Abstract
Background The Balance Evaluation Systems Test (BESTest) is a relatively new balance assessment tool. Recently, the Mini-BESTest and the Brief-BESTest, which are shortened versions of the BESTest, were developed.
Objective The purpose of this study was to estimate interrater and intrarater-interoccasion reliability, internal consistency, concurrent and convergent validity, and floor and ceiling effects of the 3 BESTests and other related measures, namely, the Berg Balance Scale (BBS), Functional Gait Assessment (FGA), and Activities-specific Balance Confidence (ABC) Scale, among patients with total knee arthroplasty (TKA).
Design This was an observational measurement study.
Methods To establish interrater reliability, the 3 BESTests were administered by 3 independent raters to 25 participants with TKA. Intrarater-interoccasion reliability was evaluated in 46 participants with TKA (including the 25 individuals who participated in the interrater reliability experiments) by repeating the 3 BESTests, BBS, and FGA within 1 week by the same rater. Internal consistency of each test also was assessed with Cronbach alpha. Validity was assessed in another 46 patients with TKA by correlating the 3 BESTests with BBS, FGA, and ABC. The floor and ceiling effects also were examined.
Results The 3 BESTests demonstrated excellent interrater reliability (intraclass correlation coefficient [ICC] [2,1]=.96–.99), intrarater-interoccasion reliability (ICC [2,1]=.92–.96), and internal consistency (Cronbach alpha=.96–.98). These values were comparable to those for the BBS and FGA. The 3 BESTests also showed moderate-to-strong correlations with the BBS, FGA, and ABC (r=.35–.81), thus demonstrating good concurrent and convergent validity. No significant floor and ceiling effects were observed, except for the BBS.
Limitations The results are generalizable only to patients with TKA due to end-stage knee osteoarthritis.
Conclusions The 3 BESTests have good reliability and validity for evaluating balance in people with TKA. The Brief-BESTest is the least time-consuming and may be more useful clinically.
Total knee arthroplasty (TKA) has become a common surgical intervention in the treatment of severe osteoarthritis (OA) of the knee joint. Both scientific research and clinical observations support the use of TKA for correction of deformity, mitigation of pain, amelioration of physical function, and symptoms of OA.1–3 However, substantial functional deficits may persist long after surgery for many patients with TKA.4–8 One important area of concern is balance impairments, which could increase fall risk in these patients.9 Indeed, the fall rate has been reported to be as high as 7% to 40% postoperatively.10–13 Understanding balance problems in patients after TKA is important.
Previous research in balance assessment among patients with TKA mainly involved advanced technology and sophisticated equipment in laboratory settings such as virtual or real obstacle avoidance,14,15 stabilogram analysis,16,17 kinematic and electromyographic analysis,9,18 and computerized dynamic posturography,19 which might not be available and feasible in real clinical situations. Although the Berg Balance Scale (BBS)20,21 is a common tool for balance assessment and can be considered as a reference standard for assessing balance in patients with TKA clinically,22,23 it is not without limitations. First, it mainly assesses static balance and has been shown to have considerable ceiling effects in various patient populations.24–26 Balance in important dynamic tasks, such as walking, is not addressed in the BBS. Second, maintaining body equilibrium involves many different balance control systems. The BBS has limited ability to identify what balance systems are impaired and thus direct treatment.27
Recently, the 36-item Balance Evaluation Systems Test (BESTest)28 was developed. It assesses the functioning of 6 balance control systems (ie, “Biomechanical Constraints,” “Stability Limits/Verticality,” “Anticipatory Postural Adjustments,” “Postural Response,” “Sensory Orientation,” and “Stability in Gait”). Good interrater reliability and intrarater-interoccasion reliability (intraclass correlation coefficient [ICC]≥.88) have been reported in a cohort of individuals with different diagnoses (eg, Parkinson disease [PD], vestibular disorders, total hip arthroplasty)28 and people with PD.29,30 However, the BESTest takes about 45 minutes to administer and might not be a practical option in daily busy clinical situations where time limitation is a serious concern. Hence, a condensed version of the BESTest, named the Mini-BESTest,31 was derived. With only 16 items, the Mini-BESTest takes only 15 minutes to complete and has demonstrated good interrater reliability and intrarater-interoccasion reliability (ICC≥.91) in different patient populations.29,32,33 However, 2 of the balance systems (“Biomechanical Constraints” and “Stability Limits/Verticality”) are omitted in the Mini-BESTest, thus contradicting the theoretical framework of the original BESTest and biasing toward dynamic balance.
In response to the drawbacks of the BESTest and Mini-BESTest, the 8-item Brief-BESTest34 was developed more recently. Unlike the Mini-BESTest, the Brief-BESTest includes items that assess all 6 balance systems. It requires less than 10 minutes to administer and could be more feasible for clinical use. Nevertheless, the psychometric properties of the BESTest, Mini-BESTest, and Brief-BESTest have not been tested in individuals with TKA specifically. Using a sample of patients with TKA, the objective of this study was to estimate the psychometric properties in terms of interrater and intrarater-interoccasion reliability, internal consistency, concurrent and convergent validity, and floor and ceiling effects of the 3 versions of the BESTest and other established balance and related measures, namely, the BBS, Functional Gait Assessment (FGA), and Activities-specific Balance Confidence (ABC) Scale.
Method
Study Design
An observational measurement study was undertaken to examine the reliability and validity of the 3 versions of the BESTest, BBS, and FGA in people with TKA. This study consisted of 2 phases. In the first phase, we sought to establish the reliability of the 3 versions of the BESTest, the BBS, and the FGA because reliability is a prerequisite to validity.35 To establish intrarater-interoccasion reliability, individuals with TKA participated in 2 assessment sessions (within a 1-week interval), during which the 3 BESTests, the BBS, and the FGA were administered by the same physical therapists. Some of these individuals were evaluated independently by 3 physical therapists in the first session to establish interrater reliability. In the second phase, we examined the concurrent validity, convergent validity, and floor and ceiling effects of the different balance tests at different stages of recovery after TKA. Another group of people with TKA were evaluated with the same balance tests 3 times (ie, at 2, 12, and 24 weeks after the operation).
Participants
Participants admitted for TKA in the Joint Replacement Centre of the Buddhist Hospital in Hong Kong from September 2012 to May 2013 and referred for rehabilitation were recruited. The inclusion criteria were: aged 50 to 85 years, having had their first TKA due to a diagnosis of knee OA, and able to follow verbal instructions and provide informed consent. Exclusion criteria were: TKA due to rheumatoid arthritis (RA) of the knee or traumatic injury, previous history of surgery in the lower limbs, and known medical diagnoses that affect balance (eg, stroke). For the first phase of the study (the reliability experiments), participants were required to have undergone TKA at least 6 months earlier. This requirement was important because we had to be able to assume stability in the response variable (balance performance) for reliability experiments. A previous study36 showed that little improvement occurred beyond 26 weeks after TKA. Prior to enrolling in the study, all participants signed a written informed consent statement that had been approved by the Human Research Ethics Subcommittee of the Hong Kong Polytechnic University and the Institutional Review Board of the Kowloon Central Cluster, Hospital Authority. All procedures were carried out according to the Declaration of Helsinki.
Sample Size Estimation
All a priori sample size calculations were performed using PASS 2011 software (NCSS Statistical Software, Kaysville, Utah). The sample size for interrater reliability analysis was based on the following assumptions: (1) 3 raters, (2) a null ICC of .75,35 (3) an expected ICC of .90,28,29 (4) a type I error of 0.025 (1-tailed), and (5) a power of 0.80.37,38 Hence, a sample of 23 patients with TKA was required for establishing interrater reliability among 3 raters.
The sample size for intrarater-interoccasion reliability analysis was based on the following assumptions: (1) 2 occasions, (2) a null ICC of .75,35 (3) an expected ICC of .90,29 (4) a type I error of 0.025 (1-tailed), and (5) a power of 0.80.37,38 A minimum of 33 patients would be needed for establishing intrarater-interoccasion reliability in 2 assessment sessions.
For establishing validity, Horak et al28 found a moderate correlation between the BESTest and the ABC in a mixed population (r=.64). A high correlation (r=.79) between the Mini-BESTest and the BBS in patients with PD was demonstrated by King et al.39 Thus, a medium-to-large effect size (r=.4) was assumed for this study, requiring a minimum sample size of 44 participants.
Qualifications of Raters
All 3 raters involved in this study are qualified physical therapists with at least 10 years of experience in working with people with TKA. They had prior experience using the BBS and ABC, but not the 3 BESTests and FGA. To ensure competency in using the BESTests and FGA, all raters were required to read the instruction manual for these tests, and they viewed the training video for the BESTest. This training was followed by a 2-week practice period during which the raters practiced administering the different tests used in this study among themselves. In addition, the raters were required to administer all tests on at least 2 patients with TKA prior to the commencement of the actual data collection period.
Measurement Tools
BESTest.
Each of the 36 items was scored on a 4-level ordinal scale from 0 to 3 (0=severely impaired balance or inability to complete a task; 3=no impairment of balance or able to perform a task successfully). The BESTest provides 6 subsection scores and a total score. The 6 subsections are: Section I–Biomechanical Constraints (5 items; score range=0–15), Section II–Stability Limits/Verticality (7 items; score range=0–21), Section III–Anticipatory Postural Adjustment (6 items; score range=0–18), Section IV–Postural Responses (6 items; score range=0–18), Section V–Sensory Orientation (5 items; score range=0–15), and Section VI–Stability in Gait (7 items; score range=0–21). The total score (range=0–108) was converted to a percentage score for subsequent analysis.28
Mini-BESTest.
The Mini-BESTest consists of 16 items from the original BESTest. Each item was scored on a 3-level ordinal scale from 0 to 2 (0=severe impairment of balance; 2=no impairment in balance), yielding a maximum score of 32.29
Brief-BESTest.
The Brief-BESTest comprises 8 items. The scoring method for each item was the same as for the full BESTest described above. The maximum possible score is 24.34
BBS.
The BBS comprises a set of 14 balance tasks. Each item was scored on a 5-level ordinal scale from 0 to 4, yielding a maximum total score of 56. Higher scores indicate better balance.20
FGA.
The FGA is a 10-item assessment used to evaluate postural stability during various walking tasks.40–42 Each item was scored on a 4-level ordinal scale from 0 to 3 (maximum total score=30). Higher total scores are indicative of better performance. The FGA has excellent interrater reliability (ICC=.93) in independently living individuals aged 40 to 89 years.43 It also has good interrater reliability (ICC≥.86) and test-retest reliability (ICC≥.74) in individuals with PD30 and vestibular disorders.40
ABC.
The ABC quantifies how confident a person feels that he or she will not lose balance while performing 16 activities of daily living on a scale from 0% (“absolutely no confidence”) to 100% (“completely confident”).44 The test was self-administered, and the score of the 16 items was averaged. The ABC had good test-retest reliability (ICC=.99) and interrater reliability (ICC=.85).44
Procedure
Demographic information was obtained from medical records and participant interview in the first session. The average pain intensity experienced on the operated knee over the past 24 hours was measured with an 11-point numeric pain rating scale ranging from 0 (“no pain”) to 10 (“worst imaginable pain”).45 Knee range of motion on both the operated side and nonoperated side was measured with a 1-degree-increment long-arm goniometer (Baseline 180° goniometer, NexGen Ergonomics Inc, Pointe-Claire, Quebec, Canada).
In the first assessment session, the balance performance of 25 participants with TKA was assessed independently by 3 raters to establish interrater reliability. As the items for the Brief-BESTest were taken from the full BESTest and the item scoring method was exactly the same, the Brief-BESTest score was computed from the BESTest. Previous research46 also used a similar method to score items on a shortened version of the BBS in patients with TKA and total hip arthroplasty. The common items for Mini-BESTest and BESTest were graded simultaneously with their respective scales (0–2 for Mini-BESTest; 0–3 for BESTest). In addition, for those items that were duplicated between the BESTest and other balance and related tests (BBS and FGA), the participants were asked to perform the items only once, and the performance was rated according to the specific scoring criteria for each test.
Each balance test was administered by any one of the 3 raters in random order, and all raters concurrently observed and rated each participant's performance. Sequence of test administration (BESTest, BBS, FGA, and ABC) and raters was randomized by a computer program (random blocks generation by Excel 2013, Microsoft Corp, Redmond, Washington). The average length of assessment session 1 was 1.5 hours. Short breaks were given as needed between tests to avoid overexertion. The assessment session took place in the afternoon to minimize the effect of morning stiffness. The raters were instructed not to discuss the scores among themselves.
A total of 46 individuals with TKA (including the 25 individuals who were involved in the interrater reliability experiments) participated in the intrarater-interoccasion reliability experiments. The procedures for the first assessment session were the same as described above. A second assessment session was conducted within 1 week after assessment session 1. No physical therapy treatment was provided during the period between assessment sessions 1 and 2. In session 2, the 46 participants were evaluated individually with the same balance and related tests by the same rater as in assessment session 1. To minimize the confounding effect of different times of testing, assessment session 2 also took place in the afternoon.
Another 46 patients with TKA participated in the validity experiments. Each participant was assessed at 3 time points: 2, 12, and 24 weeks after the operation. In each session, each participant was evaluated with the same 6 tests. The sequence of tests was randomized as described in the reliability experiments. These data also were used to examine the floor and ceiling effects.
Data Analysis
The IBM SPSS Statistics for Windows software program (version 19.0, IBM Corp, Armonk, New York) was used for all statistical analyses. The level of significance was set at P≤.05.
Reliability.
Interrater and intrarater-interoccasion reliability of the BESTest, Mini-BESTest, Brief-BESTest, BBS, and FGA total scores were assessed using ICC (2,1). Using the data from session 1 of the intrarater-interoccasion reliability tests, the internal consistency of the 5 balance tests was evaluated using Cronbach alpha. Cronbach alpha also was calculated for the subtests of the BESTest. This approach allowed us to examine individual items to determine how well they fit the subscales of the BESTest as well. A low Cronbach alpha for a particular subtest may indicate that some items in the subscale may represent a different component of balance function than the other items.35 The following criteria were used to judge the magnitude of the reliability coefficient: poor reliability=ICC<.4, fair reliability=ICC≥.4 but <.7, good reliability=ICC≥.7 but <.9, and excellent reliability=ICC≥.9.47 The minimal detectable change at 95% confidence interval (MDC95) for each balance test, which was an estimation of the smallest change in score that can be detected objectively for a participant more than measurement error, was calculated using the formulas35: MDC95=SEM × ✓2 × 1.96 and SEM=✓(MSE), where MSE is the mean square error generated from the analysis of variance model based on the intrarater-interoccasion reliability data, and SEM is the standard error of measurement.48
Validity.
The BESTest, Mini-BESTest, and Brief-BESTest scores were correlated with the BBS and FGA total scores (ie, concurrent validity) and ABC score (ie, convergent validity) using Pearson product moment correlation coefficient (r) or Spearman rho (ρ), depending on whether the assumptions for parametric statistics were fulfilled. The intercorrelations among the 3 BESTests also were examined using the same statistical methods. A correlation coefficient of .00 to .25 indicates little to no relationship, .25 to .50 means a fair correlation, .50 to .75 means a moderate correlation, and .75 to 1.00 means a high correlation.47
Floor and ceiling effects.
The skewness (γ1) of the score distribution at 2, 12, and 24 weeks post-TKA was examined. An absolute value of greater than 1.0 indicates that the distribution is highly skewed.49 Thus, a γ1 value greater than +1.0 denotes a substantial floor effect, whereas a γ1 value less than −1.0 indicates a substantial ceiling effect. The floor and ceiling effects were further examined by calculating the proportion of participants attaining the lowest and highest possible scores at the 3 time points. A proportion greater than 20% was considered to be significant.33
Results
Ninety-two individuals with TKA (n=46 for the reliability tests and n=46 for the validity tests) participated in the study. Characteristics of the participants are shown in Table 1. None of the participants required any mobility aids for indoor walking or during testing. One participant did not return for the validity experiments at the 24-week follow-up because he had moved to a different city.
Characteristics of Participantsa
Reliability
Twenty-five and 46 participants were involved in the interrater and intrarater-interoccasion reliability testing, respectively. The BESTest, Mini-BESTest, and Brief-BESTest demonstrated excellent interrater reliability (ICC [2,1]=.96–.99, P≤.001), intrarater-interoccasion reliability (ICC [2,1]=.92–.96, P≤.001), and internal consistency (Cronbach alpha=.96–.98) (Tab. 2). Good-to-excellent interrater and intrarater-interoccasion reliability and internal consistency also were established for the 6 subtests of the BESTest (Tab. 2). The MDC95 values of the BESTest, Mini-BESTest, and Brief-BESTest were 6.22, 3.71, and 3.19, respectively.
Interrater Reliability, Intrarater-Interoccasion Reliability, and Internal Consistencya
Validity
There were moderate-to-high associations of the 3 BESTests with the FGA and BBS at 2 weeks post-TKA (correlation=.72–.81, P≤.01), 12 weeks post-TKA (correlation=.58–.80, P≤.01), and 24 weeks post-TKA (correlation=.55–.73, P≤.01), demonstrating good concurrent validity (Tab. 3). Scores on the 3 BESTests also were significantly correlated with the ABC score at 2 weeks post-TKA (correlation=.34–.43, P≤.05), 12 weeks post-TKA (correlation=.40–.48, P≤.01), and 24 weeks post-TKA (correlation=.47–.50, P≤.01), showing good convergent validity. In addition, high intercorrelations were found among the 3 BESTests (correlation=.82–.93, P≤.01) at all 3 measurement time points.
Concurrent and Convergent Validitya
Score distribution and ceiling and floor effects.
None of the 6 measures had a skewness value greater than +1.0 or smaller than −1.0 at 2 weeks post-TKA (Tab. 4). At 12 and 24 weeks post-TKA, the distribution of BBS and FGA scores showed skewness values smaller than −1.0 (ie, ceiling effect). At 24 weeks post-TKA, the ABC also had a skewness value lower than −1.0. The score distribution of the 3 BESTests, BBS, FGA, and ABC at 24 weeks after TKA is shown in the Figure. When examining the proportion of people obtaining the maximum possible score, it was obvious that the BBS had the most severe ceiling effect, with 52.2% and 57.8% of the participants attaining the maximum score at 12 and 24 weeks after TKA, respectively. The 3 versions of the BESTest, in contrast, showed little ceiling effect, with only 2.2% to 8.9% of the participants reaching the top score at the same time point.
Floor and Ceiling Effectsa
Frequency distribution of scores on the (A) Balance Evaluation Systems Test (BESTest), (B) Mini-BESTest, (C) Brief-BESTest, (D) Berg Balance Scale (BBS), (E) Functional Gait Assessment (FGA), and (F) Activities-specific Balance Confidence scale (ABC) at 24 weeks postoperatively (45 participants with total knee arthroplasty).
Discussion
In the current study, the psychometric properties of different versions of the BESTest, BBS, and FGA were systematically examined for the first time in people with TKA. The study showed that the 3 BESTests have good reliability and validity to measure balance performance in individuals with TKA due to knee OA, without significant floor and ceiling effects at 2, 12, and 24 weeks post-TKA.
Reliability
The BESTest, Mini-BESTest, and Brief-BESTest had high internal consistency, indicating that the 3 BESTests measured similar underlying attributes. The interrater and intrarater-interoccasion reliability of the 3 BESTests were excellent when administered to individuals with TKA, which were comparable to reliability of the BBS and FGA. The MDC95 values of the BESTest, Mini-BESTest, and Brief-BESTest obtained in our study were 6.22, 3.71, 3.19, respectively, which represent the smallest differences that would reflect a genuine change in the total score of these tests. These values are comparable to those found in people with mixed neurological conditions (3.5)32 and in people with chronic stroke (3.0)33 using the Mini-BESTest. The MDC95 values found in the current study would be useful when interpreting the results of future clinical trials. A real change in balance ability following intervention should exceed the MDC95 value.
Validity
The high correlations between the BESTest, Mini-BESTest, and Brief-BESTest and the established balance (BBS, FGA) and related (ABC) measures indicate excellent concurrent and convergent validity. Our results are in line with previous findings in other patient populations. Strong associations of the BBS with the BESTest (r=.87)30 and Mini-BESTest (r=.79)39 have been reported among patients with PD. The BESTest also has been shown to have strong correlations with the ABC (r=.75) and FGA (r=.88) in patients with PD.30 In people with stroke, significant correlation existed between the Mini-BESTest and the BBS (ρ=.83) and ABC (ρ=.50).33 In addition, the BESTest showed a significant association with the ABC (r=.636) in a population with different balance disorders.28 The 3 BESTests showed strong intercorrelations, indicating that individuals with a less optimal score in one version of the BESTest also tended to have a less optimal score in the other 2 versions of the BESTests. This finding concurred with a previous study in patients with PD, where a strong correlation was identified between the BESTest and Mini-BESTest (r=.95)29 and between the Mini-BESTest and Brief BESTest (r=.94).50
Score Distribution and Ceiling and Floor Effects
Although none of the balance tests evaluated here had significant floor effects, the 3 versions of the BESTests, especially the full BESTest and Mini-BESTest, had the least ceiling effect, which was shown by the low degree of skewness and the small percentage of participants achieving the top scores at 2, 12, and 24 weeks post-TKA. A considerably higher ceiling effect of BBS was found at 12 and 24 weeks after TKA (52.2% and 57.8%, respectively) compared with the 3 versions of the BESTest (Tab. 4). In addition, the FGA demonstrated substantial skewness (γ1 >1.0) at 12 and 24 weeks post-TKA (Tab. 4). Other studies also showed more severe ceiling effects of the BBS and FGA compared with the BESTests. For example, King et al39 found that 1% and 13.4% of people with mild PD achieved a maximum score on the Mini-BESTest and BBS, respectively. Another study in PD showed that the proportion of people who received a perfect score on the BBS, FGA, and BESTest was 10%, 1.3%, and 0% respectively.30 In people with chronic stroke, the score distribution for the Mini-BESTest was significantly less skewed compared with the BBS. Only 0.9% achieved a maximum score on the Mini-BESTest compared with 32.1% for the BBS.33
What might explain the differences in ceiling effects between the 3 BESTests and the BBS and FGA? The BBS consists of tasks (eg, rising from a sitting to a standing position, sitting and standing without support, and turning to look over the shoulder while standing) that are relatively less challenging. On the other hand, the FGA is an ambulatory-oriented test focusing mainly on dynamic gait balance. The majority of our participants had experienced substantial recovery of their physical and functional mobility, especially at 12 and 24 weeks after the operation, leading to a ceiling effect of the BBS and FGA. The BESTests, in contrast, consist of more demanding tasks (eg, hip and trunk lateral strength, reach test, and postural responses to external perturbations). The inclusion of these more challenging tasks may have enhanced the ability of the BESTests to discriminate among participants compared with the BBS and FGA at different time points.
Upon examining the ceiling effects of the 3 BESTests, we found that none of the participants attained the maximum score at 2 weeks post-TKA. At later stages of recovery, the proportion of people who obtained a perfect score (ie, ceiling effect) was the lowest with the BESTest, followed by the Mini-BESTest and Brief-BESTest. Although the Brief-BESTest had the highest ceiling effect among the 3 BESTests, merely 8.9% of the participants achieved a perfect score at 24 weeks post-TKA. Overall, after considering the findings on reliability, validity, and ceiling effects, we conclude that the BESTest is the best balance assessment tool. However, the Mini-BESTest and Brief-BESTest are reasonable alternatives if time constraints are an important concern.
The BESTest is a relatively new balance assessment tool. According to the original authors of the BESTest, physical therapists who were naive to the BESTest should be able to learn how to administer it with prior review of the instructions.28 For the sake of safety, however, it was recommended by the original authors that the push-and-release technique to elicit automatic postural responses by suddenly releasing the individual's leans requires observation and practice with at least video demonstration.28 Our study confirmed that physical therapists who had no prior experience in using it can achieve excellent reliability after self-learning that involves reading the instruction manual, watching a demonstration video, and a brief practice period. The results of this study can be generalized to the physical therapists who have undergone similar training.
Limitations and Future Research Directions
The participants in this study had a first TKA (unilateral) due to knee OA. Therefore, the results can only be generalized to individuals with similar characteristics. Further investigation is warranted to confirm and expand the present results and generalizability to people with bilateral TKA or other conditions such as rheumatoid arthritis. Future research is warranted to evaluate the sensitivity and specificity (predictive validity) of the BESTest for predicting “fallers” among patients with TKA and the responsiveness of the BESTest in assessing change in balance ability in patients with TKA during recovery. The interrater reliability coefficients may have few implications in real clinical practice. Although 3 raters were involved in the interrater reliability experiments, only 1 rater actually administered the test. The other 2 raters simply observed the performance of the patients and provided their own ratings independently in the same visit. Such a scenario does not resemble what is typically encountered in daily clinical practice. The interrater reliability coefficients derived in this study do not include the patient variability that would exist if one clinician performs the measurement at a patient's initial visit and a second clinician performs the measurement at a reassessment, which is a situation that may be more frequently encountered in the real world. Further study also is needed to use the BESTest in directing the treatment regimen by identifying the body balance systems that are most impaired in people with TKA.
The BESTest, Mini-BESTest, and Brief-BESTest have good reliability and validity in evaluating balance in people with TKA. Although the 3 BESTests have comparable psychometric properties, the use of the Brief-BESTest is least time-consuming and could be particularly useful for clinicians and researchers in the field.
Footnotes
Both authors provided concept/idea/research design, writing, data analysis, and project management. Mr Chan provided data collection, participants, facilities/equipment, institutional liaisons, and administrative support. Professor Pang provided consultation (including review of manuscript before submission).
- Received October 29, 2014.
- Accepted April 2, 2015.
- © 2015 American Physical Therapy Association