Transforming Healthcare with Cloud-Native Services at the Edge
In this article
Cloud adoption is accelerating at an extraordinary pace. While the latest race to the cloud can be attributed to a pandemic that has forced most businesses to embrace digital transformation as a tool for survival, there is little doubt cloud is here for the long run.
Even industries like healthcare, which tend to lag behind their early adopter cousins in the consumer space, have embraced multicloud technology with impressive haste.
But as many CTOs and CIOs can attest, full migration to the cloud is not always feasible or advisable. Real-life hurdles — such as compliance with complex industry regulations, decades of incompatible infrastructure, a lack of skilled resources, remote sites, limited funding or cybersecurity concerns — mean cloud simply cannot be the entire solution for every business.
This is particularly true in healthcare, where many storage and processing workloads require greater compliance and regulatory oversight. These industry challenges to cloud adoption include:
- Compliance: Hospital systems are highly regulated environments (e.g., HIPAA) with risk postures that shy away from permitting personally identifiable information (PII) or personal health information (PHI) to be stored in the cloud.
- The right data: Medical systems want to get the right patient data in the hands of the right clinician at the right time. But identifying the small subset of data that can make everyone's jobs easier is a challenge when the system is inundated with more and more data each day.
- Latency: Within hospital walls, real-time location services for both mobile and fixed assets (e.g., think high-risk or vulnerable patients, like newborns), generally cannot afford latency, poor performance or periods of inactivity regarding changes in location.
This article highlights how Microsoft and Intel engaged WWT to develop a hybrid cloud solution that demonstrates the power of Azure Stack Edge Pro's edge compute and AI technology to solve these types of real-life hurdles to cloud adoption.
While we explore a healthcare use case below, WWT is confident our solution can be game-changing across many other industries and scenarios where an edge presence, featuring two-stage analytics capabilities, can be used to deliver the right data at the right time to the right people.
As organizations generate more and more data, they need to be able to extract actionable insights and real-time inferences that can drive decision making. To enable this, organizations need modern application architectures that can handle the incredible amounts of data generated daily.
The robust IT architectures that can handle such workloads are increasingly powered by a mix of public and private cloud data center environments. This hybrid cloud approach to developing network infrastructures evaluates and unites the best components of each cloud platform into a holistic solution.
Within this hybrid cloud model, edge computing is a distributed cloud architecture — sold under a hardware/infrastructure-as-a-service model — designed to meet the needs of data-intensive organizations. Edge computing moves the computing power of the cloud closer to the "edge" of the customer's network and data sources.
The benefits of edge computing include localized access to data sources, better privacy and data security, reduced latency and bandwidth, conservation of resources through AI and automation, quicker actionable insights, and less risk of adversely affecting near real-time operational and safety decisions.
With hybrid cloud becoming the new normal across many industries, we will continue to see the proliferation of edge computing across industries as organizations seek new ways to process and ingest data closer to the source.
Microsoft's Azure Stack Edge Pro is a cloud-managed, artificial intelligence (AI)-enabled edge computing appliance that extends compute, storage and intelligence capabilities into discrete customer environments. Released in 2019, Azure Stack Edge takes the form of a cloud-connected hardware device that can be installed locally on an organization's network.
Powered by Intel CPU and FPGA compute, Azure Stack Edge Pro enables development teams to deploy cloud-native Azure services on-premises with ease. The solution is offered turnkey from an organization's Azure portal under a hardware-as-a-service model, meaning the device does not need to be purchased upfront to use it.
But Azure Stack Edge Pro does much more than on-premises work. These intelligent appliances are designed to be an extension of an organization's existing cloud presence, seamlessly integrating back into the services deployed in Azure. This enables organizations to optimize their compute and storage footprint where it brings the most value.
In January of 2021, Microsoft and Intel challenged WWT to build a proof-of-concept system using Azure Stack Edge Pro that showcased its power and value to potential customers.
- The goal: Deliver a fully functional proof of concept by the end of February for the Microsoft Ignite conference in early March.
- The challenge: It was January, meaning WWT had a small window of time within which to develop and deliver a solution.
WWT Application Services, our custom software development arm, took up the challenge. They quickly rallied an agile team of software engineers and delivery experts — along with strategic experts across our Cloud, Data Analytics and AI, and Healthcare practices — to tackle the problem from all angles.
Our team worked closely with Microsoft and Intel to identify the right use case. The collective team determined they wanted to highlight a story from the healthcare industry, given the ongoing pandemic and the potential impact that edge computing and AI can have in the space.
One of WWT's Chief Healthcare Advisors, Dr. Sanaz Cordes, helped identify remote patient monitoring for cardiovascular disease as a highly pertinent use case. Not only is there an urgent need for remote patient monitoring across medical use cases as a short-term response to the pandemic, but the technology has the potential to significantly improve long-term patient outcomes by facilitating a shift in industry focus from "sick care" to "well care."
As an example of the potential impact edge compute and AI can have in healthcare, consider heart failure — the number one cause of death in the U.S. According to Dr. Cordes, cardiovascular disease is the most expensive condition to treat in this country, and that's within a healthcare system where the top six chronic conditions make up 60 percent of visits to the ER.
Per Erik Vesneski, WWT Business Development Manager for Healthcare and Life Sciences, remote patient monitoring medical devices provide critical two-stage analytics capabilities, measuring both real-time and historical patient data. This data, through governance frameworks (e.g., Business Associate Agreements [BAA]) and Cloud Service Providers, can live either in the cloud or in an on-premises solution.
Erik goes on to explain:
Instead of waiting for critically sick patients to visit the ER for cardiovascular emergencies, an edge computing solution built on Azure Stack Edge Pro can give providers access to patient health-indicator data in real time. Providers can use this intelligent data to identify trends faster, which can accelerate intervention and possibly reduce preventable ER visits by as much as 60 percent.
The concept of leveraging real-time and historical patient health data to improve outcomes is a rapidly growing field, especially as more and more companies begin gathering data directly from consumers through wearable IoT solutions like blood glucose monitors, blood pressure monitors and heart rate monitors.
With remote monitoring of patients with cardiovascular disease identified as the edge use case, WWT's team set out to develop a hybrid cloud solution that would demonstrate how Azure Stack Edge Pro can accelerate a development team's ability to deliver medical device applications to market while meeting the industry's unique challenges around compliance and regulatory oversight.
If the team could develop and implement a compelling solution within the two-month timeframe, it would underscore the Azure ecosystem's ability to enable other small development teams to deliver incredible value in similarly small windows.
Watch this 12-minute video for a closer look at WWT's collaboration with Microsoft and Intel in developing our solution.
Some of the requirements our team faced in developing a solution included:
- On-demand analytics: On-premises relational data storage to enable on-demand analytics of PHI data. This allows for insights into patient status to be created and presented to medical experts regardless of internet connection status or PHI cloud storage risk posture.
- FHIR: Meeting the Fast Healthcare Interoperability Resources (FHIR) standard API requirements for exchanging healthcare information. FHIR is an important connector that bridges gaps between all the various parties that might need access to patient data.
- Multiple data sources: Accounting for many different data sources, coming both from within the hospital system network and from the public network.
Several distinct components comprise our Azure Stack Edge Pro solution:
- Hardware-accelerated machine learning (ML): To accelerate ML inferencing via onboard FPGA or GPU to get results closest to the data sources.
- Edge compute: To run virtual machines (VMs), Kubernetes containers and cloud-native Azure services at the edge locations.
- Azure-managed appliance: As an Azure-managed applicable, customers can use the Azure portal to order a Stack Edge solution as well as manage the appliance and workloads.
- Cloud storage gateway: To transfer data to the Azure cloud over the network while retaining local access to blobs and files.
WWT's primary solution-building efforts centered around showcasing the computational power of the Azure Stack Edge Pro through the collection of mock patient vitals in on-site clinic and remote monitoring scenarios.
Azure services made this process surprisingly uneventful. For example, the API for FHIR server and SQL server services streamlined the implementation of easy-to-store and accessible FHIR-compliant data.
Deploying containers to host the custom apps designed to analyze vitals for clinically relevant events was likewise super simple. This allowed WWT's dev team to focus on UI features like clinician dashboards that update when triggering events are detected, adding additional patient information to the dashboard when relevant, and notifications.
Below is a diagram of the open-source hybrid cloud solution the team developed:
At the end of a fast two months and just in time for Ignite, we're happy to report that our team successfully met the challenge of developing and implementing a fully functional Azure Stack Edge Pro remote patient monitoring proof of concept for cardiovascular disease.
WWT's solution, designed to be installed via Azure Stack Edge Pro on a hospital's own network, allows providers to ingest and intelligently process the mountains of data generated from countless devices each day.
We believe this solution truly has the potential to change the game in healthcare systems in several ways:
- Proactive care: Once enough historical data is captured from the passive monitoring of patient health metrics, the solution enables predictive analysis that can intelligently accelerate the identification and triage of potential problems before they occur. Such predictive analysis has the potential to change how patient care is delivered in the U.S. and around the world.
- Expedite care, free doctors' time: With historical data on hand, plus predictive analysis dialed in to the patient level, doctors will be able to diagnose, update treatment plans and facilitate patient care much faster.
- Free IT resources: To the delight of CTOs and CIOs everywhere, the automation and aggregation capabilities of our solution can allow internal IT teams to refocus their valuable time and energy on delivering strategic wins instead of updating servers, trying to prevent them from crashing, and patching different arrangements of mismatched IT products that were never really meant to work together in the first place.
Our Azure Stack Edge Pro solution for monitoring cardiovascular disease data represents just the tip of the iceberg of applicable edge compute use cases that exist today across industries. Add in the market's ever-evolving ecosystem of cloud managed services that can further enhance security and reduce time to market, and modern organizations are well-positioned to innovate and scale faster, become more agile, and expand into new markets with multicloud and edge solutions designed to meet their unique needs.