Wearables are quickly evolving from somewhat gimmicky gadgets to active partners in our healthcare. These smart electronic devices track everything from bread-and-butter fitness indicators like steps taken and hours slept to complex data like fluctuations in heart rate. These devices can be our partners as we work toward innovative solutions to improve preventive medicine and quick response. Simply put, people increasingly want to take ownership of their own health.

(stock image)

But how reliable and secure are these devices, and can we really trust everything they say? At its core, a wearable is an Internet of Things (IoT) device, which means it can communicate its data to other devices, users and websites. The accuracy of this measured information is critical not only to your health, but also to decisions that employers and insurance companies may make about your coverage. As healthcare is expected to make up one-fifth of the economy by 2025, there is a growing clamor for solutions that can reduce costs. While annual physicals are commonly recommended, wearables are considered by some to be a complementary solution to monitor our health signals on a more continual basis and to nip any problems in the bud. Thus, making sure the information on them is safe and correct is of significant importance.

In the Dependable Computing Systems Lab (DCSL) at Purdue, our researchers conducted studies on Wear OS, the dominant open source operating system for smartwatches and other wearables. We found systemic issues with the software architecture—including inconsistent (at best) handling of test-generated errors that led to application crashes, application hangs (freezes), and uncontrollable rebooting (restarting). We identified scenarios in which errors in the wearable led the coupled mobile device to fail. Our experiments also indicated that heart rate monitoring via wearables, when compared with clinical devices, is accurate at resting but less so with motion.

The encouraging news: We’re not just hunting for problems, though—we’re also helping fix them. We’ve made a series of recommendations for improving the device software architecture; our study on Android from 2012 resulted in vendor changes and enhanced reliability. We’ve recently developed solutions to secure the Systems on Chip (SoC) embedded in wearables and mobile phones, by compartmentalizing and “chaperoning” the data and the control to fence in problems. This tightly corrals bugs in the software module so they cannot be used as springboards to hijack the entire device.

As we add even more functionality to these devices—such as more sensors, readings, outputs, and complex application communication patterns, all vying for limited display area, battery capacity, and computing memory—we can expect their reliability and security to become even more frayed. The software and hardware development practices for wearables have simply not matured enough at this stage of device evolution. A more disciplined, research- and testing-focused approach is required. Today’s efforts, the equivalent of “looking under the hood” and “kicking the tires,” are not enough for long-term extrapolations of robust and secure performance.

The design and use of wearables are in their infancy, still rooted in an “isn’t this rad” approach, rather than being seen as serious and consequential medical devices. But I expect this will change in the near term, as research-driven progress in reliability and security makes wearables dependable and omnipresent must-haves rather than occasional, nice-to-have devices.

Writer: Professor Saurabh Bagchi
Source: Wearables in Healthcare: Are They Reliable and Secure?