Research Wastage: From Trial Registration to Translation

Abstract

Background: Each year globally, there are millions of academic research papers being produced; this represents billions of dollars of research investment. It is accepted that many of these studies are either not published, not published in a timely manner or if they are published, they are not included into research syntheses. Researchers conducting clinical studies have an obligation to patients, health care providers and funders to publish their results. The failure to do so potentially could introduce bias and produce misleading results affecting the accuracy of research synthesis, and this constitutes research wastage. The effect and extent of this research waste is the focus of this thesis. Aim: This thesis aims to investigate research waste in terms of publication bias of clinical studies and then to assess the rates of published trial dissemination into systematic reviews. Methods: This thesis investigated five projects to examine research waste in each of the following areas: • A large international Cochrane methodology review (CMR) which assessed the extent of non-publication and publication bias in interventional clinical trials; • A New Zealand focused cross-sectional study that examined the publication rate and publication bias of randomised controlled trials of registered in the Australian New Zealand Clinical Trials Registry (ANZCTR); • A cross-sectional study that assessed the publication rate and publication bias of studies funded by the Health Research Council of New Zealand (HRC); • A cross-sectional project that examined the dissemination of published gynaecology trials into systematic reviews; • and another cross-sectional project that assessed the dissemination of published fertility trials into systematic reviews. The detailed methods for each of these studies can be found in chapters two through to six. Results: Cochrane methodology review: This review included 204 studies with a total of 165,135 trials Only just over half of these trials (53%) were published. The CMR reported the existence of publication bias where trials with positive results (OR 2.69 95% CI 2.02- 3.60), large sample sizes (OR 1.92, 95% CI 1.33 to 2.77), were multi-centre (adjusted OR 1.20, 95% CI 1.03 to 1.40) or had non-industry funding (adjusted OR 2.13, 95% CI 1.82 to 2.49) were more likely to be published than trials with negative/null results, small sample sizes, single centre or industry funded. Trials with favourable results (adjusted HR 1.92, 95% CI 1.51 to 2.45), were large (adjusted HR 1.41, 95% CI 1.18 to 1.68), or funded by non-industry (adjusted HR 1.46, 95% CI 1.15 to 1.86) were also published in shorter time. Cross-sectional study of the ANZCTR: This study included 1,970 trial registrations from the years 2007, 2009 and 2011. Five hundred and forty-one (27%) of these trials remained unpublished 10 to 14 years later. The average time to publish was 4.36 years and the prospective trial registration rate was 48% (952 trials). Trials funded by non-industry organisations were more likely to be published (74%, 1204/1625 trials) than the industry-funded trials (61%, 224/345 trials). Larger trials were published at a rate of 88% (85/97 trials) and on average took 5.4 years to be published. Smaller trials with fewer than 100 participants were published at a lower rate with 67% (687/1024 trials) published and these trials took 4.31 years on average to publish. Cross-sectional study of HRC grants: The HRC project included 374 studies with grants from the years 2006 to 2014 and in total this represented NZD 471,663,336, with no evidence of publication or reporting of results for 48 of these grants (13%) (representing NZD 47,095,727). Of the published or reported results there was a mean dissemination time of 4.73 years (SD 2.37). Study of inclusion of gynaecology trials in systematic reviews: Two hundred and thirty (32%) of the 740 gynaecology randomised controlled trials published in 2010 and 2011 were not included in any Cochrane systematic review and from these 21 new review titles were developed. Study of inclusion of fertility trials in systematic reviews: Seventy-two (14%) of the 564 fertility trials published in 2010 and 2011 were not matched to any existing Cochrane systematic review and from these eight new systematic review topics were developed. Conclusions: This thesis has provided conclusive and comprehensive evidence of ongoing research waste, particularly in the area of clinical trials. Three of the projects (the CMR, the ANZCTR and the HRC) confirmed that non-publication and publication bias of studies continue to exist in the global research environment. The thesis also reported that significant numbers of published gynaecology and fertility trials have never been included in Cochrane systematic reviews and were, therefore, not contributing to the body of evidence in their respective clinical topics. Research funding is generally difficult as resources are scarce, and the processes are competitive. The opportunity costs of this waste are considerable but only in the HRC project were actual costs able to be quantified; this money could have been spent on other research. It is hoped that funders of research will reflect on these findings and as a result initiate better processes that potentially could include the introduction of regulations to ensure that research efforts are not wasted.