How to crack a new market by doing what you do best

Despite the obvious differences between the biggest IT markets of today, one thing that they all have in common is the importance of data. Focusing on that common thread, it might be easier than you think to shift from one market to a totally new one.

At least, that’s what Modus Operandi, a small software company that serves the defense and intelligence communities, is doing as it pushes into the health care market.

The company’s niche is semantic technology, which helps “intelligence analysts find specific unique data and correlations of data within the multitude of databases across the intelligence agencies and the military services,” said Rick McNeight, Modus Operandi president.

In practice, an intelligence analyst can put a query into a semantic wiki, such as the name of a known terrorist, at which point software will search relevant databases and automatically fill in the wiki with information not only about the terrorist, but about his associates, notable locations, and so on, McNeight said.

“We don't just find a needle in a haystack; we find the right needle in the right haystack,” he added.

The semantic technology capabilities that McNeight described are just as useful in health care as they are in the defense and intelligence space, and that’s because of how semantic technology works.

“The key to semantics is that you create intelligent kinds of models around people's data, and the models you create form these graphs,” said Dr. Eric Little, vice president and chief scientist at Modus Operandi, said Dr. Eric Little, vice president and chief scientist at Modus Operandi.

These “knowledge” graphs, as Little called them, can be built into software systems, and analytics can translate them into understandable concepts and relationships, he said.

Despite coming from a great number of sources, the data that goes into this process can be divided into three general groups: structured, semi-structured and unstructured, Little said.

In terms of Modus Operandi’s defense and intelligence work, structured data can be database information, such as map information or people characteristics. Semi-structured data is derived from things like spreadsheets or from table information on the web. Unstructured data can be information that isn’t necessarily stored somewhere, but is rather collected from various internet sources like social media sites.

“We spend a lot of time in very large sects of this varying kind of data, and when you look at the health care space, you find a lot of striking similarities,” Little said.

Health care data has continuously been collected over the past couple of decades, and especially in the past few years, there has been a push to digitize all of it for easier storage and access.

With Affordable Care Act requirements, there has been a push for doctors to not only convert to electronic means of communications, but that the newly electronic information be portable, searchable, and able to provide meaningful use, like better analytics and a more holistic patient approach, Little said.

“To put this together in the medical field, you would need access to structured data, to semi-structured data and to unstructured data,” he added.

From a medical perspective, structured data could be things like lab results in a database, or any information about your stay in a hospital, such as what drugs were administered, whereas semi-structured data could be things like billing information. Unstructured data could be things like handwritten doctors’ notes pointing to things like your attitude, mood, family health history, and other things, Little said.

Broken up into the categories of structured, semi-structured and unstructured, it isn’t difficult to see how data could be used in similar ways despite belonging to different spaces. And it’s with that confidence in mind that Modus Operandi will continue to push into the health care space.

“The analysis of what's going on is very, very similar to what we're doing in the intelligence community; it's just a different set of data, and the models that you would build require a different set of subject matter expertise,” Little said. “Instead of inputting knowledge about potential terrorists or bombing events, you have knowledge of disease patterns and medical tests and treatment.”