Flexible skin patches and on-device AI aim to bring clinic-grade vital sign monitoring home
A review of multimodal wearable patches that track vital signs and biomarkers, with on-device machine learning for real-time telediagnosis.
Summary
This review surveys thin, flexible sensor patches that conform to the skin. The goal is to continuously track the five core vital signs (ECG/heart rate, respiration, body temperature, blood pressure, SpO2) plus skin-measured biomarkers such as lactate, electrolytes and glucose. The authors argue that the widening gap between life expectancy and healthy life expectancy (about 9.6 years in 2019) drives demand for home monitoring that could reduce clinician workload and speed triage. A detailed example is a breathable, stretchable kirigami ECG patch with Bluetooth that tracked heart rate variability during daily activities. The authors also outline respiration sensing via abdominal strain or nostril humidity, and propose on-device (edge) machine learning so feedback does not depend on the cloud. This is an engineering review, not a clinical trial. The demonstrations shown are proof-of-concept, the patches still need rigid circuit chips, and the authors note technical challenges remain before practical use.
Detailed Summary
Healthy life expectancy lags total life expectancy by about 9.6 years globally (2019), a gap that grew 13% since 2000. The authors attribute this to population aging, rising chronic illness, functional decline and insufficient social care, with rising demand for medical care and hospitalization. They propose home-use telediagnostic wearables that could pre-screen who needs medical intervention, extend healthy lifespan, reduce clinician burden and speed triage in emergencies. This is a narrative review of device engineering, not a clinical study, so the evidence is technical performance rather than health outcomes.
The authors note that watch-type wearables capture limited parameters such as heart rate and activity, which are insufficient for accurate diagnosis. To substitute for clinical diagnostics, a patch would need to measure all five vital signs plus biochemical markers such as lactate, electrolytes and glucose. Thin, conformal patches reduce noise from motion and environmental change. However, current systems remain hybrids: flexible sensors paired with rigid circuit chips for signal processing and wireless transmission. Cloud analysis of continuous data costs power and adds latency, so the authors propose edge computing with machine learning.
For ECG, the review details a kirigami-cut PET film. Chest skin deforms up to about 30%, and the kirigami pattern was designed to stretch up to 35% in multiple directions. The slits localize strain and allow out-of-plane buckling, which preserves electrode-skin contact impedance during motion. The film's water vapor transmission rate was about 17.5 kg m-2 d-2 (unit as printed in the source), versus about 12.6 for standard polyurethane bandage film. At 60% relative humidity and 28 °C, impedance stayed stable, whereas a non-kirigami patch dropped because of sweat bridging between electrodes. In a proof-of-concept demonstration, a Bluetooth Low Energy version streamed ECG and accelerometer data to a smartphone. Heart rate variability tracked changes in activity during desk work, lunch and walking, and the accelerometer data corroborated these changes.
For respiration, the authors note that patches can detect abdominal motion with stretchable strain sensors, or humidity changes near the nostrils. Respiration rate is derived from intervals between signal peaks. One example uses laser-induced graphene made by laser-irradiating polyimide film. The supplied text ends at this point, mid-way through the respiration section. Content on temperature, blood pressure, SpO2, biomarkers, integrated systems and edge machine learning was not available for this summary.
Implications: combining multiple signals in one conformal patch, with on-device analysis, could enable continuous home monitoring and lower power use and latency compared with cloud processing. Caveats: patches still rely on rigid circuit chips, the demonstrations in the supplied text are proof-of-concept with no reported sample sizes, and no health outcome data are presented.
Key Findings
- Gap between life expectancy and healthy life expectancy reached about 9.6 years in 2019, up 13% since 2000
- Kirigami ECG patch designed to stretch up to 35% in multiple directions, versus up to about 30% chest skin deformation
- Kirigami film water vapor transmission about 17.5 versus about 12.6 for standard polyurethane bandage film (units as printed: kg m-2 d-2)
- Kirigami patch impedance stayed stable at 60% RH and 28 °C; a non-kirigami patch dropped because of sweat bridging
- Proof-of-concept Bluetooth ECG patch tracked heart rate variability during desk work, lunch and walking, corroborated by accelerometer data
- Current patches are hybrid systems needing rigid chips; authors propose edge machine learning to avoid cloud latency and power costs
Methodology
This is a narrative review of flexible sensor and wireless system literature. No systematic search, pooled analysis or quality grading is described. The cited ECG demonstration is a proof-of-concept device test with no sample size reported. The supplied full text ends partway through the respiration sensor section (2.2). The temperature, blood pressure, SpO2, biomarker, integrated-system and machine learning/edge computing sections were not available for summary.
Study Limitations
Sensors still depend on rigid circuit chips, because fully flexible circuits are hard to make cheaply and stably under body motion. Motion artifacts remain a challenge. Cloud analysis adds power and latency costs. The demonstrations in the supplied text are proof-of-concept with no sample sizes reported. The authors state that several technical challenges remain before practical implementation. The supplied text is truncated and no conflicts of interest were stated in it.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.
Enter your email to subscribe:
