- About Us
- Editorial Board
- For Authors
- Submit Manuscript
1University of Nottingham, UK
2The affiliated Suzhou Science & Technology Town Hospital of Nanjing Medical University, Suzhou, China.
Received Date: 20/11/2020; Published Date: 15/12/2020
*Corresponding author:Zhiwei Yang, the affiliated Suzhou Science & Technology Town Hospital of Nanjing Medical University, Suzhou, China.
Parkinson’s disease (PD) is associated with the degeneration of the central nervous system. PD can lead to motor dysfunction which mainly manifests as bradykinesia, abnormal gait and postural instability (McNeely et al. 2012) . Several sources of evidence report that motor dysfunction can directly result in the impairment of balance ability (Kerr et al. 2010; Lhommée et al. 2012) [2,3]. Due to balance dysfunction, individuals with PD commonly experience falls, which can lead to serious injuries. Therefore, it is important for therapists to have a measurement tool which can evaluate the balance ability of individuals with PD.
The Berg Balance Scale (BBS) is a tool which can evaluate balance for individuals with various disabilities (Downs 2015) . The BBS contains 14 balance tasks and each one is scored on a Likert scale from 0-4. The total score can range from 0 to 56. A higher score stands for better balance function. If the result of the BBS is below 40, it indicates that the balance function of the individual is poor and the individual may have a higher risk of falling (Berg et al. 1992) . The 14 tasks of the BBS include sitting and standing, standing unsupported, sitting unsupported, standing and sitting, transferring, standing with eyes closed, standing with foot together, reaching forward with outstretched arm, picking up an item from the floor, turning to look behind, turning 360°, placing alternate foot on a step, standing with one foot in front, and standing on one foot (Badura and Pietka 2016) . Previous studies of the BBS have identified that it has been widely applied to assess the impairment of balance function caused by PD (Franchignoni and Velozo 2005) . However, the psychometric properties (validity, reliability, responsiveness and feasibility) of the BBS in assessing balance function of an individual with PD remain unclear.
A valid reliable measurement not only plays an important role in assessing whether the research results are acceptable but also demonstrating the effectiveness of treatment in clinical practice (Hicks 2009) . This assignment will discuss these psychometric properties of the BBS for evaluating balance function of individuals with PD, identify the need for Outcome Measures (OM) in practice and conclude as to the relevance of the BBS in clinical practice with PD.
The term “berg balance scale”, “Parkinson’s disease”, “validity”, “reliability”, “responsiveness” and “feasibility” will be used in a computer-based search strategy from the following databases: Springer Link, Ovid and Google scholar. The assignment will start by defining the terms utilized to describe an OM and use the SURE checklist to critique studies.
The term “validity” is defined by Hicks (2009)  as the degree to which a measurement tests what it attempts to test. Generally, there are four main types of validity which are face validity, construct validity, content validity and predictive validity. Face validity can be defined as the test seemingly measuring what it should measure. The function of construct validity is to reflect whether a measurement can test a theoretical concept or construct. Convergent validity is a branch of construct validity, which can be defined as whether two different tests can measure a same theoretical concept or construct. The definition of content validity can be described as the correspondence between test tasks on a scale and the symptoms of a scenario being assessed. Predictive validity refers to the degree to which an instrument can predict future performance (Hicks 2009) .
Berg et al. (1989)  test the content validity of BBS in his original study which recruited 38 elderly Canadian patients (28% of them with PD) and 32 health-related staff from different healthy setting in Canada. 38 items were initially generated from three dimensions of balance (static balance, self-dynamic balance and dynamic balance) through discussing the relevance of all tasks of the BBS. Inappropriate movements were deleted due to difficulty to complete, such as “Turning head while sitting on pillow with legs crossed”. Duplicate, unimportant and irrelevant items were deleted. Because of these rigorous screening methods, the content validity of the BBS should be excellent. However, although the background of health-related staff is different and the patients are elderly people with different balance-related diseases, they all come from two cities in Canada. Cultural bias may lead to a low generalization of research results (Wan 2001) .
70 stroke patients from acute care hospitals were recruited and tested using the BBS, Barthel Index (BI) and balance subscale of Fugl-Meyer (FM-B) (these were viewed as the gold standard at this time) in order to measure their construct validity (Berg et al. 1992). All the samples were evaluated by these three instruments at baseline, 4th, 6th and 12th week. The excellent correlation between the BBS and the BI (r=0.87-0.93), and good correlation between the BBS and the FM-B (r=0.84-0.94) could be observed in this study. The correlation between the BBS and BI has no significant difference from the correlation between the BBS and FM-B. Because the BI is for assessing the daily activity and the FM-B is for assessing the balance function, this may imply a positive correlation between balance function and daily activity function.
There are currently no literature reviews addressing the construct validity and content validity of the BBS. Although the original research on the BBS investigates its validity, very few subjects of the research have PD. Most of the subjects are stroke patients and elderly people. Nonetheless, data from several investigations establish that impaired function of balance is a common problem among stroke patients, PD patients and elderly people (Byl et al. 2015; Beghi et al. 2018) [11,12]. Hence, the outcomes of content and construct validity from the BBS on measuring stroke and elderly subjects is partly transferrable to PD subjects.
The high convergent validity (using the Pearson correlation coefficient) between the BBS and Mini-BESTest has been reported by Godi et al. (2013)  at baseline (r=0.85). After the therapy programme, moderate correlation (r=0.58) between the BBS and Mini-BESTest could be found. Godi et al. (2013)  suggest that the change of correlation between the BBS and Mini-BESTest during baseline and after therapy may result from the generally higher scores in the BBS after treatment. This phenomenon is regarded as the ceiling effect of the BBS (Schlenstedt et al. 2015) . However, only one-quarter of the total sample size is PD patients, so the results of this study do not transfer well to all PD patients. Nonetheless, King et al. (2012)  recruited a total of 97 PD patients and indicated the excellent correlation between the BBS and Mini-BESTest (r=0.79). Also, Bergström et al. (2012)  recruited a total of 9 PD patients and indicated a high correlation between the BBS and Mini-BESTest (r=0.94). The above studies demonstrate the high convergent validity between the BBS and Mini-BESTest on measuring PD population. Therefore, both would appear valid to measure validity in the PD population.
Park and Lee (2016)  attempt to assess the predictive validity of the BBS on managing patients with balance issues in their systemic review. This systematic review finds that the predictive validity of the BBS is moderate when the BBS is used to predict fall risk among populations with different neuromuscular diseases. However, the population of this systematic review was not exclusively PD patients, stroke patients were also included. Although nearly one-third of the sample (n=1690) were PD patients (n=513). The presence of patients with other diseases affects the ability to conclude on validity of the BBS for predicting balance function of PD patients. Nonetheless, the population of this study has common symptoms of impaired balance. According to the high predictive validity of BBS from the outcome of this systematic review, the BBS can effectively predict the balance function of neuromuscular patients with balanced functional impairment.
There is a lack of research testing the validity of the BBS in PD patients, although the construct of the BBS is proven to measure balance and be predictive of falls. There appears sufficient evidence by Godi et al to support its use with PD patients.
Reliability refers to whether an instrument can be administered repeatedly and achieve the same result. (i.e. is it consistent). If the measurement recorded is similar when measured by different individuals (inter-rater) and the same person on different occasions (test-retest), this indicates strong reliability of the tool or OM. To determine whether the measure is reliable or not, recordings are tested using correlation statistics (e.g. intraclass correlation (ICC) for parametric data and Spearmans or Pearsons test for nonparametric data) (Hicks 2009) . If the correlation coefficient is higher than .75, it demonstrates high reliability otherwise .50 to .75 or 0 to .50 demonstrates medium or low reliability respectively (Portney and Watkins 2013) . The BBS produces ordinal data within each section, however, the total score becomes interval ratio and therefore when critiquing the evidence, it is important to consider the score being evaluated.
The inter-rater and test-retest reliability of the BBS was assessed by Schlenstedt et al. (2015) . For inter-rater reliability, 15 PD patients and three assessors were recruited. These three assessors were trained and had experience implementing the BBS. Measurements were carried out by them who took the measures once. For test-retest reliability, 17 participants were measured. One assessor evaluated each patient by taking one measure and a repeat measure 3±1 days later. The repeat was taken by the same assessor at the same time of day for each patient. This study mainly tested the reliability of the total score of 14 items of the BBS, the total score is interval level data and therefore statistical testing can be analysed by parametric tests. Therefore, this study uses correct statistical testing i.e. (ICC). The ICC for the inter-rater reliability and test-retest reliability of the BBS was very high as the ICC was r=0.98 and r=0.95, respectively. The recruitment of three experienced and trained assessors for testing the inter-rater reliability of the BBS can guarantee the accuracy of study outcomes and minimize human errors (Vincent 2006) . However, this may question whether an inexperienced tester would also have good reliability. Also, all assessors scored the BBS independently and their results are blinded to each other. This blinding can avoid the bias in the study design (Pannucci and Wilkins 2010) . Therefore, the outcome of inter- rater reliability of the BBS in this study appears to be convincing.
However, the outcome of test-retest reliability of the BBS in the study of Schlenstedt et al. (2015)  appears questionable. Although the researchers considered the impact of evaluation time (the same time of day) and antiparkinsonian medicine dosage and ensured there was no difference in the dosage of antiparkinsonian medication between the first assessment and the second assessment, the interval between two assessments is very short (3±1 days). The impact of the memory of patient and the assessor over a short period of time may result in a false high test-retest reliability of the BBS (Schatz and Maerlender 2013) . Since PD is a chronic progressive disease, the balance function of the patient may not change significantly within three weeks (Lhommée et al. 2012) . It is recommended that future studies extend the interval between assessments.
32 patients (only eight PD) were recruited by Godi et al. (2013)  to assess the inter-rater reliability and test-retest reliability of the total score of 14 items of the BBS. Its use of ICC was suited to this study, which is a strength. This study indicates the excellent inter- rater (r=0.97) and test-retest reliability (r=0.92) of the BBS. Compared to Schlenstedt et al. (2015) , whereby the procedures for assessing the reliability is almost the same. However, the inclusion of only eight PD patients may mean that the sample is not representative of the whole Parkinson population. Hence, the outcome of this study can only be accepted with caution.
Berg et al. (1989)  demonstrate excellent inter-rater reliability (total score: r=0.98; each item: r=0.71-0.99) and test-retest reliability (total score: r=0.99; each item: r=0.71-0.99) of the BBS in the original study. All samples came from a geriatric ward and their balance function was affected by a variety of diseases. Combined with the above two studies, the good inter-rater and test-retest reliability of the BBS in assessing patients with PD and other balance disorders can be demonstrated.
Internal consistency (IC) is a kind of reliability in which all components of a measure that propose to measure the same things actually produce similar scores. Cronbach’s alpha is usually used for calculating internal consistency. The range of Cronbach’s alpha is between 0.00 to 1.00. The value of Cronbach’s alpha is required to be high and greater than 0.70 is satisfactory in demonstrating good IC (Connelly 2011) . Halsaa et al. (2007)  report the IC of the Norwegian version of the BBS (r=.87) is higher than .70 which could show relatively strong IC. The original study of the BBS’s IC is very high (r=.96), with a value >.90 which has been described as a requirement for inclusion clinically use (Berg et al. 1989). The high IC shows that all items of the BBS have an excellent function for evaluating balance.
Responsiveness of an OM reflects the ability of a measurement to demonstrate and examine change over time (Portney and Watkins 2013) . 26 PD patients were recruited by Lim et al. (2005)  for examining the responsiveness and feasibility of the BBS. Because there is very little literature on the study of responsiveness of the BBS, this study determines to apply the Smallest Detectable Difference (SDD) which is a kind of statistical data calculated by standard error (Beckerman et al. 2001; Beghi, E. et al. 2018; Bergström, M. et al. 2012; Morris, M. E. et al. 1994) [25-28]. However, the SDD from this study does not show the extent to which the variable changes of BBS. Moreover, the results from the SDD were subjectively applied for determining the responsiveness of the BBS. Therefore, the responsiveness of the BBS remains unclear. This study also focuses on the feasibility of the BBS which is determined by the length of one assessment time in a different setting. Assessment time in the clinical setting for one patient is 20-25 minutes and 20-30 minutes in the home environment, indicating that the BBS assessment is feasible in clinical and home circumstances. In addition, the equipment for assessing the BBS is just a chair, a watch, a ruler and a step. The cost of administering the BBS is not expensive. Therefore, the BBS is a feasible tool.
Taken together, importance of using valid reliable OM in physio is to prove the effectiveness of treatment. The BBS is indeed a reliable and valid instrument that can be used to assess PD patients with impaired balance. One assessment time of the BBS in the clinical and home environment is short and stable, and its cost is low. However, due to its lack of literature on investigating the responsiveness of the BBS, this remains unclear. Also, some issues from previous studies include complicated study populations (not all PD patients) and methodological flaws (short interval between assessments). Attempts to correct the above issues would help to establish a greater degree of certainty when assessing balance in PD patients using the BBS. Nonetheless, the BBS remains an appropriate assessment of balance for the clinical PD setting.