Objective: To test the impact of joint impairment on ambulation in the elderly, using a multivariate model.
Design: Cross-sectional observational study (baseline data from an ongoing longitudinal study).
Subjects: Five hundred thirty-two persons over age 60, including continuing care retirement community (CCRC) (n = 222), homebound (n = 63), and ambulatory (n = 247) respondents. Mean age at assessment = 76.6 (SD = 6.6).
Measurement: Independent variables included sociodemographics, physician measures of lower joint impairment, an index of cormorbidities derived from physical examination or chart abstract, self-assessed arthritis pain, depression, and anxiety. The dependent variable, ambulation, was measured as walk-rate, based on the time required to walk 50 feet.
Main results: For the total sample, 61% of the subjects were impaired in more than one lower joint group, with almost 50% of the homebound impaired in more than three joint groups. Demographics (particularly age and education) explained much of the variance in walk-rate. After controlling for demographics and membership in the groups purposely sampled (CCRC, homebound, ambulatory), lower joint impairment accounted for an additional 7% of the variance in walk-rate. Total amount of variance explained by the model was 56% (Adjusted R2 = .56). A 3-point change in lower joint impairment score, equivalent to the maximum impairment score for a single joint group, is associated with a 4-second change in the mean time required to walk 50 feet. The knee and lower spine joints contributed most to the impact of the lower joint impairment measure.
Conclusions: Among the elderly, age and education are strong predictors of performance on a walk-rate test. Independent of demographics and non-musculoskeletal conditions, joint impairment is associated with diminished walking ability in this population.