The wearable technology market seems to be experiencing a lull, largely due to a slowdown in sensor innovation. While emerging fields like microfluidics and stretchable electronics have shown potential in wearable research, they have yet to make a significant commercial impact.
In contrast, Klick Labs is exploring a new frontier: voice recordings as a promising source for biomarkers. The idea of using smartphone-recorded voice clips to monitor conditions like Type 2 diabetes or assess glucose levels may sound futuristic, but Klick Labs has been diligently working on it, yielding encouraging results.
Their latest innovation is an app that leverages voice recordings to detect signs of chronic high blood pressure, with AI playing a crucial role in the analysis. “Voice technology holds the potential to revolutionize healthcare by making it more accessible and affordable, particularly for underserved populations,” says Jaycee Kaufman, a research scientist at Klick Labs.
The Significance of Hypertension
High blood pressure, or hypertension, affects an estimated 1.28 billion adults globally, with nearly half unaware of their condition. According to the World Health Organization (WHO), it’s a leading cause of premature death worldwide. Left untreated, hypertension can lead to severe complications such as heart attacks, kidney failure, strokes, vision loss, and various heart diseases, as highlighted by the American Heart Association. Unfortunately, timely detection and accessible analysis of hypertension remain challenges.
Although Samsung has integrated blood pressure monitoring into its smartwatches, the high cost of these devices limits their accessibility. This is particularly problematic for individuals in remote areas or those lacking adequate healthcare infrastructure. A solution that harnesses the ubiquity of smartphones could be a game-changer, addressing multiple critical challenges simultaneously.
The Potential of Voice Recordings
For their study, Klick Labs analyzed voice samples from 245 participants, who provided recordings six times daily over two weeks. The team developed gender-specific predictive models and, for the first time, documented an innovative method of acoustic analysis.
Their research, published in the IEEE Access journal, demonstrated that the app could detect high blood pressure with 84% accuracy in women and 77% in men. According to Yan Fossat, senior vice president of Klick Labs, the team is now focusing on expanding voice-assisted analysis to detect non-chronic high blood pressure.
One of the significant advantages of their system is that it requires no calibration before use, eliminating a common obstacle. Additionally, the data was collected in an unsupervised setting, further showcasing its potential.
The app examines various vocal attributes, such as speech energy distribution, pitch variability, and the sharpness of sound changes, to correlate them with elevated blood pressure. While the app is still in the fine-tuning stage and not yet publicly available, its prospects are promising.
Fossat, who also led the study, shared that the team plans to distribute application programming interface (API) keys as they move towards commercialization. The app will be platform-agnostic, meaning it could be available on both iPhones and Android devices soon.
Regarding regulatory hurdles, Fossat mentioned that they plan to submit the app for regulatory clearance as a Software as a Medical Device (SaMD) Class I. According to the U.S. Food and Drug Administration (FDA), SaMD Class I solutions pose the lowest risk and are not integrated with hardware medical devices.
The Future of Voice-Assisted Health Monitoring
How should the average person view Klick Labs’ voice-assisted chronic blood pressure detection system and its potential? “Medical-grade tools remain the gold standard for detecting blood pressure anomalies,” Fossat told Pro Well Tech. While Klick Labs’ app isn’t a comprehensive solution for managing hypertension, it can serve as a valuable tool in guiding individuals towards timely medical intervention.
Challenges and Next Steps
Like other innovations, this system faces challenges before it can be deployed at scale. For example, the initial study was limited to a specific ethnicity, with an insufficient number of hypertensive cases in the test pool. The team is also working on reducing the number of recordings needed for accurate analysis, which will require more data to train the underlying AI models. Additionally, the need for some level of training to gather the necessary acoustic data presents another hurdle for widespread adoption.
“Our next step is to replicate the study with a larger and more diverse group, including participants from various ethnic backgrounds and a broader spectrum of hypertension symptoms,” Fossat notes. Klick Labs is exploring advanced techniques, such as deep neural networks, to address some of these challenges.
Nevertheless, the idea of a noninvasive, smartphone-based approach to monitoring chronic high blood pressure represents a significant leap forward. By eliminating the need for expensive wearables and leveraging widely available smartphones, this innovation has the potential to positively impact millions of at-risk individuals worldwide, particularly with support from health authorities and commercial adoption.