The Hospital Will See You Now

Step inside Johns Hopkins’ 5,200-square-foot Judy Reitz Capacity Command Center and it’s a bit like you’ve entered NASA mission control.

Computers hum at 24 workstations. Dispatchers field calls and direct virtual traffic. Large wall-mounted flat screens display live data dashboards. In the background, an array of artificial intelligence (AI) applications track the occupancy of every bed in every department at every second, and use predictive modeling to anticipate when the next ICU bed or operating room will become available in the next hour—or, better yet, the next minute.

Since Johns Hopkins collaborated with GE Healthcare to launch the command center several years ago, it has improved efficiency so much that it has boosted hospital capacity by the equivalent of 16 beds without extending its physical footprint by one inch.

Meet the next-generation “cognitive hospital”—a seamless integration of the physical facility and advanced technologies that optimize resources, streamline operational processes and workflow and, most importantly, improve patient outcomes. With the cognitive hospital, technology is poised to transform not just the tools that physicians use, but the hospitals where they use them, making “smart” facilities themselves strategic partners in patient care.

With the cognitive hospital, technology is poised to transform not just the tools that physicians use, but the hospitals where they use them, making “smart” facilities themselves strategic partners in patient care.

If the hospital of tomorrow might be considered cognitive, today’s facilities are merely connected. More than 60% of global hospitals have already implemented Internet of Things (IoT) in their facilities, a number that’s expected to skyrocket in the coming years, according to an Aruba Networks study. At the same time, hospitals are investing heavily in healthcare AI, with spending to reach $6.6 billion annually by 2021, according to a 2018 Accenture report. All told, AI could slash U.S. healthcare costs by $150 billion a year by 2026, Accenture analysts project.

But don’t expect to check into a fully integrated smart hospital in the immediate future, cautions Gurpreet Singh, U.S. health services leader at PricewaterhouseCoopers (PwC). For most healthcare facilities, adoption of advanced technology is occurring piecemeal and slowly.

“It’s starting in the middle and back office, where we’re seeing automation of finance and HR functions, such as the creation of bots to perform manual tasks,” Singh says. “And we’re starting to see it move into the clinical space. ‘How do I use AI and other tools to predict early organ failure?’ Hospitals are experimenting, but it’s still very early.”

Still, snapshots of those experiments—by major academic institutions like Johns Hopkins, Mayo Clinic and Cleveland Clinic, as well as innovators like Miami’s Nicklaus Children’s Health System (NCHS)—reveal how the melding of facilities with IT could shape the next-gen cognitive hospital.

A High-Tech Nerve Center For Operations

On any given morning, the Johns Hopkins command center receives about 500 messages per minute from 15 discrete IT sources within the health system. The center’s AI crunches that data to manage patient movement and volume, and trigger interventions to prioritize high-risk cases.

“It is amazing to see the streams of data that come in in real time,” says James Scheulen, chief administrative officer for emergency services and capacity management at Johns Hopkins Medicine. “Nothing on any of these 23 monitors is more than three minutes old. Our predictive models are updated constantly.”

At the moment, the command center manages 1,000 beds in the main campus, which logs more than 100,000 emergency room visits a year. According to Scheulen, results so far have been impressive: Emergency room patients are assigned a bed 30% faster; transfer delays from operating rooms have been reduced by 70%; ambulances are dispatched 63 minutes sooner to pick up patients from other hospitals; and the ability to accept patients with complex medical conditions from other regional and national hospitals has improved by 60%.

Before the command center launched, Scheulen and his team spent a day at an airport control tower to observe how it managed its operations. The visit proved eye-opening.

“Air traffic control is trying to get airplanes safely onto a runway where they need to be, and we’re trying to get patients safely into a bed where they need to be,” he says. “Air traffic controllers try to do it as efficiently as possible or there might be delays, or worse, accidents. It’s the same thing with a hospital.”

When it opened in 2016, the Johns Hopkins command center was the first of its kind in the world. Now you’ll find one at Toronto’s Humber River Hospital, and GE Healthcare plans to open as many as 10 more command centers across 30 hospitals in 2019. Other major hospitals, such as Mt. Sinai in New York City, have reduced ER wait times by 50% thanks to the use of GE’s AutoBed software.

But, in Scheulen’s view, even smaller-scale hospitals can benefit by consolidating their IT operations and using AI for predictive modeling.

“Not everybody needs what we have, but in order to be successful, hospitals need to understand how to use data differently,” he advises. “You need somebody paying attention to it every minute of every day.”

Saving Lives With Sensors

Among the greatest challenges facing the hospitals of the future will be the management of terabytes of unstructured data generated by IoT sensors, connected devices, and mobile health apps and wearables. By 2020, 646 million IoT devices will be used by healthcare providers and consumers, and the volume of medical data is expected to double every 73 days, according to a 2017 report published in IEEE Pulse magazine.

But for Edward Martinez, senior vice president and chief information officer at NCHS, sensor technology presents a golden—and potentially lifesaving—opportunity.“

Using our Insight system, I can determine if a patient falls off the bed immediately, or a patient is about to fall because they’re stumbling,” Martinez explains. “I can tell if a patient has spent too much time in their bathroom, and I should probably check and see what’s going on.”

“Insight” refers to the Insight Location Intelligence system, which Nicklaus Children’s Health System developed in-house through its incubator NESA. It combines Bluetooth low energy (BLE) technology and a proprietary AI to automate tracking of doctors, patients, and medical equipment.

Martinez points out that while radio-frequency identification (RFID) technology has been available for decades, it provides only 15-foot tracking accuracy. By contrast, BLE technology tracks a person or object within six inches; the more precise location technology can be the difference between a near tumble and a broken hip.“

The cool thing is that as it’s becoming smarter, it starts anticipating what’s happening in the hallways and in a patient room,” Martinez says.

“All of our facility equipment is managed through an iPad, basically,” Martinez says. “I can track anything from a wheelchair to a major piece of equipment anywhere in the facility. And I can create reports about how a department is utilizing a piece of equipment.”

Among the other technology developed internally at NCHS is “crash cart” monitoring, which will soon be rolled out to select hospitals across the U.S. A crash cart is the emergency equipment nurses use when a patient is having a heart attack. They’re filled with everything from scalpels to IVs that could be used to save a person’s life immediately.

“The problem is, it takes a long time to make sure the cart has everything at all times, especially after it’s used,” says Martinez. Through its sensor technology, NCHS can scan a cart for all necessary equipment almost instantly. “What we’ve been able to do is take something that used to take a whole day down to four minutes.”

Innovations like these are among the reasons that NCHS was named the “Most Wired” hospital in the U.S. by the College of Healthcare Information Management Executives (CHIME) in 2018.

Partnering For Healthcare Profits

Hospitals are tilling the fertile ground for cognitive hospitals to emerge in collaboration with both major technology companies and ambitious startups.

Like Johns Hopkins’ collaboration with GE Healthcare, the University of Pittsburgh Medical Center (UPMC) recently announced a $2 billion partnership with Microsoft to create three new, highly specialized “digital hospitals.” Microsoft has also joined with Cleveland Clinic to integrate virtual assistant Cortana into its eHospital system.

“We’re finding that many of the top academic medical centers have created enterprise divisions or innovation divisions to investigate the use of new technologies in their health systems,” says PwC’s Singh. “They’re partnering with pharma and tech companies to come up with new products and solutions that could be a new revenue stream.”

Among the products that Microsoft is piloting with UPMC is EmpowerMD, an AI that can listen into and learn from doctor conversations with patients in the hospital room.

Today, many doctors barely look at a patient during their brief hospital rounds, tending to focus their eyes on a medical chart instead. The goal of EmpowerMD is to allow a physician to engage fully with a patient during an exam while the AI analyzes their communication and makes notes and suggestions in the patient’s electronic health record. Later, the doctor can review their treatment suggestions and accept or reject them.

But will doctors be willing and able to change their habits? As more AI-powered tools like these become available, perhaps the biggest bottleneck in adoption may not be technology accessibility so much as human adaptability, Martinez muses.

“Healthcare transformation is slow,” he explains. “The technology piece is actually easiest. The challenge is the human factor. How do you acclimate a human being to new technology? That’s an art by itself, which makes this a bit more complicated than other industries, but it’s very exciting when we do get it working correctly.”