Background and purpose: The sensory and motor deficits associated with multiple sclerosis (MS) contribute to falls with the majority of persons with MS falling at least once annually. To appropriately direct treatment, accurate fall prediction measures are needed. In this study of community-dwelling individuals with MS followed for 12 months, we sought to determine frequency of falls, utility of clinical balance tests to predict falls, and accuracy of participants' retrospective recall of fall events.
Methods: Independently ambulatory persons with MS underwent 5 clinical balance tests including Activities-specific Balance Confidence, Berg Balance Scale, Functional Reach, Timed Up and Go, and Dynamic Gait Index, and one disease-specific measure of disability (Expanded Disability Status Scale) and then were followed for 1 year. Participants were queried monthly by phone to determine the number of fall events that had occurred. Accuracy of fall prediction was determined by receiver operating characteristic curve analysis and comparison of balance test performance between fallers and nonfallers.
Results: Sixty-one percent of the 38 participants were classified as fallers at 12-month follow-up. Only the Berg Balance Scale, Dynamic Gait Index, and the Activities-specific Balance Confidence demonstrated clinically useful levels of accuracy. In addition, participants were generally poor in their accurate recall of fall events relative to their monthly fall reports.
Discussion and conclusions: The majority of participants fell during a 1-year prospective follow-up. Only 2 balance performance measures and 1 balance confidence measure accurately distinguished between fallers and nonfallers as well as possessed clinically useful levels of sensitivity and specificity. These results also emphasized the inaccuracy of retrospective fall history in an MS sample.