On Thursday, June 6, 2024, the third edition of digitalks took place, an event organized by Apptitude. The event was held in Lausanne, at the Casino de Montbenon.
In 2011, Netscape’s inventor and investor Marc Andreessen predicted that “software would eat the world,” meaning that everything would become software-driven. Today, we see the accuracy of his prediction. Administrative services, transportation, and shopping now all happen through mobile or web applications, often via cloud platforms.
More surprisingly, our watches, scooters, coffee machines, and even light bulbs are now software-driven, connected to the Internet. It’s no surprise, then, that the medical field and medical devices are also being “eaten” by software. Yet, one might wonder, who is eating whom? The software industry thrives on startup mantras, often prioritizing speed to capture market share before worrying about regulations. Facebook didn’t initially consider its responsibility for user-generated content. Uber didn’t worry about the legality of its contracts with drivers. The crypto industry largely ignored financial regulations. The latest example: OpenAI, which overlooked copyright issues during ChatGPT’s training.
The Agile method, sometimes postponing fixes depending on a bug’s severity to launch version 1 faster, also embodies this mindset. But this startup mentality clashes with the medical field’s rigor. Excellence is non-negotiable in this sector; you can’t sell a product without formal regulatory approval. After all, you can’t install a pacemaker with the promise that fixes will come after the next sprint.
- Kim Rochat, Senior Advisor Quality & Compliance at Veranex
- Maël Fabien, co-founder of Biped.ai
- Didier Blanc, Senior Software QA Engineer at Tandem Diabetes Care Switzerland
shared insights on how to reconcile software development with the medical field in their talks titled “Creating a Medical Device: Challenges and Opportunities.”
The digitalks aims to bring together entrepreneurship, design, and engineering. Three speakers take the stage to share their experiences, successes, and challenges around a common theme. The goal is to present concrete and practical cases, avoiding any promotional approach. Discover and register for the next edition: www.digitalks.ch
Kim Rochat, Senior Advisor Quality & Compliance at Veranex, spoke about the fundamentals of determining whether a device qualifies as a medical device and the underlying challenges.
Veranex provides integrated services for medical technologies, offering comprehensive expertise to healthcare product companies. With examples like a toothbrush (which is not a medical device), a bread knife (inappropriate for surgical use), sunglasses or prescription glasses, and the ambiguity of Garmin or Apple smartwatches, Kim highlighted the often subtle differences. From a software perspective, merely displaying data doesn’t make a program a medical device. However, the workload required for certification and the risks and opportunities in this classification are significant. Ultimately, the real question seems to be: why isn’t my device considered a medical device?
Maël Fabien, co-founder of Biped.ai, shared how he observed a visually impaired person using both a white cane for navigation and a smartphone for receiving real-time guidance from a friend.
This led him to develop a harness equipped with cameras, analysis software, and a GPS to assist visually impaired individuals in a new way, aiming to make the white cane obsolete. However, challenges arose: how to design the device to account for different body types that might tilt the camera? How to ensure the person could easily put on the harness without sight? The image analysis software was iterated based on user feedback during tests. Balancing information volume and warning signals was crucial, especially in scenarios like detecting a hole and estimating its distance. This startup-mode development faced a wake-up call during a test when a cyclist encountered the tester, who had the white cane tucked away. Without visible indicators of visual impairment, the situation highlighted potential dangers. The pressing questions followed: is this a medical device? If so, which category—1, 2, or 3? What are the implications? Are regulations the same across countries? Maël shared his experience in developing a medical device without prior knowledge in the field.
Didier Blanc, Senior Software QA Engineer at Tandem Diabetes Care Switzerland, discussed quality management. Financial concerns may dissipate when a startup is acquired by a global player, but investor pressure to accelerate work intensifies.
Quality management, however, must be controlled from A to Z to avoid having a product pulled from the market by regulators at the first issue. According to Didier, the question isn’t if a problem will occur but when. Coding without bugs seems unattainable. Failing to prepare for problems is a grave mistake. It’s essential to trace all processes to reproduce the problem and quickly identify corrective measures. Ensuring the quality of medical devices at every stage, from design to global commercialization, is more of a mindset than a goal. The field doesn’t allow for approximation, and only by having teams fully aware of these aspects can the required quality be achieved.
Discover the theme and speakers of our upcoming digitalks conference and register online at: www.digitalks.ch.