Patient-Reported, Not Performance Based, Outcome Measures Are Correlated With Future Falls in Community Dwelling Older Adults

Background: One-third of community-dwelling older adults fall at least once a year, yet many commonly used performance-based outcome measures for falls have ceiling effects. Additionally, psychological factors such as fear of falling and avoidance behaviors may contribute to fall risk. Purpose: To determine whether balance confidence, fear of falling avoidance behaviors, or performance on high-level mobility outcome measures can predict falls in community-dwelling older adults. Methods: Eighty-nine participants (76 ± 7 years) completed the Activities-specific Balance Confidence Scale (ABC), Fear of Falling Avoidance Behaviors Questionnaire (FFABQ), Functional Gait Assessment (FGA), and Community Balance & Mobility Scale (CB&M) and then recorded their falls for 6 months. Spearman’s rho, a Point-Biserial Correlation, and Receiver Operating Curves were used to analyze the relationship of the selected outcome measures, past falls, and future falls. Results: There were statistically significant correlations between falls and both the ABC (r2=-.235, p=.022) and the FFABQ (r2=.286, p=.019). There was a positive correlation between history of falls and future falls (r=.323, p=.002). The ABC (area under curve [AUC] =.632, p=.038) and the FFABQ (AUC=.655, p=.015) were significant predictors of future falls. The optimal cutoff scores for determining individuals who experienced falls were: ABC-96% (79% sensitivity, 66% specificity) and FFABQ-1.50 (79%, 55%). Conclusion: No single outcome measure was found to be strongly correlated with falls. The ABC, FFABQ, and history of falls were weakly correlated with future falls. A cutoff score of 96% on the ABC and 1.50 on the FFABQ had adequate sensitivity, but low specificity for predicting falls