Validation of the Czech adaptation of Mini-BESTest into clinical practice

Abstract

BACKGROUND

One of the frequently used standardized tools for balance assessment is the Mini-BESTest. Creating validated versions in national languages is very important for its correct administration and usability in people with acquired brain injury.

AIM

Our intention was to verify the validity of the Czech translation Mini-BESTest_CZ and describe experiences with its implementation in clinical practice.

DESIGN

Validation study.

SETTING

Department of Rehabilitation Medicine, University Hospital.

POPULATION

Patients after stroke or traumatic brain injury.

METHODS

The Mini-BESTest was translated and cross-culturally adapted into Czech according to established guidelines. The validation of this test was performed on a group of individuals with acquired brain damage in comparison with the Berg Balance Scale (BBS) and the Timed Up and Go test (TUG). Subsequently, physiotherapists who used the Mini-BESTest_CZ in their practice were also contacted to evaluate their feedback and experiences with implementing the test into clinical practice.

RESULTS

A total of 63 people with acquired brain damage participated in the study. Internal consistency of the Czech version of the Mini-BESTest was excellent, with intra-class correlation coefficient ICC=0.99. There were very high correlations between the Mini-BESTest and the BBS (r=0.86) or the TUG (r=-0.74). These results indicate excellent convergent validity for the Mini-BESTest_CZ. Physiotherapists considered items related to reactive postural control to be the most difficult to evaluate, and together with Timed Up and Go with cognitive dual task to be the most difficult for patients to understand.

CONCLUSIONS

The Czech version of the Mini-BESTest showed great inter-rater reliability and strong convergent validity with neither floor nor ceiling effects (as compared to the Berg Balance Scale), and can be recommended for the use in research and clinical practice for patients with acquired brain damage in subacute or chronic phase.

CLINICAL REHABILITATION IMPACT

The creation of a validated language version of the Mini-BESTest will contribute to better administration and evaluation of this test, not only for people with neurological disorders. We also describe therapists’ practical experiences with administering this test.

Balance is the ability to maintain the body in a stable position, both at rest (static balance) and in motion (dynamic balance). It is ensured by complex interaction among the vestibular apparatus, proprioceptive inputs, vision, and motor control. It is crucial for coordination, effective movement in space and fall prevention. Balance disorders are associated with a wide range of diseases and lead to disability in patients, thereby reducing their quality of life. Improving balance is often one of the main goals of rehabilitation, which is why there are a number of tests that help us objectively assess it. However, it is extremely important to choose the most relevant test for a specific group of patients.¹

Since we primarily work with patients in the subacute and chronic stages of brain damage, we sought a test that would be most suitable for them.

The Mini-BESTest (The Mini-Balance Evaluation Systems Test) is a comprehensive test that assesses both the proactive and reactive components of stability and was developed as a modification of the well-established clinical BESTest (Balance Evaluation Systems Test), primarily to reduce examination time.² It was first published in 2010³ and its main author is the American physiotherapist Fay B. Horak, who works in the Department of Neurology and Biomechanical Engineering at Oregon Health and Science University. The target group of patients for whom it is intended include people after traumatic brain injury (TBI),⁴ stroke,⁵ people with Parkinson’s disease,⁶ vestibular disorders,⁷ or people with multiple sclerosis (MS).⁸

The Mini-BESTest is a 14-item balance test assessing balance performance in four subcategories: anticipatory control, reactive postural control, sensory orientation, and dynamic gait conditions. A higher score (maximum = 28) indicates greater ability to maintain balance.² The total score in the Mini-BESTest corresponds to an objective balance measurement on a force pressure platform, where postural sways of the center of pressure are evaluated.⁹ The total test score also correlates very well with patients’ subjectively perceived balance performance (assessed by the Activity-specific Balance Confidence questionnaire).⁶ Due to complexity of the test tasks it has a high ability to predict falls in people with neurological impairment or elderly.^{6, 10} The test can also be used to identify recurrent fallers.¹¹ The Mini-BESTest can also predict walking ability on various surfaces.⁷ Compared to the Berg Balance Scale which is the most widely used test, the Mini-BESTest seems to be more suitable for people with a higher level of functional mobility, as it includes more challenging items.^{8, 11, 12}

Since its introduction, the Mini-BESTest has been translated to and validated in many languages, most recently German,¹³ Persian,¹⁴ Turkish¹⁵ and Portuguese.¹⁶ Creating validated versions in national languages is very important for a correct administration of the test (correct instructions and scoring).

The psychometric properties of the Mini-BESTest have been repeatedly validated in multiple groups of adults with neurological disorders (stroke, Parkinson’s disease or multiple sclerosis), as well as in the elderly.¹⁷ It is generally observed that the test scores decrease with age, when there are significant differences between age categories. The mean reported score in age group 50-59 is 26.3 points, in the age group 60-69 is 24.7 points, in the age group 70-79 is 21.0 points, and in the age group 80-89 is 19.6 points.¹⁸

Similarly, average scores for various diagnoses were also published. For instance, the mean score was 24.7 points in visually impaired persons (compared to young adults with normal vision with an average score of 27.3).¹⁹ And in people with incomplete cervical spinal cord injury, the mean of 16.7 points was shown, with the minimal detectable change of 3.8 points.²⁰ However, for most people with neurological diseases, a description of changes over time (whether a minimum detactable or clinically significant change is achieved) is sufficient. The so-called cut-off point is used to determine the score, at which the patient is at risk of falling. For the chronic stage of stroke, a cut-off point of 17.5 points was set with a minimum detectable change of 3 points,^{21, 22} although some studies report a minimum detectable change of 2 points.¹⁵ Similarly, a cut-off score of 19.5 points or less was set in persons with MS; below this value they need assistance when walking.⁸ In this case, a change of 4.5 points²³ is considered the minimum clinically significant change. In people with Parkinson’s disease the cut-off for using walking aid was 21 points²⁴ with a minimal clinically detectable change of 4 points.²⁵ In elderly people (80 years and more), a cut-off of 21 points was determined for a higher risk of falling.¹⁰ The most recent study, which was conducted in Canada in patients in the chronic stage of stroke, confirms the minimum detectable change of 3.2 points with a 95% confidence. The minimum clinically significant change after an outpatient rehabilitation program was set to be 4 points for detecting small changes and 5 points for detecting significant changes perceived by the examiner (physician or physiotherapist) and/or the patient.⁵ Similar results, i.e. a change of 4-5 points as the minimum clinically important difference for people with stroke was also described after an inpatient rehabilitation program.²⁶

The primary aim of this study was to assess the psychometric properties of the Czech version of the Mini-BESTest (Mini-BESTest_CZ). A secondary objective was to evaluate the experiences of therapists who began using this test in clinical practice.

Materials and methods

Participants

A cross-sectional study was carried out in the period 01/2020-10/2023. A total of 63 people with acquired brain damage (stroke or traumatic brain injury) in the subacute or chronic phase were recruited into this study. All participants provided written informed consent before entering the study. This trial was approved by the Ethical Committee of the First Faculty of Medicine and General University Hospital in Prague, Czech Republic. The registration number of the study is ISRCTN11744221. The participants were patients after stroke or traumatic brain injury who participated in a specialized 4-week rehabilitation program in the Department of Rehabilitation Medicine of the First Faculty of Medicine, Charles University in Prague. Participants underwent two balance assessments during the rehabilitation program. Balance tests were conducted at the start of the program and then again one week later.

The inclusion criteria were: 1) age over 18 years and diagnosis of stroke or craniotrauma; 2) able to stand independently; 3) ability to understand and follow a three-step command; 4) no visual impairment.

The exclusion criteria were: 1) participants who had any other neurological, vascular or musculoskeletal disorder that could affect balance; 2) severe cognitive impairment limiting understanding of commands (using cut-off of minimum 10 points in Mini-Mental State Examination); 3) vertigo or pathological fainting.

Procedure

The official Czech translation of the Mini-BESTest questionnaire was created according to the official guidelines with the consent of the test author and has already been published.²⁷ Minor discrepancies in the translation and instructions for therapists were consulted with the author of the test. However, the psychometric properties of the Czech version have not yet been tested.

The balance assessment in this study consisted of the Mini-BESTest_CZ, the Berg Balance Scale (BBS)²⁸ and the Timed Up and Go test (TUG).²⁹ Each balance examination of the patients in this study was performed by two therapists. One therapist performed the examination and the other videotaped it, so that it could be evaluated by a further independent therapist. This comparison was used to evaluate the inter-rater reliability of the Mini-BESTest. The other balance tests (BBS, TUG) were used to compute and compare the convergent validity.

Physiotherapists who started using the Mini-BESTest in their clinical practice were asked to complete a short questionnaire about their experiences. As a part of this simple questionnaire, they were asked about the items they perceived as challenging and those with the highest clinical information.

Statistical analysis

Descriptive statistics were used to summarize the demographic data of the participants. The reliability of the Mini-BESTest_CZ was analyzed for the total scores considered to be (pseudo-) continuous. The inter-rater reliability (IRR) of the total scores of the Mini-BESTest_CZ was assessed by comparing scores provided by the first rater (in-person) and the second rater (via video). The intra-class correlation (ICC) was used to assess IRR, as it is a common statistic for evaluating the reliability of ordinal, interval, and ratio variables. IRR was evaluated using a two-way mixed-effects model for consistency and average measures to determine the degree of agreement among raters in their scoring of empathy across subjects.³⁰

The convergent validity was assessed using Spearman’s rank correlation coefficients.³¹ This analysis examined the relationship between the Mini-BESTest_CZ and the two other tests: the Berg Balance Scale (BBS) and the Timed Up and Go test (TUG). The Mini-BESTest_CZ scores obtained by the assessor in direct contact with the patient were used for this correlation.

The floor/ceiling effect was evaluated by calculating the percentage of patients who achieved the maximum or minimum scores. According to a Norwegian study,³² a threshold of 15% was chosen for the present study.. This effect was calculated for the Mini-BESTest scores obtained face-to-face and for the BBS scores. The results were graphically represented using a histogram with an overlay of a density curve. All analyses were performed using the RStudio version 2024.04.0.

Data availability

Data are not publicly available but can be provided from the corresponding author on reasonable request.

Results

Participants

Sixty-three people with acquired brain injury met the criteria and were enrolled. Of these individuals, there were 35 (56%) males and 28 (44%) females, the mean age was 48±14 years. The most prevalent diagnosis was ischemic stroke (65%), followed by subarachnoid hemorrhage (SAH) at 14%, traumatic brain injury (TBI) at 11%, hemorrhagic stroke at 7.9%, and thrombosis at 1.6% (Table I). The mean score Mini-BESTest in our participants was 13-14 points (Table II), most of them were thus below the cut-off point at a risk of falling.¹⁷

Table I. —Patient characteristics (N.=63).

Characteristic	N. (%)
Sex
Male	35 (56%)
Female	28 (44%)
Age
Median (IQR)	48 (40, 56)
Mean±SD	48±14
Diagnosis
Ischemic stroke	41 (65%)
SAK	9 (14%)
TBI	7 (11.5%)
Hemorrhagic stroke	5 (8%)
Thrombosis	1 (1.5%)
Mobility aids
No need	28 (45%)
Walking stick	14 (22%)
4-Leg crutch	9 (14%)
Braces/orthosis	8 (13%)
Walker	2 (3%)
Crutch	2 (3%)
Median (IQR)	12 (8, 19)
Mean (SD)	17 (16)

IQR: interquartile range; SD: standard deviation.

Table II. —Mean±SD and inter-rater reliability of Mini-BESTest.

Assessment	Rater	Mini-BESTest Mean±SD	Inter-rater reliability ICC2,1 (95% CI)
First assessment	1	13.30±7.18	0.985
	2	13.30±6.88	(0.978-0.989)
Second assessment	1	14.67±7.52	P<0.001
	2	14.41±7.42

Psychometric values of the test

Internal consistency

The Cronbach alfa of the Mini-BESTest_CZ was 0.857 which is considered good reliability. Standard Error of Measurement was 0.655.

Construct validity

The convergent validity of the Mini-BESTest_CZ in patients was assessed using the BBS and the TUG. As shown in Figure 1, there were high correlations between the total scores from the Mini-BESTest_CZ and the BBS (r=0.86), as well as between the Mini-BESTest_CZ and the TUG (r=-0.74). These results indicate excellent convergent validity for the Mini-BESTest_CZ.

Figure 1.

—A, B) Convergent validity Mini-BESTest and BBS and TUG.

Floor and ceiling effects

None of the participants achieved the highest possible score in the Mini-BESTest, indicating that no ceiling effect was observed. The floor effect was observed in two cases (1.59%). In contrast, the results for the BBS showed the opposite situation: no floor effect was observed, but a ceiling effect was noted in 12 cases (9.52%). However, neither of these effects exceeded the threshold of 15%. Figure 2 represents the density distribution of the scores.

Figure 2.

—A, B) Floor and ceiling effect.

Inter-rater reliability

Table II shows that the inter-rater reliability of the Mini-BESTest_CZ was excellent, with an ICC=0.99 (95% CI: 0.98 to 0.99).

Experience of therapists

Therapeutic experience with the implementation of this test was obtained from physiotherapists who have started to use it in their clinical practice with neurological patients. The feedback was obtained from 15 physiotherapists who evaluated their experience of implementing the test into their practice. The demographic characteristics of the therapists are shown in Table III.

Table III. —Therapist characteristics.

Therapist	Age (years)	Experience in neurorehabilitation (years)	Experience with Mini-BESTest (years)	How many patients with acquired brain injury
1. Woman	34	7	4	40%
2. Woman	33	4	3	40%
3. Woman	40	15	9	100%
4. Woman	29	5	2	50%
5. Woman	29	5	5	50%
6. Man	34	5	2	5%
7. Woman	53	23	5	50%
8. Woman	45	20	7	50%
9. Woman	33	10	5	50%
10. Woman	29	3	3	50%
11. Woman	26	3	3	25%
12. Man	32	4	3	60%
13. Man	38	4	4	60%
14. Man	34	12	6	60%
15. Woman	26	2	1	50%
Median (IQR)	33 (26, 53)	5 (2, 23)	4 (1, 9)	50 (5, 100)
Mean±SD	34.3±7.03	8.1±6.3	4.1±2	49.3±19.4

Therapists described that it took them on average about 30 minutes to administer the test. They reported item no. 1 – sit to stand (60% therapists) and item no. 7 – stance on firm surface (46% therapists) as the easiest items to evaluate. On the other hand, items connected with reactive postural control were perceived in almost half of all therapists as the most difficult to evaluate. The same reactive balance items were also assessed as the most difficult for patients to understand. Similarly, item 14 (Timed Up and Go with cognitive dual task) was also quite complicated for some patients to understand. The compensatory stepping items no. 4, 5, 6 and the item no. 14 were reported by one third of all therapists as the ones that had to be repeated most often due to misunderstanding of instructions or incorrect performance. Therapists perceived the item no. 11 (walk with head turn) and the items no. 3 (one leg stand) and 13 (step over obstacles) as the most relevant balance indicators in the Mini-BESTest. The details of therapists´ opinions are displayed in Table IV.

Table IV. —Evaluation of the Mini-BESTestest by the therapist (N.=15).

Mini-BESTest items	Easy to evaluate	Difficult to evaluate for therapist	Difficult to understand for patients	Need to be repeated	Most indicative for balance
Sit to stand	9 (60%)
Rise to toes	1 (6.6%)	1 (6.6%)
One leg stand	1 (6.6%)				4 (26.6%)
Compensatory stepping-forward		7 (46.6%)	6 (40%)	7 (46.6%)	3 (20%)
Compensatory stepping-backward		8 (53.3%)	6 (40%)	6 (40%)	2 (13.3%)
Compensatory stepping-lateral		7 (46.6%)	5 (33.3%)	5 (33.3%)	2 (13.3%)
Stance, firm surface	7 (46.6%)
Stance, eyes closed, foam surface	2 (13.3%)				2 (13.3%)
Stance, eyes closed, inline	1 (6.6%)
Change in gait speed				1 (6.6%)	2 (13.3%)
Walk with head turn		2 (13.3%)	1 (6.6%)	1 (6.6%)	6 (40%)
Walk with pivot turn			1 (6.6%)		1 (6.6%)
Step over obstacles					4 (26.6%)
Timed Up and Go with dual task	1 (6.6%)	3 (20%)	6 (40%)	5 (33.3%)

Discussion

The Czech version of Mini BESTest showed a high inter-rater reliability (with ICC=0.99) and good internal consistency (with Cronbach alfa 0.857). Thus, our findings are very similar to previous studies that reported internal consistency Cronbach alpha from 0.73 to 0.97^{9, 13, 17, 33} and with previously reported inter-rater reliability ICC 0.93-0.97.³⁴ Our results also confirm the findings of previous studies that the Mini-BESTest does not have a floor or ceiling effect.^{13, 17, 35} When comparing our scores of the Mini-BESTest_CZ with the Berg Balance Scale and the TUG functional mobility test, a high convergent validity was demonstrated. None of our participants reached the ceiling effect. A ceiling effect ranging from 0.9% to 4.3% was reported in various study populations.²¹ The floor effect was observed in our study in 1.59% (two participants), both of them with a higher impairment. In a previous review, reaching the floor effect was similarly reported in a low number of people with a severe impairment.³⁶

In addition to the psychometric properties of the Mini-BESTest_CZ, our aim was to evaluate the therapists’ experience with its administration. At the Department of Rehabilitation Medicine of the First Faculty of Medicine of the Charles University in Prague, we started to use this test officially in January 2020. It is not necessary to undergo any special training to be able to use this test, however at the beginning it is necessary to carefully study the manual including educational videos on the test’s official website.³⁷ Our department specializes in long-term comprehensive rehabilitation for patients with acquired brain injury (usually after stroke or TBI); thus, we started to use this test in these patients. Initially, we routinely examined every patient with two therapists. First, to ensure patient safety, and second, for the purpose of video documentation of the test performance. Given that the presence of two therapists is time-consuming and costly, we used it primarily during the evaluation and validation of the Czech version of the test. We later abandoned this approach because we consider the administration of the test by a single therapist to be sufficiently safe. The testing time is usually reported in the range of 10-15 minutes in literature,^{13, 29, 32} but in our clinical practice, patients with more severe neurological or cognitive deficits can be tested up to 30-35 minutes. This may be due to the inclusion of people with acquired brain injury with a more severe impairment of mobility. The Mini-BESTest therefore seems to be more beneficial for people with milder neurological impairments who are able to walk for longer periods of time (including gait modification testing). Our findings are therefore consistent with previous studies that recommend the Mini-BESTest as more suitable for detecting possible changes after a rehabilitation program, among others due to the lack of ceiling effect.⁸

After some training period, an excellent consensus between experienced and novice raters in evaluating this test in people with acquired brain injury has been demonstrated.⁴ In contrast, a meta-analysis of more than 60 publications describes a lower inter-rater agreement in therapists with a longer testing experience. This may be caused by the fact that less experienced raters are more rigorous in applying the rules and the scales. Furthermore, they describe finding that the duration of the disorder also appears to be significant for the intra-rater reliability, indicating that an increase in the number of years the patient has suffered from the disorder implies a decrease in the intra-rater agreement. This may be due to compensations adopted by people with chronic neurological or musculoskeletal disorders that may influence an assessment of balance control.³⁴

In our study, the therapists perceived items no. 4, 5, 6 (reactive postural control) as the most difficult to assess and evaluate. Based on their experience, the reactive postural control and the TUG with cognitive dual task (item no. 14) appear to be also challenging for some patients to understand instructions and perform. Often they are also these items that need to be repeated during the examination due to patient misunderstanding of instructions. This confirms findings of a previous study, in which performance on the TUG with dual task was also described as significantly negatively affected by cognitive difficulties. By analogy, balance performance in items no. 4, 6, 10, 12 and 13 may be adversely affected by the use of compensatory aids.³³

Most therapists considered items assessing dynamic gait, specifically head-turning gait and walking over obstacles, to be the most significant indicators of balance. Dynamic gait tests, along with anticipatory balance items, are the most discriminating when comparing people with visual impairments and sighted individuals.¹⁹ According to an analysis comparing the Mini-BESTest with trunk excursion measurements, the item no.5 (compensatory step back) emerged as one of the best predictors of falls with a solid balance between sensitivity and specificity.³⁸

Creating language and culture specific versions of the Mini-BESTest is important for its proper implementation in clinical practice, not only for the accurate communication of instructions to the patient, but also for the correct evaluation of performance on the test items.¹⁶ In some previous studies, there were some errors in the assessment when the maximum final score was calculated to be 32 points (instead of maximal score 28 points).^39-41 This error occurred when item no. 3 (single limb stance) and item no. 4 (compensatory side step) for both limbs and both sides were incorrectly included in the final score. According to the author, Faye Horak, only the worse score (from the more affected limb or body site) should be counted.²

Thus, based on our experience with people with acquired brain injury, it seems important to focus on precise instructions for performing and evaluating reactive balance tests. Similarly, in the case of the cognitive dual task TUG item, it is important to instruct patients correctly so that the test does not have to be repeated for an incorrect performance. This may be too challenging for those with higher levels of impairment. On the contrary, for people with Parkinson’s disease it seems to be important to focus on the assessment of items no. 2, 3, 5, 7 and 12, which seem to be the most indicative of their balance.²⁴ This may vary from person with acquired brain injury described in our study.

Limitations of the study

The first limitation of this study is the relatively small sample size. In this study, the validity of the Czech version of the Mini-BESTest was only performed in people with acquired brain injury (caused by stroke or trauma). Given the composition of patients in the department, where the study was carried out, our sample included mostly individuals with chronic stroke, thus our findings is only partially valid for patients after acquired brain injury. It may also be considered a limitation that only the therapists’ opinions on individual test items were assessed, while patients’ views were not examined.

Despite the limitations, we believe that our findings may be useful for other rehabilitation facilities considering an implementation of the Mini-BESTest into their routine rehabilitation practice.

Conclusions

The Czech version of the Mini-BESTest showed a high inter-rater reliability and strong convergent validity with neither floor nor ceiling effects (as compared to the Berg Balance Scale), and can be recommended for the use in research and clinical practice with adult individuals with acquired brain damage.