IQWiG: Law must prescribe obligation to publish all clinical trials

If the currently discussed draft on the "Law on the Reorganization of the Pharmaceutical Market” (AMNOG) is implemented, drug manufacturers will in future have to register their clinical trials at inception and publish results after completion. Two articles on the drug reboxetine, which were written by employees of the Institute for Quality and Efficiency in Health Care (IQWiG) and have just been published in the renowned British Medical Journal (BMJ), impressively prove that comprehensive legislation on the obligation to publish all clinical trials is long overdue. This not only applies to drugs, but also to medical devices and non-drug interventions.

Pfizer only provided data under massive public pressure

The authors documented in detail how laborious and time-consuming it was in their benefit assessment to initially identify studies concealed by the manufacturer of reboxetine, Pfizer, and subsequently to obtain the study results. In Germany it is still at the sponsor's discretion to decide whether trials are published or not. Pfizer only provided the data after massive public pressure.

The analysis of these data revealed a potential reason for this secretiveness. Data from trials published to date suggested that reboxetine could alleviate depression. However, when the results concealed up to now were included, the effect of reboxetine (marketed under the trade name Edronax) became so small that patients did not benefit from the drug, meaning that it is useless.

The extent of bias can now be quantified

In the articles published in the BMJ, the IQWiG employees have now for the first time exactly quantified how strongly statements on the benefit of reboxetine have been distorted by the concealment of data. In scientific language this phenomenon is referred to as bias. If all trials comparing reboxetine and placebo are analysed, positive effects are only about half the size of those in published trials. "The difference becomes so small that reboxetine has no benefit anymore," says Beate Wieseler, one of the authors of the BMJ articles and Deputy Head of the Department of Drug Assessment at IQWiG. "At the same time we see bias in the opposite direction for harms: if all trials are considered, the proportion of patients who discontinue reboxetine because of adverse effects is about twice as high."

Results of older trials should also be disclosed

Against this background, the authors defined several specific requirements: "The AMNOG draft plans an obligation to publish only for selected trials; however, we need the results of all trials," Beate Wieseler demands. "In addition, it needs to be clarified how the results of older trials can be made accessible. For example, reboxetine has been on the market since 1997, and was tested for approval in the years before 1997. Even after the reform of the law these trials can still be kept under lock and key," says Beate Wieseler.

According to Jürgen Windeler, IQWiG's Director, the planned reform is a step in the right direction. "The gain in transparency in the area of pharmaceuticals is to be welcomed. However, I do not understand why we should still have to be left in the dark in the areas of medical devices and non-drug interventions. An obligation to publish is also necessary here."

Data transparency is a precondition for good decisions

"The example of reboxetine shows once again: only if all study data are available can the benefit of a drug be estimated realistically," says Jürgen Windeler. "However, this is the precondition for actually making good decisions. Patients and physicians can only choose the best possible treatment and the self-administration of the health care system can only distribute the limited resources properly if the benefit - and potential harm - of an intervention are known. Money that is spent on useless drugs will otherwise be missing elsewhere for meaningful treatments," says Windeler.

The German regulatory authority approved reboxetine in 1997; however, it could not consider all trials that IQWiG has now analysed. Later the manufacturer also applied for approval in the United States; however, this application was declined in 2001. On the basis of the IQWiG report the Federal Joint Committee (G-BA) excluded reboxetine from reimbursement in September 2010, but this decision is not yet legally effective.

Contact: Tel. ++49(0)221-35685-0, info@iqwig.de