An additional factor that might cause variation in the reliability of the Berg Balance Scale is the underlying health conditions of subjects

whose balance is tested. Individual studies are unlikely to be able to investigate the Berg Balance Scale over the full range Selumetinib purchase of the scale and over the broad spectrum of causes of disordered balance. This review describes the range of subjects in whom the reliability of the Berg Balance Scale has been studied, reporting both their balance as well as any underlying health condition. A previous literature review of the Berg Balance Scale (Blum and Korner-Bitensky 2008) considered the relative reliability of the Berg Balance Scale in patients with stroke and found it to have strong reliability. The current S3I-201 in vivo review covers important aspects of the reliability of the Berg Balance Scale not considered by the earlier review, including absolute reliability, and the reliability of the Berg Balance Scale in patients with conditions other than stroke. Floor or ceiling effects occur when a significant proportion of a tested population

achieve the lowest or highest possible score on a test, respectively (Everitt 2010). In groups where the mean Berg Balance Scale score is close to 0 or 56, the scale is unlikely to be useful in discriminating between individuals and will exhibit floor or ceiling effects. In such cases the scale is unlikely to be able to detect a change in balance, even if there is a real change. While floor and ceiling else effects can potentially impair the clinical and research usefulness of the Berg Balance Scale, they are also likely to inflate its absolute reliability. A person with extremely poor balance is likely

to be uniformly rated at 0/4 on most elements of the Berg Balance Scale. Conversely, a person with extremely good balance is likely to be uniformly rated 4/4 on most items of the Berg Balance Scale. Floor and ceiling effects involve groups with lower variability, which in turn lead to lower estimates of relative reliability compared to groups with more variable scores. Therefore, absolute and relative reliability should be interpreted with reference to floor and ceiling effects. The specific study questions for this systematic review were: 1. What is the relative intra-rater and inter-rater reliability of the Berg Balance Scale? A literature search was undertaken to locate eligible published studies. Electronic searches of Medline, CINAHL, Embase, and the Cochrane Library from 1980 to August 2010 were conducted using ‘Berg Balance Scale’ as a search term. No search terms were used for intervention type or health condition and no methodological filter was used for study design. See Appendix 1 on the eAddenda for the detailed search strategy. All potentially relevant papers were identified from abstracts and assessed for inclusion. The reference lists of included studies were searched for additional relevant papers.