Basic design

The following features should be considered when assessing the basic design of drugs trials:

• choice of subjects - define criteria for entry, exclusion criteria. Age, sex, severity, other illness other drugs, etc
• controls - no treatment, placebo, other drug prescribed. Defined criteria for entry, comparable groups which are matched, and proper randomisation
• cross over trials - beware washing out period and nature of the illness
• historical controls can only rarely be used - e.g. paracetamol and N-acetyl cysteine.
• randomisation - eliminate bias; once allocated include in analysis even if the patient drops out - avoids bias of choice, i.e. intention to treat. Need external coordinator. Need a good mechanism for blind allocation
• blindness - single - patient, double - patient and doctor. Difficulty - drug may be a giveaway, and may unblind the study
• placebo should be ordinarily indistinguishable from the drug
• number of centres - e.g. ISIS - huge numbers
• prospective or retrospective