The efficacy of antihypertensive drugs requires evaluation in clinical trials, the design of which must minimize variability (observer or patient) and order effects, eliminate bias, and include sufficient numbers of patients (i.e., have sufficient power) to allow real differences in blood pressure between patient groups to be detected at conventional levels of statistical significance. The crossover design of trial requires the successive examination of different treatments. This may result in a lengthy trial for the participant and an increase in the dropout rate. Therefore the parallel group design is usually performed for a comparison of several active drugs or drugs and placebo. This design avoids the potential problems of order and carryover effects, and when of a long duration, takes into account short-term and longer term responses to drugs. Crossover trials with limited periods of treatment for each phase can be carried out with fewer numbers of subjects and are more economical of patient resources. However, care must be taken to correct for order effects by appropriate design and to exclude carryover effects. Usually such a design only allows the investigator to comment on short-term blood pressure responses. The extrapolation of such information for longer term use of the drugs in clinical practice requires much caution.