ABSTRACT
Nearly six decades have elapsed since the MRC Tuberculosis Research Unit conducted the first randomized controlled clinical trial [1-3]. It is widely held [4,5] that the randomization and control design aspects of the trial were the brainchildren of Sir Austin Bradford Hill, Director of the MRC Statistical Research Unit. The 1962 Kefauver-Harris (K-F) Amendments [6] to the Federal Food, Drug and Cosmetics Act of 1938 represented a watershed event in the evolution of evidence to support drug claims. This landmark legislation required that all drugs thereafter be proven effective prior to being approved by the Food and Drug Administration (FDA) for marketing in the United States. The authors and others have often referred to the K-F Amendments as the full employment act of biostatisticians in the pharmaceutical industry. Much progress has been made in strengthening evidence to support claims
deriving from clinical trials since the first randomized control clinical trial (RCCT). The FDA has been a major player in advancing the need for better quality clinical trials, as well as evolving evidentiary methodological standards to accomplish (see comments by Temple and O’Neill [6]). The double blind (DB) RCCT is now considered the gold standard for evidentiary medicine. Strengthening the evidence to support claims deriving from clinical trials is
the result of, first, recognizing the need for improvement and, second, the collective desire to improve quality in all aspects of such investigations [7-9]. Improving the experimental design of the investigation is one aspect, and this includes ensuring an adequate number of participants [10-13]. Improving the quality of reporting the investigation [14-18] is another aspect. Recognizing that clinical drug trials don’t necessarily mimic clinical prac-
tice has spawned the relatively new area of translational research [19]. ‘‘Translational medicine is a branch of medical research that attempts to more directly connect basic research to patient care. Translational medicine is growing in importance in the healthcare industry, and it is a term whose precise definition is in flux. In the case of drug discovery and development, translational medicine typically refers to the ‘translation’ of basic research
into real therapies for real patients. The emphasis is on the linkage between the laboratory and the patient’s bedside, without a real disconnect. This is often called the ‘bench to bedside’ definition.’’ [20]. All of these settings require the design of an investigational protocol,
conducting the investigation per the protocol and collecting the data, statistically analyzing the data, making valid inferential conclusions relative to the objectives of the protocol, and reporting the results. When considering the validity of statistical inferences from clinical trials, many will restrict attention to whether the statistical methodology used to analyze the data is appropriate for the type of data and whether the assumptions underlying the methodology hold for the data. This is necessary for an inference from a statistical analysis to be valid, but it is not sufficient. Valid inferences derive from well-planned, well-conducted, and properly analyzed investigations. All aspects of an investigation, whether in the clinical drug development
area, in the medical university research area, in the public health intervention area, or in the basic laboratory research area, should be documented to permit an audit of ‘‘what was to be done,’’ ‘‘what was done,’’ and how differences might affect conclusions or inferences. Planning activities culminate in a protocol [7,11] for the investigation. The protocol starts with a welldefined question or objective that requires an investigation to answer [9]. The data or endpoints needed to provide an answer are identified. The question is then formulated within a hypothesis testing framework. The number of subjects required to address the question is determined. Procedures for conducting the experimental investigation that produces the required data are developed. Methods for collecting, computerizing, and quality assuring the data are specified. Statistical methods for analyzing the data addressing the question are described.
