation coefficient (ICC 2.1).
RESULTS:
A ceiling effect was observed in 12 patients (14.6%). Internal consistency was acceptable (Cronbach’s alpha: 0.91). The reliability was r = 0.764 [CI 95%: 0.7 –0.86]). Correlations were observed between the VVAS-A and the total DHI-A score (rho = 0.571), the DHI-A physical subscale (rho: 0.578), and DHI-A functional and emotional subscales of the DHI-A (rho: 0.537 and 0.387, respectively).
CONCLUSION:
The VVA-A is a valid, reliable tool to evaluate VV in patients with vestibular disorders.
Background
Visual vertigo (VV) causes dizziness or discomfort in dynamic visual environments and negatively impacts activities of daily living [7, 29]. It can be triggered by looking at moving cars or trains, watching action movies, walking down the aisle of a supermarket, or using escalators [6, 38]. Questionnaires designed to assess VV include the Situational Characteristics Questionnaire [23, 45], the Longridge and Mallinson interview questions [28, 32], the Pediatric Visually Induced Dizziness Questionnaire [33], and the Visual Vertigo Analogue Scale (VVAS) [12, 13].
The VVAS consists of nine visual motion scenarios and the patients rate the intensity of their symptoms on a visual analogue scale [13]. The VVAS has good internal consistency, is used to document treatment results, and is utilized in vestibular rehabilitation programs [12, 50]. A moderate positive correlation has been found between the VVAS and the Dizziness Handicap Inventory (DHI) [13]. In Argentina, there is a VVAS version (VVAS-A) that has been culturally adapted but requires validation (Fig. 1) [44].

Translated questionnaires in languages and cultures other than the ones of origin need to be semantically and conceptually equivalent to the original, but also culturally adapted [2, 47]. Additionally, the psychometric properties (feasibility, reliability, and validity) of the newly developed VVAS-A must be determined [1, 43]. The objective of this study was to validate the VVAS-A by determining feasibility (time to take the test), reliability and its validity in patients with vestibular disorders. It was hypothesized that the VVAS-A has good psychometric properties.
Participants and procedures
The study included consecutive Argentine nationals (n = 82) who were greater than 18 years of age who presented to an out-patient physical therapy clinic in Buenos Aires with a vestibular diagnosis between January 2016 and January 2018. The patients were referred to the physiotherapy clinic by the local otorhinolaryngology and neurology services. All subjects were assessed and referred for rehabilitation by otoneurological specialists. In those cases where the referral did not contain a specific diagnosis (n = 17), an experienced physiotherapist evaluated each subject by means of a battery of clinical oculomotor (ocular alignment and tests of skew, ocular range of motion, smooth pursuit, volitional saccades, and vergence) and vestibular tests [vestibulo-ocular reflex cancellation test ((VORc)), head impulse test, head shake test, or mastoid vibration test], thus ensuring that the disorder was primarily of vestibular origin [13]. Those patients who had positive clinical vestibular findings were included in the study (Table 1).
Demographic characteristics patients with vestibular disorders
Demographic characteristics patients with vestibular disorders
SD: Standard Deviation. IQR: interquartile range. VVAS-A: Argentine version of the Visual Vertigo Analogue Scale.
Exclusion criteria were persons who could not read, non-Spanish speakers, those with communication and/or understanding difficulties, a history of mental illness or visual disorders (untreated diplopia, amaurosis, glaucoma, macular degeneration, cataracts, myopia or astigmatism), cardiac arrythmia, benign paroxysmal positional vertigo (BPPV) and non-vestibular central nervous system neurological disorders (stroke, Parkinson’s disease, cerebellar disease, or multiple sclerosis). The study was approved by the Ethics Committee at Carlos G. Durand Hospital of the city of Buenos Aires (DI-2015-480-HGACD; November 18, 2015). Patients signed an informed consent, and all study procedures and results were performed and reported in accordance with literature recommendations and guidelines [40, 43].
All participants (n = 82) were asked to complete the culturally adapted VVAS-A and the DHI-A at visit 1 and at visit 2 one week later in a random order [9, 24]. The VVAS scale is a self-administered questionnaire, however, the definition of VV was provided verbally to the patients prior to them completing the VVAS-A so that they fully understood the concept [44]. The definition of visual vertigo described to participants was “dizziness caused by dynamic visual scenes, such as computer screens, police lights or large crowds. It is not dizziness caused your body movement.” The Global rating of change scale was used at the beginning of the second visit.
The VVAS-A is a 9-item self-administered questionnaire developed to assess VV [13]. Items are rated by the patient on an analogue scale ranging from 0 (no dizziness) to 10 (extreme dizziness). Each item represents a daily situation that may induce VV and the severity of the visual vertigo is rated by asking patients to draw a vertical line across a 10 cm horizontal line. A person is considered to experience VV when two or more items are rated above zero. The following equation is used to estimate severity: VV severity = (sum of all the items’ analogue scale scores / number of items answered) x 10. A score of 0 means the patient does not experience VV, while a 90–100 score indicates severe VV. The VVAS is user-friendly and allows for rapid administration and grading. The scores for each item are calculated with the use of a ruler, added together and then divided by 100 for the total score of the VVAS-A based on the number of recorded responses.
The DHI-A [9] is a self-administered questionnaire that provides a rating of self-perceived handicap due to imbalance, dizziness, and/or vertigo [24]. The DHI includes 25 items that can be further subdivided into three subscales (functional, emotional, and physical). The worst possible score is 100:36 points pertain to emotional aspects (9 questions), 36 functional (9 questions) and 28 to physical aspects (7 questions). For each question there are three possible answers: yes (4 points), sometimes (2 points), and no (0 points). The higher the score, the greater the level of self-perceived handicap [24].
Global Rating of Change (GRC) scales record self-perceived change in health status (improvement, deterioration, or no change) [26, 49]. The GRC scale has been used in research and clinical practice to assess the effect of an intervention or the clinical course of disease. In this study, patients were asked the following question in one week after completing the VVAS-A and the DHI-A: “Regarding your visual vertigo, how do you feel now as compared to your first visit?”. Answers were recorded using an 11-point numerical rating scale ranging from –5 (much worse) to +5 (completely recovered) with descriptors at both ends and a mid-range descriptor (0 = “no change”) [26].
Patients were told not to begin their exercise program until after their second visit to the physical therapy clinic. During the second visit, patients were asked to complete a global rating of change score to determine if their condition was stable plus complete the VVAS-A and the DHI-A for the second time. Participants who scored less than a two-point difference on the global rating of change scale related to their visual vertigo were included in the reliability study (n = 71) [26].
Internal consistency, ceiling/floor effects, test-retest reliability and construct validity of VVAS-A were determined. Patient characteristics (i.e. age, sex, diagnosis, length of symptoms) were recorded. The time to complete the VVAS-A was recorded in seconds [44]. Patients with a < 2 point difference in their GRC scores were regarded as not having significant changes in their self-perceived health status and were included in the reliability analysis [14]. To determine if the VVAS-A was clinically useful, the data were analyzed for all persons who completed their second visit regardless of their self-reported GROC rating.
Internal consistency describes the extent to which all the items in a test measure the same concept or construct and hence it is connected to the inter-relatedness of the items within the test [41]. Internal consistency was assessed using Cronbach’s alpha for the first visit VVAS score. Internal consistency is considered acceptable when the coefficient ranges between 0.7 and 0.95 [41, 42]. Item-item and item-total correlations were assessed using Spearman’s correlation coefficient. Values between 0.15–0.85 were found acceptable for inter-item correlations. Item-total correlations were assessed to prevent redundancy. Values≥0.5 were considered acceptable for item-total correlations [10].
Test-retest reliability measures the consistency of a measurement instrument to produce the same result when repeating the test under the same conditions. Ideally a reliability test includes repeated measures obtained within a short period of time from an untreated group, patients who reported “no-change” in GRC (> –2 and < +2) during their second visit) were included [26, 30].
The Intraclass Correlation Coefficient (ICC) was calculated with the corresponding 95%confidence interval [31]. A two-way random effects model was used (ICC 2.1). Reliability was classified either as poor (< 0.50 ICC), moderate (0.50–0.75 ICC), acceptable (0.75–0.90 ICC), or excellent (> 0.90 ICC) [27].
Likewise, a Bland-Altman plot was conducted to compute 95%limits of agreement [3]. The difference of the VVAS scores at the first and second visit were plotted against the mean of the two VVAS-A scores. In the Bland-Altman plot the data from a reliable instrument will evenly scatter around the mean difference, with the variability independent of the score size [3, 19].
The VVAS-A construct validity was determined by correlating the scores from visit one on the VVAS-A and DHI-A. Pearson correlation and Spearman’s rank correlation coefficients (rho) were used. Correlation coefficients were classified as strong (> 0.50), moderate (0.35–0.50), or poor (< 0.35) [25, 48]. Validity was determined using Boer et al.'s criteria [14, 20], under which construct validity is poor, moderate, or high if < 25%, 25–50%or > 50%of the hypotheses are confirmed, respectively.
Continuous variables with normal distributions were presented as means and standard deviations (SD). Otherwise, data were reported using median and interquartile ranges (IQR). Categorical variables were reported as the number of cases or percentages. The sample distribution was examined with the Kolmogorov-Smirnov test. Continuous variables of patients who completed both visits and those who had only one clinic visit were compared using the independent samples t Test or the Mann-Whitney U test. Categorical variables were compared using chi-square test or Fisher’s exact test. A < 0.05 p value was considered statistically significant. Data analysis was performed using 24.0 Mac IBM SPSS statistic software (IBM Corp, Armonk, NY, USA).
Results
Eighty-two patients were enrolled in the study with 78 returning for their second visit (4.9%dropout rate). Seven of the 78 subjects who returned for visit 2 had a two or greater change on the GRC score with 5 patients getting better and two worse. Seventy-one of the patients were stable between test administrations. Fifty-three patients (64.6%) of the total sample were women with a mean age of the total sample of 55.2 years (SD 17.7; range 20–85) (Table 1). The most common class of diagnoses were peripheral vestibular disorders (58.5%). Patient characteristics are reported in Table 1 including gender, age, diagnosis category, length of symptoms, the time in seconds required to complete the VVAS-A, and the subjects’ educational level.
A total of 82 participants completed the VVAS-A during visit one. The median score was 15.89 (IQR 4.72 –40.33) points. The range was from 0 –80.4 with a floor effect seen in 12 (14.6%) patients who reported no symptoms on the VVAS-A.
Internal consistency
Cronbach’s alpha was 0.91 for the VVAS-A, which showed acceptable consistency. Item-total scale correlations ranged between 0.62 and 0.80. The minimum and maximum values corresponded to items 9 and 1, respectively (Table 2). Item-item correlations ranged between 0.38 (Item 3 versus Item 9) and 0.74 (Item 4 versus Item 5). Nine of 36 correlations were outside literature-recommended ranges (≥0.5) [5, 10].
Internal consistency of VVAS-A (n = 82)
Internal consistency of VVAS-A (n = 82)
Visit one and visit two VVAS-A median scores for the stable group were 20.78 points (IQR 5.33 –42.44) and 16.67 points (IQR 4 –40.22), respectively. The ICC of the stable group was 0.79 (95%CI: 0.70 –0.86), which demonstrates acceptable reliability. An ICC was also computed based on the entire sample who returned, including those who changed 2 or more points on the GRC revealing an ICC of 0.76 (CI 95%0.64 –0.86).
According to the Bland Altman plot, differences between the first and second visit suggest non-constant variance across the entire range of measurements and observed differences show rightward asymmetry (Fig. 2). Log transformation could not resolve the asymmetry. Therefore, we took a non-parametric approach based on literature recommendations [26]. Thus, the median of differences was reported with percentiles 2.5 and 97.5. The median of the visit 1 and visit 2 difference was 0 with percentiles of –43.7 and –35.5 points, respectively.

The Bland Altman Plot illustrates the difference of the VVAS-A scores at the first and second visit on the Y axis versus the mean of the two VVAS-A scores on the X axis. VVAS: Visual Vertigo Analogue Scale.
The VVAS-A and DHI-A correlation results are illustrated in Table 3. The DHI-A total and the VVAS-A were strongly correlated (rho: 0.57). Strong correlations were observed between the VVAS-A and DHI-A physical (rho: 0.58) and functional (rho: 0.54) scores. The correlation between VVAS-A and DHI-A emotional scores was moderate (rho: 0.39).
Construct Validity between VVAS-A and DHI-A
Construct Validity between VVAS-A and DHI-A
IQR: interquartile range. DHI-A: Dizziness Handicap Inventory Argentine version.
The VVAS-A is a reliable and valid measure for use in persons with vestibular disorders. The VVAS-A is a brief questionnaire that can be completed in less than 2 minutes, regardless of their level of education. This short administration time proves extremely useful in clinical practice and utilizing the test may help to focus care.
The correlation between the DHI and the VVAS in the Dannenbaum et al. study [14] was 0.67 versus 0.57 in the current study. Although similar, the relationship between the DHI-A and the VVAS-A was less strong.
A floor effect was observed in 14.6%of the people seen who had no symptoms on the VVAS-A, with one out of five people reporting that they did not experience VV symptoms after a vestibular disorder when they present to an out-patient physical therapy clinic. Our result is consistent with the symptoms reported by vestibular patients during the validation of the original questionnaire, as 23%of their vestibular patients rated either one or no items above zero on the VVAS [13].
Internal consistency was slightly below that of the original scale. Test-retest reliability was acceptable. Regarding construct validity, a strong correlation was observed between the VVAS-A scores and total DHI-A, DHI-A physical, and DHI-A functional scores, which is consistent with the original validation study [13]. There was not a strong correlation between the VVAS-A scores and DHI-A emotional scores, yet Dannenbaum et al. [13] reported a strong relationship (r = 0.65) between the emotional component of the DHI and the VVAS.
Visual vertigo, the construct tested in the VVAS-A, is often seen in persons who become visually dependent [11]. The VVAS-A took an average of 94 seconds to complete, suggesting that it not overly burdensome for patients to complete. The benefit of early identification of visual vertigo may allow the clinician to focus on the optimal type of exercises to prescribe for the patient. Persons who are visually dependent benefit from a rehabilitation program that focuses on utilizing proprioceptive and remaining vestibular function to maximize recovery [16, 46].
Persons with bilateral vestibular loss during an fMRI imaging study upregulated vision as the vestibular system was damaged [16], suggesting that the brain sensory inputs were able to substitute for the damaged vestibular sensor. It has been postulated that because of a visual/vestibular mismatch, patients may become more reliant on visual cues for orientation [11, 17]. Visual vertigo symptoms may be one of the factors that leads to the development of persistent postural perceptual dizziness (PPPD), which is a challenging condition to manage in persons after a vestibular disorder [18, 39]. Riccelli et al. [36] reported that persons with PPPD had different responses to a virtual reality roller coaster experience in their occipital cortex and insular regions of their brains compared to control subjects. The ability to record visual sensitivity with the VVAS-A can assist in identifying visual vertigo and provide information to provide targeted interventions to address the visual sensitivity in the exercise program early in the rehabilitation process.
The VVAS-A is a valid measurement instrument for clinical use. Acceptable consistency was observed, and the item-item and item-total correlation coefficients were within literature-recommended ranges, demonstrating the VVAS-A’s homogeneity and internal consistency [5, 10].
The patients were diagnosed by local neurologists or otolaryngologists; however, laboratory findings were not available to better describe the sample which is a limitation of the study. Study strengths include the use of standardized procedures, a large sample size, and inclusion of patients with various vestibular disorders. Recently, the original VVAS scale has been described as a valid measurement instrument that shows responsiveness to change, which can be used to both identify VV and assess symptom changes [12]. More studies are needed to document the psychometric properties of VVAS-A and determine its sensitivity to change. Future studies should attempt to determine if there is a relationship between anxiety and VV, since there appears to be a relationship between VV and avoidance [34].
Conclusion
The VVAS-A is a validated and reliable test that is easy to complete in patients with vestibular disorders. It can be used to assess the intensity of VV in different situations, which is useful in treatment planning and documenting visual vertigo symptoms.
Footnotes
Acknowledgments
The authors of this study would like to thank the physical therapists Valentina Urbina Jaimes and Agustina Monzón for their help in recruiting patients and administering questionnaires.
Disclosure statement
There is no conflict of interest to disclose.
