Publication bias in trials registered in the Australian New Zealand Clinical Trials Registry: Is it a problem? A cross-sectional study

Abstract

Background Timely publication of clinical trials is critical to ensure the dissemination and implementation of high-quality healthcare evidence. This study investigates the publication rate and time to publication of randomized controlled trials (RCTs) registered in the Australian New Zealand Clinical Trials Registry (ANZCTR). Materials and methods We conducted a cross-sectional study of RCTs registered with the ANZCTR in 2007, 2009, and 2011. Multiple bibliographic databases were searched until October 2021 to identify trial publications. We then calculated publication rates, proportions, and the time to publish calculated from the date of first participation enrolment to publication date. Results Of 1,970 trial registrations, 541 (27%) remained unpublished 10 to 14 years later, and the proportion of trials published decreased by 7% from 2007 to 2011. The average time to publish was 4.63 years. The prospective trial registration rate for 2007, 2009 and 2011 was 48% (952 trials) and over this time there was an increase of 19% (280 prospective trials). Trials funded by non-Industry organizations were more likely to be published (74%, 1204/1625 trials) than the industry-funded trials (61%, 224/345 trials). Larger trials with at least 1000 participants were published at a rate of 88% (85/97 trials) and on average took 5.4 years to be published. Smaller trials with less 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. Conclusions Just over a quarter of all trials on the ANZCTR for 2007, 2009, and 2011 remain unpublished over a decade later. The average time to publication of nearly five years may reflect the larger trials which will have taken longer to recruit participants. Over half of study sample trials were retrospectively registered, but prospective registration improved over time, highlighting the role of mandating trial registration.

Implementation Documentation and Process Assessment of the PharmNet Intervention: Observational Report

“…This can happen in at least 2 ways. First, results from registered trials may never be published at all (a recent study of RCTs in New Zealand found that 27% of results had not been published more than a decade after the study registration date) [ 44 ]. This can produce the mistaken impression that not much is known about a particular approach.…”

Supporting study registration to reduce research waste

“…For decades this lingered at around 50% 24 , but recent analyses show improvement. For example, of 1,970 trial registrations on ANZCTR, 541 (27%) remained unpublished 10 to 14 years later, and the proportion of trials published decreased by 7% from 2007 to 2011 66 . It should be noted that trials represent only a small fraction of health and medical research, and much of that remains unregistered.…”