In 2020, I wrote about how New Jersey-based RWJBarnabas Health (RWJBH) had begun working with a startup called Wellsheet, whose solution pulls data from the electronic health record, prioritizes clinical content through machine learning algorithms, and streamlines provider workflow. Another health system that has been working with Wellsheet is Concord Hospital in New Hampshire. Its chief quality officer, Dr. Christopher Fore, recently spoke with Innovation in health care on the impact it has had.
Newark, New Jersey-based Wellsheet’s predictive clinical workflow platform uses FHIR API standards to work within an existing EHR and surface the most relevant content to physicians in a contextualized view prioritized based on the physician’s needs. For example, Wellsheet puts risk tools front and center. If the patient has atrial fibrillation, Wellsheet puts the appropriate calculator in the EHR directly in the foreground. Fore said Concord physicians have been using it for mobile chart review, real-time notifications, care team collaboration, patient handoffs and discharge planning.
Healthcare Innovation: Were there any limitations with your EHR in Concord that made you interested in Wellsheet?
Front: We started collaborating in 2019, right before the pandemic. At first, we used Wellsheet just to create checklists for teams and customize ways to view information to make our day a little easier. As it turns out, that’s actually not the best reason to use Wellsheet. It also allows you to better manage patient flow, access, and capacity, set up notifications, and communicate better.
HCI: So this issue of managing flow, access and capacity, is that something that is up to administrators or someone who manages a hospital unit?
Front: It’s typically used by people with clinical experience in administrative roles like myself. Our utilization management and care management teams use it all the time. Our nursing unit leaders refer to it all the time and clinical teams refer to it on our whiteboards.
HCI: You are the quality director. Do you see this from a quality or patient safety perspective?
Front: Yes, I do. I will tell you that every quality review that I facilitate, I have Wellsheet open and on the screen in the room. Because if you want to trend vital signs, look at labs, read provider notes, look at imaging results, it’s all there. The way that the data is trended, the visualization, is far superior to what we can get within the electronic medical record. It makes it very, very easy. I can’t remember the last time I went to Cerner to review a case, versus going to Wellsheet to review the Cerner data on the case, because it’s just so much easier.
HCI: Can we give you an example of the ROI achieved by using Wellsheet?
Front: One of the things I’m involved with is managing length of stay, meaning reducing the amount of time that patients have to stay in a hospital bed. I think everyone recognizes that this is a very scarce resource. Being in the hospital is expensive, and sometimes it’s not even the best place to be if you don’t have to. So we’ve worked very hard because of issues of capacity constraints and access for the community to try to reduce length of stay, like many health systems, and we’ve actually achieved a reduction in length of stay of about a day in the last year. We’ve gone from 5.5 days to 4.5 days in a year, which is pretty remarkable.
We’ve found that the key is teamwork, communication between people. When a patient is in the hospital, it’s a big process with multiple people involved. It’s about getting the nursing team, the consult services, the interventional radiology, the imaging, the lab results, the care management, and the utilization management all on the same page. You can’t all be in the same place at the same time, so communication is key. We’ve developed ways to create team notes within Wellsheet that allow us to leave notes about what’s going on, and they can be handoff notes. They can also be group notes that are more related to a difficult disposition, a homeless patient, someone who’s going to a facility but hasn’t qualified for Medicaid yet, things like that.
HCI: You mentioned that you started working with Wellsheet just before the pandemic. Did that help with some of the critical issues that arose during the pandemic?
Front: Yeah, it was our go-to tool. Inside the electronic health record, you have all this information, all this discrete data. But you know, the big EHR vendors, Epic and Cerner, they can’t be agile and change the way they operate, or the way they view data in a heartbeat, but literally within a couple of weeks, we saw that we needed to make a list of all the COVID patients in the hospital. We wanted to know what the precaution status is. How long have they been here? When was the last time they tested positive? When is their next test due? What kind of oxygen supply are they on? And we created a list that we maintained, and all we had to do was just swipe patients on that list, and at a glance, you can tell how many COVID patients are in the hospital. How are they doing? How long have they been there? When did they get tested? What kind of test did they get? It was really ingenious.
HCI: Did that help you manage the resources you needed to deploy?
Front: Yes, the height of the pandemic was a real challenge from a capacity standpoint, and we had built the facility literally overnight to be able to care for 60 patients on ventilators. I think we got to about 50 patients, but we never got to the point where we didn’t have a ventilator for anybody. But you can imagine when you’re 55 and you can care for 60, being able to check at 10 o’clock how many people we have was really a game-changer for us.
HCI: I’ve been talking to a lot of health systems that are implementing generative AI tools to help generate clinical notes from patient encounters. Has Concord considered doing that for its patient encounters?
Front: Yes, we are actually looking at that. The Wellsheet team has done something that is not necessarily generative, but is definitely within the spectrum of machine learning and also saves you a lot of time. Wellsheet is really good at pulling discrete data that can be discovered from the electronic medical record. If you go to a patient’s medical record and that patient has coronary artery disease and is in the emergency department for chest pain, it may be able to calculate some of the cardiac risk scores with the available data before you even do it. I also like the way it pulls out lab data and presents it in an easy-to-digest way, exactly like you would write it by hand on a piece of paper in your pocket. We’ve had really good uptake in our cardiology group, because it brings together all the labs that they review regularly in one place.
Craig Limoli, CEO and Founder of Wellsheet, joined the conversation.
HCI: Craig, do you have to work with Cerner and Epic on these deployments? Are they happy with you or do you find it frustrating to hear someone like Dr. Fore say that this is so much better than having to go to Cerner to find this information?
Limolí: The integration itself is very straightforward. We have fully API-based integration models, so the integration is done through FHIR standards. It’s something that we can reliably access across all healthcare systems because it’s federally regulated. So it’s not like we have to bring in Cerner teams to implement our product. And because of this API-based integration model, we can also get it up and running very quickly, within a month or two. In terms of how they perceive it, I think at the end of the day, the EHR vendors that are successful are those that care about their customers getting the best outcomes, and this is clearly something that’s contributing to that goal. So we’ve been pretty well received across the board.
By the way, one of the capabilities that large language models have is to classify unstructured text in documents, extract key data elements, and display them to doctors as well. That can really enable accelerated review of the most relevant information. Pre-filling calculators and the like can be accelerated by that kind of application of these new AI tools. So it’s not just note generation that large language models enable, although we’re doing that too.
