CEO Punit Soni on Why Suki Is Creating a Nursing Consortium

One of the companies vying in the ambient clinical intelligence space is Redwood City, Calif.-based Suki, which supports more than 350 health systems. The company’s CEO, former Google and Motorola exec Punit Soni, recently spoke with Healthcare Innovation about why his company has formed a nursing consortium – a group of health system nursing execs who are helping the company develop its Suki for Nurses solution. 

Suki’s nursing consortium is comprised of health systems across all major EHRs, including McLeod Health, Boone Health, Citizens Memorial Hospital and Fisher Titus Medical Center.

Healthcare Innovation: Could you talk about why you decided to create a nursing consortium and some of the issues you are working through with them? 

Soni: The Suki app for clinicians was product number one. Suki app for Suki for nursing is product number two. With the Suki app for clinicians, it was pretty clear to us that we had to work very closely with doctors and clinicians of all types and shapes to understand exactly how to build it. But we didn't really have the heft when we were building it to actually have a consortium of health systems who would want to work with us. But when you get to a point where you have a massive footprint, people know you. There’s a reputation, which means now you can find a small group of health systems who want to collaborate and co-develop this. 

The other issue is that nursing is a problem that everybody talks about, but it's not clear to anybody how people are going to make money. If you don't figure out the economic setup of nursing, how are you going to build and scale a product that can serve nursing? The best way to do that is to have the right kind of people, administrators, who are actually thinking about this problem and the cost of it. 

We decided to pick a couple from each EHR ecosystem. We ended up getting McLeod, which is an Epic center of excellence for us. We got rural systems like Citizen Memorial and Boone, which are on Meditech. We got Fisher Titus, which is an Oracle-based system. There are at least six more who we haven't announced yet, but which are large, substantial systems of their own that are going to join. We needed to get different EHR ecosystems and different economic stratas to work with us to figure out not just the core product, but most importantly, a sustainable business model so that this can scale.

HCI: Did you have to find nursing informatics execs inside each organization to work with with you on it?

Soni: Yes. We have had to build a nursing team internally within Suki. We have to hire nurses. We have to work with chief nursing informatics officers and chief nursing officers across these places to learn how to actually do this stuff. It's about drawing expertise from a lot of different sources to create solutions. But I also don’t want to spend four years iterating on the product. We should be able to get something out within a year. That's why we had to be thoughtful about how we built this consortium.

HCI: Are there things about the nursing workflow that are more challenging to capture in an ambient way for documentation than a physician sitting in an office with a patient? 

Soni: It's different. Whether it's less challenging or more challenging is probably best said by the people who are in the middle of that stuff. A nurse is not doing as much narrative, and they're sometimes doing a lot of forms. They're doing hand-offs; they’re doing discharge summaries. They're writing these long physical exam forms, which have hundreds of fields in them, and they're doing it every day. That's a very different skill set from having a conversation with somebody, and wanting to summarize that into a clinical artifact. If you think about our ambient clinical intelligence platform, we built the form-filling AI into the product. Then we said, OK, now let's start working with nurses to figure out how to use that AI to do that. 

The other aspect is that there's less control over the form factor in nursing. At Suki, we decided there's no point trying to just build one form factor and hoping that people will use it. You have to meet the doctor where they are. So we built an iOS app, an Android app, a web app, a Windows app, a Mac app, a Chrome extension. But when it comes to nursing, many hospitals don't even allow nurses to bring their own phones or devices. Sometimes they hand devices to them. Sometimes they have Zebra devices. Sometimes they have these workstation on wheels or WOWS.

HCI: I think there's a sense of urgency on the part of health systems because they're dealing with shortages of nurses, and the idea that this technology could increase the efficiency of the nurses they have and maybe cut down on their burnout factor must be highly appealing to them.

Soni: Exactly. The health systems are spending billions of dollars trying to circumvent nursing shortage issues. If they did not see an increase in efficiency from this, then they will not invest on top of all the other money they are already spending. Therefore you have to have upfront an efficiency gain argument to make; otherwise this product is not going to be deployed. 

Another issue is that nurses are a tough constituency. They are not going to have the patience to use something and hope that it will get better. It has to have an immediate impact in terms of the quality of the work that they have to do. 

Also, the AI artifacts that are captured have to be super high-quality. The risk is too high. You can’t capture it and put it in the wrong place, or have a little bit of a mistake in a dosage. You can’t do that in this fast-moving environment. So the combination of this efficiency and a commercial argument with this satisfaction and happiness argument, combined with a quality argument makes it a really important but also a very tough problem to solve.

We also have a very significant platform footprint. There are a lot of healthcare tech partners who work with us. So we chose some of them who are leaders in the nursing area and said let's work together. An example of that is Avasure. It's a leader in virtual nursing. Why should I spend time trying to figure out how to get to these nurses when there are companies that have already done it and they can work with us and have our ambient clinical intelligence built into their products?

HCI: What about ambient clinical intelligence in the emergency department? Have you worked on that?

Soni: We have. You have to have a very different way of looking at scheduling, because folks are coming in and out. The second thing is, the notes structure is very different from what it is usually for other kinds of specialties. The third thing is, there is a lot of ambient noise. These are places that are busy, so you have to be very careful about the kind of device you use. We have noise cancelation software baked into Suki to help with this.

HCI: I understand you are also working on revenue cycle solutions as well. 

Soni: If you can create a note, you should be able to extract all the financial information from it. If you can extract all the financial information and create a note, you should be able to give the doctor pertinent information contextually in the encounter, so they can have all the information they need at their fingertips to do the right thing. If you can do all of those things, and you create a plan, and it said, “reschedule David for a visit three months from now,” why can't we send out an agent that will call David, set up the thing, and then automatically put it into the right system?

Now in this arc, a lot of work has been done on clinical documentation. That was the wedge. Now a lot of work is being done in revenue cycle. What we're doing is focusing on making sure that we capture everything that's happening and assigning appropriate coding in real time. We are in the room. The doctor is talking to the patient. We should be able to do a pretty good job of collecting all this with diagnostic specificity, so you know exactly the diagnosis the doctors were talking about. We have launched a suite of features around ICD, CPT, EM and this diagnostic specificity together. You can call it assisted revenue cycle. So that's what we have launched, and right now it's actually scaling up through our users and being used more and more over time.