HIStalk Interviews Terry Edwards

Terry Edwards is president and CEO of PerfectServe. 

Tell me what PerfectServe does.

At a high level, it’s all about making it easy for clinicians in a hospital environment — nurses and doctors — to connect with the right doctor at the right time and in the right way. It’s about enabling communication processes that are accurate and simple.

Then from the perspective of the doctor, it’s about enabling processes that gives physicians the ability to really selectively filter and control the communications that they receive throughout their work day. When I use those terms, ‘the ability to selectively filter and control,’ it’s really about enabling them to mold the flow of communication that a hospital directs at them around their own personal unique workflows.

Give me an example of two or three problems that this would solve.

Let’s say that we are three physicians, and maybe we’re cardiologists. Our workflow rules might be that we each take our own calls during the week beginning at 6:30 a.m. until 4:30 p.m. except for the days that we’re off.

John is off on Tuesdays, Bill is off on Wednesdays, and I’m off on Thursdays. On those days that we’re off, one of the other two covers for the person who’s off. Then we have our consults during that same weekday time period, handled by our PA. Our PA will route those to one of us based on the things that are unique to those consults. In the evening hours and on the weekend hours, everything is handled by one doctor on call and we change that schedule daily. One guy would cover the whole weekend, like Friday night through Monday morning.

In addition to that, maybe John likes to have everything just route to his cell phone in real time. He doesn’t mind being interrupted; he like being really accessible. Maybe Bill likes only the urgent things routed to his cell phone, and for everything else, he wants a text message to show up on his phone. Me, maybe I’m just a numeric pager kind of guy — I just want a call-back number and that’s it, but beginning at 9:00 p.m., I want you to call me at home.

Instead of all those instructions living in pieces of paper, Rolodexes, and call schedules in three-ring binders and things in and around the hospital that nurses and other doctors need to refer to and interpret every time they need to call one of us, we build all of those rules into the PerfectServe system. We give the hospital one number to dial.

This is where the ease comes in. If I’m trying to reach Bill, I just call on the system and I say, “Bill Smith.” The system will confirm that, it will interpret the rules, and then route the call either to Bill or maybe the PA or one of us according to the workflow rules, and then our individual instructions for that specific moment in time. It does it automatically.

That sounds something a little bit similar to what Vocera offers. Who would you consider to be your competitors?

Our competitor is the status quo, just in terms of all the manual processes that are in place in health systems today. Vocera is not really a competitor. They’re really a complementary technology in that Vocera is more like a device or a node on PerfectServe’s network or PerfectServe’s platform.

We can answer calls and messages that would originate on Vocera devices and we can send calls and messages to clinicians on Vocera devices. But Vocera, for the most part, is an ‘in the four walls of the hospital’ solution; whereas PerfectServe includes the communication to physicians regardless of whether they’re in the hospital or out in their practices, or mobile — wherever they are. We’re device agnostic.

Someone comes to you and they decide what end-user devices they need, whether it’s voice-over-IP or plain old telephones, or whatever. Then you’re what rides in the middle to decide, when those calls come in, how to route them?

Exactly. It’s about really intelligent workflow-based, rules-based routing according to the rules of the recipient. Physicians are the biggest users in terms of receivers of calls and messages, and then those who might be around them as part of their workflow — like a nurse practitioner, physician assistant, or some other sender.

Like in the example that I gave, the consults during the weekdays would route to a PA. If, for example, I was the nurse on the floor, I might not know that those are the rules, but when I called in and said, “Bill Smith,” I might hear prompt, “Are you calling about a consult?” Yes. Then, that would route that call to the PA, and maybe the outcome would be a voice message where I can leave all the detail, and the PA picks those up throughout the day and then determines where to forward it to.

So if I’m calling a doctor and their preferred method at that moment is to roll to a pager, do I then get prompted for their pager number or does it just say, “Your message will be delivered,” and then it logs you off and it just does what it does in the background?

Yes, it does what it does. When we configure the system in a hospital, we configure it so that we’re pulling all the appropriate Caller ID information. We map that back to departments so that we greatly reduce the need for nurses to key in numbers and make mistakes and things like that, as well as speed up the system.

Everything that happens on the platform is documented automatically. We’ve developed an analytical tool that then allows, say, a chief medical officer at a hospital to look and see how many calls and messages we are initiating to which doctors and from which departments, and drill down into that by specialty and contact method. When you have that data, then you can begin to use it as a tool to enhance process improvement for those clinical processes in which communication are really a fast, and/or efficient communication is central.

What about patients? Would physicians’ practices use this as the front-end for patient calls?

Yes. There is another module of the service that is designed to replace the conventional answering service. That provides a portal, or an access point for patient calls coming into the practice. The caller experience is different than the hospital portal, which is optimized for communication that originates from clinicians at a given hospital site.

What would be the compelling reason that they would do that? I know a lot of them are really tied to their answering services.

It’s the elimination of human error. In the current communication processes, whether they’re those that originate at the practice, or those that are being done manually in the hospital environment, the complexity of the physician workflows coupled to the manual processes — you’re basically dealing with a high volume of highly variable processes handled in manual ways, so you have a high degree of communication breakdown.

When you have a breakdown, you have situations where the wrong doctor, for example, is paged or called when he should be off. Or maybe a doctor, when he is on call in the middle of the night, his rule is that he wants his pager to go off and not his home phone to ring. The reason why is if the home phone rings, it wakes up everybody in the house. Whereas another person, when they’re on call and they’re at home, they might sleep in a different room from their spouse and they want their home phone to ring.

With doctors, those breakdowns really result in a high degree of frustration; and then the overall process inefficiencies cost them time. Time is the only thing that the doctors have, really. That’s their key asset.

Looking at it from the hospital perspective, tell me about what the advantages would be to hospitals and to patients for physicians being able to directly access physicians and nurses to make their communications with physicians.

Those benefits, at the highest level, really accrue up into the coordination of care, the reduction of risk, and tighter physician alignment. When I talk about tighter physician alignment, I mean really being able to integrate the care delivery processes of the hospital with the communication workflow of the physicians. These things are important because of the high volume of communication that takes place just in the coordination of care.

For example, a typical 300-bed hospital will probably initiate some 15,000 voice-related calls to doctors every single month, or 180,000 a year. These are interactions where it’s about: there’s been a change of status on this patient; or we’ve got these results and I need to act, but I need to know what order you’re going to write.

Those events are all related to patient care. When you improve the accuracy of those processes, then nurses aren’t wasting time making phone calls to the wrong people. You reduce the delays in communication and decisions can be made faster. That has an impact on not only reduction of risk and quality, but faster throughput; and it can make a contribution to reducing length of stay and everything related to improved efficiency.

You mentioned the measurement of length of stay. Have any users done any quality measures or satisfaction measures?

We had started a study with one of our small hospital accounts and we didn’t complete it because of some changes in their operation, but that was a study that suggested that we had been able to impact length of stay when physicians used the tool in the right way; but some of that data was inconclusive.

I was in a meeting actually yesterday with one of our clients at the St. John Health System where we were talking about value. It was a value analysis discussion with some of the executives. They had indicated that their length of stay had gone down with PerfectServe. There were many things that they had done to drive that, but they felt that PerfectServe was a major contributor.

Has anybody ever asked you to do anything with emergency communication to an entire group, where one call blasts information out in different ways to more than one person?

We have what we call a ‘team alert’ function, and it’s really designed for smaller clinical teams; although I think we have some applications of maybe 50-100 people. We’re not designed for mass emergency notification. There are other companies that do that.

Where our team alert applications work is when we have processes — such as a process around mobilizing an open-heart team or maybe a cath lab team where there are multiple people who might serve in one of the handful of roles — we need to know who’s covering which role at what time. Then we need to have fail-safe or measurable processes, where we can get a message out and know and receive confirmation that it has been retrieved or acknowledged within a certain time period, and if not, it escalates.

We have a lot of those applications that are out there and running in hospitals for not only what I mentioned in terms of cath teams and open-heart teams, but rapid response teams and other code teams as well.

I had read the case study on your site of Fairfield Medical Center and the cardiac cath lab, which I thought was pretty cool. Can you describe that and think about other ways hospitals might do something like they did?

In that particular application, they had … it was again, manual processes, and they literally had one person who would go down and manually call through a list. It just took time. They were able to basically replace multiple calls, and the time and the acknowledgements back and forth, with a single phone call.

There’s another case study on there around Henry Ford Macomb Hospital, and that’s an interesting story. They had an environment where there was a lot of dependence on paging. The call-backs and the repeat calls were creating satisfaction and quality problems between the ED physicians and a lot of the primary care doctors.

The ED physicians got together with the primary care doctors and they said, “Look, we’ve got in PerfectServe the ability to filter communications that go to you from the ED. Here’s the deal. If you guys will enable rules to route calls from the ED to your cell phones, we’ll make sure we’re on the other end of the line when we call you.”

They put that process in place and saw significant increase in real-time calls to physicians, which had a major impact on helping them improve flow, as well as physician satisfaction with both the ED and the primary care doctors.

What are the time and the requirements and the cost of implementing?

The costs depend on the size of the hospital and the medical staff, but the time requirements are … I mean, there are certain things that can change this a little bit, but typically, in about a three- to four-month time period from the project kickoff.

In fact, we just went live at Monroe Regional Medical Center just about two or three weeks ago. I guess it was about three weeks ago because it was in May. That project kicked off the end of February, and in addition to the planning, the big job of implementation is gathering the workflow rules of all the members of the medical staff and educating them on what the platform can do. That starts with some high-level awareness, workshops, one-on-one meetings with heavy admitters and other key doctors and practices; then gathering and collecting all this information and configuring it, and then we have the go-live.

In Monroe’s case, we went live on one day with the entire medical staff, 552 doctors, and they were immediately, out of the gate, processing at a rate that was equivalent to probably about 20,000 interactions a month. It will just go up from there.

On that particular thought, one of the things that has really amazed me about their organization, even to this day, is that it’s the kind of project that a hospital can do and actually succeed. I mean, it has a major impact on nursing and a major impact on physicians. They can actually drive a process change in a very short amount of time and realize and see those benefits just immediately.