New approaches to trials of antirheumatic drugs in rheumatoid arthritis.

Abstract

While drug trials can be performed in different designs and sequences, the paramount consideration is careful definition of limited trial objectives such as efficacy, tolerance, and safety. Efficacy variables might include tests for analgesic effect, antiinflammatory activity, or specific antirheumatic action. Unless a drug has remarkable therapeutic effect, double-blind comparison with placebo or a standard reference drug such as aspirin is mandatory. Random assignment of patients to treatment groups must be assured and patients should not receive other antirheumatic drugs unless the trial is specifically designed to study drug interactions. Drug dosage poses a problem because average doses are usually employed. Compliance is monitored by several techniques including measurement of blood levels of test drugs in all study patients. Multicenter trials have several advantages: (1) averaging patient selection, (2) minimizing observer bias, (3) shortening time of trial by rapid entry of patients, (4) entering large number of patients, which can permit examination of "responders" versus "non-responders." Data analysis should employ proper statistical evaluations (nonparametric tests for nonnormally distributed variables). Duration of study period, trials to evaluate drug interactions, and variables used to measure drug effects will be discussed.