Evidence-based medicine is fine in principle, but needs to demonstrate tangible benefits to clinicians and their patients in practice. The main obstacles to its implementation are: (a) identifying information needs; (b) delivering reliable and clinically useful information to the ward/clinic; and (c) ensuring the information is regularly updated. How can this be achieved in psychiatry?

IDENTIFYING INFORMATION NEEDS

The first step is to establish what the information needs for psychiatrists are. To this end, we recently surveyed all senior psychiatrists in south-east Scotland and 93 (76%) replied (Reference Lawrie, Scott and SharpeLawrie et al, 2000). Respondents identified over 150 clinical questions they would like answered, mainly about treatment. They reported that the greatest barrier to practising evidence-based psychiatry was time — particularly the time required to search and appraise the literature. The five most frequently asked clinical questions were identified. These concerned three main topics: (a) the use of mood-stabilising drugs in bipolar disorder; (b) the place of new generation antipsychotic drugs in the management of schizophrenia; and (c) evidence-based guidelines for the management of depressive illness that had not responded to antidepressant drug treatment. As an exercise, we set about finding answers to these questions. Systematic reviews were not available for three of the five most asked questions and for only one question was a published meta-analysis available. Consequently, we had to rely on our skills of critical appraisal to assess a poorly executed meta-analysis and numerous randomised controlled trials and other studies. It took three senior psychiatrists working together, all of whom were experienced researchers, trained in the skills of critical appraisal, and with excellent facilities, 15-60 minutes to answer each question. We sent our answers to the relevant psychiatrists who had asked these questions. They reported the information to be of value, and suggested that such a ‘question answering’ service would be useful.

A previous commentary included the plea that research in psychiatry should be firmly rooted in the everyday needs of practitioners (Reference LewisLewis, 1997). Four years later, we are not sure that the profession is any nearer to achieving this laudable aim. Apart from our own survey, we are not aware of any systematic attempts to assess the information needs of today's practitioners. It is not, therefore, known to what extent the needs of senior psychiatrists in south-east Scotland are representative of the needs throughout the rest of the country, but if they are, it is disconcerting that none of their most frequently asked questions appear among the mental health topics prioritised by the National Health Service's (NHS) Health Technology Assessment Programme (Reference Stein and MilneStein & Milne, 1999). Individual users of such research have been invited to comment on priorities for research via a website or through their local NHS Research and Development Directorate. If the senior psychiatrists that we surveyed are typical, not enough individual users of the research have provided feedback to influence the priorities.

DELIVERING RELIABLE AND CLINICALLY USEFUL INFORMATION TO THE USER

Another of the pleas made 4 years ago was for the technology to bring the best available evidence to the practitioner, including a mechanism to continually revise research evidence of effectiveness (Reference Sheldon and GilbodySheldon & Gilbody, 1997). This is still only an aspiration. A survey of senior psychiatrists in the west of Scotland in 1998 found that only 27% of consultant and 10% of senior trainees had access to an office-based electronic database such as Medline, albeit the vast majority had access to a library with an electronic database (Reference Carey and HallCarey & Hall, 1999). The results of this survey were similar to one carried out among general practitioners in the Wessex region in 1997 (Reference McColl, Smith and WhiteMcColl et al, 1998). Access to electronic databases may have improved, but these surveys also illustrated the lack of consensus among doctors about how best to implement evidence-based medicine. Carey & Hall (Reference Carey and Hall1999) recommended that individual practitioners increase their skills for data searching and critical appraisal; in contrast, McColl et al (Reference McColl, Smith and White1998) and Guyatt et al (Reference Guyatt, Meade and Jaeschke2000) have criticised the idea of expecting all doctors to be proficient in these skills, but recommended the development of relevant easy-to-access summaries of evidence. Our survey found that psychiatrists preferred easy access to up-to-date summaries.

Reliable evidence is, of course, not necessarily used (Reference Guyatt, Meade and JaeschkeGuyatt et al, 2000). We think this is because most of it is unusable in the ward or clinic. The National Institute for Clinical Excellence will need years to cover anything like the full array of therapeutic options for a given patient. The Cochrane Library is an ever-expanding source of reliable reviews but each is lengthy and demands careful reading. The journals Evidence-Based Mental Health and Best Evidence summarise scientifically valid and clinically relevant articles for clinicians, but coverage is necessarily dictated by what is published. Even Clinical Evidence, which does summarise all the current evidence for specific clinical questions and is updated every 6 months, has as yet rather patchy coverage. A variety of websites are accessible but usually provide too little information, too slowly. We are not aware of any source that synthesises all the available evidence into brief ‘pros and cons’ of sensible interventions in particular clinical scenarios. These problems are typified by the current strategy document for NHS research and development (Department of Health, 2000), which mentions the importance of ‘support(ing) evidence-based policy and practice by improving access to findings [sic] and evidence' ’ but does not suggest any mechanisms for this.

ENSURING THAT INFORMATION IS REGULARLY UPDATED

If there is no consensus about how to deliver clinically useful information to the practitioner, then there can be no satisfactory solution about how best to update the information. We believe it is helpful to conceptualise the solution as partly technological because of the need to update information flexibly as new research findings emerge. This requirement can be achieved through electronic publication or an electronic information retrieval system. This could be configured to record commonly asked questions that are not answered to the clinician's satisfaction, thus establishing priorities for clinically relevant research, and to make patient- and clinician-specific prompts to encourage use of reliable evidence.

CONCLUSIONS

A few conclusions seem obvious to us. There is no robust system to identify the information needs of today's psychiatrists in the UK. This may be one reason that the gap between research and practice in psychiatry is not closing. It is simply not feasible to expect individual practitioners to be proficient in the skills of critical appraisal and apply these to each of the questions they will face in their daily practice. Easy access to summaries of evidence seems essential. Electronic information retrieval systems are fast, accessible and can easily be updated as required. In a relatively affluent country like the UK, they are the ideal technological solution to deliver useful information to the ward or clinic.

There are, however, many unresolved issues, and we hope this editorial will stimulate debate. Who will be responsible for the implementation of an evidence-based psychiatry? In particular, whose responsibility ought it to be to survey the information needs of practitioners and to ensure that this informs research priorities in this country? Is it the sole responsibility of the individual practitioner as a member of the medical profession to ensure that his or her practice is evidence-based? Can the practitioner expect support from his or her employer at local level or through central government? Who are the people who will be trained in the skills of critical appraisal, appraise the research literature, and make their findings available to practitioners? How can local and national initiatives be harmonised? Do professional bodies such as the Royal College of Psychiatrists have a role, either directly or through journals such as Psychiatric Bulletin or the British Journal of Psychiatry? Is private enterprise the answer? Above all, we need to know what the practitioners think.