For years health systems have recognized that incidental findings in imaging studies often fall through cracks in the system and they have sought to develop workflows to ensure evidence-based follow-up happens. Healthcare Innovation recently spoke with Chris Frost, M.D., chief medical officer and chief quality officer at 60-community hospital Lifepoint Health, and Aki Al-Zubaidi, M.D., founder and CEO of a company called Eon, about Lifepoint’s deployment of Eon Breast, a platform designed to identify, track, and manage patients requiring follow-up across both routine mammography screening and incidental findings detected on non-breast imaging.

The partners said that the Eon Breast solution enabled Lifepoint to diagnose patients at risk of breast cancer early as well as gain operational efficiencies. In addition, after implementation Lifepoint found that more than half of incidental cancer diagnoses were among patients not traditionally eligible for routine screenings based on clinical guidelines, and, therefore, would have been missed for identification and follow-up in a screening-only model. 

Overall, Lifepoint’s Healthy Person Program is an enterprise initiative across multiple disease areas including lung, pancreas, kidney, liver, thyroid and aortic aneurysms powered by Eon’s AI platform. To date, Lifepoint’s incidental findings management program has identified more than 100,000 high-risk abnormalities and delivered a 36% improvement in early-stage cancer diagnoses. 

Healthcare Innovation: Before we talk about your work together, Dr. Al-Zubaidi, could you give a brief background about yourself and the launch of the company? 

Al-Zubaidi: I'm an interventional pulmonologist. I had recently started my practice and I saw some patients who were presenting to me with late-stage lung cancer diagnoses, but they had diagnostics that were done years earlier and just missed follow-up. I wanted to solve that problem for my own clinic and went out there looking for a solution, but there wasn't one, so I decided to create one on my own. That was in 2013, and then in 2015 the company officially started.

HCI: In cases where a radiologist has noted an incidental finding but it doesn't get followed up on, is it because they've put it in a note in the EHR but it is not written as an action for someone to take? 

Al-Zubaidi: A patient comes in for something else, right? Maybe they have a CHF exacerbation or they had a motor vehicle accident. What’s stuffed in this big discharge summary is this needle in the haystack that says in a year, you should probably get another CT scan or you should get a mammogram as a follow-up for this abnormality. When they go follow up with their physician, they’re focused on the initial issue or other comorbidities, so it gets buried. The system is just set up where these needles in the haystack are very hard to identify and get to the guideline-appropriate next step.

HCI: Is there something about Eon’s platform on the imaging side that’s doing a better job of identifying potential issues using AI — or is it more that you're creating this structure so that when the incidental finding happens, it's followed up on within the health system?

Al-Zubaidi: This is a really important point that I think there's a lot of confusion about in the market. There are two problems. There’s the potential that a radiologist might actually miss something that's on the image. But that actually doesn't happen that often. Radiologists are quite good at documenting the abnormalities. The second problem is that 70% of the time when they document an incidental, the patients don't get the appropriate follow-up or any follow-up at all. That's the specific problem that Eon’s addressing. 

HCI: What were some initial challenges in scaling the company up and finding external customers for it?

Al-Zubaidi: I found that we had created something that actually was more efficient within my own clinic, and we transitioned it to a couple of colleagues. Early on, Cleveland Clinic was the third place that we actually installed. They had a scale that was probably 100 times more than my clinic had. HCA was install number five. They drove us to really focus on that enterprise scale. We knew early on from that Cleveland Clinic experience what to attack to make this thing make financial sense for a client like Lifepoint, which had 80 hospitals at the time. 

HCI: Dr. Frost, had Lifepoint tried other approaches to dealing with incidental findings. Had it recognized this was an issue and wanted to find an enterprise solution?

Frost: We look for anything that we can do to learn a best practice at a local level and where we see the opportunity to scale across the organization to have a broader impact for our communities. We're going to look for opportunities to standardize where applicable, and then scale that solution. The Eon solution was a perfect example of that. 

HCI: Did you start with one pilot hospital and then expand from there? Are there some issues that have to be worked through or lessons learned as Eon and the health system work together?

Frost: We always start with a pilot process and frequently we'll learn that there are technology issues around the connectivity between any solution that we would partner with and our electronic health records. We have a variety of electronic health records, so when you solve for one, you may solve for the critical mass of our facilities, but there are still opportunities in terms of further integration. 

Also, with any type of system, whether that's Cleveland Clinic, HCA or Lifepoint, there's some type of variation in clinical processes and clinical practice. So the alignment around the technology integration is one of the crucial components, but integrating around a clinical process of care is something that we also use to learn from our pilots. Once we identify with a partner like Eon what that standard of care should look like, we look to quickly scale it, allowing for customization where absolutely necessary. We've now scaled Eon across around 53 of our facilities.

Al-Zubaidi: To be clear, it is Eon Breast that is in all these hospitals. They did start out in Lung, and they did a pilot that was at one site versus another vendor actually, when they were looking to solve incidental lung. During that implementation process, we ended up actually going out with Lifepoint to each one of their hospitals prior to go-live. It was a five-phase rollout of Lung across all their hospitals. Since then, we’ve also rolled out other programs, including the pancreas solution, liver, other cancers, as well as cardiovascular and now the breast.

HCI: A press release I saw quoted a Lifepoint exec, Bart Daugherty, talking about some efficiency improvements.  He said, “We started with 53 different breast programs, each with its own workflows and practices…We can now consistently manage both screening and incidental findings in one system across our facilities.” He said this solution’s adoption helped manage more patients and save the mammography teams four hours per site per week. So besides the potential for clinical improvements for the patients, are there efficiency gains as well? 

Al-Zubaidi: Definitely. There's a supply and a demand issue, right? There's increasing demand for mammography. One of the highest rates of screening is breast cancer screening. The issue is can you get those patients into the machine, get them scanned and have a high-quality, appropriate follow-up. Before they initiated this program, their teams were spending about 30 minutes per patient on administrative tasks — mailing letters or extracting data from the healthcare record. Post-Eon, that really went down to 15 minutes per slot. So they were able to double the amount of slots for mammography by implementing the solution.

Frost: We talk about supply-demand mismatch as it relates to physicians and nursing staff. The importance of ancillary staff, to include technologists, especially in the space of radiology, is very poignant. These efficiency improvements really address a need that if not addressed either results in lapses in care or delays in care, because the queue for the patients builds as it relates to evaluation.

HCI: Lifepoint and Eon did a study that found that patients with incidental findings were 6.2 times more likely to result in a cancer diagnosis than patients undergoing routine screening. Could you explain why that is and the significance? 

Al-Zubaidi: Of the incidental patients, more than half of those patients were not eligible for traditional, routine screenings. What that really shows is that we're finding people who traditionally aren't in a program where they're being monitored every single year, or they're not getting that preventive care on a regular basis. 

If you're on that path of screening and you're adhering every single year, when something does show up, it's likely to be early. But if you're looking at an incidental — this is the same for other things outside of breast, too — you're likely going to find somebody who has cancer more often than when you’re screening at a higher rate.

Frost: Aki is an interventional pulmonologist. I was a hospitalist. We would take care of the sickest of the sick that came through the ER. And it was not an uncommon scenario in taking care of a patient with advanced lung cancer or advanced breast cancer that when you followed the thread back in terms of prior imaging studies, the answer was, ‘Aha, there was something there.’ If there had been a program like Eon, there was the possibility that it would have been identified. That level of awareness would have prompted follow-up including the care navigation that ties technology and clinical processes together, so that we wouldn't have had to address this patient at such a late stage.

HCI: Anything else you want to add about your work together?

Al-Zubaidi: One of the reasons Lifepoint has been successful with this implementation is that they really involve the local markets. That means not just the radiology service line leader, but the person who is running radiology at that specific hospital, the CEO at the local hospital and all the stakeholders. They did that at every one of our implementations, and I think they have crushed it because of that.

Frost: To me, the magic of Eon is that they found that nexus, that sweet spot of people, process and technology. You have asked about some of the technology considerations, but the people and the processes are just as important. There’s a clinical algorithm that says this particular finding warrants follow-up attention. Then the people follow through with the corresponding communication cascade to ensure that it doesn't get lost to follow-up.