Inspired by hackers: push for open source clinical trials

Information wants to be free*, or so they say. Yet much of the data generated from clinical trials is locked away in proprietary databases never to see the light of day.

Now a group of Australian researchers are calling for an overhaul to the way clinical trials are conducted modelled on the open source software movement in an attempt to see clinical trial data spread more widely to those who could make best use of it.

Writing in Science Translational Medicine, a team of researchers, led by Dr Adam Dunn from the University of New South Wales, say that physicians are hampered in their ability to make effective clinical decisions because much information is locked away.

They draw an analogy with the close software industry, where code is kept under lock and yet, preventing other programmers from learning from it or improving on it.

“Similar roadblocks plague the clinical evidence domain where, despite a rapid increase in the volume of published research, physicians still make decisions without access to the synthesised evidence they need,” said Dunn.

Inspired by the hacker movement that spearheaded the open source software movement, they go on to argue that the same principles of transparency and collaboration could improve the clinical trials process.

“If the same principles were applied to medical research, bottlenecks, biases and self-interest would be largely removed,” said Professor Enrico Coiera, a co-author on the paper along with UNSW Professor Ric Day, and Professor Kenneth Mandl from the Harvard Medical School.

“Clinical trial data is a potential goldmine. If researchers, doctors and patients were able to re-analyse and pool this data, there would be a host of questions that could start to be answered. But these meta-analyses are very uncommon because researchers and companies don’t like to share data,” Coiera said. They propose a three main reforms to the clinical trial process:

(i) Clinical trialists, other scientific researchers, and clinicians should be able to access and contribute to a repository of interoperable patient-level data that sits alongside the mandatory registration and provision of summary-level clinical trial results.

(ii) Infrastructure for repositories, data standards, and interaction among clinical trialists must be provided to foster the growth of a decentralised community, whose main aim is to rapidly identify and address gaps in clinical evidence.

(iii) Improved dialogue between trialists and physicians, which would allow physicians to routinely ask new questions of the clinical trial community and follow and discuss the aggregated clinical trial data to answer existing questions.

They also acknowledge the challenges that would have to be overcome to make open source clinical trials possible, including privacy issues, the implementation of new data standards and providing the infrastructure to foster information exchange and collaboration.

They envisage a set of tools similar to those provided by GitHub or SourceForge, two popular open source software community websites, to allow researchers to submit and have access to deidentified patient-level outcome data along with other metadata provided when registering for a clinical trial.

This would enable patient-level meta-analyses to be conducted using vastly greater databases than are generally available today.

Another challenge is overcoming social and financial barriers. “Most researchers want to hold their data as long as they can as the basis for publications,” Dunn said.

“And unfortunately, pharmaceutical companies want to control the messages that are delivered to doctors and maximise profits rather than facilitate the cost-effective delivery of care.”

The commentary piece was published in Science Translational Medicine today.

* The phrase “information wants to be free” was originally meant to imply that the cost of disseminating information is continually dropping, thus making it cheaper and cheaper to spread information. It was not meant to imply that information ought to be free.