Problems in publication of clinical trial methodology

Abstract

The literature on analgesic testing in man reveals a series of informational gaps. These include a failure to document the occurrence of "trials that failed", the problems associated with recruiting patients and in dealing with dropouts, the relative advantages and disadvantages of different methods of assessing relief from pain, the importance of baseline variables, and the utility of global and comparative judgments by patients. A trial is described in which only 100 subjects of a total of over 8,000 patients theoretically available for study proved suitable. Significant differences between consenters and nonconsenters and selection factors that are used in choosing an experimental population have implications for the generalization of a study. Such conclusions have been largely ignored both in and out of the scientific literature.