Activities-specific performance frequency can accurately detect fallers in elderly populations: an alternative method for quantifying activity restrictions

Background The high prevalence of falling among older adults constitutes a major public and clinical health concern. Many elderly persons may develop activities-specific restriction due to the risk of falling. This highlights the need for relevant evaluative tools. Methods This cross-sectional study used activities-specific performance frequency indicators to quantify activity restrictions in elderly participants, with all measures based on items from the Activities-Specific Balance Confidence (ABC) scale. Specifically, we tested for correlations between activities-specific performance frequency and balance confidence, functional balance/mobility, and fall history. There were 88 elderly participants, including 28 with stroke, 30 with Parkinson’s disease, and 30 with no neurological diseases. In addition to their activities-specific performance frequency measures, we collected a series of demographic and health-related characteristics from each participant. We analyzed between-group differences in activities-specific performance frequency and other demographic and health-related characteristics via the one-way analysis of variance and Kruskal-Wallis test. Next, we used the Spearman’s rank correlation test and binary logistic regression to investigate the correlations between activities-specific performance frequency and demographic/other health-related characteristics. Results There were significant group differences in performance frequency for all ABC activity items except for walking around the house, average ABC scores, and functional balance/mobility among normal older adults, participants with strokes and those with Parkinson's disease. Activities-specific performance frequency showed stronger correlations with activities-relevant functional mobility ( r =0.250-0.713 for 15 items with significant correlations, 13 activity items with r ≧0.4) than with balance confidence ( r =0.279-0.668 for 13 items with significant correlations, 10 activity items with r ≧0.4). The performance frequency of walking in crowds/bumped was the most sensitive measure for predicting fallers (odd ratio=3.310, p <0.05). Conclusions This study proposed and validated the usage of activities-specific performance frequency as an alternative method for quantifying activity restrictions among older adults.