Digital health, medical devices and OTA software updatesOpinion
The Device Chronicle interviewed IoT expert and staff systems engineer Suresh Alla to learn about the latest trends for OTA in digital health, medical devices and connected diagnostics.
Suresh has a broad understanding of the digital health landscape, seeing both the directions taken by the large medical device providers and the start-ups who are looking to disrupt and innovate. Suresh outlines broadly three categories of medical products where robust and secure OTA software updates are needed.
IoT devices in digital health
The first category is that of home kit devices. These are devices used by patients in home settings. An example would be a wearable device providing migraine treatment through wearable plugs into the ear. An OTA capability is needed to improve the UI/UX or update the software to improve the product based on learnings from the customer.
The second category is medical devices and this can span from low to high end capital equipment. Suresh says that in this context, OTA software updates and a secure remote connection to the device is especially important for failure investigation and correction. Several key strategic questions come into play: what would be the cadence for small feature releases compared to a full blown release? How is the OTA platform to be leveraged in these contexts? You would want to get visibility and remote control over the fleet of devices as opposed to a field force having to update and maintain these devices locally.
In clinical trials, IoT adds a new dynamism to clinical trials where devices can now provide real world data to send to the cloud for analysis.
Digital health, cybersecurity and patient data protection
Suresh also reflects on the strategic connection of IoT device management in medical to cybersecurity and patient data protection. He points out that when a device is acting as a gateway to the cloud, then the security aspects become really relevant. “Even where the amount of data is small, it can be easy for a mal-actor to hack into a device via the cloud. To counter this threat, you should not have bi-directional data exchange if it is at all possible. Data should only go up. Dual access is important for OTA software updates but it must be key encrypted. The user should enter a key and should get an upgrade to open the downward link. You need to be sure of the authenticity of the upgrade. In this regard, RSA encryption and encryption keys providing 3 layers of authentication are becoming the norm for security protection of devices. This type of validation is also needed for OTA software updates.
Importance of control over the IoT device fleet in digital health
Device planners are also looking for control over a medical device fleet. Suresh says that a statistical overview of the devices that have been upgraded is very important as “you would not want one set of users on one software version and the other set of users on another version. You need to understand what are the limits or constraints on the OTA software update.”
This kind of control and oversight is critically important for someone who is managing the scalability. Suresh adds “You need a robust and secure OTA framework to push the upgrades. The importance of a dashboard to see the timeliness of the update and the number of the devices updated. How many versions of the software should I keep in my repository? This all affects how you prioritise the work of a software development team.”
There are a plethora of applications emerging in digital health and Suresh believes that OTA software updates will be the key to keeping these applications in “good health” and future-proofed. “OTA is critical in the service and repair of devices in diagnostics space, there is the flexibility of OTA as a feature to capture certain logs. Is this device ready for an OTA update? Or should we hold on given they are on a critical timeline of finishing some tests at this point. So can we hold off on doing the update?”
Market ready now for OTA software updates
Often, Suresh admits that the IT infrastructure of the companies are not ready to handle such device management tasks. They often rely on the service partners to know what kind of framework should be put in place to achieve this. 60 or 70% devices out there are not designed for OTA, often it was not planned for, even by the global life science players. OTA software updates are now being thought about and being planned at the start of a new medical device or diagnostics product development project. It is now a strategic imperative whereas before it was often not seen as that.
Suresh also believes that medical device and digital health providers need to build in the process and to this point, there has been a lack of infrastructure to push code into an OTA framework, a lack of one end to end provider in the segment is a challenge. Can Mender be a complete “go to” service provider? Suresh asks. Trusted relationships are needed, can the OTA solution also integrate well with Azure?
AI and OTA software updates
What is the interface between AI and OTA software updates? Suresh explains that models are trained on the cloud, but the model inference happens on the edge device so again very important to be able handle OTA software updates here. Suresh continues to say that continual learning in machine learning and deep learning means that new models will be deployed from the cloud, and pushed towards the edge devices. The cadence needs to be frequent. The model will drift if you do not take care of it and have OTA updates to constantly refresh the models.
Use cases for OTA software updates in medical devices
In terms of where OTA software updates could take a powerful role in medical devices, Suresh points to the example of ECG cardiac monitors in the home. Traditionally, none of these devices have had OTA updates as it was mostly about the physical patches and there were no HMI screens. But through innovation, these architectures will be revamped to gain more insights from the ECG. Point of care ultrasound is another area, they are running machine learning models on these devices. This type of application will also need to be updated if machine learning is being run on the device. Capital equipment in hospitals and clinics will also see more data analysis on the edge which would see the need for OTA software updates.
In the end for Suresh, OTA software updates have too many benefits to ignore: “Take OTA (software updates) acting as an encryption framework for bidirectional communication, that OTA (software updates) already work with the cloud partners, and that OTA (software updates) can help securely render analytics that are shown on a device HMI.”
We wish Suresh well on his journey in digital health device management and engineering.