Dizziness is a common complaint (1, 2) with 16% reporting dizziness or balance problem in a recent Norwegian survey (3). It is a common feature in acute vestibular diseases where an abrupt start of symptoms, such as dizziness, nauseousness, reduced balance, and visual problems occurs (4). The symptoms usually subside within a couple of weeks; however, ~30% of patients develop persistent symptoms (5, 6). When dizziness persists, it is important to also assess symptoms beyond those defined as vestibular (7). These signs and symptoms include reduced gait velocity (8), postural misalignment (9, 10), increased body sway (11, 12), pain (13), and rigid body movements (14), in addition to anxiety and depression (15–19), avoidance behavior (20), and reduced quality of life (QoL) (21). Fatigue has also been reported in patients with vestibular diagnoses (22). These symptoms may influence everyday life, but to our knowledge, a combination of physical tests and self-reported measures on the perception of dizziness-related handicap has not been investigated extensively.

A commonly used outcome measure related to dizziness is the Dizziness Handicap Inventory (DHI) (23), which was developed to evaluate perceived dizziness-related handicap (24). There seems to be less relationship between diagnostic vestibular function tests and DHI (25, 26), suggesting that dizziness-related handicap may be more associated with the perception of signs and symptoms that have a direct impact on everyday life. Previous studies have found relationships between DHI and functional balance tests involving locomotion (27), slower walking velocity and reduced step and stride lengths (28), and an increasing number of self-reported dizziness triggers (e.g., loud sounds and stress) (29). Severe scores on the DHI have also been associated with poorer psychological outcomes such as anxiety, depression, avoidance behavior, fear of bodily sensations, illness perception, and cognitive responses (5, 30). However, there seems to be a lack of knowledge regarding the relationship between DHI and other outcomes, such as muscle strength, body flexibility, movement-induced dizziness, dizziness-severity, and health-related QoL.

Dizziness Handicap Inventory is measured on a continuous scale (0–100 points). Attempts have been made to establish severity levels of the DHI that links to functional abilities on several occasions (31), but less seems to have been published apart from one study categorizing the DHI into mild (0–30 points), moderate (31–60 points), and severe handicap (61–100 points) (32). Patients with a severe DHI level have shown larger functional impairment on balance confidence and number of falls, compared with patients presenting with mild DHI level scores (32), implying that patients in the severe DHI level may need more clinical attention.

Several studies have reported correlations between DHI and a variety of signs and symptoms. However, there seems to be less scientific evidence examining the relationship between DHI and aspects of function which includes both physical tests and self-reported outcomes. This is an important knowledge in the clinical assessment and treatment of these patients. Therefore, the objectives of this study were to describe both physical tests and self-reported outcomes in people with persistent dizziness in relation to DHI severity levels and to investigate the associations between these outcomes and increasing severity of the DHI total score.

A total of 107 participants were included, with a mean age of 49 years (20–70 years), most of the participants were women (76%), and 50% of the participants reported a dizziness duration of at least 2 years (Table 1). The mean DHI score was 49 points (Table 2). Based on the DHI scores, 18% of the participants were classified as having mild dizziness-related handicap, 56% had moderate dizziness-related handicap, and 26% had severe dizziness-related handicap. Participants in the mild category tended to be older (mean age = 54 years) with a higher percentage of women (84%). A similar proportion of participants in the mild and moderate categories experienced dizziness for more than 24 months (53%), but almost two times as many in the moderate category (41%) were currently on sick leave compared with those in the mild category (21%), slightly over a third of the participants in the severe category presented with dizziness for more than 24 months; and 61% of these reported being on sick leave.

Participants presented with poorer scores in all groups of measures with increasing severity levels of DHI (Table 2). In Group 1, there was a tendency of higher preferred and fast gait velocities for those in the mild and moderate DHI categories compared with the severe category, and the participants in the mild DHI severity level demonstrated a trend toward lower NRS_dizziness scores compared with the severe group. The Group 2 measures all presented with low scores in the mild DHI severity level, which increased with DHI severity levels. The measures in Group 3 showed a similar trend of worsening symptoms with increasing severity levels of the DHI (increasing scores in CFQ and VSS-SF, decreasing scores on the EQ-VAS).

A total of three separate backward stepping multivariable models, with additional analyses including age and sex, were fitted to determine which measures within each group were associated with DHI (Table 3). After adjusting for age and sex, we found a significantly negative association between fast gait velocity and the DHI, and a significantly positive association between head movement-induced dizziness and the DHI (Model 4). For the measures in Group 2, the adjusted model showed significantly positive associations between both the MI-A and HADS scores and the DHI (Model 5). In Group 3, there was a significantly positive association between the VSS-SF scores and the DHI, and a significantly negative association between the EQ-VAS and the DHI after including age and sex in the model (Model 6). Model 4 could explain 13% of the variance in the DHI scores, whereas Models 5 and 6 explained 43 and 41% of the variance.

The independent variables associated with changes in DHI scores are presented as scatterplots (Figure 1). The plots showed a small negative association between DHI and fast gait, whereas NRS_dizziness presented with a larger variety of scores, with a small positive association. The MI-A, HADS, and VSS-SF all had a clear positive association, whereas the EQ5D-Vas illustrated a negative association with the DHI.

The independent variables in the final model (Model 7) could explain 58% of the variance (Table 4). After adjusting for age and sex, the MI-A, HADS, VSS-SF, and EQ-VAS (Model 8) remained significantly associated, and together, they explained 56% of the variance in the DHI scores.

There was a clear trend toward worse scores for the different outcomes in each of the DHI levels. In Group 1, gait velocity and head movement-induced dizziness (NRS_dizziness) became poorer with increasing severity level of the DHI. This also applied to avoidance behavior (MI-A), fear of bodily sensations (BSQ), fear itself (ACQ), and psychological distress (HADS) in Group 2, and fatigue (CFQ), dizziness severity (VSS-SF), and health-related QoL (EQ-VAS) in Group 3. In the adjusted multiple regression analyses, significant associations were found between increasing DHI severity and two measures within each of the three groups; fast gait velocity and NRS_dizziness in Group 1, MI-A and HADS in Group 2, and VSS-SF and EQ-VAS in Group 3. In the final adjusted combined model, only the MI-A, HADS, VSS-SF, and EQ-VAS remained significantly associated with increasing DHI severity, explaining 56% of the variance in the DHI score.

The participants scoring in the severe DHI category walked slower than the individuals in both the mild and moderate categories, at both preferred and fast velocities. The current results are in conflict with a previous study that demonstrated no significant difference in gait velocity between severity levels of DHI (32). The participants in this study were generally younger than those in the study by Whitney et al. (32) which could be one explanation for the different findings. However, different analysis methods and description of pace hamper direct comparisons. In this study, the adjusted regression analyses found that fast, but not preferred, gait velocity was significantly associated with DHI. This is in contrast to a previous study (28) and possibly due to the differences in versions of the DHI and statistical methods used. Since both studies suggest gait to influence DHI scores, results are considered not to be conflicting. The adjusted model also found significant associations between DHI and NRS_dizziness, but the lack of previous studies using this measure hampers comparison. Overall, less than half of the physical tests were significantly associated with the DHI, and the model explained only 13% of the DHI variance. These results indicate that the selected physical tests were unable to capture the current population's perception of dizziness-related handicap as measured by the DHI.

There was a trend for worse scores with increasing severity level of the DHI in the psychological measures. However, since ACQ and BSQ scores were within normal range (equal to a community sample) (54) across all DHI severity levels, the symptoms of panic and anxiety were of limited importance in this study population. Also, the MI-A (avoidance) and HADS (psychological distress) scores increased with each DHI level, in line with the previous studies on populations with dizziness (30, 55, 56). Although only participants in the severe DHI level presented with abnormal scores for MI-A and HADS, [MI-A 2.30 points (cutoff value; 1.65 points (54), HADS 15.5 (cutoff value; 12 points (57)], the regression model including the psychological outcomes (Group 2) explained a relatively large part of the DHI variance (43%), with half of the selected measures being significantly associated with the DHI. This indicates that psychological outcomes may play an important role when dizziness-related handicap increase. This is in line with previous studies that have documented avoidance behavior in these patients (5, 30, 55), and that avoidance correlates with DHI scores (30). The same applies to HADS and association with DHI scores (56, 58) and was not surprising given the reported link between vestibular diseases and psychological complaints (18).

The outcomes in Group 3 (fatigue, dizziness severity, and QoL) also presented with worse scores with increasing DHI severity levels. It was surprising that the participants in the severe DHI level had CFQ scores similar to patients with chronic fatigue syndrome (mean; 24 points) (59), and that all three current levels of the DHI presented with severe dizziness according to VSS-SF (from 12.53 to 27.0, mean 20.58; cutoff 12 points) (60). The adjusted regression model showed significant associations between DHI and VSS-SF, in addition to EQ-VAS, in line with previous studies (29, 30, 38), whereas CFQ was not associated with DHI scores. This was somewhat surprising and may be explained by the fact that none of the items in DHI directly address questions related to fatigue.

The final adjusted model including significant measures from the previous models indicated that the self-reported outcomes covering avoidance (MI-A), anxiety, and depression (HADS) in addition to dizziness severity (VSS-SF) and QoL (EQ-VAS) and had a strong relationship with the DHI, whereas the physical tests did not. The model explained 56% of the variance in the DHI scores which was considered to be rather high, and in line with another study (30) who found that a model of different psychological outcomes could explain 62.7% of the variance in the DHI scores. Another study evaluating the relationship between demographic characteristics, mental health and dizziness-related characteristics also found significant associations between DHI and several outcomes which explained 63% of the variance (29). As far as we are aware of, this study is the only study to use multivariate regression analyses including both physical tests and self-reported outcomes in a stepwise backward model that at each step gradually eliminates variables from the regression model to find a reduced model that best explains the data. This represents a rather novel approach. As none of the physical tests were significant in the final model, it could indicate that the DHI is unable to capture problems associated with everyday physical activities using gait as an example. Diverging results are reported by others concerning such associations (27, 28, 32, 61). Studies using different static balance and functional tests have found moderate to strong correlations with the DHI, like for instance, single-leg stance and Dynamic Gait Index (27, 32, 61), whereas other outcomes such as Romberg test and single-leg stand test had weak or no correlation (27, 32). Whether other physical measures would have rendered the significant associations to remain in this study is, however, unclear. From our point of view, the DHI seems to be useful in collecting self-reported information concerning psychological problems, dizziness severity, and QoL. However, as people with dizziness often present with a variety of complaints, it is necessary to use a combination of tests and measures in the assessment of individual patients with persistent dizziness.

The study has several limitations. As data were collected in relation to an ongoing RCT, power calculations were not performed for this study. The study has a cross-sectional design, and data from this study do not allow for inferring conclusions regarding direction and causality. Even though there are risks of confounders that may influence the association in cross-sectional studies, this was not likely, since this study tested for assumptions for regression analyses, and no sign of multicollinearity was found.

A weakness of this study could conceivably be that diagnoses are not included as a confounder in the association analysis. However, we recruited persons with prolonged dizziness (average 24 months) without paying attention to diagnoses as they have shown to be of less importance for function when dizziness persists (13). We rather aimed to explore what functional challenges the participants had and if these problems influenced scores on the DHI. Another possible limitation is the use of three severity levels of the DHI (32), which is neither based on evidence nor consensus. However, we opted to use the categorization due to the clinical impression that scoring in the severe DHI level could be an indication for directing special attention toward these patients. Jacobson and Newman (24) also included subscales (physical, functional, and emotional) of the DHI. However, the validity of these subscales has been questioned (62) and the use of the total DHI score has been recommended (31, 35, 36, 38, 63), and thus, only the relationship between the total DHI and selected outcomes was examined in this study.

Strengths of the study include the inclusion of a relatively large population with persistent dizziness. Data were collected systematically following a published protocol (33) and trained testers were used. It is further a strength that the current age and sex distribution equal previous studies on populations with vestibular populations (13, 29, 30), and patients attending an oto-neurology clinic due to dizziness (30, 32), thereby increasing generalizability.

The use of a regression analysis including groups of physical tests and self-reported measures, to reveal the most important factors to explain the variance in DHI scores, represents a new approach in this field. The findings indicate that DHI seems to cover self-reported aspects such as avoidance, behavior, psychological distress, dizziness severity, and QoL. However, the relationship between DHI and physical tests was not established. From this study, it seems that DHI should be complemented with physical tests to establish a more complete picture of the patients‘ complaints. However, further studies are needed to establish which physical test will be the most appropriate to use together with DHI.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by the Regional Committee for Medical Research Ethics. The patients/participants provided their written informed consent to participate in this study.

The primary investigator for the study was LK. She developed the study in collaboration with LM, SM, KW, SHN, RC, and BJ-K. LK was responsible for the screening and testing of participants, with assistance from LM. LK completed the statistical analysis in collaboration with EB. LK drafted the manuscript with contributions from all authors with critical revision. All authors have read and approved the final manuscript.

This study was funded by the Norwegian Fund for Postgraduate Training in Physiotherapy. The funding body had no active part in planning and conducting the study.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.