Objective: The objective of this research was to determine if a personalized music intervention reduced the frequency of agitated behaviors as measured by structured observations of nursing home (NH) residents with dementia.

Design: The design was a parallel, cluster-randomized, controlled trial.

Setting: The setting was 54 NH (27 intervention, 27 control) from four geographically-diverse, multifacility NH corporations.

Participants: The participants were 976 NH residents (483 intervention, 493 control) with Alzheimer's disease or related dementias (66% with moderate to severe symptoms); average age 80.3 years (SD: 12.3) and 25.1% were Black.

Intervention: The intervention was individuals' preferred music delivered via a personalized music device. MEASUREMENT: The measurement tool was the Agitated Behavior Mapping Instrument, which captures the frequency of 13 agitated behaviors and five mood states during 3-minute observations.

Results: The results show that no verbally agitated behaviors were reported in a higher proportion of observations among residents in NHs randomized to receive the intervention compared to similar residents in NHs randomized to usual care (marginal interaction effect (MIE): 0.061, 95% CI: 0.028-0.061). Residents in NHs randomized to receive the intervention were also more likely to be observed experiencing pleasure compared to residents in usual care NHs (MIE: 0.038; 95% CI: 0.008-0.073)). There was no significant effect of the intervention on physically agitated behaviors, anger, fear, alertness, or sadness.

Conclusions: The conclusions are that personalized music may be effective at reducing verbally-agitated behaviors. Using structured observations to measure behaviors may avoid biases of staff-reported measures.