The relationship of study findings to publication outcome in anaesthesia research post implementation of mandatory trial registration

This project is part of a Doctor of Medical Science research higher degree. There are three main parts to this project, of which this is part two. Part Two will involve a further systematic review of all randomised controlled trial abstracts presented at the ASA Annual Meetings 2010-2013. This time period has been chosen as it is five years after the implementation of mandatory trial registration in 2004. From these abstracts, the odds ratio for abstracts with positive results proceeding to journal publication will be calculated. The primary endpoint will be a comparison between the odds ratios from Part one and from Part two. We define a 33% decrease in odds ratio as clinically significant. This corresponds to a new odds ratio of 1.33.Purpose: Publication bias of positive studies has been demonstrated in many areas of medicine. In order to estimate publication bias in the anaesthesia literature and the effect of mandatory trial registration, we will pose the following research questions: Part One: What amount of publication bias is present in the anaesthesia literature, as seen in the abstracts presented at the 2001-2004 American Society of Anesthesiologists (ASA) annual meetings? Part Two: Has the introduction of compulsory trial registration in 2004 resulted in a decrease in publication bias, as seen in the abstracts presented at the 2010-2013 ASA meetings? Part Three: What is the proportion of prospective trial registration in abstracts presented at the 2010-2013 ASA annual meetings? What is the proportion of major discrepancies between abstract trial registration entries and the corresponding published papers? Hypotheses: That there is significant publication bias present in the anaesthesia literature. That the introduction of mandatory trial registration has not resulted in a decrease in publication bias in the anaesthesia literature. That trials that are not prospectively registered (but have a reported outcome) are more likely to have a positive outcome. Methods: All abstracts from 2001-2004 and 2010-2013 ASA annual meetings performed as randomized-controlled trials in humans will be reviewed. Their outcome results will be scored as positive or negative and the abstract quality will be assessed using a 13-point scoring system. A systematic literature search to identify any subsequent publication of the studies will be performed and the odds ratio for journal publication calculated by comparing positive vs negative studies. The odds ratio from the 2010-2013 abstracts period will be compared to the 2001-2004 period. We define a 33% decrease in the odds ratio as clinically significant. A systematic search for trial registration of the 2010-2013 abstracts will also be conducted, and any major discrepancies between registry entries and published papers recorded. We will perform a comparison between the proportion of positive outcome trials which did not have prospective registration versus the proportion of positive outcome trials with prospective registration. We define a 20% reduction in positive outcomes as clinically significant when comparing trials without prospective registration to trials with prospective registration