- Semantic analysis
- High performance
When researching a particular disease or condition, clinicians often turn to specialized knowledge databases that categorize information according to some standard medical taxonomy or codes. But such databases are only as good as the effort that went into creating and maintaining them. Data that is buried within clinical notes may never be “coded,” either because no one has the time to wade through reams of text, or because the information is not couched in standard terms.
CysNET Software, a healthcare application provider in Spain, has leveraged InterSystems iKnow™ semantic analysis technology to expand the amount of medical information available to clinicians, hospital management, and knowledge workers. CysNET’s application, called Badakit Healthcare (Badakit means “I know” in the Basque language), automatically analyzes clinicians’ notes and other sources of free text, flagging terms that may correspond with a medical taxonomy such as SNOMED CT or other medical dictionaries. It gives clinicians, hospital management, and knowledge workers real-time access to all the records pertaining to a given condition, resulting in better, faster diagnoses and reporting.
“InterSystems iKnow™ technology is unique in that it starts with the text itself, taking a ‘bottom up’ approach to analysis,” says Marcos Sabourdin, CEO of CysNET. “And it doesn’t just search for words. iKnow can find concepts and relationships within free text. For example, it doesn’t just find ‘cancer.’ It will automatically find entire concepts such as ‘metastatic cancer’ or ‘suspected cancer’ as they appear in the text, rather than only finding words that match an external taxonomy. iKnow even understands negation, and, of course, there’s a huge difference between ‘a tumor was found’ and ‘no tumor was found’!”
We analyzed and indexed approximately 500,000 historical patient records in five hours. That would have taken days using other methods.
CEO, CysNET Software
CysNET partnered with a well-respected teaching hospital, the Clínica Universidad de Navarra (CUN), to create Badakit, and the first installation was in CUN’s oncology department. “Since then,” Sabourdin continues, “CUN has found they can use Badakit for more than analyzing clinical records. By using Badakit Healthcare to index their huge website, CUN enables visitors to research diseases using synonyms. And there are plans to combine Badakit with voice recognition software to create a breakthrough dictation application.”
“When developing Badakit, we evaluated other semantic analysis tools, but none could match iKnow for performance and cost-effectiveness,” Sabourdin says. “iKnow is built into the InterSystems Caché® data platform, which is known for providing very high performance even when running on minimal hardware. As we implemented Badakit at CUN, we analyzed and indexed approximately 500,000 historical patient records in five hours. That would have taken days using other methods.
Sabourdin continues, “Another advantage to Caché is that it works absolutely seamlessly with InterSystems’ other products. We leverage InterSystems DeepSee® in Badakit’s Analytics Portal, which provides dashboards and timelines that incorporate information found by iKnow. And with the InterSystems Ensemble® integration platform, we can easily integrate Badakit with all the various data sources at a hospital or pharmaceutical company.”
Interoperability will be extremely important as CysNET starts selling Badakit to other healthcare organizations. “We have already demonstrated Badakit at a hospital in northern Spain,” Sabourdin says, “and they were very surprised by what it can do. We are also working with the marketing department of a Spanish pharmaceutical company, using Badakit Healthcare to leverage social media in a new iPad and iPhone app.”
Sabourdin thinks the future looks bright for Badakit. “For now, Badakit works in Spanish, so our target market includes healthcare organizations in Spain, Colombia, and Mexico,” he explains, “but iKnow can analyze a number of different languages, so we have the potential to expand geographically. The possibilities for leveraging such powerful semantic analysis technology are practically endless.”