- Posted by Jim Jordan
- On February 26, 2021
- 0 Comments
It is undeniable that the Internet of Things (IoT) has single-handedly transformed the way we live and work. While reaping the benefits brought forth by the ubiquity and cross-industry appeal of IoT-based technologies in recent years, let us remind ourselves that a venture as awe-inspiring as IoT integration can also be met with skepticism. As IoT continues to evolve, so does the production and accumulation of data, thus leaving systems open to vulnerabilities. The IoT security and privacy risks we have previously discussed may be especially damaging in the healthcare industry, with hacking and system breaches potentially compromising confidential patient data. In relation to these issues, healthcare delivery organizations (HDOs) may want to take into account where and how data is being processed and stored within their IT system. This is important because it can actually affect workflow efficiency– like their response time when dealing with emergency situations. To avoid delays that could disrupt hospital workflows and hamper care delivery, HDOs can strengthen their IoT infrastructure with edge computing.
Edge computing: What it is and how it works
Edge computing means moving storage, compute, and networking resources closer to the data source and away from the centralized data center. Gartner offers a more basic definition, referring to edge computing as solutions that enable data analysis at or close to where data was generated. As a result of denser and more sophisticated IoT systems, higher volumes of data are generated and consumed which can be too much for the cloud — a centralized data center– to process. When the cloud is swarming with data amassed from a surplus of interconnected IoT devices, issues with congestion, latency, and bandwidth may follow. In an article published for TechTarget, Stephen Bigelow explains that the continued relevance of edge computing lies with its ability to address emerging network issues which not only involves reallocating massive amounts of data produced from thousands of sources but also improving response time, thus eliciting urgent action to applications or workflows whose efficiency greatly depends on real-time data exchange. Furthermore, edge technology can also act as a data filter that only sends relevant data to the cloud. This in turn reduces data transmission and storage costs.
Exor International has published a blog post that enumerates different types of edge devices. Examples of these devices include (but are not limited to) the following:
- a router establishing a link between a public network and the internet
- a firewall as a network accessibility detector, functioning as the point of entry to the said network
- IoT/smart devices
- bigger machines and equipment
A commonly cited use case to illustrate edge computing functionality is the system that empowers self-driving cars. Autonomous, driverless cars and cloud computing are believed to be incompatible. Because these cars rely heavily on data (apart from AI) in order to fulfill the role of an actual driver, there is no room for lag times as it could put the lives of passengers in danger. If the cloud is congested with data, network connectivity may suffer as a result; and if the internet connection is down or unstable, how on earth can a driverless car drive itself under normal circumstances? This is where edge computing comes in and saves the day. The edge is better suited for self-driving cars because it is decentralized and localized– meaning, it performs data processing around the vicinity in which the data source is located or even on the edge device itself. This is what distinguishes the edge from the cloud.
For the purpose of comparison, let’s do a quick run-through of what the cloud is and does. Cloud computing is “a huge, highly scalable deployment of computing and storage resources at one of several distributed global locations (regions).” The Forbes article Cloud Computing vs. Edge Computing: Friends or Foes explains that the cloud utilizes remote servers or data centers to analyze, process, and store high volume data. While the edge is close to where data is generated, the cloud data center could be several hundred miles away from the data’s point of origin. With this kind of infrastructure, the back-and-forth transmission of enormous amounts of data– from the source to the cloud then finally, to the end-user– can take a toll on bandwidth and internet connectivity; not to mention, data processing in this scenario is more time-consuming. When it comes to technologies like autonomous cars whose efficiency is dictated by real-time insights, these issues can be disastrous. There is literally no time to waste.
I would like to point out, however, that in certain situations, cloud computing is a much more suitable option, given its breadth and elasticity (e.g. big data analytics). But when an enterprise’s service productivity is determined by time-sensitive tasks requiring immediate responses (e.g. driverless cars and the Internet of Medical Things), edge computing is the more viable solution.
The advantages of edge computing in healthcare
Like self-driving cars, IoMT can tackle latency and connectivity problems that could negatively impact care delivery via the implementation of edge computing. Decentralizing data processing at the edge could convey to HDOs the following benefits:
Enabling rapid data transmission. Edge computing at a network edge, coupled with the velocity of 5G connectivity, can optimize clinical and operational workflows, encouraging faster decision-making and access to patient information and records. An article published by STL Partners on edge use cases in healthcare expounds on how HDOs can leverage edge computing capabilities in a “connected ambulance.” Emergency situations are a matter of life and death, and the edge can significantly accelerate HDOs’ responses through its low latency and high-speed connectivity, both of which contribute to speeding up data exchange without worrying about network bandwidth usage. HDOs can also leverage edge computing in remotely facilitating emergency workflows. For instance, edge technology allows paramedics to live stream incoming patient’s status to alert hospital staff as well as analyze patient’s vitals to determine diagnostic procedures among other things.
Imagine how much time can be saved if paramedics and first responders had the means to relay vital patient information to ER doctors in real-time as opposed to waiting until they reach the hospital to do so. When human lives are involved, time is of the essence. This statement alone should be enough motivation for HDOs to seek the appropriate technological solutions that can help them save more lives.
Enhancing patient data security and privacy. Apart from aiming for optimal clinical and operational outcomes, HDOs should also ascertain the security of patient data. STL Partners provides insight on the potential cause of security vulnerabilities in HDOs, highlighting the current state of in-hospital monitoring devices– most of which are either unconnected or carry high-volume data which utilizes a third-party cloud storage system. Cloud infrastructure is typically built to withstand data influxes, but for IoMT systems that generate and store sensitive patient data (e.g. personal information, records, and medical history), uploading everything to one expansive data center may not be the wisest move.
With all these things considered, it becomes even more apparent that edge computing can be an asset in curtailing IoT security and privacy risks. Decentralizing patient data processing can be valuable mainly because edge technology is capable of establishing smaller networks and data centers to help manage patient data transfer, analysis, and storage instead of having all patient data contained in one centralized cloud data center. In addition, processing data at the edge means the data being transferred to the cloud can be secured via encryption, hence fortifying the system’s defense against cyber threats. In terms of safeguarding data privacy, edge computing supports security and privacy regulations that demand certain data points to be processed on-site or close to their source.
Promoting a positive patient experience. Apart from improving clinical and operational workflows and averting the risk of system infiltration, analysis of patient data at the edge promotes convenient and more accessible healthcare. In an interview with HealthTech magazine, Dr. Shafiq Rab, senior vice president, and CIO at Rush University Medical Center, firmly asserts that edge computing will greatly transform chronic disease treatment (e.g. diabetes and cardiac diseases) since IoMT devices, combined with 5G, can augment remote patient monitoring and at-home care. IoMT devices and wearables with edge computing capabilities can feed real-time insights to clinicians regarding a patient’s current status by setting up alerts and notifications to signal missed medications or a spike in blood glucose levels to name a few. This is particularly helpful for rural communities where the nearest hospital or medical facility is, in fact, a hundred miles away. Even more so, a dynamic edge-enabled telehealth infrastructure is vital in improving healthcare accessibility to the elderly.
Edge technology limitations
The case in support of edge computing as a potential feature of a smart hospital may appear convincing, but there are certain challenges to its deployment that begs to be discussed.
Security. Although on-site edge computing can keep sensitive patient data secure since the said data is kept close to the source, it still doesn’t guarantee full protection of that data. The sad truth is IoT devices will always be vulnerable to hacking and breaches. Therefore, it is essential to adopt a feasible edge computing framework anchored in proper device configuration and management that also complies with data security regulations. This means regularly updating software as well as encrypting data regardless of its mobility.
I think it’s safe to assume that completely eradicating security risks is a pipe dream. What HDOs can do, however, is to have a comprehensive risk management plan that, at the very least, reduces the probability of classified patient data being exposed. The process may be laborious, but as care providers, HDOs have the responsibility to uphold patient confidence in how their personal data is being managed during all stages of care delivery. Bear in mind that improving patient experience is the driving impetus that has propelled this RTHS vision forward in the first place, so it is only reasonable for HDOs to stay true to that pledge by also securing patient data at all costs.
Bandwidth availability. In HealthTech’s roundtable interview with experts, Dr. Rab seems convinced that one of the principal barriers to edge deployment in HDOs is the uncertainty of having enough bandwidth at all times. This is exactly why extensive integration of 5G networks is necessary; to get the most out of the edge, a high-speed connection is obligatory, and that’s exactly what 5G is.
Cost. Expanding an edge solution to accommodate more devices involves investing in new hardware and software which can be quite costly. Edge computing technology is not really as flexible, scalable, and adaptable as the cloud. If HDOs were to deploy the edge, they will have to be prepared to allocate extra financing for edge infrastructure expansion and device upgrades.
Data management. Knowing what really happens to our data at the edge is a valid concern. Let’s face it– in general, most of us are not really aware of how our data is being used by the entities to which we surrender information. For this reason, HDOs should provide clear provisions about the nature of patient data being processed, its storage location as well as the management of connectivity surrounding the data as they take steps in mapping out an enterprise-specific edge deployment strategy.
Edge as the future of AI
Today’s digital landscape is a reflection of decades-long persistence which further emphasizes how technology is never static; it bends and evolves alongside our ever-changing problems. Like any tech innovation, edge computing welcomes even broader possibilities particularly in our quest towards fully integrating a real-time health system. The future looks auspicious for the edge, with its role in AI evolution taking center stage. Lionbridge describes Edge AI as AI algorithms with localized data processing capabilities at the edge, devoid of an internet connection. In other words, Edge AI enables data production offline; there is no need to download data from and upload data into the cloud. In healthcare, Edge AI can boost clinical workflows significantly. This combined Edge and AI infrastructure can provide ancillary support to radiologists as it allows them to input, analyze and store composite imaging data (e.g. MRI and CT scans) locally, expediting diagnoses and saving time in the process. In relation to telehealth, AI-based technologies can personalize reminders for medication based on unique patient activity and facilitate remote patient monitoring through autonomous telehealth robots.
To keep up with the growing complexity of IoMT, HDOs should best be creative in their efforts to integrate a real-time health system (RTHS). Consequently, it is imperative for HDO CIOs to consider all possible solutions that can help concretize their vision, and at the same time, ensure that such technologies are embedded with security. The approach to making our RTHS vision a reality may actually differ from one HDO to another. Regardless of how you’re doing to do it or where you’re at right now, one thing’s for sure: preparation is key. Arrange an IT infrastructure strategy that reinforces your objectives. Above all, let it be motivated by purpose.