I did not find too much new in this book. It is composed of twelve presentations on biopsychosocial medicine given at a conference in London in 2002 to which international experts were invited. The edited transcripts of the discussions after each talk are included, together with a final discussion chapter. The editor of the book, Peter White, Professor of Psychological Medicine at Queen Mary, University of London, provides a concluding chapter. I wish he had attempted to incorporate the discussions with the rest of the book, but he preferred to try to keep them as ‘spontaneous as they were on the day’.

I also think there were lost opportunities to contribute to progress in the field. For example, Francis Creed, Professor of Psychological Medicine at Manchester University, discusses whether the patientcentred and biopsychosocial approaches are compatible. In his chapter, he juxtaposes them, whereas, as pointed out in the discussion after his paper, Moira Stewart, Ian McWhinney and others, who have developed the patient-centred method at the University of Western Ontario, would look for integration rather than distance between the approaches.

The contributions of George Davey Smith, Professor of Clinical Epidemiology at Bristol University, redeemed the book for me. As he points out in discussion, George Engel's work in the 1970s, which of course is seminal for the understanding of the biopsychosocial approach, became influential in the context of the acknowledgement of the limits of biomedicine by, for example, Thomas McKeown and Ivan Illich. I found this a useful insight. Davey Smith's own chapter argues that there is, in fact, little evidence that psychosocial factors have a direct aetiological effect on physical illness and biological processes. Correlation of stress, for example, with outcome may be explained by confounding, rather than reflecting a causal explanation. Bias is also introduced into observational studies through an increased reporting tendency of stressed individuals. The number of experimental studies of psychosocial interventions for physical disease has been relatively few, and any significant effects, if found, are small and may not be specific.

Davey Smith's critique should not be taken to imply that physical symptoms cannot be psychogenic in origin. Doctors fail to recognise the emotional and psychological nature of too many patients' complaints. So-called medically unexplained symptoms are common. For this reason alone, more emphasis should be placed on comprehending the biopsychosocial approach. This book is a useful contribution to that aim.