IRPA AI Founder, Frank Casale, Interviews David Ledbetter, a Senior Data Scientist at Children’s Hospital Los Angeles where he works with the Laura P. and Leland K. Whittier Virtual Pediatric Intensive Care Unit (vPICU) to discuss his AI journey.
Frank: What prompted your department to investigate AI?
David: Our department is the brainchild of Dr. Randall Wetzel, who is the chair of anesthesiology and critical care medicine at CHLA. As a clinician practicing in the ICU, he saw firsthand the overwhelming amount of data that was being captured, but under-utilized. He also understood that by storing that data a reservoir of experience was being created that spanned not just his own clinical practice, but the collective experience of hundreds, if not thousands, of other doctors. Advances in machine learning [and now deep learning] have provided tools to help us glean from those experiences to provide the best possible treatment for every child that comes into the ICU.
Frank: How did you get all the stakeholders onboard?
David: It’s an iterative process. Starting small with proof of concepts for particular groups, demonstrating value and utility, and slowly building a coalition. There’s never a silver bullet that solves all the problems for all the people. Instead it’s important to take small steps with objective milestones to enable a slow and steady march of progress.
Frank: What has been the most significant benefit to date?
David: We’ve built some exciting algorithms to help assess severity of illness to identify kids who need increased observation in the ICU or provide earlier warnings for patients that are at risk of critical decompensation. But I think the biggest benefit has been to help expose the value data. Children’s Hospital Los Angeles was one of the leaders in transitioning to Electronic Medical Records (EMR) so we have one of the richest Pediatric ICU datasets available in the world. With that data we’ve been able to demonstrate how our algorithms improve with access to more and more ICU encounters. We’re getting close to a tipping point where healthcare institutions start to break down their data silos and help create collaborative datasets with hundreds of thousands of ICU episodes to learn from. We’re about to see some really amazing advances start happening.
Are you interested in how artificial intelligence or robotic process automation can bring benefits to your company? Email [email protected] to take advantage of our 30-minute intelligent automation briefing to learn more.