Brief Observations on Research Methods in Clinical Trials

Abstract

This paper will address some of the current issues concerning the conduct of phase III clinical trials. The issues to be discussed fall into two areas – trial design and mensuration of target variable change. The former includes the determination of sample size; selection bias in sample populations due to particular inclusion and exclusion criteria; ensuring trials are randomised, double blind controlled studies; and designing the trial so that the two arms are balanced in every way possible, with especial reference to matching treatment variables. The mensuration of target variable change includes the selection of appropriate rating instruments which are (i) sensitive to target symptoms, (ii) have been adequately profiled psychometrically in a similar population, and (iii) are able to detect change within the time frame of the study. Instruments should be cross-culturally robust if used in multi-country studies and should have validated local versions where possible. Adequate training in the use of the rating instruments should be clearly reported and a clear description of inter-rater reliability assessment and maintenance should be provided. Brief mention will be made of the Cochrane collaboration and the role of evidence-based medicine in clinical trials.