written by: doc. dr. sc. Ivo Dumić-Čule
The development of wearable technologies in healthcare over the past few years shows a clear trend: devices are no longer just auxiliary tools for counting steps and tracking sleep but are also part of the medical ecosystem. The most interesting and promising segment is precisely the devices that combine two key features – regulatory approval from relevant institutions, such as the U.S. Food and Drug Administration (FDA), and the introduction of artificial intelligence (AI).
Such a combination gives devices the status of not only technological novelties but also serious medical tools that can contribute to diagnostics, early disease detection, and monitoring of chronic conditions. Why is this combination important? The mere presence of AI algorithms in a gadget does not mean that the device is safe for drawing conclusions about a medical condition that a doctor may later use to create a complete picture of a patient’s health. When these two dimensions come together, what emerges can be termed ‘true transformation of healthcare’: smart, reliable, medically validated devices that use the power of artificial intelligence for better decisions and timely interventions.
Smartwatches and Blood Pressure Monitors
One of the most prominent examples is smartwatches with regulatory approval for detecting atrial fibrillation. As early as 2018, the FDA approved an electrocardiogram (ECG) application that allows users to measure a single-channel ECG directly from their wrist. At that point, the smartwatch crossed the boundary from a fitness gadget to a medical tool for the first time. What initially seemed like an experimental feature soon proved to be extremely useful in everyday practice. Subsequent developments brought algorithms capable of analyzing long-term pulse patterns and alerting users to possible signs of elevated blood pressure, employing sophisticated AI methods to filter noise, recognize anomalies, and distinguish benign fluctuations from clinically significant problems.
Another important example comes from the segment of home devices for measuring blood pressure – smart blood pressure monitors – which, in addition to standard blood pressure measurement, contain an integrated machine learning-based algorithm for detecting atrial fibrillation. The FDA has given the green light to this functionality as part of the de novo procedure, meaning it is an innovation that has no direct predecessor. The algorithm analyzes hundreds of mathematical parameters of pulse waves generated during blood pressure measurement and reliably identifies irregularities characteristic of atrial fibrillation. This function allows users to receive alerts about potentially serious heart rhythm disturbances through routine blood pressure measurements in their own homes, significantly shortening the time to diagnosis and increasing the chances of timely treatment.
Predictive Capability
In both cases, artificial intelligence is not just a marketing addition but a functionality that carries real clinical value. A third example worthy of attention relates to the further development of algorithms in smartwatches, this time focused on monitoring long-term trends and predicting hypertension. In 2025, the FDA approved a high blood pressure alert function that does not rely on a one-time measurement but on continuous analysis of signals obtained from optical sensors. This approach opens up a whole new dimension: it is no longer about detecting already occurred episodes of arrhythmia, but about the possibility of predicting other arrhythmias before they clinically manifest. If these algorithms prove to be accurate and reliable enough, they could become a sort of early warning system for millions of users.
