1 Boring Old Man

Pfizer Peels Back The Curtain On Clinical Trial Data

Pharmalot

by Ed Silverman

12/04/2013

In response to the growing debate over disclosure of clinical trial data, Pfizer has announced an updated policy and web site that will publish large amounts of information and offer a procedure for independent researchers to obtain study data for analysis. And the drugmaker boasts that its effort exceeds the voluntary aims recently articulated by industry trade groups in the US and Europe. “Pfizer’s expanded policy, which you can read here, is part of a larger and evolving effort by those who create and use clinical data to arrive at a transparent, harmonized process to expand access in ways that protect patient privacy, respect the regulatory process and maintain incentives to conduct new research,” says Fred Lewis-Hall, a Pfizer executive vice president and chief medical officer, in a statement. The move, which met with some criticism [see below], comes amid increasing turmoil over clinical trial data transparency.  A growing chorus of researchers, regulators and patient advocates has been pressuring drugmakers to release study data in the wake of scandals involving different medicines. The push includes an online petition created by the AllTrials campaign…

The Pfizer effort comes a year after GlaxoSmithKline committed to creating a new procedure for disclosing data, a position that, until recently, other drugmakers have not taken [read this]. Glaxo has been largely praised for its plan, which got under way last spring, although has already encountered push back from some researchers who claim the drugmaker is too tightly controlling the release of data [look here]. Meanwhile, the European Medicines Agency has delayed implementing a proposal to proactively make trial data, which is submitted in connection with marketing approval for a drug, available to researchers once an approval decision has been made [back story]. The delay comes amid a squabble with two drugmakers that have gone to court to prevent the agency from releasing data concerning several drugs [see this]…

However, drugmakers have increasingly found themselves on the defensive over the perception that they are hiding data that could either harm patients or lead to better health outcomes. Aside from concerns that more regulators will push for more disclosure, this central issue of trust is prompting industry to rethink its position and find ways to compromise over releasing data. Toward that end, Pfizer says it will publish synopses of clinical study reports filed with regulatory agencies for approved products for which basic results are posted in the ClinicalTrials.gov registry, but only dating to September 2007. These CSR synopses will include summary results for all primary and secondary endpoints, but any data that could be used to identify individual patients will be removed.

The issue of releasing clinical study reports is a separate debate within the larger debate. There appears to be general agreement about avoiding the release of data that can specifically identify and, possibly, compromise the privacy of individual trial parricipants. But some critics argue that certain material can and should be released in order to detect safety signals and can be accomplished with breaching privacy. "There is a real risk to be mitigated," says Hans Eichler, senior medical officer at the EMA, who was speaking today at a conference on clinical trial data disclosure issues at the Harvard University School of Public Health.

Pfizer, meanwhile, is also creating an external independent review panel to consider all requests for trial data that is denied or only partially approved by Pfizer and make a final decision [make a request here]. This is the same approach taken by Glaxo, although one of its panel members has done consulting work for the drugmaker [read here]… In addition, Pfizer plans to produce and distribute lay-language summaries of trial results to participants who wish to receive them, starting with trials that begin enrolling in 2014, in countries where regulations permit, according to its statement. And the drugmaker is piloting the use of technology to enable its trial participants to download their own electronic clinical data collected in the trial.

"I’m not impressed for the following reasons. This policy does not cover all Pfizer trials, just more recent ones and… also does not cover trials in which Pfizer medicines were used in off-label, FDA-unapproved indications," writes Peter Doshi, a senior scientist at Johns Hopkins University who is among several researchers calling for a system for restoring invisible and abandoned trials, or RIAT [see this], and is advising researchers negotiating with Glaxo. "And for the [newer, on-label] trials that are potentially available, the policy does not allow third party access to patient-level data for a minimum of six months following publication of the trial by Pfizer [so no independent checking of the journal publication’s accuracy for a minimum of half a year]."

bor _{[beginning of rant]}… In most Clinical Trials, the article could actually be written in advance of the un·blinding [and maybe it should be]. Pre·program which tables to compile, graphs to generate, statistics to run, and conclusions to be drawn based on the outcome cranked out by the computer. Such automation doesn’t happen and probably never will, but an alternative would be to make that outcome data available so that if there are questions of interpretation, someone else can have a shot at analyzing the data themselves. Many clinical trials have been strategically unpublished. And the published articles are filled with examples of selective publication bias – the data analyzed to fit the desired conclusion rather than to illustrate what the data actually says. That was never okay. Just because people got away with it in the past is no precedent for the future – in fact, the opposite. It’s a call to arms.

The pharmaceutical companies are making offers like this one from Pfizer that have roadblocks to total access. They’re fighting to maintain control over the data release citing patient/subject confidentiality, competitive advantage in the marketplace,  the right to control access to the data, control its use, etc. One gets the idea that they think that the requested full data transparency will play havoc with their business plans. I expect they’re right about that, but that tells us something up front. Why would it, if everything is on the up and up? The published article is only a proxy for the actual results of the clinical trial [proxies…]. That space between the un·blinding and the publication [or non-publication] has been treated as a sacred cow, used for creative interpretations that have often made a mockery of the science of data analysis.