The changing face of clinical trials.

We have now seen 20 years of clinical trials testing the efficacy and safety of antihypertensive drugs, involving more than 40,000 patients internationally. Much has been learned and should be applied to future studies. First, patient selection: will selection of patients from different races and nationalities permit pooling of data? There should be a sufficient number of female patients and elderly patients above 80 years of age to draw unequivocal conclusions about the effectiveness of antihypertensive treatment in that population. Further, will the patient population reflect the population cared for by individual community practitioners? End-points must be carefully selected, and it will be important to prove not only that a specific drug can lower blood pressure, but also how far blood pressure should be lowered. Lowering of blood pressure per se is not enough, and end-points should involve all-cause mortality and specific mortalities. Data analysis may revolve around intention to treat analysis or on-treatment analysis. Regarding side effects of the study drug, consideration should be given to survival statistics. Finally, decision analysis is needed. After the trial, postmarketing surveillance is important, and patient selection is just as important in that instance as it was in the original trial. Once the trial has been completed, careful meta-analysis is needed to answer the question still unanswered: can small multiple-centre trials replace large, international trials? One should not consider a trial completed unless measures have been suggested or taken to inform the practising clinician of the inferential statistics applied to the data and the meaning of the data analysis for the clinician in managing the individual patient.