In 1988 a random sample of 310 unemployed aged 16 to 63 years who had been registered for more than 12 weeks and lived in the four municipalities of Greenland was examined by a doctor and a nurse. Self-reported health and psychometric tests produced higher prevalences of disease than in a reference population of employed persons. On the basis of both clinical examination and the testing, 22% of the unemployed, were judged to be depressed, as against 2% in the reference group. 17% of the unemployed were scored as "morbidly anxious" and 17% as psychosomatically sick, as against 2% and 4% of the references. The health and social services which the group had received were regarded as unsatisfactory, both by the doctor and by the unemployed themselves. A special health maintenance programme is proposed for long-term unemployed.