In a Utah warehouse that serves as Intermountain Healthcare’s supply chain center sits the nerve center of the 22-hospital system’s tele-critical care program. Connected by audio and video to more than 260 intensive care unit beds, a team of 22 doctors and 20 nurses monitor patient conditions, provide real-time clinical support and ensure consistency of care.
I recently spoke with Bill Beninati, M.D., medical director for critical care telehealth at Intermountain, about both progress made and the challenges they have faced since launching the program in 2014.
Beninati said that although Intermountain had paid attention to the developing national experience and literature in tele-critical care, its executives hadn’t felt the need to start their own program until a CEO of one of its own community hospitals wanted to open an intensive care unit (ICU) and asked Intermountain to consider telemedicine to help support them.
It is difficult for a community hospital that doesn’t have much volume to support the level of expertise needed for critical care nurses and physicians 24x7, he said. “The most common serious problems that we are involved with are sepsis and septic shock from a variety of sources that can be managed well in a community hospital if the right things are done,” Beninati said. Other common cases they help with are respiratory failure and overdoses.
That request from one of Intermountain’s community hospitals came in late 2011. What followed was an in-depth system-wide discussion to analyze the benefits of critical care telemedicine and then a decision about how they were going to implement it. Part of that analysis was whether to build a homegrown system vs. buying a commercial system. “We started in earnest in 2012, and it took us nearly a year to build a consensus around a system and decide we wanted to build our own system rather than buy a product,” Beninati said. By May 2014 they opened the center with a nurse-only model through August. (Initially it was easier to hire nurses than physicians, he said.) Synchronizing the hiring of additional physicians and figuring out how there were going to integrate into the system was a time-consuming process. By December 2014, they had 24x7 coverage.
Opening the center in a location separate from a hospital was a key recommendation from consultants, he said. “We heard loud and clear that we should not locate this at a hospital because of the fear that these clinicians could be absorbed or distracted. By locating in the supply chain center, there is no concern about that. When we are on duty for tele-critical care, we have no other role. The nurses have on average over 20 years of experience as bedside nurses, but with few exceptions they do this now as their sole occupation. A few still moonlight back at the bedside. The physicians, however, all practice more time at the bedside than in tele-critical care.”
I asked Beninati what made Intermountain come down on the build side of the build vs. buy decision. “Fundamentally it was flexibility,” he said. “There was really only one commercial system we were considering. The vendor offered a very well developed product that has capabilities that to this day we have not replicated in our homegrown system,” he explained. “But in the end, more than that capability, we valued the flexibility of our own internal system and at this point I feel that it was a good decision. That flexibility has been a benefit to us, and that missing capability has not been a significant burden.”
Although all the hospitals in the Intermountain system share an EHR, that system does not have the capability to do true online order entry. “So if I want to write an order, I write it in the computer, but it is only visible to me and other physicians until I print it and fax it to the bedside,” he said. But the progress notes are immediately visible to the bedside or anybody who accesses the record.
“One of the challenges we have faced is that around the time we made the decision to do tele-critical medicine, Intermountain also made the decision to get a new EHR. They basically had a contest between four different possibilities. One was a major upgrade of our own homegrown EHR. Interestingly enough, I was on a team working on inpatient order entry. All of a sudden that project stopped when we made a decision to go to a vendor EHR solution [in partnership with Cerner]. That brings amazing capability, but has been a setback for some things. I mentioned that we have not replicated some things that are in a commercial product. That is in part because our medical informatics resources have been heavily involved in the transition of the EHR, so someday that will be a benefit to us, but right now it is a challenge for us.”
Over 18 months, Intermountain reports that its community hospitals have cared for more complex patients and still achieved a 33 percent decrease in hospital mortality (2.96 to 2.06 percent), and a 34 percent decrease in ICU mortality, while seeing a net million-dollar cost savings. Beninati spoke about some of the limitations of this study: “We did this as a pre/post implementation analysis. We would rather have used randomization, but the timing of the physician hires and the planned implementation of our Cerner product overtook our ability to do a randomization. So our ability to draw a causal inference was weakened by our inability to do a randomization, but our effect size was large enough that if you don’t conclude that the effect we saw was due to the tele-critical care implementation you would have to conclude that the results were due just to general progress in critical care was responsible,” he said. “Any reasonable person would know that although we continue to get better, it is not that much better in that short a period of time.”
Beninati said Intermountain would continue to do more research. “From an operational perspective, we want to be able to know what things we are doing that are of little added value and stop doing those things, and extend what we are doing that is effective.” He said they don’t need a data analyst to tell them that in case after case of diabetic ketoacidosis, they had nothing meaningful to add, and the patients cleared up very quickly and left the ICU with no complications. “That was an easy one. But we do need more advanced analytics to help refine other things.”
Another area they are investigating is the benefit of pushing farther and farther from the ICU. “We are seeing in our smaller hospitals and even hospitals that don’t have an ICU, in the course of their operations they encounter patients who are critically ill or injured,” he said. “For instance, we have four hospitals in central Utah that are near major interstates and there are accidents on those interstates. We are working in concert with our own air transport program to push critical care support as early in the episode of critical illness as possible. That is our most active area of research — to look at very early notification that someone very ill or injured has appeared at our smallest facilities and we would focus immediately, with the family physicians at our rural critical access hospitals to begin optimizing care of those patients as the transport team arrives at the bedside to take care of them. That is our biggest growth area.”