Thought Leaders in Healthcare IT: William King, CEO of Zephyr Health (Part 2)

Sramana Mitra: What are the other buckets of data?

William King: We take the public domain data and co-mingle it with proprietary data from our customers. This could be anything that they’re generating from things like CRM systems, ERP, or any of the data capture system that they’ve built and are leveraging internally. We can co-mingle that with public domain data.

The final category is purchase data. It’s very exciting to see how the data vendor landscape continues to evolve. Of course, there are the big behemoths that have been in the category for decades and continue to be very relevant and important. Increasingly, we have things like specialty pharma, which helps with drug distribution. These therapies are being distributed by much smaller and more nimble companies. Being able to co-mingle them with all of those other sources is really important.

Sramana Mitra: Does that cover all the different data sources?

William King: Yes. Those are the three big categories. At the end of the day, we do two very important things that are, from a computer science standpoint, quite tricky. We’re able to bring the data together and disambiguate the data and deduplicate. In effect, we run what we call record linkage.

The machines and the algorithms are able to sift through the data, identify commonalities between the datasets, create keys, and ultimately link them and put them in a structure that is easily queryable. Data is continuing to grow. Healthcare is growing at about 50% year on year in terms of new data generation. It’s one of the fastest, if not the fastest, category.

Because we have this great technology in the backend, we’re able to be source and structure agnostically and continue to keep pace with the velocity of all this information that’s coming in. We’re really excited.

Sramana Mitra: Could you actually take us through maybe two or three customer use cases? You can choose whatever types and styles of customers where your technology is particularly well highlighted. Walk us through what kinds of scenarios are coming up and how you are addressing those use cases.

William King: I’ll address all of our customer types. I’ll start with a large pharmaceutical company that’s headquartered in the United States. We’ve spoken with their IT and business groups. They were trying to solve some critical business problems that relied on connectivity of datasets in their ecosystem that they just couldn’t connect. They’re having trouble doing it, and they gave us a call.

When we got in and investigated, we found that they were trying to connect 75 different sources of data. They’ve been working for over four years on the problem. They have spent many millions of dollars on the problem and weren’t anywhere close to the 75 being connected. The root of the challenge that they have is the fact that datasets were connected by humans. That process has been human-led. There’s great tooling and great rules-based software that enables humans to connect data.

When you’re talking about 75 datasets, that’s a lot of time, people, and expense. We were able to come in and were able to link up their 75 datasets, but we ended up adding many sources from the public domain. It ended up being well over 150 datasets. We pulled it together in about 45 days.

This segment is part 2 in the series : Thought Leaders in Healthcare IT: William King, CEO of Zephyr Health
1 2 3 4 5 6 7