IS YOUR NEXT MEDICAL DEVICE AN EDGE DEVICE?
Looking at the regulations and standards, that must be followed when designing a Medical Device, it may seem impossible to use Cloud Services and IoT for this. After all, IEC 62304, IEC 82304-1 and ISO 13485 are all very rigid when it comes to configuration management, version control, management and a range of other areas. The whole concept behind the software standards 62304/82304-1 is a stand-alone device with explicit functionality limited to measurement of something, e.g. a vital parameter, such as ECG. Moreover, the standards not only cover the functionality of the Medical Device, but also the design and development process.
And yes, it is indeed a challenge to use Cloud Services and IoT in a Medical Device.
But what exactly does a Medical Edge Device involve?
There are at least 3 possible scenarios for a Medical Edge Device. It may
• Use the Cloud as a Device Management Platform
• Use the Cloud as a Data Collection System
• Use the Cloud as an Integrated Part of the Software System
Let me explain this further.
Using the Cloud as Device Management Platform
The architecture for this, using Microsoft Azure IoT as example, is called Azure IoT Edge. Microsoft is focusing a lot of resources on this and has a system, that works very well (and yes, we have it running).
Off course, you need to build your infrastructure to support it, but it is possible even on small embedded devices running variants of Linux.
The advantage of this is primarily three things:
• You have a housekeeping infrastructure for software updates, patches and recalls
• You have access to your Medical Edge Device usage and performance (part of your post marketing surveillance)
• Analysis of collected data from the Edge will give insights on how to improve the software
On the other hand, you need to control mainly two things:
• Configuration Control of development tools
• Configuration Control of SOUP components
Control of your configuration of SOUP components will require focus. Your architecture and segregation strategy has to support this.
And you also have new challenges, which should not be underestimated
• Registering, provisioning, of new Medical Edge Devices on the Cloud Platform
• Secure data from measuring on the Device to deletion in the Cloud
Off course, all major Cloud IoT providers have solutions for provisioning – some more mature than others. Azure IoT Hub Device Provisioning Service is an example of a powerful solution.
Depending on your volume of Medical Edge Devices and level of security, securing data requires different, mostly established and proven strategies. For example, securing safe transmission from sensor to cloud has well known solutions from the industrial area.
Using the Cloud as a Data Collection System
Collection of results in a Cloud database is not particularly interesting in the context of this article, as the “intended use” of this typically does not fall under the definition of a medical device.
Using the Cloud as an Integrated Part of the Software System
This is where it becomes difficult. But why would you want to do this?
• You may have an analysis algorithm, that requires more resources than your device provides
• You may want to utilize the Interfaces of the Cloud Architecture
• You may have a business model that requires online access
In the world of real-time units like ECG and SpO2 monitors, pacemakers, Cardiopulmonary bypass machines (heart-lung machines) or anesthesia workstations, this does not make any sense. The speed and reliability of internet access to the Cloud is not guaranteed. Possible delays conflict with the need for fast response time.
However, for devices that measure, analyze and report (diagnose) in a sequence that takes minutes, the Cloud Architecture does provide some obvious advantages. Examples of this are blood analysis and heart and lung abnormality detection devices. Typically, In Vitro Diagnostics (IVDs) devices fall in this category.
But are you able to move the very heart of this diagnostic device to the Cloud. Has it been done? Is it possible?
Yes – it does seem possible. And yes – it has been done.
Figure 1, Monitoring using Azure IoT, click for info.
Microsoft references a monitoring case (digilog) where Azure IoT is used as basis for the off-line analysis.
Figure 2, Spectrometer using AWS, click for info.
The RND Group references a spectrometer solution using Amazon Web Services.
Figure 3, Flex and Google collaboration, click here info.
Even Google has a reference with a Medical Edge Device (or at least the architecture for it).
Also, the vast range of healthcare apps could be argued to fall in this range (although most of the have not been registered as Medical Devices).
And yes, it is possible to get a Cloud Based Architecture FDA cleared.
Figure 4, Cloud DX, click for info.
“Pulsewave is a unique “pulse acquisition device” that records up to 4,000 data points from your radial artery pulse, then securely transmits the raw pulse signal to our Cloud Diagnostics servers, which display nearly instant results: Heart Rate, Blood Pressure, Pulse Variability, Average Breathing Rate*”.
Having the fundamental issues regarding configuration control of Development Tools and SOUP components and provisioning of Medical Edge Devices addressed, the remaining issues can be related to risk management. Examples of these issues are:
• What offline functionality is required?
• How do we ensure data integrity in communication of the result?
• How do we mitigate unknown behavior of the Cloud platform?
It will be interesting to follow how well the Medical Device Community will adapt all the new exciting possibilities and how this will impact the traditional business in the Medical Device market.
Glaze Partner Flemming von Holck, firstname.lastname@example.org, +45 30 66 30 61