Decode Health on Fighting Chronic Disease and Capturing Nuanced Data: More Intelligent Tomorrow, Episode #13
The COVID-19 pandemic has catalyzed the healthcare sector to reconsider how it performs medical tests, uses data, and decodes the nuance of respective diseases that occur in varying geographic areas and among different demographic groups. Decode Health’s CEO Chase Spurlock and Chief Strategy Officer Julia Polk recently joined the DataRobot More Intelligent Tomorrow podcast, hosted by Ari Kaplan, Director of AI Evangelism and Strategy at DataRobot, to explore the data and AI strategies healthcare professionals are using to determine medical treatments and outcomes.
Julia Polk, who worked on the first AIDS test back in the1980s, underscored the importance of delivering logical testing while understanding possible further implications. “Chase and I spend a lot of time understanding what good testing looks like—scientifically-based good testing—but one of the things we started talking a lot about at Decode is: Is COVID the next chronic disease? Is COVID going to drive more chronic disease? Is COVID going to create chronic disease that didn’t exist before?”
This is particularly prescient when recounting the toll chronic disease takes on individuals and the economy:
- 90% of U.S. healthcare expenditures goes toward treating chronic and mental health conditions, according to the Centers for Disease Control and Prevention (CDC).
- More than 868,000 Americans die of heart disease or stroke every year, costing the U.S. healthcare system $214 billion per year and $138 billion in loss of job productivity, according to the CDC.
- Cancer care costs continue to rise in the U.S. and are expected to reach almost $174 billion in 2020, the CDC reports.
A One-Size-Fits-All Approach Is No Longer Applicable to Managing Disease Trends and Personas
Chase Spurlock warns against a rigid, uniform approach when dealing with disease. “Our experience delivering solutions to partners has taught us a lot. We’ve learned how to be flexible with data. But one of the key takeaways for us—and this is where DataRobot has come into play—is there really is no one-size-fits-all approach. There’s no national model. We have to have an understanding of the target population that we’re serving and there are regional, geographic, even time differences associated with disease trends and disease personas, and we have to be able to capture the subtlety and nuance of disease that might occur in Tennessee versus California.”
AutoML Use Case Demonstrates How Real-Time Models and Guardrails Deliver Proactive Clinical Treatment
Charles Spurlock went on to emphasize how much AutoML has helped Decode, a Nashville-based AI solution provider that can predict and monitor chronic diseases and help to deliver better care. “This is really a use case for AutoML: if you’re starting to build hundreds of models, being able to take care of those models in real-time would be a challenge if you were building this alone. But being able to have an infrastructure that has some guardrails, and is taking a lot of that pressure off, allows you to capture more of the clinical persona, the clinical pattern that leads to actionable behavior at the point of care—it leads to that proactive treatment we’re gunning for.”
Spurlock drew on culinary metaphors to explain how DataRobot helps Decode’s data scientists. “We have AutoML solutions like DataRobot that rough it in. The data scientist’s job is really to take what the DataRobot sous-chef has produced and taste the soup at the end of the process—maybe add some seasoning to it—and then send it out the door.”
To hear more about how AI, Decode Health, and DataRobot are predicting and monitoring chronic diseases better, check out Datarobot.com/podcast or http://datarobot.buzzsprout.com/. You can also listen everywhere you already enjoy podcasts, including Apple Podcasts, Spotify, Stitcher, and Google Podcasts.