By Godwin Maben, Synopsys Fellow; Piyush Sancheti, VP of Engineering, Synopsys Strategy & System Architects Group; and Hany Elhak, Group Director of Product Management and Solutions Engineering; Synopsys
Innovation in the medical device industry is booming as demand for personalized, robust medical devices increases along with an aging population. In fact, the United Nations predicts that by 2050, one in six people in the world will be over the age of 65. The growing senior population, need for earlier diagnoses, and value that real-time patient data can provide for all age groups has led to pioneering medical device designs that can detect diseases, provide treatments, assist those with disabilities, and more.
For instance, devices that utilize artificial intelligence (AI) and virtual reality (VR) technology are being used by medical professionals to advance diagnostics, support robotic surgery, train surgeons, and even treat depression. The global market for AI in healthcare is expected to reach $120 billion by 2028 due to the success that AI tech has already seen in hospitals, wearables, and routine doctor’s office visits.
This innovation is being enabled by semiconductor technology that allows devices to be smaller than ever before and support these new functions. Designing chips for medical applications requires different planning than for other markets, even other mission-critical markets like autonomous vehicles or aerospace. Read on to learn about the different categories of medical devices and the three key challenges for medical chip design: power, security, and reliability.
The term “medical device” is one that can be used to describe a variety of different items, from things as simple as a BAND-AID to as complex as an MRI machine.
The World Health Organization (WHO) defines a medical device as “an article, instrument, apparatus or machine that is used in the prevention, diagnosis or treatment of illness or disease, or for detecting, measuring, restoring, correcting or modifying the structure or function of the body for some health purpose.”
Experts estimate that there are over 2 million different kinds of medical devices on the market, which could be categorized into more than 7,000 groups. Fortunately, this discussion doesn’t need to get quite so granular. At Synopsys, we like to say that medical devices fall into three broad categories:
While each of these categories has its own specific requirements and considerations that designers need to examine, all three of them require careful analysis of power (specifically low power), reliability (especially for devices that are expected to last for 10 years or more), and security (in terms of both protecting private medical data and preventing breaches).
We’ll dive into each of these medical device design considerations with more detail in the coming sections.
Ensuring that medical devices consume low power is one of the largest concerns that engineers need to consider when developing semiconductors used in healthcare, especially for implantable devices which need to be ultra-low power. That’s because those devices require surgery to place in and remove from the body; doctors don’t want to be doing surgeries on patients and end up changing the battery every few years. In general, doctors and patients expect implantable medical devices to last between 10 and 20 years.
Even medical devices that are not implantable mostly require ultra-low power designs because they run off batteries in most cases (like the fitness tracker on your wrist). Designers need to consider techniques that reduce active and standby power, including low leakage process, voltage domains, and switchable power domains.
Reliability is the probability that a chip will perform its required function in a certain environment (in the human body, on a wrist, etc.) for a designated time period, which varies depending on the use of the medical device. You’ll see a reliability “bathtub curve” in the chart below which shows the proportion of devices that have failed (y axis) by a given time (x axis).
Most failures happen during the manufacturing phase or toward the end-of-life, which varies by specific product. For example, the laptop or mobile device you are using to read this article has a lifespan of around three years.
End-of-life failures are mostly due to transistor aging and electromigration. Aging is the process of transistors gradually degrading in performance over time as they are used, which eventually leads to the failure of the overall device. Electromigration, or the unwanted movement of atoms due to current density, plays a significant role in the failure of the interconnects between the transistors. The higher the density of the current passing through the wire, the greater chance it’s going to break down in the shorter term.
Lower geometry silicon technologies that may be used in other innovative applications simply aren’t being used in medical device silicon designs because the industry doesn’t fully understand the impact of aging at low geometry. Instead, engineers typically opt for more mature silicon technologies (28nm and larger) that have been widely used and studied, a fact that is comforting for anyone who currently has an implantable device helping them to survive and thrive.
The proper operation of medical devices is mission-critical, which is why reliability needs to be integrated at the very beginning design stages and throughout the process. Reduction of variation in the production stage has also become essential. Synopsys offers a whole suite of analysis tools, broadly called PrimeSim™ Reliability Analysis, that includes electrical rule checks, fault simulation, variability analysis, electromigration analysis, and transistor aging analysis.
Finally, security is paramount to silicon design for medical devices in two major ways.
First, the confidential medical data that medical devices collect needs to be secure so that unauthorized people are unable to access private medical information.
Second, designers need to make sure that the medical device is not susceptible to sabotage in any way, such as a pacemaker being hacked into by a bad actor in a way that could harm the patient. Due to the COVID-19 pandemic, connected devices are being used increasingly in the medical field to reduce contact exposure to patients and for convenience’s sake. The more remote connections that exist, the more opportunities there are for both data breaches and other cyberattacks.
Both the security of the device operation and the security of the data that these devices are collecting or transmitting over wireless or wired networks are vital in this industry. While there is no security solution that will work for all devices, security needs to be baked in at the design level both in hardware and software to truly prioritize the safety of the patient.
Overall, designers who focus on medical devices use many of the same tools engineers use for other applications, including IP blocks, reliability analysis tools, and more. These EDA tools help designers plan around the ultra-low power, increased reliability, space constraints, and security considerations that are critical for patient health, convenience, and safety.
The healthcare industry needs cutting-edge EDA tools to overcome these challenges in today’s competitive landscape. These EDA tools offer solutions such as allowing real-time data processing power at the hardware level vs. the software level, system integration (i.e., integrating as many components as possible into a single chip platform), assessing the impact of low-power designs on thermal and battery life, and more.
The medical device industry is rapidly evolving and must ask itself how it can learn from other regulated mission-critical fields, including automotive, as silicon designers help create and power the next generation of implantable devices, hospital medical equipment, and healthcare wearables.