Video
Using open source to accelerate patient cohorting
Apr 27, 2022

It's an exciting time right now for open-source developers, I think, because there are amazing tools out there now that are completely free, that anyone can download, anyone can start building applications with. These tools, because they're all open source, have been democratized. You can actually, you know, hand that software to the community and then the community can enhance it. It can live on beyond the purpose that you built it for. So, it's not just that you have to wait for something to come to you, you can actually be proactive and be part of the solution yourself.

The volumes of data that used to be challenging in healthcare, there’s a lot of really neat and interesting new tools out there to process it quicker, and you can get going very quickly.

I run a Research and Development group, and that's just very gratifying. And we get to solve the hardest problems of a company. The way my development group changed the game using open-source, and being in life sciences, is that we were able to take a lot of existing technologies off the shelf and blend them up into a really interesting way, all using open- source. We built an application called CLASP, which is basically patient cohorting technology. If you imagine a specific example, A patient who took a smoking cessation drug in 2017, took a statin in 2018, and then in 2019 they had a diagnosis of anxiety. And then, maybe they are males, aged from 45 to 65.

That's an incredibly complicated series of attributes to match. What CLASP can do is take that description and produce the patients with that definition in real time. So, there's 90 billion events that it can sift through in 4 or 5 seconds and produce that patient population.

CLASP is different in its ability to answer these questions quickly, to basically perform attrition on a huge amount of data and get to that answer very quickly. So, the speed is important because if you have to wait 10, 15, 20 minutes, every single time, you have to change your definition slightly like: “Oh, I want another, I want to add a drug to this definition”. Maybe I want to see patients that took this other smoking cessation drug. You don't always know what you're looking for. You may be trying to determine what the size of your cohort is. It isn't big enough to even do a study on it. So, there's a lot of what I call “what if” analysis that you need to do.

The iteration process takes a long time. And to be able to produce that in seconds, you can whittle down, your definition very quickly and then produces your result and give it and hand that information off through treating technology to other systems.

Data drives all of the systems that we build at IQVIA. IQVIA is uniquely suited to solve the biggest challenges in healthcare, simply by its rich richness of its data set, its huge infrastructure that it has in place, and the many, many years it has in the healthcare dataspace.

Related solutions

Contact Us