Complementary Therapiesin Medicine (1996)4, 254-257 © PearsonProfessional Ltd 1996 FAMILY MEDICINE Evaluating complementary therapy on the National Health Service: a critique of reports from three pilot projects R.W. Rees Research Council for Complementary Medicine, London, UK SUMMARY. Complementary medicine has always been available in a limited form to National Health Service patients. Referral to several NHS homoeopathic hospitals has been possible since the formation of the NHS. Certain forms of treatment are also available from doctors, nurses and physiotherapists who use complementary therapy techniques as part of their everyday practice. NHS referral to specialist complementary therapy units and to so-called 'independent complementary therapists' has recently been piloted on a trial basis by a number of health authorities. This article is a review of the reports produced by three of these programmes: the Liverpool Centre for Health, the Complementary Therapy Centre of Lewisham Hospital NHS Trust and West Yorkshire Health Authority's introduction of independent complementary therapists in Huddersfield and Dewsbury. It will introduce the issues surrounding quality in health programme evaluation methods and then present a brief overview of the projects, their evaluation methods and recommendations. The projects vary considerably in their methodological quality. It is impossible here to do justice to the considerable amount of work and goodwill that has gone into establishing and assessing these innovative schemes. For further details and a fuller understanding of the complexities involved, readers are directed to the original reports. 1-3 WHAT SHOULD WE EXPECT FROM AN EVALUATION REPORT? The objectives of evaluation research. In health service evaluation research, the administration and effects of health care programmes are studied systematically to investigate their impact in the 'real world'. For exam- ple, one such study might look at how satisfied patients in a physiotherapy unit are with various aspects of their treatment, another might look at the costs and benefits of a new well-woman clinic within a general practice. Their main purpose is to inform pol- icy makers and others involved in these programmes. There is no single 'correct' way to do evaluation research. Given that so little is known about what hap- pens when independent complementary practitioners are introduced into the NHS, any well-designed project is likely to produce useful information for future research. But what constitutes a well designed evaluation project? A limited number of points will be mentioned here: a more detailed discussion of evaluation methodology can be found in several standard texts.4,s Rebecca W. Rees BA, MSc, RCCM Audit and Outcomes, Research Council for Complementary Medicine, 60 Great Ormond Street, London WC1N 3JF, UK. To be useful an evaluation needs to be critical and con- struetive. Some 'evaluation' studies are merely descriptions, for example, of how many patients are seen, by whom and why. Such accounts are necessary if a service is to be developed. Good-quality evalua- tion also involves an appraisal of whether a project's objectives are being met. An evaluation report should include recommendations for action based upon the report's findings. The evaluation itself needs to have clearly stated objec- tives. In an evaluation of therapy provision, finding out what happens to the patients who use the thera- pies is often the highest priority: many evaluations, as a result, are effectively uncontrolled clinical trials. Other questions of interest include: how much does the project cost in comparison with other types of health care? Is it reaching its target population? Are any aspects of the programme not running as well as they could? How satisfied are patients and health pro- fessionals with the day to day running of the project? has the programme been fully implemented? Different questions will be appropriate at different stages of project development. Trying to answer too many questions at one time will render an evaluation unmanageable. Evaluation methods. Like any form of research, the methods used in an evaluation need to be valid, simple 254