SAVE THE DATE! KMWORLD 2019 in Washington DC NOVEMBER 5 - 7, 2019

 

Healthcare: The importance of making connections

This article appears in the issue May/June 2019 [Volume 28, Issue 3]
<< back Page 2 of 2

Accessibility

In order to make this diverse information accessible to end users, Cambridge Semantics developed the Anzo platform, which is a data discovery and integration layer that blends data sources to allow users to navigate across the entire spectrum of enterprise information. It provides automated ingestion of data, data layers for modeling, interactive query, and both exploratory and automated analytics. “A good metaphor for how this works is a traditional map,” explained Powers. “When you look at a map, you can see how to get from one place to another, and you can also see all the surrounding territory and how the places are connected and related to each other.”

Graph database applications are still relatively rare, but are beginning to emerge, particularly in the realm of recommendation engines. At one leading healthcare institution, for example, efforts are underway to develop a single source of collective intelligence for treatment of advanced melanoma. With this disease, at some point, patients run out of options. The facility is developing a system that is designed to capture information from many different data sources, including doctors from throughout the country. The model is based on abductive reasoning, in which the outcome is identified and then the simplest hypothesis proposed for how this outcome occurred. Data sources can include notes, emails, and also transcriptions or highlights of case review meetings, among others.

Healthcare thrives on connections

“The graph database can find relationships in a very user-friendly and scalable format,” noted Powers. “The starting place is often the EHR, but it can be blended with other documents, including research literature and publications.” The structure of the graph database lends itself to establishing profiles of other patients, quantifying their characteristics, and then analyzing likely outcomes for particular patient. “A knowledge graph can quantify patient characteristics as well as finding the related documents, so it becomes a very versatile tool.”

Healthcare thrives on creating connections, whether among patients, physicians, in R&D, or in predicting medical outcomes. Graph database products are proliferating because of their ability to discern relationships and accommodate rapidly changing data sources. For example, The Zephyr Health platform, acquired in 2018 by Anju Software, is a data discovery and analytics solution built on a graph database from Neo4j. It complements Anju’s products for clinical trials and post-approval feedback on pharmaceuticals.

For many years, graph databases were used mainly in the academic community. Although relatively small now, the commercial market for graph databases is predicted to grow 25%–40% per year over the next 5 years. The versatility of this technology in handling diverse data sources and its ability to be presented

in an accessible format are likely to make it a good match for many healthcare applications as well as others in which relational databases and document repositories have not provided the desired capabilities.

<< back Page 2 of 2

Search KMWorld

Connect

Buyers' Guide
Learn More in the Buyers' Guide!