This study tested the effectiveness of the Serial Trial Intervention (STI), an innovative clinical protocol for assessment and management of unmet needs in people with late-stage dementia. A double-blinded randomized experiment was conducted in 14 nursing homes with 114 subjects. The treatment group had significantly less discomfort than the control group at posttesting and more frequently had behavioral symptoms return to baseline. The group of nurses using the STI also showed more persistence in assessing and intervening than control group nurses did. There was a statistically significant difference between the groups in the use of pharmacological, but not nonpharmacological, comfort treatments. Results suggest that the STI is effective and that effective treatment of discomfort is possible for people with late-stage dementia.