The study tested the effect of strength and endurance training on gait, balance, physical health status, fall risk, and health services use in older adults. The study was a single-blinded, randomized controlled trial with intention-to-treat analysis. Adults (n = 105) age 68-85 with at least mild deficits in strength and balance were selected from a random sample of enrollees in a health maintenance organization. The intervention was supervised exercise (1-h sessions, three per week, for 24-26 weeks), followed by self-supervised exercise. Exercise groups included strength training using weight machines (n = 25), endurance training using bicycles (n = 25), and strength and endurance training (n = 25). Study outcomes included gait tests, balance tests, physical health status measures, self-reported falls (up to 25 months of follow-up), and inpatient and outpatient use and costs. There were no effects of exercise on gait, balance, or physical health status. Exercise had a protective effect on risk of falling (relative hazard = .53, 95% CI = .30-.91). Between 7 and 18 months after randomization, control subjects had more outpatient clinic visits (p < .06) and were more likely to sustain hospital costs over $5000 (p < .05). Exercise may have beneficial effects on fall rates and health care use in some subgroups of older adults. In community-living adults with mainly mild impairments in gait, balance, and physical health status, short-term exercise may not have a restorative effect on these impairments.