An attempt was made to compare the diagnostic efficiency and overall usefulness (for the patient and for his primary care physician) of an investigation unit in internal medicine. 100 patients seen successively were studied. At the end of the investigation the patients were divided into 3 groups: (A) organic disease diagnosed (n = 44); (B) unexplained somatic complaint (n = 16); (C) unexplained functional complaint (n = 40). One year later, an inquiry about the patient's outcome was made by telephone to the patient and/or his physician. Follow-up of 99 patients did not disclose diagnostic errors in patients of groups A and C; in group B, 2 diagnoses were missed. The investigation was considered "useful" in 31 patients (disease cured or improved, patient reassured); 57% of primary-care physicians were aided despite a negative investigation.