From Complexity to Confidence: The Smart Approach to Testing Smart Devices
REGISTER NOW
Pharma’s 6 Crucial Digital Health Challenges (and How to Overcome Them)
Pharma companies need to deal with several challenges that all affect one another and yet, at the end of the day, need to be addressed holistically in order to be solved efficiently, quickly and effectively.
- Jan 21, 2022
- 11 min
In addition to developing innovative and effective digital health products, pharma & medical device companies need to deal with added crucial responsibility towards their customers, challenges that revolve around functionality, effectiveness, and customer experience as well as many compliance issues that need to be addressed thoroughly.
While all these challenges are substantial, years of helping pharma & medical device companies to increase efficiency, improve quality as well as gain regulatory compliance, enable us to be in a unique position to successfully guide you through them.
These are the challenges you’re likely to be encountering and how to best address them.
1. Building the right Quality Management System (QMS)
A QMS is a procedural approach in its essence, which determines the set of processes and work methods that an organization utilizes to meet regulations and quality standards. In many cases, the cost is in the ability to innovate as well as, unfortunately, long release cycles.
The two main challenges here are:
- Balancing velocity and regulation
- Finding an optimal method of operation
Balancing velocity and regulation
It’s critical for pharma companies to build a QMS that can balance correctly between the high velocity and short cycles of the technology world and the regulatory constraints and demands of the pharmaceutical world.
Finding an optimal method of operation
There are differences when it comes to an effective QMS for pharma companies compared to medical devices or other industries, since a QMS for a pharmaceutical company that produces medications differs from that of companies that focus on software/hardware. The procedures that pharma companies need to adopt while entering the Digital Health world are sometimes in contradiction or vastly different from those they regularly use.
It’s essential to find a balance between the pharma quality system (Current Good Manufacturing Practice (CGMP) Regulations) and the software quality system (Software as a Medical Device (SaaMD) as well as the medical devices quality systems (such as ISO 13485).
The Solution
Finding an ideal solution for these challenges requires deep understanding and experience with both worlds: the speedy technology and development world and conservative world of pharmaceuticals and regulations.
By finding the golden path that can utilize the former while accommodating the latter is not an easy feat, but it can be achieved by looking at current procedures and processes and building the right QMS. It’s often the case that pharma companies, who in most cases lack the proper experience, will need to approach this project using an expert external advisor.
2. Agile development and methodologies
This is a hot topic and in the pharma world. On the one hand, pharma is trying to create innovative and effective development methodologies such as agile, but on the other hand, it’s still bound by regulatory requirements and very strict QA demands.
The solution
Bring together the best of both worlds. It’s recommended to build a general QA methodology that combines both agile and waterfall methodologies in a balanced way. Usually that means starting with exploratory testing, followed by functional testing as part of the development sprints. Towards the more advanced stages of the SDLC, you switch to a waterfall methodology and conduct end-to-end testing in a stable environment of code and features freeze.
This enables the shortest possible release cycles with the least amount of effort and resources, while achieving the best quality. As is always with this triangle of cost-speed-quality, it’s a balance that’s hard to achieve and maintain, but a combination of agile and waterfall methodologies (employing different methods of testing) is an effective way to do that.
3. Testing of connectivity
A considerable part of the architecture of pharma and digital health solutions is based on multi-component systems and the successful communication between all those parts. significant challenges for pharma companies that have to do with connectivity, and this is true in general for IoT, are due to the reliance on BLE (Bluetooth Low Energy).
There are several things to consider and to test to see how systems handle different real-life scenarios and situations, both in the physical aspect and the logical aspect:
- Essential functionality regardless of communication. The critical elements need to be able to function without communication (a medical device has to be able to work independently) and the systems need to be able to handle poor or no communication.
- Slow or uneven communication. For example, if you have a wearable device and you’ve entered an elevator where there’s no connectivity or if you’ve moved the device further away – and then moved back and reconnected.
- Correct identification and pairing of devices. For example, you may have several smart inhalers, in which case the systems or apps need to be able to identify the relevant device for each user and not to mix them up with other similar inhalers that might be in range.
The solution
In order to test these situations and scenarios well, it’s necessary to use actual, real medical devices as well as mobile phones instead of virtual mobile devices. This is important because it is essential to ensure tests are conducted in an environment in which the hardware used to test is identical to the hardware used in real-life situations, and there are a lot of different configurations of situations, devices and functions that need to be tested. Working on projects with clients, we’ve seen actual differences in the way virtual devices behave compared to real devices, so it’s vital to be able to test how they behave in actuality.
But not all devices are created equal; not all of them can communicate properly using the BLE standards. For example, some mobile phones use a chipset that either doesn’t support communication well or supports communication in a different way. Old phones are different form new phones, Android phones are different from iPhones etc.
Once you gain an understanding of what isn’t working and why, this knowledge can and should be used to check how these issues might also affect other scenarios and/or devices. For example, Samsung devices use two different platforms for processors: Qualcomm’s Snapdragon and Samsung’s Exynos processors – and that affects how devices interact and operate and may change in different scenarios and situations when connected to other devices with similar or different processors or chipsets.
Another important capability that’s needed is to be able to create smart testing scripts to cover many different real-life possible scenarios and situations: overloading a phone’s CPU, increasing distance between devices and checking connectivity while several other devices are transmitting between them, checking devices in elevators, changing time-zones, etc.
Since some tests are very time-consuming and also physically difficult to do,test automation should be used to test a large number of devices simultaneously. For example, if there are five inhalers in a room, all connected to devices, we need to be able to simulate this type of usage.
We use hardware-based simulators that simulate the device we want to test for connectivity as well as functional testing by automatically or manually creating “events” for the medical device using the simulator. Since in some situations it’s very difficult to operate and test the actual device (like in the case of very large and complex equipment where it’s highly unlikely to have multiple devices in a room to test them), hardware simulators actually broadcast Bluetooth in the same way that the actual device would, so the application responds in the same way it would to an actual device.
4. UI/UX
Good UI/UX for pharma applications is essential, since correct operating ability is crucial. It’s imperative to ensure that the system works and is designed well, is simple, not confusing, and is highly intuitive to prevent human error. This is true not just in terms of how the product or device is used in general, but also for testing the small things that might have a big impact.
For example, it’s important to ensure that the messages, notifications, colors, buttons and other elements the user interacts with are clear, precise, not ambiguous or in any way make usage confusing or difficult.
Much like the connectivity issue above, the challenge here is partly the fact that you’re dealing with different devices, sizes and configurations, which can create UI issues that need to be discovered and dealt with.
Performance and response times are another important aspect of UI/UX. It’s critical to ensure, in multiple situations and scenarios, that the system works well and doesn’t cause delays, which can lead to both misuse of the system/device as well affecting the user’s willingness to actually use it.
Languages and localization also play a big part in this. For example, an application that supports dozens of languages would need to ensure that the messages, notifications and other copy to be accurate in each language and for each country. A possible issue, for example, is the amount of space available or needed for the same messages in different languages, since that might change due to different length of the copy for each language.
The solution
This is where AI comes in. First of all, AI enables to plan tests much more easily, compared to regular automation. With AI you can also visually identify the different elements that comprise the UI and check for issues or changes that might affect it. Just one example of that would be that AI can identify and point to post-localization changes in the login screen’s buttons micro-copy (or other elements) that might affect the UI/UX.
The advantages of AI in this context is that it can look for any errors or unwanted changes (visual or textual) due to a failed test script. Automation alone cannot do this.
5. Identity management and data privacy
We live in reality in which people feel exposed, and it’s vital for them to be able to protect their privacy, data and online identity. This is especially true when it comes to their very personal and sensitive medical information.
That’s why the identity of users in the pharma and medical world isn’t kept alongside the data itself. The data stored in systems is virtually anonymous, so in order to connect the data to the user you also need access to the identities database, which is usually managed by a third-party identity management platform. So, in order to work with the system for different functions, for example, logging in to an application or analyzing user data, you need to be authenticated by that third party, which is a potential failure point that might cause issues in this process.
The solution
Automation comes to the rescue once again here. What we do to help this is to build an automation script that runs a relatively simple scenario that performs authentication and a data transfer process. This script can be run frequently (in some cases we even automatically ran it every five minutes), which ensures that the identity management platform is alive and kicking.
Other methods include data tests to ensure the data is not associated with a user identity in any way.
Adhering to privacy regulations, such as GDPR, as well as cyber security (since the communication needs to be encrypted and secure) also needs to be baked into the testing. Cyber security analysis includes pen testing, hardware testing, checking for backdoors, monitoring communication between the medical device and other devices as well as the cloud, and to ensure it can’t be hacked (you wouldn’t want a pacemaker that its settings can be messed with).
6. V&V (Validation & Verification)
V&V is a formal testing stage that plays a crucial role for regulatory purposes. V&V processes can be very complex and lengthy, which has a big impact on release times and the ability to react quickly to the demands of the market. In cases where tests are not done and/or documented correctly and fully, the V&V becomes a major blocker that can hinder launching or updating products.
The Solution
If you have the ability to confidently rely on the entire quality operation that we’ve discussed above, and these tests are very organized and consistently well-documented, then the scope and extent of V&V become appropriately smaller and more efficient.
This provides the business confidence even with tests that were done and documented well, for example, a year ago, that they will pass the V&V processes successfully and quickly.
Key takeaways
Pharma companies need to deal with several challenges that all affect one another and yet, at the end of the day, need to be addressed holistically in order to be solved efficiently, quickly and effectively.
The weakest link in this chain can affect critical aspects of a product, which means that meticulous and expert pharma quality engineering is essential to ensure a flawless execution, achieve regulatory compliance and assure quality peace of mind.
We’ve listed the key challenges for pharma companies: effective QMS, reliable connectivity, intuitive UI/UX, working in agile, assuring IDM and data privacy and V&V. There are no shortcuts to addressing these challenges, except by ensuring thorough, detailed quality engineering and testing that, combined, will enable efficient quality assurance, improve processes and the SDLC, and ensure regulatory compliance as well.
If you need support of pharma quality engineering experts, we’re here to support you. We have years of experience in helping world-leading pharma companies overcome these (and other) challenges and we can do the same for you. Contact our experts today.