Physical performance measures in the clinical setting.

OBJECTIVES To assess the ability of gait speed alone and a three-item lower extremity performance battery to predict 12-month rates of hospitalization, decline in health, and decline in function in primary care settings serving older adults. DESIGN Prospective cohort study. SETTING Primary care programs of a Medicare health maintenance organization (HMO) and Veterans Affairs (VA) system. PARTICIPANTS Four hundred eighty-seven persons aged 65 and older. MEASUREMENTS Lower extremity performance Established Population for Epidemiologic Studies of the Elderly (EPESE) battery including gait speed, chair stands, and tandem balance tests; demographics; health care use; health status; functional status; probability of repeated admission scale (Pra); and primary physician's hospitalization risk estimate. RESULTS Veterans had poorer health and higher use than HMO members. Gait speed alone and the EPESE battery predicted hospitalization; 41% (21/51) of slow walkers (gait speed <0.6 m/s) were hospitalized at least once, compared with 26% (70/266) of intermediate walkers (0.6-1.0 m/s) and 11% (15/136) of fast walkers (>1.0 m/s) (P <.0001). The relationship was stronger in the HMO than in the VA. Both performance measures remained independent predictors after accounting for Pra. The EPESE battery was superior to gait speed when both Pra and primary physician's risk estimate were included. Both performance measures predicted decline in function and health status in both health systems. Performance measures, alone or in combination with self-report measures, were more able to predict outcomes than self-report alone. CONCLUSION Gait speed and a physical performance battery are brief, quantitative estimates of future risk for hospitalization and decline in health and function in clinical populations of older adults. Physical performance measures might serve as easily accessible "vital signs" to screen older adults in clinical settings.