Edge computing—which you likely already use—drives workflow efficiency, reduces costs, and protects data. Here is how.
By Christophe Dore and Robert Cohen
Imagine a device at the point of care able to capture crucial data about the patient, analyze it, and then perform actions in real time based on parameters previously entered by a clinician—all without a direct connection to the hospital’s cloud or on-premise server. At the same time, the device captures important patient data so it can be automatically added later to the electronic health record (EHR).
The chances are your hospital already has several types of these “edge” computing devices in numerous locations, including inside a patient. Edge computing—which brings data processing as close as possible to the end user to optimize remote resources dependency, network latency, and bandwidth demands while increasing security—is not a new concept. Thanks to advances in computer processing speed and decreases in technology costs, edge computing has numerous use cases for hospitals that want faster responsiveness, greater data resiliency, and stronger data security.
Edge computing is already a complement to traditional cloud or on-premise data center-based models. As medical devices in the hospital grow in number and the complexity of those devices continues to increase, however, hospitals may want to expand their edge computing footprint in order to maintain the connection to the patient at the bedside and improve outcomes through more timely actionable insight.
If the same devices that allow for that connection can also improve patient safety by securing patient and device data both at the bedside and in contact with the larger hospital network, edge computing devices become even more essential.
Early Space-Age Technology Paradigm in Healthcare
Edge computing is a term that became common recently. However, the concept has been used for much longer without a name, including in famous endeavors such as the 1969 Apollo 11 mission to take U.S. astronauts to the moon. Astronauts needed immediate location data for navigation, so the lag time transmitting data to Mission Control in Houston and back would have been too long and too high of a risk to the astronauts’ safety. Those computations needed to be performed in real time within the spacecraft, each major module onboarding a dedicated computer.
Fast-forward several decades, and the rapid improvements of processing speed and battery power have pushed the edge computing concept beyond space travel to many other industries, including healthcare.
In hospitals, the edge is mainly the point of care, which is a busy place where primary critical workflows are executed, requiring critical decisions to be made based on a lot of local data which needs to be processed to provide actionable insight. Edge computing enables these workflows to run smoothly because, while edge devices are often in constant communication with a centralized server, they do not rely on data or instructions from a centralized data center to perform functions or to analyze data from the patient. On the edge, these devices can better inform the clinicians making care decisions, providing the right information at the right time and in the right place.
Edge computing devices also offer greater data resiliency. Just as the astronauts on the Apollo mission could not pull the ship over if they lost a connection to Mission Control and wait for it to return, clinicians cannot stop treating a patient if a medical device loses its connection. By processing and caching data at the edge, edge computing devices allow clinicians to continue their work uninterrupted even with signal loss during patient transport or a network outage, receiving real-time, intelligent alerting and information about that patient.
Moreover, any data collected during this time would be cached and shared once a network connection is regained, preventing data loss. As hospitals continue to think through business continuity, including care continuity, requiring data resiliency, this ability is an essential complement to the elements already included in those plans.
Edge Everywhere You Look
One of the most common examples of edge computing’s use in these environments is the smart infusion pump. When a regimen of drugs is prescribed for a patient, the nurse enters the patient and drug information into the pump, often through a barcode scan. It is up to the clinician to check the right dosage is administered at the right time, at the right place, for the right patient and the right route, and to ensure all this information is captured. But the pump, after having collected the prescription, can provide a substantial help.
The pump does not need to be connected to a cloud or on-premise server to perform these functions or to offer a double-check on “the five Rs” of medication administration. This not only allows for better patient monitoring, but also improves patient safety by delivering timely and automatic treatment when necessary.
Other examples of edge computing are implantable cardioverter defibrillators (ICD) to detect and treat arrhythmia. The ICD implanted in the patient’s body is connected directly to the heart and constantly monitoring the rate, rhythm, and morphology of that signal to determine how it needs to act. The ICD performs this function without sending data to a centralized location for analysis or instruction.
If the ICD detects a treatable event, it has the computational power to identify and correct the arrhythmia with a small shock to the heart in as little as a few seconds. While primarily operating without input, the ICD can share data with a server or database for functions such as maintenance, reprogramming, or sending alerts.
In the two cases above, a single device is connected to a single patient, operating autonomously. A newer utilization of the edge computing model is integrating the patient medical devices at the bedside or in an operating room with hospital information systems through a “medical device integration” (MDI) hub. By displaying data from multiple devices though a single hub, clinicians rounding or operating on these patients can better contextualize that data to receive a more complete picture of the patient’s condition, allowing them to more easily and quickly identify and even predict health deterioration.
Using edge computing for such a medical device integration initiative significantly reduces data latency and can translate to immediate analysis of key vital signs and other physiological data used to formulate early warning scores. When clinicians at the bedside are alerted to meaningful information to enable prompt, high-quality care, costs are reduced.
Moreover, creating an edge computing hub enables devices that otherwise would or could not be connected to the network to prioritize patient data at the point of care, trim it down to what is really needed for immediate review, and send the less timely or non-urgent information to the centralized server later for further analysis. Again, this saves bandwidth by limiting data transfer, but also reduces the number of non-actionable alerts delivered to clinicians.
Security “Off the Grid”
Edge computing used for MDI can also offer significant efficiency and cost benefits from a security perspective. Considering that an edge computing MDI hub can be autonomous in capturing and processing data from the medical devices prior to sending the refined data to various clinical applications, the medical devices themselves can be left off the hospital network. In this configuration, multiple medical devices are connected to the edge computing hub, but only the secure edge computing hub is connected to the network, keeping all of the devices connected to it protected from a malicious actor crawling on the network.
Medical devices, after all, are not immune to such attacks. In a survey of 232 security decision-makers in healthcare, 82% reported they have experienced an Internet-of-Things (IoT) focused cyberattack in the past year. Of those organizations attacked, 30% reported experiencing compromised end-user safety, 43% operational downtime, 42% compromised patient data, and 31% brand or reputational damage.
To protect devices with traditional client-server architectures, hospitals could segment the network, partitioning the devices from other information systems that do not require direct access. Network partitioning, however, is time-consuming and complex, requiring valuable and limited IT staff resources. Considering the hospital medical fleet is constantly changing—adding or decommissioning devices, and devices being more mobile across the network—keeping up with the fleet’s changes is a constant challenge.
Instead, edge computing nodes sitting between medical devices and the hospital network can be easily added one per bedside to this network, in a segmentation built once for all. As they keep their connected medical devices isolated from the network—still providing their crucial data—those connected devices can be added, removed, moved, and updated, all while remaining invisible to the network, removing the need for constant re-segmentation.
A Trend Worth Following
While edge computing may not be new—and is likely already more prevalent in your hospital than you may realize—its numerous clinical and technical advantages make it a model worth expanding. As hospitals continue to look for ways to improve data accessibility and enable informed clinical decision-making while reducing IT costs and protecting data, edge computing should be at the forefront of any medical technology conversation.
Given these benefits, hospitals should expand their edge computing footprint, creating an infrastructure that, alongside cloud or on-premise client-server architectures, is fully enabled to deliver the best of both worlds for patients and clinicians.
Christophe Dore is the cybersecurity manager at Capsule Technologies; Robert Cohen is the senior product manager for Edge Computing at Capsule Technologies. Questions and comments can be directed to 24×7 Magazine chief editor Keri Forsythe-Stephens at [email protected].
