Children’s Health of Orange County, a California-based pediatric healthcare system, is working with other medical institutions to gather clinical images — MRIs, echocardiograms, and the like. The goal: to create a repository of images to analyze for insights that could train algorithms capable of aiding clinicians with diagnoses.
Although the initiative is in its early stages, CHOC vice president and CIO John Henderson is already thinking about the technologies his organization would need to enable such an intelligent system. Edge computing, with its ability to deliver real-time analysis of large files, would be a key component to making that work, he says.
Edge computing is increasingly seen as a critical component in the data ecosystem created by connected devices. It’s generally defined as a distributed computing paradigm that puts computation and data storage physically close to where the data is generated and used.
In a typical edge computing setup, an edge server or gateway located at the edge of the network stores and processes the data collected by end devices such as sensors, cameras, or smartphones, and often sends some or all data to the cloud for additional processing. In some cases the initial processing is performed on the end devices themselves, but either way, the initial data processing and analysis happens close to where the data is collected, rather than at a distant cloud server or corporate data center.
In healthcare, the technique allows data collected by sensors and other medical devices to be analyzed where patients are. It’s this proximity of computing and storage capabilities that many health IT experts predict will help transform healthcare by delivering near-instantaneous processing.
“Edge is in the early phases of application in healthcare, but it’s something we’re looking at,” Henderson says. He sees edge computing as a critical enabling technology in multiple healthcare scenarios.
Edge could support advanced remote-patient monitoring by processing data from medical devices such as glucose monitors and blood pressure cuffs and then alerting clinicians to problematic readings. It could enable real-time management of medical equipment as the various pieces move through hospital facilities. And it could deliver on-demand content for augmented reality and virtual reality training sessions, with the proximity of the edge computing devices ensuring there’s no lag in the experience.
“Edge allows healthcare organizations to roll out new features and to be more responsive,” says Dave McCarthy, a vice president with research firm IDC, where he leads a team of analysts covering shared (public) cloud, dedicated (private) cloud, and edge strategies.
Everywhere and anywhere healthcare
A confluence of healthcare trends is driving the interest in — and need for — edge computing, according to experts.
The exponential growth of endpoint devices generating data is one of the big factors fueling edge computing needs. Lynne Dunbrack, group vice president for public sector at IDC, uses American Hospital Association estimates of connected medical devices per bed along with AHA’s tally of 920,000 staffed hospital beds to calculate the number of total bedside devices at 9.2 to 13.8 million.
At the same time, the healthcare sector is accelerating its adoption of algorithms, machine learning, and artificial intelligence. The sector is also adopting more augmented reality and virtual reality for clinical care and training. All of those uses require significant real-time data-processing capabilities to effectively deliver usable results in a healthcare setting.
Meanwhile, healthcare is expanding beyond hospital walls and physician exam rooms, with virtual visits, remote-patient monitoring, and consumer wearables creating even more data that clinicians want to capture and analyze so they can provide care anywhere and everywhere.
Figures from IDC Futurescape: Worldwide Health Industry 2021 Predictions underscore these trends. According to the report:
By the end of 2021, 7 of the 10 leading wrist-worn wearables companies will have released algorithms capable of early detection of potential signs of infectious diseases including COVID-19 and the flu.
By 2024, the proliferation of data will result in 60% of healthcare organizations’ IT infrastructure being built on a data platform that will use AI to improve process automation and decision making.
To enable immersive training for healthcare professionals and enhance customer experience, 60% of providers will move from proof of concept to full deployment of AR/VR technologies by 2025.
By 2026, 65% of medical imaging workflows will use AI to detect underlying disease and guide clinical intervention, while 50% will use teleradiology to share studies and improve access to radiologists.
Although much of that data will end up in centralized servers (whether in the cloud or a hospital’s on-premises data center), health IT experts say a good chunk of it will need to be collected, processed, and stored closer to where it’s generated and/or consumed.
Experts say healthcare needs that proximity of processing power due to the high costs of moving large volumes of data to the cloud, networking limitations, and latency concerns.
“One of the main reasons that edge computing has become so important over the past couple of years is due to the limitations that people have experienced with the cloud,” McCarthy says. “You need edge computing anytime you’re generating a lot of data and need real-time analysis of it.”
Mike Angelakos, chief technology officer at Geisinger, a healthcare system based in Danville, Pennsylvania, says he’s seeing those issues at play as he and his executive colleagues advance connected healthcare and other digital initiatives within their organization.
He notes that some medical images are 3GB to 6GB each, with clinicians often reviewing multiple images of a patient at any given time.
“And because they’re so large, and because of how they’re read and rendered on the screen, that’s not something you can move to the cloud,” he says. Transmitting such material to centralized servers where applications could analyze them would not only cost too much money but also be too slow in emergency situations. “When we back up radiology images for trauma, those extra seconds mean something to those patients.”
Edge in the early stages, but growth ahead
According to IDC’s June 2021 Edge Spending Guide, healthcare provider spending on edge computing (hardware, software, and services) will reach $10.3 billion in 2025, with a five-year compound annual growth rate of 17%.
Despite that seemingly large dollar amount, health IT leaders say the use of edge computing in healthcare is still in its early stages. Furthermore, they note that much of the computational power that exists on the edge in healthcare today is embedded in the end devices themselves.
“If you walk into a hospital room, there are four or five devices such as cardiac monitors that are edge computing devices. They all take patient data and visualize it at the endpoint. We’ve blurred the lines between medical equipment and IT equipment,” says Steve Hess, CIO of UCHealth, a Colorado-based network of hospitals, clinics, and healthcare providers.
Health IT experts expect investments in edge technology to grow in coming years — and with that growth, some say edge will help transform how, where, and how quickly care can be delivered.
Angelakos points to one pilot project as a case in point. Geisinger is testing a platform that uses automation and facial recognition technologies so that patients can register for clinical visits via their smartphones and receive messages from clinical staff in advance of their exams.
“In this case, the patient’s smartphone would become the edge computing device. It will help streamline the experience from the patient perspective — and streamline it from the operational standpoint, too,” Angelakos says, adding that the platform will help limit physical touchpoints within the registration process (which reduces the spread of germs) and cut back on paper use (which saves resources).
Others envision even more transformative opportunities for edge computing in healthcare.
Weisong Shi, a Wayne State University computer science professor and an expert on edge computing and connected health, points to the future of emergency services as an example. Medics working in an ambulance could take pictures that could be collated with a patient’s biometric data from onboard medical devices and analyzed by an edge device. Results then could guide the medics on treatments to give en route and could alert emergency room clinicians on how best to prepare for the patient’s arrival.
Or, Shi says, edge devices performing in-field real-time data computations could let medics evaluate and even effectively treat patients on site, helping cut down on unnecessary hospital visits.
Others cite the potential of edge computing in enabling more effective virtual care and remote-patient monitoring as evidence of the technology’s transformational capacity.
Like many other healthcare institutions, UCHealth has expanded its use of virtual visits during the Covid-19 pandemic, Hess says. At the same time, the healthcare system has expanded its capacities to remotely monitor patients. To help better serve patients in both those scenarios, Hess says he envisions having digital tools enabled by edge computing and other technologies such as AI that would gather and analyze patient-generated data to produce insights for the patients and clinicians alike.
“The way we think about patient care, it’s really going to be in the future about keeping you healthy,” he continues. “If we can figure out what your normal healthy is, and where you’re deviating, we can intervene appropriately. All of that requires some edge computing — we need data from you, data that’s going to an intelligence layer on the edge.”
Hess says edge computing — whether it’s taking place on a patient’s mobile phone or a purpose-built device — will be particularly important for enabling data analysis in rural areas with slow or even no internet access.
Barriers to expanding edge computing in healthcare
Wide-scale adoption of edge computing in healthcare is years off, perhaps even a decade or more away, according to CIOs and health IT researchers. As CHOC’s Henderson says, “We’re still really early; this is a 10-year and beyond arc.”
Multiple factors are keeping edge deployments in check, experts say.
To start, healthcare institutions generally have other investment priorities. CIOs say funding for medical devices and diagnostic equipment that deliver immediate impacts on patient care typically have precedent.
“It takes time before new technology becomes mainstream; it takes time before it really becomes accepted, because there’s an investment involved, and when you look at the various investments that healthcare has to make, you want to make sure they’re in the right places,” Henderson explains.
For example, Shi says he and some colleagues developed a prototype for connected emergency medical services (EMS) that they labeled STREMS, for smart real-time prehospital communication system for EMS — but the proposal has yet to receive funding.
On a related note, many institutions have yet to invest in the other technology components that drive the need for computing power at the edge. Many of the medical devices at patient bedsides come with embedded computing power and, thus, are already delivering value. But there’s an emerging class of AI-enabled tools for analyzing patient-generated data that has yet to be widely deployed. The same goes for many endpoint and internet of things (IoT) devices.
In fact, many medical institutions are still working to fully integrate all the applications and systems they have had deployed for years — integration that is essential for leveraging patient data on the edge, Shi notes.
Moreover, healthcare institutions must develop strong data policies and procedures for initiatives that use edge computing, not only to protect sensitive patient data but also to identify the right data needed for each use case. Otherwise they could end up with too much data creating noise instead of value.
“The challenge is creating the whole ecosystem; the chain is so long. You may make an advancement [in edge computing], but that alone won’t really change the whole thing,” says Shi, a founding steering committee chair of the ACM/IEEE Symposium on Edge Computing and IEEE/ACM Conference on Connected Health: Applications, Systems and Engineering.
In fact, Michael Minear, senior vice president and CIO at Lehigh Valley Health Network in Allentown, Pennsylvania, says he doesn’t consider edge computing a critical need. “We’ve put a lot of focus on the clinical devices [such as] CT scanners and digital infusion pumps. To us, those are highly strategic and highly important in terms of patient care,” he says.
Furthermore, Minear says he already has the IT capacity to handle his institution’s data. He says the clinical devices within his healthcare system upload nearly 27 million data points to its electronic health records (EHR) system a day. But LVHN runs those devices on local networks and local servers. Given that, Minear says he doesn’t experience the cloud-related latency issues that would necessitate investments in edge computing to resolve.
Edge’s value for connected care
Dunbrack acknowledges that edge computing isn’t critical for organizations operating a lot of their systems locally, but she still sees edge as a valuable and transformative component in modern healthcare.
Edge reduces latency, she says. It saves on the need to transmit all locally generated data to the cloud (or on-premises data centers), thereby preserving needed bandwidth and keeping transmission and centralized storage costs in check. And it can bring security benefits by limiting the movement of data to local devices, as well as limiting the range of data being collected in one spot.
But edge computing’s biggest value, Dunbrack says, is really in its capacity to process and analyze the data physically close to the patient without any lag time due to network speed; edge computing delivers even if the external network is down.
That, she says, is essential for enabling the transformative concepts of connected care as well as healthcare anywhere and healthcare everywhere.