The object of this study was to assess the change towards a biopsychosocial health concept among medical researchers in the last two decades, after the explicit criticism of the biomedical model in the late 1970s because of its somatic reductionism. The concepts of 'health' or 'healthy status of an individual' as reported as variable in empirical articles published in the journal The Lancet over the years 1978-1982 (period a) and 1996-2000 (period b) were searched by means of Medline and compared for their definition of these variables. None of the 52 examined papers set out a positive and replicable definition of 'health' (seven papers) or 'healthy status' (45). No difference was found between the two periods studied except for the failure of reports to describe 'healthy status' at all (65.5% in a, 19% in b). Most articles do it in an indirect way, namely through exclusion conditions of subjects taking part in treatment or control groups. Only three studies include psychological dimensions in their measures of 'healthy status' (two in a, one in b). Concerning 'health', all seven examined papers include psychological or both psychological and social dimensions. Although a change towards a more holistic concept of health has occurred in academic and institutional contexts over the last few decades, there does not appear to have been a parallel change in the practical domains of medicine. Possible reasons are discussed, specially the difficulty of applying the biopsychosocial model in medical care and the difficulty of competing with the traditional biomedical concept of health, which has proved fruitful and dominant in medicine over the past three centuries.