Rogers HL¹, Cromwell RL¹, Newton RA²; ¹University of Texas Medical Branch at Galveston, Galveston, Texas and ²Temple University, Philadelphia, PA. hrogers@utmb.edu.

PURPOSE: Gait is a function requiring ongoing dynamic balance adjustments. The physiological effects of aging often result in decreased balance, stability and confidence with gait for older adults. Older adults often adapt gait to minimize instability. One adaptation commonly measured is a slowing of average gait speed. Gait speed is related to measures of balance, mobility and functional independence. The purpose of this study was to investigate the association between age, one functional balance measure (Berg Balance Scores), two self-report measures of fear of falling (Activities-Specific Balance Confidence (ABC) Scale and Fear of Falling) and gait speed in young and older adults. SUBJECTS: Thirty seven healthy, community dwelling adults were assigned to an older (N = 17, range = 67-90 ( = 76.2 ± 6.9 years) or younger (N = 20, range = 23-35 with = 26.0 ± 3.4 years) age group. The older group consisted of 6 male and 11 females and the younger of 9 males and 11 females. METHODS: Each subject completed the ABC scale and answered the dichotomous question "Are you afraid that you will fall?" A Berg Balance Scale (BBS) score was determined for each subject. The MacReflex motion analysis system was used to collect average walking speed for each subject via calculation of the average horizontal linear velocity of a reflective marker placed at the lumbosacral interspace. ANALYSIS: Analysis was performed with SAS statistical software using a linear regression model at alpha =.05. Gait speed was compared to the four variables of age group, fear of falling, BBS, and ABC score using an all possible regression models process with R² and Mallow’s C_p as model selection criteria. RESULTS: The four variable model was best for maximizing R² without increasing bias (C_p value). Combined, the four variables accounted for 26% of the variance in gait speed. Backward regression, performed to investigate variable significance, revealed only ABC score as significantly associated with gait speed (p = .034); as ABC score increased by one point, gait speed increased by .0235 m/sec. CONCLUSIONS: Twenty-six percent of variation in gait speed with age can be explained by a model with these four balance and confidence variables. This suggests that regression analysis modeling can be a useful way to investigate the association of performance variables to function. ABC score alone was significant in this study. Further investigation of other specific variables or variable combinations may increase the potential to predict function based on performance variables. FUNDING SOURCE: National Institute on Aging, grant # 1-RO3-AG16884-01.

Copyright 2004 by the American Physical Therapy Association

Reprint Information
Requests for reprints should be directed to the corresponding author of the article. Students and other academic customers may receive permission to reprint copyrighted material from Physical Therapy by contacting the Copyright Clearance Center Inc, 222 Rosewood Dr, Danvers, MA 01923. Similar inquiries by all others should be made to the APTA Editorial Office, Attn: Physical Therapy.