Tag Archives: TCPS-2

Update on withdrawn CIHR trials policy

In an only somewhat-overdue update (thanks to conference season interrupting my regular blogging activities – I do write on the road, but need to get sleep & give a read over before I can push “publish” on a post) the Canadian Instutites of Health Research (kind of the Canadian NIH, for US American readers) has put out a new message regarding the missing CIHR trials policy that we’ve been following since late March.

To backtrack a bit, while I was on the road and having fun with research and colleagues over the past month, there was more coverage of the CIHR trials policy disappearance, including Michael Geist’s blog and the National Post. Additionally, “rapid response” letters from around the world continued to roll in to the BMJ related to their article, some with great titles such as, “Canadians step back from the well of transparency while the World is thirsting for it” and “CIHR decides it must compromise my interests as a patient.”

Then, right around June 15 or so (I saw it on the 17th), CIHR President Dr. Alain Beaudet issued a “Message from the President – Policy on ethical conduct of research involving humans.” Go read it: http://www.cihr-irsc.gc.ca/e/43756.html. If you’re anything like me, you may need to read it a few times over, because it’s not the clearest statement ever made.

Here’s what I think it’s saying:

  1. That the March removal of the trials policy (“Registration and Results Disclosure of Controlled and Uncontrolled Trials”) was about “harmonization” and deference to the TCPS-2
  2. The TCPS-2 has some requirements for trial registration and public disclosure
  3. While the trials disclosure requirements in the TCPS-2 and the former CIHR trials registration policy were in the same spirit, the CIHR policy had more specific directions about what needed to be done
  4. CIHR will (at some unspecified point) be  integrating certain of these more stringent operational requirements as part of the terms and conditions of its “relevant programs.” These include: a) publication of the systematic review used to justify the trial, b) registration and compliance with WHO requirements for minimum data disclosure, and c) submission of final reports in CONSORT format.
  5. CIHR will propose 4 revisions to the TCPS-2 for “prompt consideration” (not clear on how soon this can/will happen): a) applying to all trials, not just clinical trials, b) requirement to update trial registration when the trial protocol changes, c) requiring that serious adverse events be reported in post-trial publications, and d) and a requirement to deposit aggregate data in an unbiased, publicly accessible database.
  6. In the interim, CIHR will specify that researchers have to “comply with all the requirements mentioned above” (not sure whether this means 4a-c or is also inclusive of 5a-d).

So, what does this mean? Are we all good now?

Well, we’re better than we were before the press coverage, I think. We’re not as better as we would have been, had the trials policy never been pulled.

Remaining questions:

  1. Really??? I’m still kind of skeptical that 3 months after the trials policy and the TCPS-2 came out, both of which had been in development for a looong time, someone suddenly just went, “Oh, gosh, you know what? It’s not okay to have both of these policy statements!” Why am I skeptical? Well, because it just doesn’t make sense. CIHR had a trials policy that wasn’t 100% the same as the TCPS-1. Tri-council funders have all sorts of different policies that are more stringent than the TCPS, and it’s not a problem (e.g., the beloved CIHR Access to Research Outputs policy). It’s just not adding up, and at this point the message seems to be that it doesn’t matter if it’s not adding up, CIHR is sticking to their story.
  2. When will the CIHR be implementing these new requirements for relevant grant programs, and how will this implementation be different from the trial policy rules?
  3. What’s the process for revisions to the TCPS, and how long does it take?
  4. When are we going to see the requirement to make individual-level/micro/”raw” data publicly available? The item listed above in 5d, which currently has no teeth, only requires aggregate data deposit. What does this mean? How aggregate? Does this have to include all adverse events? We need to be able to reanalyse this data to look for harms to specific groups. This is one of the most important parts of the scrapped trials policy, and there is no mention of it in the new statement from CIHR.

I think the international attention and public pressure on CIHR over the withdrawal of the new trials policy likely contributed to these developments, which seem like a step back in the right direction. However, without teeth in the current requirements, and a return of the publicly-accessibly micro-level data archivng requirement, it seems like 3 steps forward, 2 steps back at this point.

-Greyson

Previous posts on this topic, in case you haven’t been following along:

Advertisements

Leave a comment

Filed under democracy, ethics, funding, government, Health, OA, research

Details: TCPS-2 vs the CIHR trials policy of 2010

Thanks to a few days’ time and some help from people with more experience reading science policy, I now feel that I can expand on my previous post about the TCPS-2 “superseding” the Dec 2010 CIHR trials policy.

First of all, I note that I am putting “supersede” in quotes not merely because it is a really cool word to which I wish to draw attention (although it is) but because the TCPS-2 was officially launched before (not, as I had previously thought, simultaneously with) the new Trials Policy, and as such, would not likely have been intended as a replacement for another policy that had not yet been unveiled.

Non-exhaustive list of differences between the TCPS-2 and the CIHR trials policy of 2010:

  • Unlike the TCPS-2, the CIHR policy required that the systematic review that justifies doing the trial be publicly cited.
  • TCPS-2 is specifically about clinical trials, whereas CIHR could be interpreted to apply to a wider scope of “controlled and uncontrolled trials.”
  • The CIHR policy required following the WHO international standards, whereas the TCPS-2 only requires that trials be registeres in registries that meet the criteria of WHO/ICMJE. Seems like a small distinction, but the difference is the minimum dataset required.
  • This one is kind of nitpicking/pointing out an oversight, but the CIHR required the name of the trial registry as well as the registation #, whereas the TCPS-2 only asks for the # (kind of useless without the name, and presumably they want the name too, but didn’t specify).
  • The CIHR policy took steps to prevent duplicate/multiple registration, whereas the TCPS-2 does not address this potential issue.
  • The TCPS-2 says that “Researchers should also promptly share new information about an intervention with other researchers or clinicians administering it to participants or patients, and with the scientific community – to the extent that it may be relevant to the general public’s welfare,” which does not require public disclosure. The CIHR policy required the data to be made publicly available to all.
  • Research design amendments (changes) require ethical approval under the TCPS-2. The CIHR policy required that amendments  to the research design be reported within 30 days after the ethics approval.
  • The TCPS-2 asks for “new risks” or “unanticipated issues that have possible health or safety consequences for participants” during the trial. CIHR policy did not ask for adverse event info until after the trial.
  • TCPS-2 asks for info that might merit or lead to early stopping of  a trial. CIHR policy wanted notification and public disclosure within 30 days of stopping a trial early.
  • The TCPS-2 tells researchers “to make reasonable efforts to publicly disseminate the findings of clinical trials in a timely manner by publications and by the inclusion of raw data and results in appropriate databases,” whereas the CIHR policy specified reporting guidelines (e.g., CONSORT for RCTs) and required reporting and public disclosure within 12 months of the end of the trial, building on the existing CIHR research access policy.
  • The TCPS-2 encourages researchers and institutions to publish results in a timely manner. The CIHR policy required public disclosure within 12 months and reserved the right to disclose the final report themselves within 18 months.
  • The TCPS-2 asks for “the inclusion of raw data and results in appropriate databases” whereas the CIHR specified what appropriate databases are and that both micro (aka “raw” aka “participant level”) as well as macro (aka “aggregate” aka “summary”) level data are necessary.
  • The TCPS-2 talks about ethics with regard to confidentiality clauses and PI access to trial data, whereas these issues were not addressed by the CIHR trials policy.
  • The CIHR policy talks about after-trial follow-up, including submission of any “severe adverse events or harm” to the publicly-available trial registry and a requirement to “retain all trial information including original micro-level data and metadata data for twenty five years unless they are deposited in a freely accessible data repository (to align with the Health Canada requirements).”

Generally speaking, major differences are:

1) Insider vs public disclosure of information

TCPS-2 is concerned with sharing info with the research ethics board and “to other researchers or clinicians administering it to participants or patients, and with the scientific community.”

CIHR was concerned with reporting and disclosing info “to CIHR and the trial registry” – and the trial registry had to be openly accessible to the public.

2) Whether to specify how much should be reported, and how soon to report it

The TCPS-2 advises on ethical matters across disciplines. It makes few specific mandates as to timeliness or data specifics.

The CIHR policy was intended specifically for trials funded by CIHR, and similar to other funding policies could and did specify details of what kind of data should be reported (all, macro and micro), to whom (CIHR and a publicly accessible registry), and when (within 12 months of trial completion, or with some types of data 30 days of early stoppage of a trial).

TCPS-2  on its own vs. as part of a comprehensive approach

I’m not saying the TCPS-2 is bad. It’s pretty good, overall, and after years and years of collaboartive work and revision seems to do an admirable job of doing what it’s supposed to do – which is to set out general, interdisciplinary, national ethical guidelines, not details of practice requirements. Without specific procedural policies designed to instruct researchers in various disciplines, the TCPS-2’s appropriately limited scope leaves us with somewhat vague directions.

The two policies, the TCPS-2 and the CIHR trials policy of 2010, are clearly intended to complement with each other. As I noted before, the TCPS-2 specifically states that:

“[Trial] registries, in addition to agency policies, editorial policies, ethical policy reforms, and revised national and institutional ethics policies and results disclosure requirements, contribute to a multi-faceted approach to eliminate non-disclosure.” (emphasis mine)

I continue to be very disappointed that the CIHR has apparently seen fit to retroactively withdraw their facet of this approach, and I do think that as a public agency they do owe the public a decent explanation of what happened here. What has been made publicly available thus far does not add up. As I have stated previously, there may well be valid reasons for killing the new policy on clinical trial data, but the lack of transparency around the policy retraction continues to be troubling.

-Greyson


2 Comments

Filed under censorship, democracy, ethics, funding, government, Health, OA