Tomorrow’s ICU is a place wherein patients and caregivers see very little technology. It will be about making AI, devices and data on the provider side invisible to enable clinicians to focus on delivering care.
Alarms disappear. Nurses can seamlessly move in and out as needed. Physicians are called to attention when that moment of need arrives — and intensivists, even when they are in short supply — share expertise virtually such that everyone is practicing at the top of their license.
Houston Methodist was headed in that direction well before COVID-19, according to EVP and Chief Innovation Officer Roberta Schwartz.
“That’s the journey we’re on,” Schwartz said. “There are many steps along the way to virtual ICUs but going down this road has been phenomenal.”
Schwartz and Emma Fauss, CEO of Medical Informatics Corp. (MIC), shared their experience and insights during the Health Evolution virtual gathering, Scaling up the ICU.
The experts discussed what the ICU of the future might look like and the requisite steps to accomplishing that work:
- Deploying consumer- and clinician-centric tools
- Understanding what ICU data means
- Preparing to scale virtual ICU capacity
While many health care organizations began virtualizing the ICU prior to the pandemic, COVID-19 has essentially offered a glimpse into the ICU of the future.
Deploying consumer- and clinician-centric tools
An intensive care unit might be the last place that comes to mind when thinking about consumerism. But the need for remote monitoring to leverage data at the bedside is a necessary advancement to improve efficiency and reduce costs.
“It’s a natural movement toward consumer technologies and a next step of saying, ‘How do we actually take what we’re collecting from the patient? How do we better incorporate that into the care practices to drive care management of patients in a way that we would expect in a world where we have supercomputers, AI, technology, all augmenting our daily activities?” Fauss said.
On the business and operations side, those technologies also help clinicians understand what to pay close attention to for planning and care management purposes.
“We’re seeing a lot of hospitals and innovators in this space really start to think about how to impact their management of staff, and better augment coverage of areas that they might have problems staffing at scale,” Fauss said. “We’re also seeing a lot of interest in predictive analytics, automated event detection.”
Specifically, Fauss said that health systems are looking to start building analytics and bringing in data at a higher quality and normalizing it, so it can be used by downstream systems. The intention is to reduce the burden of medicine by automating as much as possible, so clinicians and employees can be more productive.
At Houston Methodist, for instance, doctors are responsible for 30 to 60 patients each night, and they cannot be everywhere at once. Nurses often need to monitor two or more rooms simultaneously.
Schwartz said Houston Methodist is working toward a nirvana in which algorithms not only enable clinicians to monitor patients from afar but also more effectively identify adverse health effects before they occur.
“There are really two pieces of virtual,” Schwartz said. “One is the monitoring, algorithms and risk scoring. The second is the cameras.”
Houston Methodist and MIC built the first 17 algorithms and, in the time since, Schwartz said the cardiology team is asking for another 10.
As for the cameras, options include using a tablet or iPad or installing more robust systems in the room. The latter does require closing the room for 24-48 hours to put the cameras in.
“You have to do those side-by-side,” Schwartz said. “The solution does not have to be from the same people know. In fact, all of the other vendors still use a secondary camera vendor to put that in.”
Understanding what ICU data means
Fauss added that learning what ICU data reveals about care quality, especially in critical care environments, with human beings connected to very expensive machines tracking more than temperature, heart rate and blood pressure will be a foundational element of the future ICU.
“There’s a lot of rich information coming from those devices, waveforms, alarms, device settings,” Fauss explained. “This is the next generation of analytics and AI in this space, to help augment risk scores when you have high quality data coming from across devices, synchronized and normalized.”
Providers, in turn, can more effectively manage blood pressure and brain trauma patients when receiving data about blood flow in the brain and cardiac monitor data.
“If you’re not collecting that data properly, if you’re not building the scalability to use that data, that is something that you can’t overlook,” Fauss said.
To help augment patient risk scores, Fauss said the next generation of analytics and AI in this area will include high quality data coming from across devices, synchronized, and normalized.
Fauss pointed to the example of a night nurse picking up the phone to call an intensivist or doctor in a different part of the health system and describing what’s happening.
“It’s a very different conversation when you’re actually looking at the same information because now instead of ‘he said she said’ or ‘I don’t understand what you’re saying,’ barriers, they are both focused on the care of the patient,” Fauss said. “It changes the speed of responsiveness.”