Letters

Clinical trials

Simple megatrials are not sufficient

Randomised database studies could serve as new strategy

Simple megatrials are not sufficient

EDITORNo one should argue with Peto and Baigent about the dangers of basing clinical practice on flawed "outcomes research" or isolated, undersized, randomised trials,¹ but the simple megatrial may not be the ideal model for the next 50 years. Although such trials have taught us not to expect a predictable physiological response from our treatments but merely an improvement in the odds of successful outcome, the homogenising effect of large numbers may still disguise important variations in response. Attempts to explore these variations through subgroup analysis are widely condemned because of the perceived association with retrospective data dredging, so that the theoretical knowledge of clinicians and the individuality of patients are seen as increasingly less relevant to decisions about treatment.

The trials that Peto and Baigent advocate focus on hard end points, such as deaths, which may be rare in some patient groups and of limited relevance in others. The ageing of the world's population may be the greatest global challenge to be faced in the next 50 years, and we will become less concerned with the length of survival in old age than with its quality. Many treatments will be evaluated in terms of their effects on "healthy active life expectancy" or on the duration and severity of "terminal dependency" rather than on mere survival. Comparatively simple measures of dependency are available for this purpose, but trial follow up will mean more than just a body count.

Older people are physiologically diverse, and we will need to look for qualitative as well as quantitative differences in treatment response in different subgroups. Trials will need to estimate not just the average effect of treatments but the benefits and risks in all important subgroups, so, despite high event rates, even large trials will be unlikely to answer all the relevant questions about a particular treatment.

Over the next 50 years the monolithic megatrial should therefore be replaced by planned collaborations between smaller studies, addressing different aspects of the same broad research question and using an agreed system for classifying patient subgroups, interventions and outcomes. ^{2 3} This would overcome the limitations of retrospective meta-analysis, and clinicians would participate to a greater extent in defining the research questions and agreeing on classifications. The risks and benefits of treatment would become clear, and the focus of uncertainty would gradually change. Clinicians could then use their clinical skills, as well as the latest accumulated research evidence, to select the best treatment for individual patients.

David Barer, Professor of stroke medicine. 
Queen Elizabeth Hospital, Gateshead, Tyne and Wear NE9 6SX d.h.barer{at}ncl.ac.uk

1. Peto R, Baigent C. Trials: the next 50 years. Large scale randomised evidence of moderate benefits. BMJ 1998; 317: 1170-1171[Free Full Text]. (31 October.)
2. Barer D, Ellul J. From meta-analysis to epi-analysis: the European Stroke Database Project. In: Fracchia G, Haavisto K, eds. European medicines research, perspectives in clinical trials. Cambridge: European Conference Publications, 1996:157-164.
3. Gladman J, Barer D, Langhorne P. Specialist rehabilitation after stroke. BMJ 1996; 312: 1623-1624[Free Full Text].

Randomised database studies could serve as new strategy

EDITORThe issue of the BMJ on 50 years of randomised controlled trials¹ clarifies future challenges to clinical trialsfor example, the need to find ways of increasing the size of randomised studies,² the increasing demand for high quality research in primary care, and the need to translate evidence into practice. The main limitation of randomised clinical trials is the fact that while they do show whether new drugs work (efficacy) they do not show if they really work in clinical practice (effectiveness).

Although the need to conduct "naturalistic" studies was postulated 20 years ago,³ an adequate methodology to conduct them has not been developed yet. Pragmatic clinical trials and database analyses have been the two main methods proposed to assess the effectiveness of treatments. Their inherent limitationsproblems of external and internal validity, respectivelyhave, however, not been completely resolved. For Peto and Baigent, the key question is how a really large number of patients can be randomised in practice.² We believe that the main challenge is in the development of a new strategy capable of maintaining the main advantages of clinical trials (randomisation) and database analyses (capturing the full range of treatments and outcomes that occur in the normal course of medical practice). Our group has recently suggested including randomisation modules in computer based patient records, if possible.⁴ This would help to conduct "randomised database studies" that can simultaneously use experimental as well as observational methods to assess the effectiveness of drugs.

Randomised database studies could be used to conduct large, long term studies and investigate outcomes of importance to clinicians and patients. We agree with Peto and Baigent that simplicity and flexibility should be two of the main aspects influencing the design of these studies. The progressive implementation of clinical practice guidelines and new computer support systems for prescribing might help to identify the best diagnostic and therapeutic options in every clinical situation. The application of randomisation using computer based clinical reports (the integration of research into daily clinical practice) could contribute to studying the problems where they come up. This would help to improve the quality of health care and enable quicker acceptance and incorporation of research results into clinical practice.

José Antonio Sacristán, Clinical pharmacologist. 
Inès Galende, Clinical pharmacologist. 
Spanish Group for the Study of Methodology in Clinical research, E-28230 Madrid, Spain

1. BMJ , 1998:1167-1172; 1177-1212; 1217-48. (31 October.)
2. Peto R, Baigent C. Trials: the next 50 years. BMJ 1998; 317: 1170-1171. (31 October.)
3. Lasagna L. A plea for the "naturalistic" study of medicines. Eur J Clin Pharmacol 1974; 7: 153-154[Medline].
4. Sacristán JA, Soto J, Galende I, Hylan TR. Randomised database studies: a new method to assess drugs' effectiveness? J Clin Epidemiol 1998; 51: 713-715[Medline].