Objectives: To examine the effect of integrated emotion-oriented care on nursing home residents with dementia and nursing assistants. Design: A multi-site randomized clinical trial with matched groups, and measurements at baseline and after seven months. Setting: Sixteen psychogeriatric wards in fourteen nursing homes in the Netherlands. Participants: One hundred and forty-six elderly residents with the diagnosis dementia of the Alzheimer (DAT) type, mixed DAT and vascular dementia, and dementia syndrome (NAO) and 99 nursing assistants. Interventions: Integrated emotion-oriented care and usual care. Measurements: Demented elderly: Behaviour and mood related to adaptation to the illness and the institutionalization. Nursing assistants: General health as measured by feelings of stress, stress reactions, feeling of competence and illness. Results: Positive effects in favour of the integrated emotion-oriented care were found in mild to moderately demented residents on two adaptive tasks: maintaining an emotional balance (less anxiety) and preserving a positive self-image (less dissatisfaction). In the trained group of nursing assistants fewer stress reactions were found only in those who perceived improvement in their emotion-oriented care skills after training. Conclusion: Emotion-oriented care is more effective with regard to the emotional adaptation in nursing homes of persons with a mild to moderate dementia. For the severely demented elderly we did not find this surplus value. This outcome is of clinical importance for elderly persons with dementia who are cared for in nursing homes. With respect to the nursing assistants it is concluded that emotion-oriented care has a positive influence on stress reactions in some of them.