Responsible data-sharing calls for guiding principles to maximize benefits, lower risks

The sharing, integration and analysis of data from clinical trials among academic and commercial institutions benefits future research efforts and improves on clinical care across all health care specialties.

According to a recent report issued by the Institute of Medicine (IOM), Sharing Clinical Trial Data, while there are significant benefits of data-sharing, there is also is a need for policies to validate analyses and protect the privacy of patients, investors and the academic recognition of researchers.

“The system is currently set up for us to not share data and we need to change the way that we are doing things,” Ida Sim, MD, PhD, professor of medicine and co-director of biomedical informatics at the Clinical and Translational Science Institute at the University of California, San Francisco, said in an interview with HemOnc Today.

Guiding principles

The committee offers guiding principles for the responsible sharing of clinical trial data:

• respect the sharing of study participants’ individual data;

• share trial data in a fair manner; and

• increase public trust in clinical trials and the sharing of trial data.

These principles should be balanced in the context of specific trials and stakeholder needs, including concerns about the potential harms and costs of data sharing, according to the report.

“We do not currently have the data to say whether or not the benefits outweigh the costs and time it takes to compile this information,” Sim said. “However, our IOM report is trying to make the case that the benefits do outweigh costs. Data is of wide public interest. It is evidence upon which we can learn from and the more people who can use this data the more we will get out of it.”

Two types of data-sharing

While there has been a significant amount of effort put forth on the sharing of summary results data, Sim, also one of the committee members who wrote the report, said the sharing of individual patient-level data is important to verify that the summary results are true.

“We need to be very clear about the different kinds of data-sharing that there are,” she said. “There are summary results data and there are individual patient-level data. While the efforts towards summary results data have set the stage for now considering the individual patient-level data pooling, we need to double-check the summary data. We need to be able to combine studies — it is much better to combine the individual observations and do the analysis directly from the initial observation than it is to pull the average of the summary results. Information is lost, in some sense, when going from individual data to the summary of data.”

According to Frank Rockhold, PhD, senior vice-president of global clinical safety and pharmacovigilance at GlaxoSmithKline, even though individual level patient data sharing is a complex and resource intensive process.

“It is vital to fulfill the clinical research enterprise’s commitment to patients,” he told HemOnc Today.

Costs vs. benefits

In the report, the committee summarizes a business model for sustainable and equitable data-sharing. The committee recommends for all stakeholders to contribute to the costs of building a data-sharing infrastructure, as current platforms are insufficient and cannot contain and manage the substantial amount of trial data.

Moreover, the system lacks an adequate workforce to manage the operational and technical aspects of data-sharing, according to the report.

“If we do this right, design good infrastructures and good policies, then the more we do it and the more capacity we build out, then the less the cost is for sharing every single study,” Sim said. “Once data-sharing becomes a part of the way clinical trials are conducted, then the cost per trial of supporting data-sharing will decrease.”

Live databases

Despite the challenges, there are a number of successful data-sharing databases up and running today.

As previously reported in HemOnc Today, one such database, the Project Data Sphere initiative, was launched in 2014 by the Life Sciences Consortium of the CEO Roundtable on Cancer. The purpose of this initiative is to facilitate the sharing, integration and analysis of data from phase 3, comparator arm cancer trials.

The initiative was launched with nine total datasets, including 4,000 patients, with an anticipated 25 additional datasets to be added.

Another active database is the Clinical Study Data Request site. Here, researchers are able to request access to anonymous patient-level data from previously published clinical trials.

“We have been live with our system since May 2013 and during this time have posted more than 1,000 trials including patient-level data from nearly 750,000 patients,” Rockhold said. “These numbers are expected to double by the end of 2015.” – by Jennifer Southall

Reference:

Institute of Medicine of the National Academies. Sharing Clinical Trial Data: Maximizing Benefits, Minimizing Risk. Available at: http://www.iom.edu/Reports/2015/Sharing-Clinical-Trial-Data.aspx. Accessed March 19, 2015.

For more information:

Frank Rockhold, PhD, can be reached at Global Clinical Safety and Pharmacovigilance, GlaxoSmithKline, 5 Moore Drive, RTP, NC; email: frank.w.rockhold@gsk.com.

Ida Sim, MD, PhD, can be reached at the Clinical and Translational Science Institute at the University of California, San Francisco, UCSF CTSI, Box 0558, 550 16th St., Floor 6, San Francisco, CA 94143.

Disclosures: Rockhold reports employment with GlaxoSmithKline. Sim reports no relevant financial disclosures.