The international classification of mental disorders (ICD-8) presented by the WHO has not been accepted in all countries and where it is used, local adjustments are made and sometimes parallel classification models are used. The diagnostic system has also been criticized as lacking in exactness and consistency and the reliability between diagnosticians has been shown to be low. As a consequence, international communication is made difficult and research is hampered. This problem is particularly relevant in the field of biological research. In fact, it can be suspected that most inconsistencies as regards results obtained in different places might depend upon an inconsistent use of the current diagnostic labels. A possible solution of this problem can be the use of a multiaspect classification model. Such a multiaspect model (MACM) including four variables – symptomatology, severity, course and supposed etiopathogenesis – has been tested for several years at Umeå. MACM is shown to be easy to communicate both in undergraduate training and in international communication. The reliability between diagnosticians is found to range from 56 to 82% as compared to 22-36% as concerns ICD-8. It is also shown that fairly homogeneous groups, both regarding course, supposed etiopathogenesis and biological basis, can be formed and that MACM seems to bear temporal stability. Computer programming of MACM diagnoses is as possible as with ICD-8 diagnoses. Furthermore, when MACM is used in clinical routine work, much more information valid in administrative routines is stored than what is possible when ICD-8 is used.