Context: The French Alzheimer plan 2008-2012 includes the implementation of "disease managers" or social workers, from the onset of the disease, for every patient with a diagnosis of Alzheimer disease or associated disorder. However, previous trials evaluating the efficacy of a systematic intervention for each case from the onset of the disease, including non complex cases, have shown little evidence.

Objective: To evaluate the efficacy of a social worker to prevent behavioural problem for patient with a new diagnosis of Alzheimer disease or associated disorder.

Design and setting: Multicentric cluster randomized control trial. Patients from several memory consultations and specialists (neurologist, psychiatrist or geriatrician) in France are included.

Intervention: In addition to usual care, patients in the intervention group have systematic home visits from a social worker and regular follow-ups by phone over a one-year period. This social worker informs the patient and caregiver, evaluates the needs, gives advices to prevent complications and is a link between the patient and the memory consultation or the specialist. Patient in the control group have absolutely no intervention and are followed-up as usually. In order to not modify the usual care, they have no information on their participation in this trial before the one-year evaluation.

Main outcome measure: NeuroPsychiatric Inventory (NPI) administrated at one year, which evaluates frequency, severity and repercussion of several behavioural problems. Secondary outcomes included disability, cognition, depression, health, quality of life, institutionalization, and resource use.

Inclusion Criteria:

• Clinical diagnosis of Alzheimer's Disease or other dementia (Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition [DSM IV] criteria of dementia)
• Presence of an informal caregiver
• Mild to moderately severe dementia (MMSE [10-28])
• Diagnosis of dementia made by a specialist ≤ 6 months
• Patient affiliated to the national health insurance system

Exclusion Criteria:

• Institutionalized patient
• Patient on legal guardianship
• Behavioural problem with important clinical repercussion
• Psychotic syndrome
• Severe and unstable general pathology