Unstructured Data Analysis – the Hidden Need in Health Care

The ‘Hood

I recently had the opportunity to attend the HIMSS 2016 conference in Las Vegas, one of the largest annual conferences in the field of health care technology.

As I walked through the main level of the Exhibit Hall, I was amazed at the size of some vendor booths and displays. Some were the size of a house. With walls, couches, and fireplaces, they really seemed like you could move in!

I was working at the OpenText booth on the Exhibit Hall level below the main one, which was actually a converted parking garage floor. Some exhibitors called this lower level “The ‘Hood”.  I loved it, though; people were friendly, the displays were great, and there were fresh-baked cookies every afternoon. I don’t know how many of the nearly 42,000 conference attendees I talked to, but I had great conversations with all of them. It’s an awesome conference to meet a diverse mix of health care professionals and learn more about the challenges they face.

The Trick

Half of the people I talked to asked me about solutions for analyzing unstructured data. When I asked them what kind of unstructured data they were looking to analyze, 70% of them said claim forms and medical coding.

This actually surprised me. As a software developer with a data analysis background, I admit to not being totally up on health care needs. Claim forms and medical coding to me have always seemed very structured. Certain form fields get filled in on the claim form and rigid medical codes get assigned to particular diagnoses and treatments. Seems straightforward, no?

What I learned from my discussions was that claims data requires a series of value judgments to improve the data quality.  I also learned that while medical coding is the transformation of health care diagnosis, procedures, medical services, and equipment into universal medical codes, this information is taken from transcriptions of physicians’ notes and lab results. This unstructured information is an area where data analysis can help immensely. The trick now is “How do we derive value from this kind of data?”

The Reveal

OpenText has an unstructured data analysis methodology that accesses and harvests data from unstructured sources. Its InfoFusion and Analytics products deliver a powerful knockout combination.

The OpenText InfoFusion product trawls documents and extracts entities like providers, diagnoses, treatments and topics. It can then apply various analyses, such as determining sentiment.

The OpenText Analytics product line can then provide advanced exploration and analysis, dashboards, and reports on the extracted data and sentiment. It then provides secure access throughout the organization through deployment on the OpenText Information Hub (iHub). Users will enjoy interactive analytic visualizations that will allow them to gain unique insights from the unstructured data sources.

The Leave-Behind

If you’re interested in learning more about our solution for unstructured data analytics, you can see it in action in this application, While this is not a health care solution, it demonstrates the power of unstructured data analysis that allows users to visually monitor, compare, and discover interesting facts.

If you’re interested in helping me develop an example using claim forms or medical coding data, please contact me at I definitely want to demonstrate this powerful story next year at the HIMSS conference.

See you next year in the ‘Hood!

About Daniel Melcher

Daniel is a Solution Architect for OpenText Analytics in California.