We have investigated prospectively the diagnostic accuracy, specialist satisfaction and patient-specialist rapport of a low-cost audio-visual link between a junior doctor with a patient and a consultant rheumatologist. Using a telephone link and subsequently a video-phone link, 20 patients, with various rheumatological problems, were presented by a junior doctor to the consultant rheumatologist for provisional diagnosis. All patients were then seen face to face by the consultant, when a final diagnosis was made. An independent consultant rheumatologist made a 'gold standard' diagnosis. Thirty-five per cent of diagnoses were made correctly over the telephone and 40% over the video-phone--there was no significant difference in the diagnostic accuracy between these two methods of communication. Rapport over the video-phone was universally poor. Where it was important, clinical signs could not be visualized over the video-phone and in more than 85% of cases small joint swellings could not be seen clearly.