Objective: To investigate the effectiveness of snoezelen, integrated in 24-hour care, on the communication of Certified Nursing Assistants (CNAs) and demented nursing home residents during morning care. Methods: A quasi-experimental pre-and post-test design was conducted, comparing sic psychogeriatric wards, that implemented snoezelen, to six control wards, that continued in giving usual care. Measurements were performed at baseline and 18 months after a training 'snoezelen for caregivers'. Independent assessors analysed 250 video-recordings directly from the computer, using an adapted version of the Roter Interaction Analysis System (RIAS) and non-verbal measurements. Results: Trained CNAs showed a significant increase of resident-directed gaze, affective touch and smiling. The total number of verbal utterances also increased (more social conversation, agreement, talking about sensory stimuli, information and autonomy). Regarding residents, a significant treatment effect was found for smiling, CNA-directed gaze, negative verbal behaviours (less disapproval and anger) and verbal expressed autonomy. Conclusion: The implementation of snoezelen improved the actual communication during morning care. Practice implications: Teaching CNAs to provide snoezelen has added value for the quality of care. Morning care by trained CNAs appeared to take more time. This suggests that (some) time investment might be required to achieve positive effects on CNA-resident communication.