We developed a telemedicine protocol for diagnosis of Alzheimer's Disease (AD). Assessments by video-conferencing (remote) were compared with face to face (direct) assessments. Eight physicians performed direct assessments and two physicians conducted remote assessments. There was alternate allocation of direct or remote initial assessment. The participants were 20 subjects over 65 years living in a rural area and referred by general practitioners (GPs) because of cognitive impairment. Each assessment included a Standardised Mini Mental State Examination, Geriatric Depression Scale, Katz assessment of Activities of Daily Living, Instrumental ADL assessment, and the Informant Questionnaire for Cognitive Decline in the Elderly. Laboratory results and radiological imaging were available from referring GPs. There was good agreement for diagnosing Alzheimer's disease between telemedicine and direct assessment, kappa = 0.8 (P<0.0001). However, because of the small sample size, the presence of systematic bias could not be completely excluded. We conclude that it is possible to diagnose AD at a distance using telemedicine, but this requires validation with a larger study.