How Ochsner Health Radiology Is Scaling Up Incidental Findings Program

One of Healthcare Innovation’s award program winners in 2020 was Kaiser Permanente Colorado (KPCO) for creating a tracking system for incidental radiological findings.

Since then, other health systems have sought to improve appropriate follow-up workflows for incidentally detected findings to improve the percentage of exams that receive timely follow-up. Kernesha Weatherly, D.H.A., vice president of imaging services and provider practice at Ochsner Health System in Louisiana, recently spoke with Healthcare Innovation about how she and her team led a full-scale deployment of a follow-up solution across Ochsner Health’s radiology departments—activating a comprehensive imaging follow-up solution throughout the entire system in less than three months.

Healthcare Innovation: Kernesha, could you describe your role at Ochsner? 

Weatherly: I essentially wear two hats, in dealing with the operations aspect of all things across the organization, but also the provider practice, the physician aspect of it, including radiology staffing. We have over 40 hospitals, and we have a complement of imaging in just about all of them, but we also have a series of ambulatory clinics. We also have freestanding emergency departments. We are all over Louisiana as well as Mississippi, and we have a small footprint in Alabama.

HCI: Is there a radiologist who is your partner in decision-making?

Weatherly: Yes. Ochsner follows what we call a dyad approach. Every service line has such a dyad, where someone such as myself will work hand in hand with a physician. I work with Cecelia Brewington, M.D., system and academic chair of imaging at Ochsner Health System.

HCI: Could you explain why the organization wanted to work on follow-up of incidental findings?

Weatherly: We partnered on this with a company called Inflo Health. We were looking to answer the question: Are we addressing every incidental finding that comes out? We know that patients show up and sometimes are not quite sure what's going on, and they think it's one thing and other things are found. We weren't sure we had a clear way of essentially closing the loop on that. When we started digging through it, we saw that we had a very manual process. If the radiologist hit a series of buttons, it made it into this work queue. People were keeping things in Excel files. When you realize how clunky a process it was for the radiologist to do it, and how it slowed them down from moving on to the next exam, we looked to try to solve that.

HCI: Had Ochsner execs seen Inflo in use at other health systems, or how did you hear about them? 

Weatherly: We had not. We started the conversation around wanting to address incidental findings. We looked to see which vendors are working on this. We are a big Epic house, so of course we first started with Epic, and at the time Epic didn’t have a solution. They've since come up with something they say is on a road map. If you have any knowledge of Epic, those road maps can be extremely long. 

So we started asking: what is out there now? We know there are a lot of conversations around lung nodules and lung cancer screenings. But one of the conversations that our physicians had was around: are we just concerned about lung nodules or are we concerned about any potential incidental finding in patients? So when we started interacting with different vendors, we saw ones that just looked at the lungs, and then we saw those that looked at the lungs and beyond.

HCI: Did you also have to involve the IT team or the EHR team about integrations?

Weatherly: Absolutely. We had a huge team, including our IT department, risk management, our physicians, and we also had some of our referrals as well, because we're about to turn on a platform that has the propensity to send out a lot of cases around lung nodules. We had to make sure that our pulmonary team was on board to receive those. 

It was a large group involved in the conversation and decision making to make sure that we were in agreement with that. If an incidental finding is found, what we can't do is notify the patient and then say we may not be able to get to you for six weeks. That just causes them unnecessary anxiety. So from a scheduling perspective, we had to make sure we could say we will get you back in within seven days.

HCI: What do you think were some of the keys to the success of rolling this out across multiple hospital radiology departments in a relatively short timeframe?

Weatherly: I think a few things helped us. One, from an operational standpoint, our team has a regular cadence of meetings, and we're always talking about different initiatives. So it wasn't that I had to go find who's the person in this division, who's the person in that division. We are constantly in meetings, and we have those touch points. When we had the initial conversations about this, the whole team heard it at one time, and we all walked through the process together. Also, the risk management team covers all the divisions. The IT team covers all the divisions. Our radiologists read at multiple facilities. So again, it wasn't a series of conversations. It was one conversation.

HCI: What are some of the biggest impacts of the Inflo Health solution? 

Weatherly: The biggest thing is Inflo has allowed us to convert manual processes. Inflo now gives us a report with data. We can say, filter it from this date, this time, this facility, this position, to see what it is that we're looking for. We can automatically generate the report. In addition to that, we now have the ability to do things like looking to see out of the recommendations that our radiologist dictated, how many of those did the clinical team want to move forward with? Because they don't want to move forward with everything. But we now have actionable data that we can work through to help us make the decisions to ensure that we're closing a loop on these patients.

HCI: I also read that in terms of advancing AI implementations, you're building off an American College of Radiology Learning Network framework with a Center of Excellence model. Could you talk about that a little bit?

Weatherly: The American College of Radiology has a process improvement program. When we were looking at implementing this, we wanted to make sure that we were aligned with industry standards. What are our peers saying? What are other organizations doing? When we knew that Inflo had a partnership with the ACR, and ACR is kind of giving them a seal of approval and helping us build out the performance improvements and processes through this, then we were clearly on the right track with that. There are a lot of questions around tools that use AI, especially now, and we wanted to be confident as we were answering those. 

HCI: When I've interviewed folks about AI tools in radiology, one of the key ones described is helping radiologists prioritize which things to read first in their queue. Are there tools like that in place there at Ochsner, or other ones that that are coming on board now to help the radiologists?

Weatherly: What I can tell you is that I get no less than three or four e-mails a week from some AI group that's reaching it out to help solve a problem that they just know that we have. We have a really good data governance team in place to filter through what is true, what is not, and where do we go with this? We're having a lot of conversations around implementation and working through what vendors can truly follow up with. Because right now, everybody is saying that they have the ultimate thing that we need, and we need to get in a 10-year contract with them, but we realize that that's not the truth. 

HCI: Are there metrics around this Inflo implementation that you will track over time to see how well you're doing at following up?

Weatherly: We know that we have a very large imaging population. We are getting in almost 2 million exams annually. Out of this population of exams, how many of them require an incidental follow-up? What is the imaging itself telling us? Are we closing the loop on the patients? And for the patients who do not want to close the loop and follow up on the incidental finding, are there any trends we can see with that? Are there trends by service lines? So right now it's almost like we're the little kid watching with our big glasses on, trying to see what's to come. Because the data is just now starting to roll in. We've partnered with our population health team to let them know that we're doing this. And when we get enough robust data, we really want to sit down and talk with them about what we're looking at.