by Jennifer Zaino

November saw the awards handed out for the 2009 Semantic Web Challenge, sponsored by Elsevier. Thinking back to the earlier years of the challenge , one of the exciting things is how pragmatic and sophisticated some entries in this year’s batch of winners are as consumer-oriented offerings. The winners of this challenge in the past have always been innovative and inspiring, but sometimes nichy and, yes, the y could be geeky too. But it’s a real validation of the semantic web and its potential to see winning applications emerge that leverage the technology backbone to deliver practical and easy to use consumer services with real market potential.

TrialX one of the winners of the Open Track challenge-- is a great example of that. With advisors including Peter Frishauf, the founder of Medscape, and medical and business executives in the thick of clinical research efforts, the site is seeking to solve a real world problem that dramatically can affect people’s lives: the difficulty in recruiting patients to participate in clinical trials. The delays, according to the submission for the awards authored by Applied Informatics’ Chintan Patel, Sharib Khan, and Karthik Gomadam, can lead to postponements of new and potentially life-saving drugs to the marketplace. Worse yet, patients with very serious conditions who don’t know about trials for which they could be eligible might die without having even had a chance at a miracle. There’s an impact for the pharma companies behind these trials, the clock is ticking on their patents, which means less time to recoup their costs and realize profits before generic versions of the drug are launched.

The goal of this site is to more efficiently match patients to clinical trials that are relevant to their particular conditions, by empowering the patients themselves with the user-friendly tools they need to leverage the details of their conditions without having to actually understand medical jargon to find the trials for which they are likely to be optimal candidates. The technology can integrate with personal health records platforms – Microsoft HealthVault, Google Health and Indivo – so patients onboard with those services can let TrialX access the data, convert their health record information into semantic models, and then use reasoning and graphing algorithms to discover matching clinical trials for the patient record. Patients can add data directly, as well.

The TrialX platform for which the service is named relies on Its Columbus Matching Technology, which uses semantic and natural language processing technologies to analyze aspects of the patient’s health records and match it against eligibility criteria for trials. As the developers describe it, doing the matching at a semantic concept level leads to more meaningful results. For example, a clinical trial might be underway for patients taking antibiotics, so TrialX would present a match with the health record of a patient taking the drug Vancomycin; it infers using semantic conceptual knowledge that that drug is indeed an antibiotic. The service lets consumers use a large variety of synonyms – hypertension or high blood pressure, for example – to describe health factors; their queries are dynamically mapped to a common concept within its Unified Medical Language System so that either description used delivers the same results to the query. From the patient point of view, however, there’s nothing that screams “semantic web”— it’s a clean interface that is as easy to navigate as any web page out there.

The semantic web techniques within the service also help those organizations conducting the trials, letting them add new trials to its database to generate form interfaces and automatically identifying the minimum information required to generate patient matches, based on previously mined trial data and external ontologies. Its database currently has about 25,000+ ongoing clinical trials, which is continually updated, according to the site. Most data comes from the clinical trials government database at clinicaltrials.gov , which provides the latest clinical trials approved by the US FDA, as well as listings provided by clinical research investigators for the ongoing clinical trials at their respective institutions.

TrialX has recently released an update of its API so that other web sites—such as Health Edge News, a site on promising research, clinical trials, and new treatments--can add its clinical trial matching technology. An iPhone app is now available too.

The two other winners of the Open Track challenge were Digital Enterprise Research Institute’s (DERI) Sig.Ma [search engine and mashup generator, and VisiNav, an application for searching and navigating RDF datasets collected from the open web. They’re cool, too, and fairly accessible by general users. But they still feel a bit “research-y” in comparison to the intuitive consumer look and feel TrialX exhibits. Of course, the heart of the semantic web is a technical one. To that point, the Billion Triples Track was won by "Scalable Reduction," an entry from Gregory Todd Williams, Jesse Weaver, Medha Atre, and James A. Hendler (Rensselaer Polytechnic Institute, USA). It shows how massive parallelization can be applied to quickly clean and filter large amounts of RDF data.