BACKGROUND: Our purpose was to examine the use of classification and diagnostic systems in the field of psychiatry (CDSP) from a bibliometric perspective, over the period 1980-2005. METHODS: We selected (in EMBASE and MEDLINE databases) documents that contained, in any of their sections, the descriptors 'psychiatr*', 'DSM*', 'ICD*', or 'diagnostic criteria',as well as other more specific descriptors. As a bibliometric indicator of production we applied Price's law. We also calculated the national participation index (PI) and correlated it with overall PI in biomedical and health sciences, and with PI in the discipline of psychiatry. RESULTS: We obtained 20,564 original documents; 15,743 referred to the Diagnostic and Statistical Manual of Mental Disorders (DSM) and 3,106 to the International Classification of Diseases (ICD). Our results indicate non-fulfilment of Price's law, since scientific production on CDSP does not undergo exponential growth (correlation coefficient r = 0.9651, vs. r = 0.9927 after linear adjustment). Of the 10 journals with the highest impact factor in the field of psychiatry, the Journal of Clinical Psychiatry has the highest PI in the DSM subgroup (PI = 14.77), and the British Journal of Psychiatry in the ICD subgroup (PI = 1.54). The principal producer country is the United States (PI = 37.9), though in proportion to its production in the psychiatric field the ranking is headed by Finland. Only 10 countries, of the 20 major producers in health sciences, surpass their own PI in the field of psychiatry (Brazil, Italy, Japan, Austria, Spain, Germany, France, India, Switzerland, and China). CONCLUSIONS: Over recent years, the use of CDSP (basically the DSM or ICD) in the scientific literature has increased. Nevertheless, the abstracts to these studies, included in the principal databases, should always specify the diagnostic criteria employed, with a view to increasing information levels and reliability for the reader.