Clinical dilemmas and the Cochrane Collaboration

Controlled trials, and specifically randomised controlled trials (RCTs) are the most powerful research design to evaluate effects of mental health care (World Health Organization Scientific Group on Treatment of Psychiatric Disorders, 1991). There are, however, too many RCTs published in too many journals for anyone to keep up-to-date (Sackett & Rosenberg, 1995). In order to decrease the potential for bias or the play of random error (Chalmers, 1989). it is desirable to produce an overview of research findings. Frequently, those interested in the effectiveness of care depend on reviews in journals, textbooks or guidelines to direct practice. Generally speaking there are two sorts of reviews, the systematic and the traditional/ subjective.


Randomised controlled trials and reviews
Controlled trials, and specifically randomised controlled trials (RCTs) are the most powerful research design to evaluate effects of mental health care (World Health Organization Scientific Group on Treatment of Psychiatric Disorders, 1991). There are, however, too many RCTs published in too many journals for anyone to keep up-to-date (Sackett & Rosenberg, 1995). In order to decrease the potential for bias or the play of random error (Chalmers, 1989). it is desirable to produce an overview of research findings. Frequently, those interested in the effectiveness of care depend on reviews in journals, textbooks or guidelines to direct practice. Generally speaking there are two sorts of reviews, the systematic and the traditional/ subjective.

Systematic v. traditional reviews
Systematic and traditional reviews are very different. The former will have a methods section, the latter may not. In a systematic review the means by which data are identified, selected and, if appropriate, assimilated is made explicit (Sackett et al 1991). These methods are open to scrutiny and valid criticism.
The recommendations of systematically con ducted reviews and traditional reviews may be quite contradictory. For example, for the man agement of those with myocardial infarction, Antman et al (1992) compared the recommenda tions of leading textbooks and journals to the results of what systematically conducted reviews would have said using the RCT data of the day. They found that leading traditional reviews, by omission, recommended interventions that were harmful or lethal up to 10 years after generally acceptable proof to the contrary was available. Similar examples are just beginning to emerge from within mental health. Up to now, some reviewers were recommending the use of vitamin E to treat neuroleptic-induced tardive dyskinesia (Lloyd, 1992;Jeste & Caligiuri, 1993). A recently completed systematic review of the best available evidence suggests that vitamin E could be any thing from moderately helpful to very harmful (Soares & McGrath, 1997). It certainly is an intervention worthy of full evaluation but, cur rently, there is little evidence to recommend its

The Cochrane
Collaboration The Cochrane Collaboration was launched in 1993 (Chalmers et al, 1992) with a view to the production, maintenance and dissemination of systematic reviews of health care. It consists of a global network of people methodically seeking every published or unpublished, complete or incomplete, controlled trial of health care. Groups of people with similar interests are forming to systematically review these studies.
These groups are open to anyone wishing to invest effort.
At present, within the Cochrane Collaboration, there are five groups with a specific interest in mental health. The first, the Cochrane Schizo phrenia Group, has been working for four years. It is composed of clinicians, researchers, occupa tional therapists, economists, nurses and con sumers of care, widely dispersed across the world. The Cochrane Depression, Anxiety and Neurosis Group is focusing on affective and eating dis orders, somatisation problems and deliberate selfharm. The Cochrane Dementia and Cognitive Impairment Group is focusing on the care of those with any type of illness that primarily effects cognitive functioning. These groups have a regis ter of relevant clinical trials. The Cochrane Addiction Group and Cochrane Developmental, Psychosocial and Learning Problems Group are starting to build a register of trials and reviews. For example, the Cochrane Schizophrenia Group has undertaken a comprehensive and methodical search strategy to build its register and make it available to anyone interested in doing review within its scope. The Group is already producing, and updating reviews within the electronic output of the Collaboration, the Cochrane Library. Currently there are only 27 systematic reviews in the Cochrane Library directly related to people with severe mental illness. These reviews deal with not only pharmacological treatments (beta-blockers, clozapine, fiuphenazine risperidone and zuclopentixol for schizophrenia; anticholinergics, benzodiazepines. calcium channel blockers, cholinergics, gamma-aminobutyric acid agonists, vitamin E and miscellaneous treatment for neuro-lepÃ¼c-induced tardive dyskinesia; antipsychotics for learning disability), but also with other forms of interventions such as case management, community mental health team management, family intervention, electroconvulsive therapy, intercessory prayer and long versus short hospi talisation for those with schizophrenia. In addi tion, nine protocols for schizophrenia and 13 for depression are available in this version of the Cochrane Library (Issue 1, 1998).

Cochrane Library
The Cochrane Library is an inexpensive electro nic database, currently published every three months and distributed by Update Software (see Appendix). It contains several databases. The Cochrane Controlled Trials Register holds refer ences to approximately 180 000 randomised or quasi-randomised trials identified by the mem bers of the Collaboration. The Database of Abstracts of Reviews of Effectiveness is a register of already published systematic reviews that have been identified by methodical searches of journals (currently with 1852 references). The Cochrane Database of Systematic Reviews is the flagship of the Library, it contains all reviews undertaken and maintained by those within the Cochrane Collaboration. The Cochrane Database of Systematic Reviews has been supplied to all National Health Service libraries in the UK and abstracts of reviews are also available on the Internet (http://cochrane.co.uk/info/abstracts/ abidx.htm). Mental health professionals and students can find this database and search for information regarding specific interventions for clinical problems. The Cochrane Database of Systematic Reviews is a new publication and is filling up with regularly maintained reviews and already is a powerful teaching tool for those interested in the evaluation of care.

Back to the dilemma
So, currently, when you try and help people with mental illnesses or problems what would you like to guide your practice? The application of up-todate evidence, along with intuition, common sense and wisdom, is increasingly desirable, and now possible.
The Cochrane Collaboration has been likened, for better or worse, to the Human Genome Project (Naylor, 1995). If the current expansion and effort of this organisation continues, the Cochrane Database of Systematic Reviews will soon contain hundreds of reviews relevant to all aspects of health care. It will clarify what is known, and what is not known, solve some dilemmas and make others more acute.