World’s Largest Coded Reference Database in the Making

To the surprise and delight of Peter O’Blenis, vice president of systematic review technology at Ottawa-based TrialStat Corp., even the most prominent research organizations are willing to recycle their data collection work “to accommodate future research” by everyone else. The altruists include members of the Campbell Collaboration, the University of Central Florida, McMaster University, Stanford University, and the Norwegian Knowledge Center for Health Services.

Their gift to science is data extracted from journal references during the conduct of systematic reviews using TrialStat’s Web-based SRS tool, says O’Blenis. The information is being fed into a publicly searchable online database, ESRNexus, designed to facilitate and promote the reuse of data that has been manually extracted from reference materials.

During a typical SRS-based systematic review, research organizations mine data from thousands of medical references, says O’Blenis. “It’s a shame if it never gets reused by anyone else.” Systematic reviews are an efficient mechanism for deriving high-quality answers to clinical questions by systematically reviewing published studies and meta-analyzing key data elements within them. “With ESRNexus, data that has been extracted by one review group is available for others to use…accelerating the overall review process.”

Ultimately, ESRNexus will aggregate multiple data sources to become a one-stop shop of references for systematic reviews and other research purposes, says O’Blenis. The database already contains about 300,000 article abstracts, together with code fields for all the data manually distilled from them, such as study design and the age and gender of participating human subjects. ESRNexus also incorporates approximately 16 million article citations from the U.S. National Library of Medicine’s PubMed database, and utilizes a similar search methodology. Other reference databases will likely be added.

It is unknown if commercial SRS users will want to contribute data to ESRNexus, says O’Blenis. Private pharmaceutical and biotechnology companies will nonetheless benefit from the database, since “high-caliber” organizations have screened the reference material. “This is different than a pure Wikipedia-type thing. Not just anyone can contribute data.” 

The idea was conceived last November. In January, a dozen of TrialStat’s largest customers were approached about sharing their summary reference data, and every last one agreed, says O’Blenis. The group “represents an interesting cross-section of users, including [university-affiliated] evidence-based practice centers [EPCs] funded by the Agency for Healthcare Research and Quality and overseas folks in the [not-for-profit] Cochrane Collaboration.”

ESRNexus made its debut four months later, on May 10. “This has been one of the fastest software projects I have ever seen,” says O’Blenis. Beta testing is expected to conclude by the end of August. Any needed adaptations will happen in real time.

To date, ESRNexus has had close to 163,000 queries. The top searches have been for information on back pain, lithotripsy, colon cancer, and breast cancer. Although O’Blenis is doing speaking tours and webinars about the database, word is also spreading “virally,” he says.

The users include “a vast number of people from the general public” who land on the site after Googling with a medical search term, says O’Blenis. ERSNexus soon will include Medical Subject Headings (MeSH), the National Library’s thesaurus of current biomedical terminology, making the database even more accessible to the public. MeSH converts common search terms into medical terminology.

Starting in July, people using ERSNexus to search for information on certain medical conditions will be offered linkage to to search for recruiting trials taking place in their area, says O’Blenis. Traffic to the site thus far indicates many visitors are “genuinely sick” or know someone who is.

Over the summer, O’Blenis says he plans to index a multitude of open access publications within ESRNexus, including the Public Library of Science.

Not-for-profit groups that do “good works,” including the Norwegian Knowledge Center, members of the Campbell Collaboration, and a number of Cochrane Review Groups, are using the TrialStat’s SRS tool at no charge as part of a reciprocal agreement that obliges them to contribute the data collected during systematic reviews to the open ESRNexus project, says O’Blenis. The licensing fee would otherwise cost them between $10,000 and $20,000. Not surprisingly, the offer is getting an “enthusiastic” reception.

Because many SRS users compile summary data for research that is not explicitly medical – i.e., the effect of certain policies on high school dropout rates -- ESRNexus contains a small amount of reference material valuable to organizations like the Campbell Collaboration, says O’Blenis. The largely voluntary organization focuses on education, social welfare, and crime and justice.

As the world’s most widely used online tool for conducting systematic reviews, TrialStat’s SRS has about 1,000 users that could potentially publish data on ESRNexus with a few button pushes, says O’Blenis. So far, 49 of them have done so.

SRS eliminates most of the paper from the systematic review process, reducing errors and making it possible for research organizations to “collaborate across geographically separated participants, and in real time,” says O’Blenis. It also reduces by 40 to 60 percent the time required to screen out irrelevant studies and extract data elements. Further, an electronic systematic review is “auditable, easy to drill into, and more transparent and credible” than the paper-based alternative.

To date, O’Blenis says, TrialStat has invested about one million dollars in human time and equipment to develop ESRNexus.

Subscribe to Bio-IT World  magazine.

Click here to login and leave a comment.  


Add Comment

Text Only 2000 character limit

Page 1 of 1