Dave DeCaprio (left) and Andrew Eye (right) co-founded ClosedLoop.ai in 2017. The company recently won an artificial intelligence competition for the Medicare and Medicaid Service Center.
When ClosedLoop.ai participated in an artificial intelligence competition hosted by the Center for Medicare and Medicaid Services (CMS), it faced several major competitors: IBM, Mayo Clinic, and Geisinger. But in the end, the startup from Austin won.
What is their secret?
CEO Andrew Eye said that focusing on transparent and interpretable algorithms helped ClosedLoop win in April. Since then, the startup has recently completed a $34 million financing led by Telstra Ventures.
Many AI competitions only focus on accuracy, which means that the winner may be a company with a higher algorithm accuracy. But for the competition of CMS, The company is also responsible for enabling their solutions to explain to doctors and address potential sources of bias.
ClosedLoop competes with other finalists to develop an algorithm that can predict unexpected hospitalizations and adverse events. Its solution was chosen not only because of its accuracy, but also because of its interpretability scored by a group of clinicians.
“If you are predicting the stock market-if I am an artificial intelligence company in the fintech field, if my prediction is correct, then no one cares how I did it right,” Eye said in an interview with Zoom. “In the field of healthcare, there is a decision maker between the insight you show and the actual impact in the real world. And her name is a doctor. If they can’t understand how the algorithm came to its conclusions, they will Never take action.”
From Eye’s point of view, the problem in the healthcare field is that so many vendors sell proprietary algorithms, it is difficult to accurately understand the content. This also makes it difficult for hospitals to review their prejudice or fairness, or to ensure that they can provide good services to people in the community.
For example, researchers at the University of Michigan School of Medicine recently discovered that the sepsis prediction model developed by Epic Systems Missed several undiagnosed cases of sepsis At the same time, the alarm is still triggered for nearly one-fifth of the hospitalized patients.
One way ClosedLoop solves this problem is to provide the health system with tools to build its own algorithms. But for those who don’t have a team of technicians to do this, they can still outsource this work to ClosedLoop without getting a “black box” in exchange.
The startup has established techniques that can explain what an algorithm “learns” and which variables have the most influence on a given prediction. It can take this information all the way to the patient level. For example, if someone jumps from the 80th percentile of risk to the 95th percentile, what happens?
“This is what the doctor wants to see,” Eye said. “What we heard over and over again was showing me codes…show me which diagnostic codes, laboratory values, prescriptions you saw made you jump from 80th to 95th.”
He co-founded the company in 2017 based on his experience in healthcare. A few months after he worked with Chief Technology Officer Dave DeCaprio (Dave DeCaprio), his daughter became seriously ill with liver inflammation.
After three weeks of diagnosis Odyssey, they learned that this is autoimmune hepatitis, a disease in which immune cells attack the liver.
“When the correct answer is to take $5 worth of prednisone, we are still two weeks away from the liver transplant,” Eye said.
He later learned that half of all pediatric liver failure cases have never been diagnosed, and 15% of them have never been tested for autoimmune hepatitis. No one uses past data to improve clinical decision-making, and this is exactly where Eye hopes his company is building to make a difference.
“It’s not just about chronic diseases and avoidable hospitalizations, although these are the places that have the biggest impact on the dollar,” he said. “This is for the long tail problem, for rare diseases such as autoimmune hepatitis. This is where I am passionate about this.”
