My first contribution to the Journal under this title (Reference MayouMayou, 1983) coincided with the foundation of the College's Group (now Section) for Liaison Psychiatry, which was one of the first, perhaps the first, national organisation for consultation—liaison. Even at this moment of achievement, the paper bears the mark of a scepticism deeply inculcated by my education in institutions that have believed firmly in British empiricism. A second paper, in 1990 (Reference MayouMayou, 1990), recorded my continuing concern about the lack of high-quality evidence (even compared with psychiatry as a whole) as well as an alleged lack of understanding by general hospital colleagues. I also noted signs of a change of clinical emphasis from traditional in-patient consultation—liaison to working with out-patients with major physical disorders and with ‘functional syndromes’ and emphasised the need for new and effective psychological treatments in routine care.

Returning now, I am pleased that from a more than 20-year perspective liaison (or consultation—liaison) psychiatry has become firmly established and increasingly well understood within medicine. Sadly, we have been less successful with our psychiatric colleagues. Their relative neglect of liaison psychiatry is part of a wider lack of interest in the treatment of the very common anxiety and depressive disorders that make up a large part of everyday medical practice. There is increasing evidence that other professions, and (infuriatingly) alternative and complementary medicine, are taking on a role that should be ours. The lack of understanding between psychiatry and the rest of medicine is now substantially a self-inflicted problem.

Liaison psychiatry has changed its focus over the years. It is now evident that, although consultation is important for conspicuous problems among a very small minority of admissions, our main clinical and research energies should be directed towards the needs of the much greater numbers of patients who present to emergency departments and out-patient clinics and of those with similar clinical problems in primary care. Overall, these psychiatric and behavioural aspects of physical disorder and physical symptoms present a very large public health problem that is as important in the developing world as it is in Western countries.

Identifying key or seminal reading is not easy. The answer is not to seek literature about the practice of liaison psychiatry within in-patient services or in other settings, but to identify a wide range of books and papers concerned with major clinical issues, particularly effective treatment. It is perhaps a last opportunity to rely on traditional reading of books and papers, since the widely dispersed literature relevant to liaison psychiatry will, in future, require considerable electronic searching skills.

Recommendations in this paper bear relatively little resemblance to previous lists. However, I begin with an author whom I mentioned in 1983 and who himself contributed the very first paper with this title (Reference LipowskiLipowski, 1979). Bish Lipowski was an indefatigable reviewer, collator of information and propagandist for consultation—liaison psychiatry for more than 20 years. His early papers are still of interest, as well as describing a historical context that helps our understanding of the present circumstances of liaison psychiatry (Reference LipowskiLipowski, 1985). A historical perspective is valuable. An account of the history of the College's Section of Liaison Psychiatry (Reference LloydLloyd, 2001) shows how, in a short life, it has contributed to the development of theory and practice. Medical history, especially when written by professional social historians, is also informative (and entertaining): see, for example, the books by Shorter (Reference Shorter1992), Hacking (Reference Hacking1998), Cohn (Reference Cohn1970), Thomas (Reference Thomas1973), Stone (Reference Stone1979) and any of Roy Porter's prodigious output. Shepherd's (Reference Shepherd2000) account of psychiatry and war, from First World War shell-shock victims to Gulf War veterans, is an impressively well informed documentation of the many psychological consequences of combat and the ways in which they have been managed by doctors and by the military. It has many implications for present understanding of issues that are still poorly comprehended.

There remains a lack of good standard texts. The very large book edited by Stoudemire et al (Reference Stoudemire, Fogel and Greenberg2000) is excellent in parts and is a useful starting point, as is a textbook from the USA by Rundell & Wise (Reference Rundell and Wise2001), now also available as a weighty abbreviation of a very large book. The College's own publication (Reference Guthrie and CreedGuthrie & Creed, 1996) is much smaller but includes several excellent chapters. Shepherd et al's (Reference Shepherd, Brown and Cooper1966) publication on psychiatric disorder in primary care remains a landmark but should be supplemented by a later book by Goldberg & Huxley (Reference Goldberg and Huxley1980). Significant papers include those by Kroenke (Reference Kroenke and MangelsdorffKroenke & Mangelsdorff, 1989) and von Korff (Reference Von Korff, Dworkin and Le Reschevon Korff et al, 1988). The textbook of Organic Psychiatry by Lishman (Reference Lishman1998), now in its third edition, still justifies the recommendation given in my two previous papers. I would also want to renew my recommendation for Holland's textbook of psychooncology (Reference HollandHolland, 1998), a new edition of a basic work. In addition to the use of an up-to-date textbook, liaison psychiatry requires the reading of editorials and reviews in major medical journals.

Epidemiology is a fundamental. The New Oxford Textbook of Psychiatry (Reference Gelder, López-Ibor and AndreasenGelder et al, 2000) contains a number of excellent chapters on neuropsychiatry, the psychiatry of medical disorders, suicide and attempted suicide and other liaison psychiatry issues. I have found the continuing series of major papers describing findings from the World Health Organization (WHO) primary care study to be highly informative (Reference Gureje, Üstün and SimonGureje et al, 1997; Reference Kisely, Goldberg and SimonKisely et al, 1997; Reference Gureje and SimonGureje & Simon, 1999). The large size of the total sample from 14 countries enabled examination of and association between major medical disorder and various types of psychiatric disorder and also the psychiatric associations of pain and other unexplained symptoms, together with evaluations of categories of somatoform disorder. At last we have a substantial evidence base that is not distorted by cultural solecisms or the use of highly selected populations.

The literature on psychological and psychiatric aspects of major physical illnesses is now very large, with the predominant interest being in cancer (Reference HollandHolland, 1998). Several notable papers have been concerned with the role of psychological factors as contributors to the aetiology of ischaemic heart disease, and well-designed studies provide increasingly convincing evidence that psychosocial factors are determinants of physical course, mortality and psychosocial outcome following myocardial infarction, cancer and other conditions (Reference Frasure-Smith, Lespérance and TalajicFrasure-Smith et al, 1995; Reference Hemingway and MarmotHemingway & Marmot, 1999; Reference Lespérance and Frasure-SmithLespérance & Frasure-Smith, 2000). Among many other topics, there has been good work on trauma, postnatal depression (Reference Cooper and MurrayCooper & Murray, 2000), transplant surgery and ethics.

While the patterns of psychological response to acute and chronic physical illnesses and to major treatments are now relatively clear, the most important theoretical and clinical developments have been in understanding the significance of beliefs and cognitions (Reference Petrie and WeinmannPetrie & Weinmann, 1997). It is unfortunate that health psychology, clinical psychology and even psychiatric literature are widely dispersed and that accessibility is further limited by the variety of theoretical models, each with its own terminology.

Understanding of patients' beliefs and wishes is fundamental to the design of interventions that are medically appropriate, that meet patients' needs and that are acceptable to them. Research on well-designed psychological interventions remains limited (Reference Fawzy, Fawzy and ArndtFawzy et al, 2000) and interest in effective delivery to large numbers of people is minimal. Even so, there are some signs that psychological and behavioural issues are becoming more prominent in treatment guidelines and that liaison psychiatry is now being offered the opportunity for major collaborative treatment trials. Will there be enough of us in a position to respond?

The central importance to liaison psychiatry of what are variously (but unsatisfactorily) referred to as medically unexplained symptoms, somatisation symptoms or functional symptoms has rightly become much more evident. The WHO primary care project has shown the scale of the public health problem; at the same time, it has provided much support for those of us who believe that DSM and ICD somatoform disorders are an unnecessary and inadequate category that has hindered both understanding and treatment. Many groupings of unexplained symptoms have attracted attention, and researchers have continued to concentrate on their own particular interests rather than considering a broader perspective. An impressive amount of evidence relates to the epidemiology, characteristics, aetiology and treatment of these problems. A multi-causal and interactive aetiology is apparent, in which psychological variables are important, whether or not psychiatric disorder can be diagnosed (Reference Mayou, Bass and SharpeMayou et al, 1995). There have been important papers on chronic pain and on the treatment of hypochondriasis (Reference Clark, Salkovskis and HackmanClark et al, 1998). Chronic fatigue has been authoritatively reviewed by Wessely et al (Reference Wessely, Hotopf and Sharpe1997) in a book that combines now old-fashioned scholarship with a broad and critical review of the scientific evidence and of its implications. It also discusses several major themes that are more widely applicable. Halligan et al's (Reference Halligan, Bass and Marshall2001) book on hysteria is valuable, even though it fails to give full weight to the social factors that influence social disorder.

In Britain the management of attempted suicide has been a major task for liaison psychiatry, but this is not the case for consultation—liaison in countries where it is the responsibility of others working in emergency departments. However, it is a major topic for many and the current state of knowledge is both comprehensively and critically reviewed by Hawton & van Heeringen (Reference Hawton and van Heeringen2000).

We must not forget the need to be aware of older publications which, although flawed or mistaken, continue to have wide influence. Examples are Miller on accident neurosis (Reference MillerMiller, 1961), Slater on hysteria (Reference SlaterSlater, 1965), Guze on Briquet's syndrome (Reference Guze, Cloninger and MartinGuze et al, 1986), Alexander on psychosomatic aetiology (Reference AlexanderAlexander, 1950) and almost all early cross-cultural writings on the expression of psychological distress as physical symptoms. I would also put a major health warning on the somatoform sections of DSM—IV and ICD—10 and emphasise their authors' often neglected reservations about the lack of evidence and the speculative nature of the classifications.

Working as a liaison psychiatrist with the whole range of psychiatric disorder in those with every form of medical disorder and treatment demands wide reading and good electronic searching skills in both psychiatry and medicine. We still lack really good introductory texts for trainees, but we do now have the opportunity to choose from a substantial evidence base. Although the next paper with this title can be expected to regard turning pages as outdated, good liaison psychiatrists will still need to read widely in psychiatry, medicine and other disciplines.