Open Data

Hidden clinical trial data are systematically undermining doctors’ abilities to prescribe treatment with confidence. A whole range of widely used drugs across all fields of medicine have been represented as safer and more effective than they are, endangering people’s lives and wasting public money. As of January 2013, the BMJ will no longer publish any trial of drugs or devices where the authors do not commit to making the relevant anonymised patient level data available, upon reasonable request. On this page we are documenting some of the BMJ’s coverage of adverse outcomes associated with hidden clinical trial data. We are also highlighting the extent of the problem, as shown in our hidden data special issue, published in 2012. We are also asking you to help us catalogue drugs, devices, and treatments for which a lack of complete clinical trial data has resulted in a skewed evidence base. Fill in our online form to tell us where and when you have seen this reported.

Latest developments

Members of the European Parliament have voted overwhelmingly (594 to 7) in favour of introducing a raft of legal measures to increase the transparency of clinical trials in Europe. From 2016, when the new clinical trials law is expected to come into force, all trials will have to be registered on a publicly accessible EU clinical trials register before they can begin, and a summary of the trial results will be required to be posted within a year of the end of the trial, along with a lay summary for the general public. This will apply prospectively, so will not solve the problem of unpublished data on drugs in use now.

Concurrently, another step toward retrospective access to data was made as AbbVie, one of two US drug companies that took legal action against the EMA to try to stop it releasing clinical trial data, has dropped its lawsuits. The other case, brought by the biotechnology company InterMune, continues - for background see Jim Murray's BMJ blogs, where he covers data transparency in Europe.

Key articles

The updated version of the Cohrane collaboration’s meta-analysis of neuraminidse inhibitors for the prevention and treatment of influenza in health adults and children has been published. This is the first meta-analysis to scrutinise the full clinical trial data, and to make those data, plus a full pre-publication history of the final review, available for public scrutiny. The full research, plus a raft of articles explaining the process of obtaining the data, what this means for open science, are available from

New research from the University of North Carolina have estimated the frequency with which results of large randomised clinical trials registered with are not available to the public.. The researchers concluded that around 29% of trials registered on are still unpublished five years later.

Elizabeth Loder, associate editor at The BMJ and associate professor of neurology at Harvard Medical School, wrote a feature "Sharing data from clinical trials: where we are and what lies ahead" outling the promises and progress towards open-data that have been made so far.

Peter Doshi, post doctoral fellow at Johns Hopkins University School of Medicine, in Baltimore, announced a new initiative Restoring invisible and abandoned trials: a call for people to publish the findings. In this analysis article, Doshi explains how provision of access to data released through requests under the Freedom of Information Act, legal proceedings, and from regulators, will be reassessed to correct the publication record.



Following an editorial “Statins and The BMJ” by Fiona Godlee in August 2014 , we said we would contact the principal investigators of all the relevant trials and publish their responses. The BMJ has identified 34 trials and we want to see how easy it is to find the data on harms.


Our first open data campaign initiative relates to a public promise Roche made in 2009 to release full clinical trial reports for Tamiflu, the drug recommended for treatment of influenza.

This page documents correspondence with Roche, and the various bodies around the world which licence or recommend drugs. This open correspondence of letters offers readers the chance to witness attempts to compel greater accountability and responsibility in public health decision making and policy.


Relenza, developed by Glaxo Wellcome in 1989, was the first neuraminidase inhibitor. While carrying out its 2009 systematic review and meta-analysis, the Cochrane Acute Respiratory Infections Group requested the full set of clinical trial data from GSK, as it did for Tamiflu and Roche.

This page documents the group’s correspondence with GSK leading up to receipt of a redacted version of those data in February 2013.

Incretin mimetics

Incretin therapies—glucagon-like peptide-1 (GLP-1) agonists and dipeptidylpeptidase-4 (DPP-4) inhibitors—looked as if they might change the face of type 2 diabetes. Their dual action of switching on insulin and suppressing glucagon to help control blood glucose has been hailed as the biggest breakthrough since the discovery of insulin. And they may soon also be licensed for treating obesity.


Rofecoxib (Vioxx)

Rofecoxib was introduced in 1999 as an effective, safer alternative to NSAIDs for the treatment of pain associated with osteoarthritis. Safety concerns emerged about an increased risk of cardiovascular events in some patients, and the VIGOR study was set up to measure this potential problem. The results reported were negative, but subsequent litigation exposed the data manipulation that took place.


Rosiglitazone (Avandia)

Rosiglitazone, a diabetes treatment approved in 2000, was withdrawn by the European Medicines Agency in 2010 in relation to increased rates of myocardial infarctions in some patients receiving the drug. The UK’s Medicines and Healthcare products Regulatory Agency has confirmed that evidence now suggests that the risks associated with rosiglitazone outweigh the benefits and “that it no longer has a place on the UK market.” Yet it has not released the clinical trial data to explain its change of heart.

hidden data

Hidden data special issue

Publication is mandatory for publically funded trials in the United States. However, in this special issue, a series of research papers confirmed that a large proportion of evidence from human trials is unreported, and much of what is reported is done so inadequately.

The authors and researchers who contributed to the special issue discuss their findings and the implications in a round table on the BMJ podcast

Open Data News