Q&A with Jason Smith, CTO at Within3

Jason Smith discusses his work with AI-based data platforms.

More Pharma companies are adopting AI-based platforms to collect and analyze data. Jason Smith, CTO, AI & Analytics at Within3, spoke with Pharm Exec about how the technology works and how it can be implemented.

Pharm Exec: What hurdles have you encountered working with AI?

Smith: One of the big ones is access to accurate data. Our primary customers are biotech and pharmaceutical companies, who very much have control of their data. It’s an isolated system, and within those organizations, there’s multiple isolated systems. One of the bigger challenges we’ve worked with is getting access to that data in ways that allows us to train effective AI models. That’s done through lots of partnerships with our customers and trying to understand how to bring a data centric approach to building AI where we’re looking for great sets of data that are clean, but not necessarily giant sets of an entire enterprise’s data lane. Those challenges allowed us to refocus on becoming more data centric, but the challenge was how to get access to the data to build the model when access is highly regulated and sometimes not accessible, even within the organization or the customer we’re talking to.

Pharm Exec: Have you had to deal with any regulatory issues?

Smith: We work with our customers and partner with them where the data lives and for the use of that data. So, when data is collected from different means, is it allowed to be used to train models or even be shared with other vendors? How does that all interplay? Those are some of the regulations. We also must deal with The General Data Protection Regulation (GDPR) in Europe. So, how do we share and collect information in the EU and properly analyze it in a way that stays within the GDPR compliance? We work with all of those pieces when building AI, most importantly when gathering data to train the model and also where those AI models live. We deploy globally to stay compliant in different regions for those customers.

Pharm Exec: How does AI impact the patient experience?

Smith: Somewhat indirectly. Where we really helped accelerate the patient is in their voice. There are patient advocacy boards that come in and speak for the patients and what their needs are and what the effects are. Also, because we’re tapped into social media, we’re able to capture what patients are saying. What are they Tweeting about, what are they posting about on Reddit, what are they saying their needs are? Especially in the rare disease community where it’s important to get a hold of that information. Then, using our natural language processes, we’re able to summarize and understand that voice and amplify that back through our reporting and into the Pharma customer. We think between those channels, while it’s not direct to the patient application, it’s still helping enhance the patient’s voice.

Pharm Exec: Have you noticed a difference between the way patients talk to their doctors and the way they discuss their condition on social media?

Smith: It’s all anecdotal, but I’d say that’s fair. When you look at the doctor’s side, it seems to be very focused, structured, and scientific. On social media, people seem to be a little bit looser and little more free floating. I think you can find a lot there. The challenge we have from an AI perspective is when people say something on social media, it can be fraught with lots of additives that change the sentiment of it and deter the valuable scientific data or what’s really happening, so that does provide a challenge. On the surface, if you take it as raw data, it may skew what the actual feelings are or what the actual science supports. There’s some balance there when we look at this data and how it gets weighted into different models.

It’s not that people aren’t being truthful, it’s that there’s a lot more emotion on social media. There tends to be a lot more adjectives describing that emotion to filter and we’re just trying to get to the meat of the point and remove all the noise and the sentiment that is emotional parts of it because we want to get to the scientific facts. Moving through that does provide a challenge. People tend to use a lot of terminology that they may not use in their doctor’s office to describe something.

Pharm Exec: What are the sources for your information?

Smith: We focus on the information that is exchanged between Pharma and healthcare providers, researchers, or patient advocacy groups. It’s those conversations that we’re able to capture and analyze and look through a long thread of conversations of a group of conversations on a topic and dig out the correct scientific points that they need to make strategic decisions. Whether it’s how is the drug progressing in a clinical trial or what patients are saying about the latest drug and how they’re reacting to it, or what are researchers thinking, it’s finding all of those conversations and then being able to analyze them en masse. This gives us a very clear understanding of how the science is occurring all the way through to the delivery to the patient. It reveals how that’s occurring and where they can strategically lean in to help patients or healthcare providers understand their products better through education or guidance.