: We recently demonstrated that the major threat to research integrity in clinical trials may be due to a violation of the equipoise or """"""""the uncertainty principle,"""""""" the fundamental principle on which nearly the entire system of human experimentation stands. This principle states that the patient should be enrolled in a randomized controlled trial (RCT) only if there is substantial uncertainty about which of the trial treatments would benefit a patient most. We hypothesize that there is a relationship between equipoise and trials outcomes. If the investigators do not know in advance what they are going to discover, and if the uncertainty principle is observed and the literature on experimental therapies is fairly complete, we would expect, over time, to find no significant difference between the proportion of published results that favor experimental treatments and those that favor comparison treatments. We have performed an earlier investigation on this issue and found that this expected distribution of outcomes was observed among those published trials that were funded by public resources, but that the uncertainty principle appeared to be violated among those published trials funded by pharmaceutical companies. However, we could not exclude publication bias as the explanation for the higher proportion of positive results in the literature. Based on other studies, failure to publish could be as high as 51%. Furthermore, we could not contrast our data with expected distributions of outcomes in Clinical trials, since this has not been determined. Therefore, the real relationship between the uncertainty principle and outcomes of RCTs remains unsettled. To elucidate this relationship, we propose to study a comprehensive population of initiated RCTs from a unique funder (using an inventory of the NCI-sponsored trials). Our hypothesis will be addressed through the following specific aims: 1. All National Cancer Institute (NCI)-sponsored RCTs trials funded in the years 1975 through 2001 will be identified using the NCI registry of clinical trials in cancer. 2. e will identify the primary and additional outcomes selected for study by the investigators. 3. We will review the trials (published and unpublished) to classify their primary and other outcomes, specifically whether the innovative or comparison treatment was preferred. 4. We will perform analyses for evidence of investigators' equipoise, and for the relationship of outcome to study quality, type of comparison treatment, investigator characteristics, and funding source Since violation of the principle of equipoise and publication bias represents two of the most serious threats to the integrity of the research process in RCTs, it is of long-term significance to understand the extent to which these factors are evident in the study of RCTs. By understanding these relationships, we will be in a position to contribute to the preservation of a system of high quality clinical trials in medicine. This proposal will answer both the question """"""""what do trials do for us?"""""""" and assess the reliability of the research in which public has invested so much.
Showing the most recent 10 out of 16 publications
