Clinical trial registries promote transparency and integrity in medical research in two ways:
Trial registration allows researchers to specify in advance what they are looking for, thus preventing scientifically unsound data-fishing at a later stage.
Results reporting on trial registries makes trial outcomes publicly available and easy to find.
Prospective trial registration and the reporting of results are mandatory under global rules governing medical research in human participants set out by the Declaration of Helsinki.
So how does New Zealand stack up against global standards? Badly, it turns out.
Trial registration: Over half of trials violate medical research ethics
As part of my PhD, I looked at a cohort of older trials conducted in in New Zealand and found that only 42% of clinical trials had been registered in advance. In 2014, new rules were introduced: New Zealand’s Health and Disability Ethics Committees required prospective trial registration prior to ethics approval. At that time, around 65-70% of trials were being registered in advance.
So what has changed as a result of the new rules? Earlier this month, I looked at 311 clinical trials that were registered during 2015-2018. According to the new requirements, each and every one of them should have been registered in advance. In reality, only 155 trials complied with the new rules. In other words, just over half of these clinical trials violated medical research ethics. While compliance rates might be somewhat higher if a wider range of trials were included in the sample, it is clear that the new rules are not being effectively enforced.
Results reporting: 92% of trials missing results on the regional registry
So how does New Zealand fare on results reporting? Are trial outcomes made public on the regional trial registry? A 2015 report by the registry itself does not address this question, so I looked for myself. Out of the 311 trials mentioned above, I selected the 113 that had been registered during 2015. Most of these are likely to have been completed, but only nine of them had posted their results onto the registry. The other 104 trials – 92% of the cohort – were missing results on the registry. Sure, some may have published results elsewhere, but that is not enough.
National ethics guidelines in New Zealand still do not mandate the full reporting of trial outcomes. Instead, they merely recommend that results are reported. Compliance is not monitored, and lack of compliance is not sanctioned.
(In comparison, in both Europe and the United States, reporting rates are steadily increasing. Both the United States and the European Union have adopted rules that make results reporting mandatory, and there is growing momentum towards monitoring compliance and imposing sanctions for non-compliance on both continents.)
Need for more transparency in New Zealand
Update, May 2019:
Key actors in New Zealand must act on clinical trial transparency. Transparency promotes safety, efficiency, public trust and much more. Specific action points:
The government of New Zealand should make prospective clinical trial registration and summary results reporting within 12 months a legal requirement, in line with World Health Organization best practices, as recommended by Transparency International, Cochrane, and TranspariMED. The government should actively monitor compliance and impose sanctions, including fines, for those who break the rules. The forthcoming Therapeutic Products Bill offers an opportunity to incorporate these provisions into law. Current efforts in the UK in this regard may provide a useful model.
The Health Research Council of New Zealand should immediately sign up to the WHO Joint Statement, and fully implement its provisions within one year. As well as accelerating medical progress, this would prevent tax money going to waste. Clinical trials that do not report results make no contribution to science.
Health and Disability Ethics Committees (HDECs) should require not only a clinical trial registration number at the point of review, but the full clinical trial registration document, as part of the ethics application process. This would avoid the current duplication of work involved in registering trials and seeking ethics approval, which researchers would no doubt welcome. It would also ensure not only prospective registration, but complete prospective registration (rather than just entering a minimum of data to get a registration number). All ethics committees should do this, but it would require a revision of their application forms. In addition, HDECs should review a trial’s registry entry 12 months after trial completion to check whether results have been shared on the registry, as per WHO guidelines. Where applicants to the ethics committees are not in compliance, the ethics committees should report this to both the applicant’s institution and to the trial funders. Institutions and research funders should support ethics committees in this work, and sanction researchers who break the rules.
The Australian New Zealand Clinical Trial Registry (ANZCTR) should strengthen the registry in general, and facilitate data mining in particular. At present, data on clinical trial activity in New Zealand is fragmented and hard to find (see here and here). Creating solutions to export data in a usable manner would help researchers and policy makers to access valuable data about the national health research landscape.
This blog was jointly written by Emma Tumilty, a postdoctoral fellow in Bioethics at the Institute for Translational Science, University of Texas Medical Branch, and TranspariMED founder Till Bruckner. Emma Tumilty completed her doctoral work in New Zealand exploring equity in the health research environment using multiple public data sources including trial registries. She contributes this guest blog in a private capacity.
Note on methodology: Recent trial registry data analysis is based on a search of the Australian New Zealand Clinical Trial Registry (ANZCTR) conducted on 6 March 2019 by Emma Tumilty. Selection methodology for the 311 trials was as follows: all trials listed on ANZCTR with completed enrolment that included a New Zealand enrolment site. Out of the cohort of 113 trials registered during 2015, nine had published results on the registry: 2 had plain English summaries, 4 had publication listings, 3 had both a publication listing and a plain English summary. None of these 113 trials had shared data, or stated that they would be willing to share data. (Please note that TranspariMED did not check the accuracy of the data cited in this blog.)