Reporting standards and research strategies for controlled trials

Abstract

From a study in Part I of many papers using controlled clinical trials in cancer research, we summarize quantitatively the frequency of reporting important statistical and procedural matters: randomization, statistical method, blindness, power, sample size, survival, and informed consent. We recommend that editors set standards for such reporting to improve readers' ability to evaluate trials. New research is needed to provide strategies for planning and analyzing collections of trials. We suggest in Part II the areas of generalizability, design, including size and allocation of sample and of volume of material collected, methods of reporting and analyzing imperfections in execution, and new archival work to take advantage, through reanalysis of trials and through analyzing collections of trials, of the substantial bodies of data already available.