How I Built Software to Save Patients -- and Hospitals
By: Graham Winfrey
In 2012, I was the director of the pulmonary department at a small hospital in Colorado, and I was pissed off. I had trained my whole life to become a physician, only to discover that health care fails at something that should be really simple. It happens all the time in medicine that when you have a test administered at a hospital, you find something abnormal that isn't what you were looking for. But hospitals don't ensure that patients with documented abnormalities receive the appropriate follow-up tests. Studies show that the right test gets done on time only up to 30 percent of the time.
That can have catastrophic downstream effects. For example, let's say you go to the hospital because you got into a car wreck. They're looking for a blood clot, but they find a pulmonary nodule. Most of these are benign, but they can potentially become lung cancer, which could progress to the point of being basically a death sentence. You would expect that a physician would be assigned to a patient with a pulmonary nodule abnormality as a result, and that they would order the appropriate test at the appropriate time. That doesn't always happen.
When I began working at this small mountain hospital, I assumed that there would be about 10 patients per month with the abnormal finding of a pulmonary nodule that's actionable. Then we started reviewing radiology reports. We found 10 times that amount--100 people per month who needed a follow-up, and the hospital had no idea. We're talking about an 80-bed hospital.
Once we began following up with these patients, I started diagnosing lung cancer much earlier compared with my peers. That helped patients live longer, and all we were doing was analyzing this unstructured data that was in a dark abyss.
In 2014, I hired a data scientist from Wall Street and put together some models to help drive better care management of patients with these incidental findings. We created a mechanism to identify those findings with superior precision that doesn't require a human to review each positive finding manually in a report. But I still needed help with business development, so in 2015, I asked Christine Spraker, from Medtronic, to join the company. I was her client and she was interested in our results. She's now my co-CEO.
We were slogging along, cold-calling doctors and selling one hospital at a time, when Hospital Corporation of America was looking for a partner to help it with lung cancer screening and incidental pulmonary nodule management. There's a lot of complexity around getting data-driven software to behave properly in health care across different environments. If HCA chooses a vendor and the vendor has success, everybody else will use that vendor. I said to Christine, "Let's throw a hail mary, stop selling anybody else, and focus on getting HCA." We won HCA as a client and sold our software in a big chunk: 100 hospitals. It was the first time we signed up more than one hospital at a time.
I had emailed Mark Cuban trying to get him to invest in the company at least 15 times, but he had always declined. Then, in 2017, we got an investment from Chancy and Zach Love, whose extended family is a part owner of the Colorado Rockies baseball team. They've had family members who were affected by lung disease, so they have a passion around making sure pulmonary patients get identified and treated appropriately. Cuban followed on that investment and invested himself.
Once Mark and Chancy and Zach got involved, it gave us a lot of credibility. At the time, a competitor was telling our potential clients we were a mom-and-pop shop that would go broke within one to two years. When a billionaire backed our company, they couldn't use that against us anymore.
We've shown that we can improve patient care, so now we're expanding our software to work with every abnormality that needs a follow-up. Eon literally brings in more patients who show up for their appropriate tests. It's the best initiative for hospitals to grow financially. All they have to do to make more money is do the right thing for the patient. We need to find more solutions like that in health care. If we can figure that out, patients are actually going to have a chance.