Use of the Berg balance test to predict falls in elderly persons

Background and Purpose. The purpose of this study was to determine whether the Berg balance test could be used to predict an elderly person's risk of falling. Subjects. Sixty-six residents of two independent life-care communities, aged 69 to 94 years (X = 79.2, SD = 6.2), participated. Methods. Subjects completed a questionnaire pertaining to their fall history and activity level. The Berg balance test, consisting of 14 functional subtests, was then administered. Six months later, subjects again completed the questionnaire. Results. Performance of activities of daily living predicted 43% of the subjects' scores. There was a difference between the subjects who were prone to falling and those who were not prone to falling, but the test demonstrated poor sensitivity for predicting who would fall. The specificity of the test was very strong. The use of an assistive device was a strong predictor of performance on the Berg balance test. No relationship was noted between increasing age and decreasing performance on the Berg balance test. Conclusion and Discussion. Although the Berg balance test demonstrated only 53% sensitivity, the results support the test developers' sue of 45 (out of 56) as a generalized cutoff score on the test were less likely to fall than were those adults who scored below the cutoff score. Decreased scores, however, did not predict increased frequency of falls. Results must be viewed cautiously because self-report was the sole means of documenting fall history.