The Berg balance scale has high intra- and inter-rater reliability but absolute reliability varies across the scale: a systematic review

Questions: What is the intra-rater and inter-rater relative reliability of the Berg Balance Scale? What is the absolute reliability of the Berg Balance Scale? Does the absolute reliability of the Berg Balance Scale vary across the scale? Design: Systematic review with meta-analysis of reliability studies. Participants: Any clinical population that has undergone assessment with the Berg Balance Scale. Outcome measures: Relative intra-rater reliability, relative inter-rater reliability, and absolute reliability. Results: Eleven studies involving 668 participants were included in the review. The relative intrarater reliability of the Berg Balance Scale was high, with a pooled estimate of 0.98 (95% CI 0.97 to 0.99). Relative inter-rater reliability was also high, with a pooled estimate of 0.97 (95% CI 0.96 to 0.98). A ceiling effect of the Berg Balance Scale was evident for some participants. In the analysis of absolute reliability, all of the relevant studies had an average score of 20 or above on the 0 to 56 point Berg Balance Scale. The absolute reliability across this part of the scale, as measured by the minimal detectable change with 95% confidence, varied between 2.8 points and 6.6 points. The Berg Balance Scale has a higher absolute reliability when close to 56 points due to the ceiling effect. We identified no data that estimated the absolute reliability of the Berg Balance Scale among participants with a mean score below 20 out of 56. Conclusion: The Berg Balance Scale has acceptable reliability, although it might not detect modest, clinically important changes in balance in individual subjects. The review was only able to comment on the absolute reliability of the Berg Balance Scale among people with moderately poor to normal balance.