Like any other piece of technology, DeepIntent Outcomes™ began with a problem. Pharmaceutical marketers didn’t have a solution for quickly measuring and optimizing their campaigns based on real-world clinical data, rather than front-end metrics such as clicks and impressions. So we set out to build one ourselves.
Clinical Data and HIPAA-Compliance
We got the idea for DeepIntent Outcomes last summer. For the product to work, everything had to click together: mockups, architecture and high-level data flow, security and privacy considerations, and offline and real-time business logic.
Starting from a concept, we envisioned how the product should look in order to be useful for marketers while also being completely privacy-safe. DeepIntent Outcomes is integrated with the most comprehensive, timely, and actionable clinical data available today. Compared with other healthcare DSPs, we have triple the depth of clinical data, including 8B+ pharmacy claims, 3B+ medical encounters, and 150+ payer-complete datasets. Once we got access to this data, we began to generate reports using only the back-end tools as we hadn’t yet built the user interface. This helped us get a sense of the value DeepIntent Outcomes could provide marketers. Next, we put a user interface on top of that data to help clients visualize it.
We spent four months building DeepIntent Outcomes while securing certifications for HIPAA compliance took about three months. Data privacy is the number one factor that drives the complexity of our operations. Building DeepIntent Outcomes, we had to tokenize two different datasets and then merge them. Optimizing toward real-world prescriptions adds an additional layer of complexity to the privacy architecture.
Of course, the devil is in the details. You need granularity, but not so much granularity that individual patients can be identified. All data we upload is HIPAA-compliant, and we even have a separate certification that covers internal reporting to ensure our analyses are HIPAA-compliant.
Building Technology Starts With Building a Team
Building DeepIntent Outcomes truly took a village. I might be biased as DeepIntent’s Chief Technology Officer, but I believe my team of engineers is the best in the business. Contextualizing clinical data for marketers is challenging, and our engineers were not only able to build the product but continue to improve upon it and make it stronger. While the period from idea to implementation took about six months, building DeepIntent Outcomes truly could have taken years.
As part of the village, it’s important to consider not just our internal team, but our clients. As beta partners, they were instrumental in the creation of DeepIntent Outcomes as they provided early feedback that shaped what the product eventually became. Because most competitive solutions take weeks or even months to report campaign data, we initially thought weekly data refreshes would suffice as a differentiator. However, our clients said a higher frequency would be more valuable and as a result, DeepIntent Outcomes offers a daily refresh of clinical data.
Given the highly sensitive nature of clinical data, our legal counsel was an invaluable partner in building this product. Bringing DeepIntent Outcomes to market, I may have actually spent more time talking to lawyers than engineers!
Understanding Clinical Data
Think of raw data as an observation. It’s simply a record of what happened. An insight is a bigger-picture interpretation of that data and what it means. For example, let’s say we have zero records for diabetes type 2 visits for a certain product. To explain lulls and gaps in data is what separates an observation from an insight, the difference between having data and understanding data.
Every step of the process inevitably has some data loss. There’s data loss when you tokenize datasets and merge them. The key is to choose data partners who understand clinical data not just from a schema perspective, but who live and breathe this data.
We couldn’t have brought DeepIntent Outcomes to life without the right data partners. The product was built atop an incredible volume of clinical data, none of which is valuable without a strong understanding of how to interpret it.
How DeepIntent Outcomes Fits In
Expensive and highly regulated, real-world clinical data is notoriously difficult to come by. Fortunately, we had a head start, having built the DeepIntent Health Advertising Platform integrated with clinical data. Because our platform was built purposefully for healthcare marketers, we’ve long had partnerships with the best clinical data providers available.
This has always been among our strongest differentiators as we enable our clients to deliver privacy-safe ad targeting. With the ability to leverage that data in real time, marketers now have far more insight into how those campaigns are performing in-flight and can optimize toward real-world business outcomes now — not months later.
DeepIntent Outcomes has been proven in beta to drive efficiencies and increase script lift by up to 35%. It’s the latest innovation in our mission to build marketing technology that measurably improves the lives of patients, and I’m very proud of what we’ve built.