Continuous, connected healthcare is becoming a reality. Sensors are central to this transformation, not only algorithms, connectivity, or cloud platforms. Medical patches worn on the skin and implantable devices inside the body convert physiological signals into electrical data streams that can be captured, interpreted, and acted upon. For engineers, these devices present nearly every challenge in electronics design: ultralow-power sensing, harsh environments, strict regulation, small form factors, demanding safety requirements, and increasingly, on-device intelligence.
Modern medical patches are thin, flexible, and often disposable devices that adhere to the skin. Although they appear simple, they integrate complex sensing chains, including electrodes or photodiodes on the surface, MEMS inertial sensors embedded in the substrate, analog front ends (AFEs), microcontrollers, memory, and wireless transceivers.
Typical applications include long-term ECG monitoring for arrhythmia detection, postoperative and post-discharge vital sign monitoring, continuous or episodic glucose and metabolic monitoring, activity and posture tracking, fall detection in elderly or fragile patients, and additional use cases. These applications must be delivered under strict design constraints: ultralow power to ensure days to weeks of autonomy on a small battery; stringent form factor requirements, with millimeter-scale thickness, flexible substrates, and soft interconnects; demanding signal quality conditions, since microvolt-level biopotentials and low-amplitude optical signals must be acquired in the presence of motion, sweat, variable skin impedance, and environmental noise; and cost and manufacturability constraints, because many patches are single-use or limited-use devices, requiring cost-optimized and highly integrated designs. In this context, sensors are a primary differentiator. Sensors define which physiological variables can be measured, at what accuracy and resolution, and under which real-world conditions.
Biopotential sensors that use surface electrodes form the basis of many cardiac patches. These sensors
Motion sensors, particularly MEMS accelerometers and gyroscopes, are central components in modern medical patches. A 3-axis accelerometer is now standard, and in some applications, a full 6-axis inertial measurement unit (IMU) comprising an accelerometer and gyroscope is used. These inertial sensors serve two primary roles.
The first role is clinical feature extraction. The sensors enable the patch to quantify activity levels, classify posture (for example, standing, sitting, lying supine, or lying on the side), characterize gait patterns, detect and characterize falls, and quantify tremor amplitude and frequency in movement disorders. Motion data is used to generate metrics such as step count, cadence, gait symmetry, sit-to-stand transitions, time spent in various postures, and overall mobility patterns. These metrics have direct clinical relevance in cardiology, neurology, rehabilitation, and geriatrics.
The second role is artifact rejection and contextualization of other physiological signals. Movement is a major noise source for electrocardiogram (ECG), photoplethysmogram (PPG), and respiration measurements. Accelerometer data provides an independent reference for body motion that is used to distinguish motion artifacts from genuine physiological changes. Algorithms correlate spikes or baseline shifts in ECG with simultaneous acceleration peaks and classify them as motion-induced. Accelerometer and gyroscope signals are used to correct or discard corrupted PPG segments. In respiration monitoring, inertial signals help separate respiratory-related chest motion from large postural changes.
From a design perspective, motion sensors are also key components for power optimization. For multi-day or multi-week use, engineers must tune duty cycles, sampling rates, and digital filtering to keep energy consumption under control. Many medical patches run accelerometers in ultralow-power mode at a modest output data rate, for example, 12.5 to 25 Hz, to provide continuous activity and posture monitoring. The system switches to higher sampling rates only when needed, such as during suspected falls, arrhythmic events, or user-triggered episodes.
Modern MEMS accelerometers often integrate embedded functions such as wake-up interrupts on motion, orientation change detection, step counting, and FIFO buffering. These features allow the microcontroller to remain in deep-sleep mode for longer periods, waking only when meaningful motion patterns occur, which extends battery life. For medical applications, the accelerometer must offer low noise density to resolve subtle movements, stable offset and sensitivity across temperature and aging, robust behavior under mechanical shocks, and self-test capabilities so that latent sensor failures can be detected by the system.
Implantable medical devices operate in a distinct design space compared with patches. These devices benefit from a relatively stable environment regarding temperature and large-scale body motion. However, they must function reliably for many years in ionic fluids, under constant mechanical load and cyclic stress, and without the possibility of routine servicing or battery replacement in most cases. Key implant types include cardiac devices such as pacemakers, implantable cardioverter defibrillators, and cardiac resynchronization devices; implantable loop recorders for long-term rhythm monitoring; and neurostimulators for deep brain, spinal cord, or vagus nerve stimulation. Across all categories, sensors determine the clinical value of the system. If sensing fails, saturates, drifts excessively, or becomes unreliable, therapy becomes ineffective or unsafe.
Intracardiac and intravascular sensing are core capabilities of many cardiac implants. Unlike surface electrocardiogram (ECG), intracardiac electrograms are larger and have higher fidelity, enabling precise arrhythmia detection and discrimination. In addition to electrical activity, impedance measurements between electrodes are used as indirect sensors of thoracic fluid status in heart failure management, lead integrity, and respiratory rate. From a hardware perspective, protection structures, fast recovery after large transients, robust electrostatic discharge (ESD) and electromagnetic interference (EMI) design, and long-term stability in a corrosive environment drive the choice of semiconductor technology, packaging, and encapsulation strategies.
Motion sensors play an important role in implants. MEMS accelerometers are standard in rate-responsive pacemakers, where they provide an estimate of patient activity that is used to adjust the pacing rate. In neurostimulators and spinal cord stimulators, accelerometers can detect posture and movement, enabling automatic selection of stimulation programs depending on whether the patient is standing, sitting, walking, or lying down. This improves comfort and efficacy. The requirements are stringent. These sensors must operate continuously for five to ten years or more on a single battery, with very low current consumption, predictable behavior under mechanical shocks such as falls, stable sensitivity and offset over long periods, and minimal risk of mechanical or packaging failure.
ST recently launched the MIS2DU12 accelerometer, which offers three key advantages for battery-powered and space-constrained devices. First, it is extremely energy efficient, consuming minimal power even during normal operation. This feature benefits products such as patches and implantable devices. Second, it includes a built-in anti-aliasing filter that removes unwanted vibrations and noise. As a result, the motion data is cleaner and requires less processing by the main electronics, saving both time and energy. Finally, the MIS2DU12 is available in a very compact package, making it one of the smallest motion sensors with such low power consumption. This allows engineers to integrate advanced sensing capabilities into increasingly smaller and slimmer products.
Looking ahead, the future of medical patches and implants leads toward more intelligent sensors and systems. Over the next decade, multimodal patches are likely to combine ECG, PPG, motion, temperature, and biochemical sensing in a single low-profile device, with motion sensors serving as a central hub for context and artifact suppression. Artificial intelligence (AI)-assisted sensing systems will not only filter and classify data but also orchestrate sensing by determining when, where, and how to measure for maximum information with minimal energy. These systems will use motion cues to trigger high-resolution acquisition or to switch modes. In every case, the starting point remains the same: what can be sensed, and how accurately? For engineers working in medical devices, this question is central.
The choice, design, and integration of sensors—especially motion sensors that provide context, robustness, and safety—continue to be decisive factors between devices that only collect data and those that improve patient outcomes.