Previous work suggests that serial functional assessment can focus treatment and anticipate outcome, reducing geriatric inpatient length of stay (LOS) and improving the care provided. Two hundred forty-five serial geriatric admissions to a 424-bed community teaching hospital in New York City were randomized in a 9-week case-controlled study comparing a treatment group with age, comorbidity, and DRG-matched controls and identical DRG-matched patients from the previous year. Comparison with DRG-specific LOSs showed an average decrease of 30% for eight combined target DRGs (P < 0.05) and 22% overall. Multiple regression analysis controlling for age and DRG found treatment a significant factor in reducing LOS within eight target DRGs (P < 0.01). Treatment group discharged functional scores were higher than control group in four of six DRGs despite the treatment group's significantly earlier discharge. Although prehospitalization comorbidity was equal, inhospital morbidity was significantly lower in treatment group patients (P < 0.01). Twelve treatment group patients (8.5%) and 25 control group patients died during the study. Hospital discharges increased by more than 1,400 for the year of the project, yielding a greater than $7,000,000 rise in efficiency. The program cost was under $120,000 for the year.