Study objective: Elderly emergency department patients have complex medical needs and limited social support. A transitional model of care adapted from hospitals was tested for its effectiveness in the ED in reducing subsequent service use. Methods: A randomized clinical trial was conducted at 2 urban, academically affiliated hospitals. Participants were 650 community-residing individuals 65 years or older who were discharged home after an ED visit. Main outcomes were service use rates, defined as repeat ED visits, hospitalizations, or nursing home admissions, and health care costs at 30 and 120 days. Intervention consisted of comprehensive geriatric assessment in the ED by an advanced practice nurse and subsequent referral to a community or social agency, primary care provider, and/or geriatric clinic for unmet health, social, and medical needs. Control group participants received usual and customary ED care. Results: The intervention had no effect on overall service use rates at 30 or 120 days. However, the intervention was effective in lowering nursing home admissions at 30 days (0.7% versus 3%; odds ratio 0.21; 95% confidence interval [CI] 0.05 to 0.99) and in increasing patient satisfaction with ED discharge care (3.41 versus 3.03; mean difference 0.37; 95% CI 0.13 to 0.62). The intervention was more effective for high-risk than low-risk elders. Conclusion: An ED-based transitional model of care reduced subsequent nursing home admissions but did not decrease overall service use for older ED patients. Further studies are needed to determine the best models of care for this setting and for at-risk patients. [Ann Emerg Med. 2003;41:57-68.]