Introduction

The vestibular system, located in the inner ear, is essential for regulating balance and perceiving spatial orientation. Dysfunction within this complex network can lead to vestibular disorders, which significantly impair balance and mobility [1]. Epidemiological studies report varying prevalence rates, including 6.5% in Germany [2], 2.98% in the U.S. [3], and a significant 2.83% in Taiwan [4], with the elderly population experiencing rates as high as 18.6% [5].

Vestibular disorders are broadly categorized into peripheral and central types. Peripheral vestibular disorders result from damage to the vestibular apparatus itself, while central vestibular disorders originate within the central nervous system and are often accompanied by additional neurological symptoms [1]. The clinical manifestations are diverse and may include vertigo, dizziness, nausea, hearing loss, tinnitus, aural fullness, oscillopsia, photophobia, phonophobia, and visual aura [6–8]. These symptoms disrupt postural control and heighten fall risk, with serious consequences for quality of life and safety [9–11].

Accurate fall risk assessment is essential in this population. Common clinical tools used to evaluate balance include the Berg Balance Scale, Dynamic Gait Index, and Timed Up and Go (TUG) Test. However, these assessments often lack sensitivity to critical components of postural control, such as sensory integration and reactive balance responses. For example, the TUG test evaluates functional mobility but does not capture adaptive or sensory strategies [12]. In contrast, the original Balance Evaluation Systems Test (BESTest) was developed to assess all systems contributing to balance control. Despite its comprehensiveness, the BESTest’s length and complexity limit its routine clinical use [13].

To address these limitations, the Mini-Balance Evaluation Systems Test (Mini-BESTest) was created as a shortened version of the BESTest. It includes 14 items that examine four domains of balance: anticipatory postural adjustments, reactive postural control, sensory orientation, and dynamic balance during gait. This format allows for efficient yet comprehensive assessment of both static and dynamic stability and advanced sensorimotor functions [14]. The Mini-BESTest has demonstrated good psychometric properties and has been used in older adults [15], individuals with Parkinson’s disease [16], and post-stroke populations [17].

While validated Mini-BESTest cut-off scores for fall risk have been established in other neurological and geriatric populations, no study to date has established a definitive diagnostic cut-off score optimized specifically for predicting fall risk within individuals with vestibular disorders. Previous vestibular studies have restricted their use of the Mini-BESTest to characterizing balance impairments or tracking longitudinal rehabilitation outcomes [18,19]. For instance, Zhu et al. (2023) mapped exploratory ROC curves comparing patients with bilateral vestibulopathy to healthy controls, reporting raw sensitivity and specificity coordinates across the entire score spectrum without designating or recommending a specific clinical cut-off score [18].

Accordingly, this study aimed to assess the diagnostic accuracy of the Mini-BESTest in distinguishing between low and intermediate-to-high fall risk among individuals with vestibular disorders. While the Mini-BESTest is widely recognized as a unidimensional scale for assessing dynamic balance [20], an exploratory analysis of its subsections, i.e., anticipatory postural adjustments, reactive postural control, sensory orientation, and dynamic gait, was also conducted to identify the specific postural control domains most predictive of fall risk in this population. Such insights could ultimately enhance clinical decision-making through more targeted rehabilitation assessments.

Materials and methods

Results

Fifty-one potential participants were screened for this study, of whom forty-two met the inclusion criteria and were enrolled. Nine participants were excluded for not meeting the eligibility requirements. One participant experienced dizziness during testing and was excluded from further assessment for safety reasons. No other adverse events occurred during administration of the Mini-BESTest or reference standard assessments. The flow of participants is illustrated in Fig 1. The characteristics of the participants and the results of all outcome measures are summarized below.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Flow of participants in the study.

https://doi.org/10.1371/journal.pone.0353230.g001

Characteristic data of the participants

Participant characteristics stratified by fall risk group are summarized in Table 1. No significant differences were found between the low and intermediate-to-high fall risk groups in terms of age, sex, BMI, vestibular diagnosis, symptom duration, MoCA score, physical activity level, or physical performance tests (4MWT and TUG).

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Participant characteristics stratified by low fall risk and intermediate-to-high fall risk groups.

https://doi.org/10.1371/journal.pone.0353230.t001

However, fall history differed significantly between the groups, with all individuals in the intermediate-to-high risk group reporting at least one fall in the past year (p < 0.001). Additionally, significant differences were observed in the Mini-BESTest total score (p = 0.049) and the reactive postural control subscore (p = 0.042), suggesting potential value in differentiating fall risk levels using these balance measures (Table 1).

Receiver Operating Characteristic (ROC) curve

The Mini-BESTest demonstrated a significant ability to discriminate between low and intermediate-to-high fall risk groups among individuals with vestibular disorders, with an AUC of 0.69 (95% CI: 0.530–0.824) (Fig 2 and Table 2). The optimal cut-off score was ≤ 23, yielding a high sensitivity of 92.31% (95% CI: 64.0–99.8) but a relatively low specificity of 44.83% (95% CI: 26.4–64.3). The positive predictive value (PPV) was 42.9%, and the negative predictive value (NPV) was 92.9%. The diagnostic odds ratio (DOR) for the total score was 9.82. The positive likelihood ratio (LR+) was 1.67 (95% CI: 1.16–2.41), and the negative likelihood ratio (LR–) was 0.17 (95% CI: 0.03–1.18), as shown in Table 2.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. AUC, cut-off score, sensitivity, specificity, LR + , and LR– of total score and subscores (anticipatory postural adjustments, reactive postural control, sensory orientation, and dynamic gait) for predicting fall risk in persons with vestibular disorders classified into low fall risk and intermediate-to-high fall risk groups.

https://doi.org/10.1371/journal.pone.0353230.t002

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Receiver operating characteristic (ROC) curve analysis evaluating the diagnostic accuracy of the Mini-BESTest total score in identifying fall risk.

The red shaded area represents the 95% confidence interval for the area under the curve (AUC = 0.69, p = 0.036). The diagonal reference line indicates the line of no discrimination (AUC = 0.50).

https://doi.org/10.1371/journal.pone.0353230.g002

Among the four Mini-BESTest subscores, anticipatory postural adjustments (p = 0.131), sensory orientation (p = 0.187), and dynamic gait (p = 0.103) did not demonstrate statistically significant discrimination between low and intermediate-to-high fall risk groups (Fig 3 and Table 2). In contrast, the reactive postural control subscore showed moderate diagnostic accuracy, with an AUC of 0.69 (95% CI: 0.528–0.823; p < 0.05), indicating its utility in fall risk classification for individuals with vestibular disorders.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Receiver operating characteristic (ROC) curve analysis for the four domains of the Mini-BESTest: (A) Anticipatory Postural Adjustments, (B) Reactive Postural Control, (C) Sensory Orientation, and (D) Dynamic Gait.

The Reactive Postural Control subscore (B) was the only domain to demonstrate statistically significant discriminatory power (AUC = 0.690, p = 0.018) with an optimal cut-off of ≤ 4. Blue shaded areas represent the 95% confidence intervals for each area under the curve.

https://doi.org/10.1371/journal.pone.0353230.g003

Discussion

This study aimed to evaluate the accuracy of the Mini-BESTest in identifying fall risk among individuals with vestibular disorders. The findings indicated a significant difference between groups with low and intermediate-to-high fall risk as determined by the Mini-BESTest. Of the subscores, only reactive postural control demonstrated significant discriminatory ability, whereas anticipatory postural adjustments, sensory orientation, and dynamic gait did not. Notably, while both the total score and the reactive postural control subscore exhibited similar accuracy, only the reactive postural control subscore remained statistically significant when assessed using the bootstrap method. These results suggest that while the comprehensive total score remains the standard for clinical risk stratification due to the scale’s fundamental unidimensionality, reactive postural control performance may be a key underlying physiological driver of balance failure in this population. This underscores its potential clinical utility for guiding targeted therapeutic interventions rather than functioning as a standalone diagnostic threshold.

The general characteristics of the low fall risk and intermediate-to-high fall risk groups were largely comparable, with the exception of fall history. As anticipated, a significantly greater proportion of individuals in the intermediate-to-high fall risk group reported a history of previous falls, which is consistent with existing evidence that fall history is a strong predictor of future risk [28]. Additionally, in this study, fall risk groups were classified based in part on fall history. As a result, participants with no history of falls were automatically assigned to the low fall risk group. Therefore, it is not surprising that there was a clear difference in fall history between the low and intermediate-to-high fall risk groups. Differences in characteristics such as MoCA scores and symptom duration were likely indicative of the heterogeneous nature of the vestibular disorder population. Additionally, differences in time since diagnosis may have affected balance performance, as the sample included both newly diagnosed individuals and those with chronic symptoms. This diversity is representative of recruitment from routine outpatient care, where patients are at varying stages of diagnosis and recovery.

Individuals with vestibular disorders exhibited reduced balance performance compared to healthy individuals [1,9–11,18]. This can be attributed to the vestibular system, which is one of the key systems involved in regulating postural stability [29–32]. The vestibular system is essential for sensing head movements and spatial orientation, as well as integrating sensory information necessary for maintaining balance [1]. When this system becomes dysfunctional, the brain receives distorted or inadequate input regarding body position and movement, thereby hindering the ability to execute appropriate postural adjustments [1,31]. Consequently, these individuals experience impaired postural control and face a higher risk of imbalance and falls [31,33].

The Mini-BESTest scores differed significantly between the low and intermediate-to-high fall risk groups. Specifically, the total Mini-BESTest score demonstrated poor discriminatory ability in distinguishing between these groups among individuals with vestibular disorders, with an area under the curve (AUC) of 0.69. Although this finding was statistically significant, the level of accuracy was lower than initially anticipated. The optimal cut-off score was identified as ≤ 23, resulting in high sensitivity (92.31%) but low specificity (44.83%). This high sensitivity suggests the tool may be useful for ruling out fall risk; however, the low specificity indicates a high rate of false positives in this population. To our knowledge, this was the first study to evaluate the accuracy of the Mini-BESTest for detecting fall risk specifically in individuals with vestibular disorders. Previous research in other populations has shown variable accuracy. For example, among individuals with chronic stroke, the Mini-BESTest demonstrated similarly low accuracy (AUC = 0.64; cut-off = 17.5; sensitivity = 64.0%; specificity = 64.2%) [17]. Additionally, moderate accuracy was observed in older adults (AUC = 0.71; cut-off = 18; sensitivity = 91%, specificity = 62%) [15], individuals with Parkinson’s disease (AUC = 0.75; cut-off = 19; sensitivity = 79%, specificity = 67%) [16], and nursing home residents (AUC = 0.71; cut-off = 9.0; sensitivity = 77.8%, specificity = 70.9%) [34]. These findings suggested that, although the Mini-BESTest could be useful for initial fall risk screening in individuals with vestibular disorders, its modest accuracy and limited specificity indicate that it should not be used as a standalone definitive assessment. Instead, it should be supplemented with additional clinical information to achieve more reliable risk stratification.

It is important to note that structural psychometric evaluations of the Mini-BESTest strongly support its unidimensional nature, proving that the cumulative total score reflects a single underlying construct of dynamic balance control [20]. While individual subsections target clinically distinct physical mechanisms, evaluating independent cut-offs for them can be psychometrically problematic due to this overarching unidimensionality. Nevertheless, our exploratory analysis suggests that specific domains may drive their underlying clinical utility in certain tracking environments. Specifically, the reactive postural control subscore of the Mini-BESTest demonstrated substantial utility as a descriptive measure of an individual’s capacity to promptly reestablish equilibrium following unexpected external disturbances through rapid, compensatory actions [32]. Optimal recovery of stability in response to unexpected perturbations depends on the seamless integration of visual, somatosensory, and vestibular sensory inputs within the reactive postural control system [32]. This function was frequently compromised in individuals with vestibular disorders as a result of disrupted sensory input from the vestibular system, which impaired the body’s ability to perceive and appropriately respond to alterations in spatial orientation [1,35]. Consequently, these individuals exhibited a reduced capacity to perform timely and effective balance corrections, which contributed to an elevated risk of falling [36,37]. For instance, fallers typically required a greater number of recovery steps following perturbations compared to non-fallers [36]. These findings underscored the clinical significance of the reactive postural control subscore in fall risk assessment for this population. In contrast, the other subscores, anticipatory postural adjustments, sensory orientation, and dynamic gait, did not effectively discriminate between fall risk categories in current study.

The Mini-BESTest total score demonstrated high sensitivity (92.31%), which is highly desirable for primary screening purposes, as it enables the identification of most individuals at risk of falling. Conversely, the low specificity (44.83%) indicates that the tool is less effective for confirming or diagnosing true cases of fall risk in isolation. Clinically, this suggests that while the Mini-BESTest can be valuable tool for initial fall risk screening in individuals with vestibular disorders, it should be complemented by additional assessments to improve diagnostic accuracy and guide targeted interventions.

Beyond sensitivity and specificity, the clinical utility of the Mini-BESTest is best illustrated by its Predictive Values and the DOR. In our cohort, the total score (≤ 23) yielded a high NPV of 92.9%. This indicates that the Mini-BESTest is an excellent ‘rule-out’ tool; a patient scoring above 23 has less than an 8% probability of being at high risk for falls. Conversely, the modest PPV of 42.9%—driven by the tool’s lower specificity—suggests that a positive screen (≤23) should be interpreted as a ‘red flag’ requiring further clinical validation rather than a definitive diagnosis of fall risk.

Furthermore, the DOR of 9.82 for the total score demonstrates a strong overall diagnostic power, as the odds of testing positive are nearly ten times higher in true ‘high-risk’ individuals than in ‘low-risk’ individuals.

The present study found that the LR+ for the Mini-BESTest total score was 1.67, while the LR- was 0.17. Generally, an LR+ greater than 1 indicates that a positive test result is more likely among individuals at high risk of falling than those at low risk; however, values less than 2 are typically regarded as having limited diagnostic value [27]. Likewise, LR− values below 1 suggest that a negative test result is associated with a lower probability of high fall risk, but values above 0.1–0.2 offer only minimal confidence in ruling out risk [27]. Accordingly, the Mini-BESTest total score exhibited only modest discriminatory power in differentiating between individuals at low and intermediate-to-high fall risk. These findings suggest that, while the Mini-BESTest can contribute to fall risk assessment, its overall clinical utility as a standalone screening tool in individuals with vestibular disorders remains limited.

This was the first study to examine the accuracy of the Mini-BESTest in identifying fall risk among individuals with vestibular disorders. Few previous investigations have assessed the diagnostic accuracy of balance assessments for predicting fall risk in this population. Prior research indicated that the TUG test, using a cut-off score of 11.1 seconds, achieved moderate accuracy with a sensitivity of 80.0%, specificity of 55.6%, and an area under the curve (AUC) of 0.71 [25]. Additionally, the Dynamic Gait Index demonstrated low accuracy, with a cut-off score of 18, a sensitivity of 64.5%, a specificity of 63.4%, and an AUC of 0.67 [25]. Furthermore, while Zhu et al. (2023) mapped exploratory ROC curves for the Mini-BESTest in a vestibular cohort, their work was restricted to diagnostic discrimination against healthy controls rather than defining a prognostic fall-risk cut-off [18]. Although the Mini-BESTest in this study demonstrated slightly lower overall accuracy (AUC = 0.69) compared to the TUG, it provided a broader scope of assessment. While tools like the TUG or 4MWT primarily capture basic functional mobility or gait speed, the Mini-BESTest systematically challenges distinct motor control systems, including anticipatory adjustments, sensory orientation, and dynamic gait adaptability. Notably, our exploratory bootstrap analysis revealed that the reactive postural control subscore was the sole domain to maintain statistical significance, underscoring its unique physiological relevance in vestibular instability. Despite a modest overall threshold accuracy, the comprehensive nature of the complete Mini-BESTest enables clinicians to identify specific system impairments to better tailor targeted, individualized interventions in clinical settings.

A major strength of this study was that it was the first to evaluate the accuracy of the Mini-BESTest in detecting fall risk among individuals with vestibular disorders, thereby providing a clinically relevant cut-off score for balance assessment in this population. Furthermore, this study investigated the individual subscores of the Mini-BESTest providing comprehensive clinical profile data regarding distinct balance domains. Rather than suggesting that subscores be used as independent diagnostic metrics in isolation—which would conflict with the scale’s validated unidimensional structure—identifying that reactive postural control performance acts as a primary physiological driver of variance allows clinicians to better understand the nature of vestibular instability. This dual focus on a total screening threshold alongside exploratory subscore characterization provides a valuable framework for both initial risk screening and subsequent targeted rehabilitation planning.

Despite these promising findings, several limitations should be noted. The overrepresentation of female participants may have limited the generalizability of the results to male populations; however, this gender imbalance was consistent with established epidemiological patterns observed in vestibular disorders [38,39]. Furthermore, this study did not incorporate symptom-specific measures. Previous research demonstrated that individuals who reported higher levels of dizziness-related distress tended to walk more slowly and had lower confidence in their balance [40], both of which could have affected their performance on the Mini-BESTest and influenced its predictive accuracy. Methodologically, our two-step reference standard included the TUG and 4MWT, which share movement components with the Mini-BESTest. This potential incorporation bias may lead to an overestimation of the reported AUC. However, it is important to distinguish the underlying constructs: while the TUG and 4MWT measure functional mobility and walking speed, the Mini-BESTest challenges higher-level vestibular integration through tasks like head turns and pivot turns. These items specifically evaluate the adaptability of gait under sensory challenge, whereas the reference standard assesses basic performance. Therefore, while the mode of activity (walking) overlaps, the sensorimotor demands differ significantly between the index test and the reference standard. Additionally, because this was a cross-sectional study, the ‘fall risk’ classification serves as a surrogate for actual falling events. While this model successfully captures functional decline, reliance on self-reported fall history may have introduced recall bias. Finally, while the sample size (n = 42) is relatively small, we employed Bootstrap resampling (1,000 iterations) as a robust method for internal validation. This approach ensures the stability and reliability of our diagnostic estimates (AUC, sensitivity, and specificity) and serves as a statistically rigorous alternative to split-sample cross-validation, which is often unfeasible in smaller clinical cohorts. This approach mitigates the risks associated with small sample distributions. Nevertheless, the resulting AUC of 0.69 suggests that the Mini-BESTest possesses poor-to-modest discriminatory ability in this population, and future large-scale studies are encouraged to further validate these cut-off scores.

Future studies should focus on validating these findings in more diverse populations and examining whether integrating symptom-specific and patient-reported instruments, such as the Dizziness Handicap Inventory, enhances the accuracy of fall risk assessment. To address the limitations of the current reference standard, implementing prospective tracking of actual falls over a 6-to-12-month period is essential to establish true predictive validity. Lastly, future studies comparing the Mini-BESTest with other balance assessments, while incorporating both objective and subjective measures, may provide a more comprehensive framework for evaluating fall risk in individuals with vestibular disorders.

For clinical implications, these findings indicate that clinicians should exercise caution when relying exclusively on the Mini-BESTest to assess fall risk in this population, due to its limited accuracy. Furthermore, given limitations such as the gender imbalance, absence of symptom-specific measures, and use of fall history data obtained from participant self-report, it is advisable to adopt a comprehensive and multifaceted strategy for evaluating fall risk. Such an approach should encompass objective balance assessments, patient-reported outcome measures, and consideration of additional risk factors, including frailty and psychological well-being. Considering the demonstrated importance of reactive postural control in fall risk, future investigations should examine whether interventions specifically targeting this domain can effectively decrease the incidence of falls among individuals with vestibular disorders.

Conclusions

This study found that while the Mini-BESTest could significantly distinguish between low and intermediate-to-high fall risk groups among individuals with vestibular disorders, its overall diagnostic accuracy was poor-to-modest. Subscores for anticipatory postural adjustments, sensory orientation, and dynamic gait did not show significant discriminatory ability. In contrast, the reactive postural control subscore was the sole domain to maintain statistical significance under bootstrap analysis, highlighting its distinct physiological relevance to instability in this cohort. Therefore, while the comprehensive total score remains the necessary standard for primary clinical risk stratification due to the scale’s validated unidimensional design, clinicians should pay close attention to reactive postural control performance to guide targeted therapeutic interventions. Despite their limited specificity, the Mini-BESTest remains a valuable clinical tool for initial screening due to its high sensitivity in detecting individuals at risk of falling.

Supporting information