The european medicines agency is trying to pull the trigger and impose useless access to important study protocols
For years, independent researchers have been trying to get access to the original protocols of clinical trials on drug addiction. The evidence is too clear that the data provided by the manufacturers and discussed in scientific journals cannot be used for any meaningful analysis. The companies not only withhold studies that are negative for them, they are not afraid to manipulate the data to their liking. This is undisputed, the only discussion is about the outcome of the catastrophe, which has led to the fact that for a large number of drugs the spectrum of effects and side effects is uncertain, let alone whether they have advantages over older substances.
The european medicines agency (ema), which is responsible for the approval of new drugs, has been conspicuously unresponsive for a long time. The lobbying was too good, the prere too strong, the jobs too important. Only public prere led to more transparency in the last years, but, as it seems, not to a rethinking in the authority.
Last year, the ema presented a draft guideline for the publication of study data, which provided for far-reaching transparency. The experts were satisfied. The reaction was all the more emotive when it became known only at the end of may this year that a fundamentally revised draft was to be put to the vote in june.
According to this, scientists were only allowed to view the study data on the screen. Downloading, saving, editing, photographing, printing, distributing and transmitting the information was prohibited. An absurd procedure, from a "publication" cannot be called a publication.
The ema was fully aware that under these conditions any scientific evaluation of clinical trial data is impossible. The handling of the protocols is already laborious, the data are often available in different formats, and agreement on a standard is a long way off.
The caving in of the ema to the interests of the industry goes even further. In future, manufacturers will be required to submit two versions of their study reports as part of the application for marketing authorization. One is a full version, on the basis of which the ema decides on the approval, but which remains secret. The other is an abbreviated, incomplete version for the professional public. In addition, the requirements for blacking out study results and reports that allow conclusions to be drawn about individuals are so vaguely worded that rough sections could be rendered illegible.
Jurgen windeler, head of the german iqwig (institute for quality and efficiency in health care), can hardly conceal his horror: "in view of the experience we have had with industry in recent years, this procedure is alarming." the experience from the early benefit assessment shows how valuable complete study data are for the discussion of new drugs. "We are therefore surprised by this sudden step backwards, which is simply incomprehensible from our perspective."
There is speculation about the reasons for the ema’s sudden turnaround. The european ombudsman emily o’reilly suspects in an interview a connection between a court case that the ema has fought with the drug manufacturer abbvie over the publication of trial data. Abbvie had recently withdrawn the case, but part of the settlement was that the company would be allowed to edit the protocols.
Between 2011 and 2013, the ema had given scientists limited access to raw data, two manufacturers had filed a lawsuit against this, among them abbvie, since then the source has dried up. Researchers from the independent cochrane network had asked in time, their recently published meta-analysis of several study protocols on the antidepressant duloxetine shows the importance of access to raw data.
In painstaking detail work the authors have extracted from over 13.The inconsistencies out of more than 13,000 pages. In the raw data of the nine placebo-controlled studies examined, there were not only descriptions of severe side effects that did not appear in the journal articles, but also some of the effects were presented in a more positive light. In some cases, the effects were presented more positively than the data indicated. Two of the nine studies were never published. In precisely these two, no significant effects of the active ingredient were found. But it gets even better: in a third study, the protocol shows no significant effect of the tested drug, but in the published journal article this is claimed.
The examples can be continued: earlier this month, the british medical journal published an article in which a group of researchers was only able to show that another antidepressant, reboxetine, had no advantages over placebo after a thorough analysis of the study protocols. And in october 2013, iqwig researchers were able to show, on the basis of 101 evaluated study protocols, that these contain significantly more important information than the specialist publications.
In the run-up to the vote on 12. June, there were numerous protests against the new directive. The ema relented and will now publish a new revised version in july, which will probably again allow a limited download of the data. It is unclear whether the vaguely worded pas about the black tongues will also be revised.
The fact remains that the ema is under strong prere from the industry to allow as little transparency as possible. In order to conduct valid meta-analyses of drug effects, however, researchers are dependent on the study protocols. They alone provide the necessary raw data. Public access is essential. A global standard is needed to keep these data machine-readable.