Calibration and Accuracy of a Wearable Pressure Sensor.

Introduction: For the Wound and Lymphedema communities, a Wearable Pressure sensor (WPS) -- to evaluate dressings and teach practitioners best wrapping techniques -- could be transformational.  To adopt such an innovation, the proposed WPS must be clinically accurate.  The goal of this poster is to describe the accuracy of such a Wearable Pressure Sensor.Methods: The wearable pressure sensor (WPS) includes a proprietary capacitive-based sensor (Reference 1).  Sensor information is read by a system on a chip (SOC), which transmits pressure information by Bluetooth to an iPhone.  This sensor is thin, flexible, and allows excellent stability and battery life ( >2 months).  The WPS has a dedicated iPhone beta test App.  This App has a novel pressure gauge display and in addition shows pressures from multiple sensors simultaneously (Reference 2).   Calibration:  We used a validated (Reference 3) pneumatic sensor (2 cm dia. balloon) as the gold standard (GLD-STD)^.  The WPS is calibrated against the GLD-STD.  A large blood pressure cuff applies pressure of 0-110 mm Hg low-to-high in 10 mm Hg increments.  An WPS algorithm calculates pressure curves from the GLD-STD relative to WPS.  Leg model:  Empirically, a human leg provides the best mechanical stress-strain characteristics for calibration and garment testing.  The first author used his left leg for all measurements.   Garment testing: There are 4 levels of compression:  Stocking (5-10- or 10-20-mm Hg), plus 0, 1,2, or 3 “E” size tubular bandages.  Deviation of a WPS reading from known pressure is described as +/- % ERROR, as follows:  +/-% ERROR = 100*absolute value (WPS PRESSURE - GLD-STD PRESSURE)/GLD-STD PRESSURE We used 9 calibrated wearable sensors: Three sensors for 2 tests; one for 3 tests; 5 for 4 tests.  Results: %ERROR of the WPS relative to GLD-STD is as follows:  Compression level (mm Hg) +/- % ERROR Standard deviation n 0-9 22% 0.3% 2 10-19 14% 3.5% 4 20-29 9% 3.0% 7 30-39 8% 3.0% 3 Discussion: For this work, accuracy is defined as +/-%ERROR of a test sensor off-target, with the “bulls eye” being the reading of a “gold standard” sensor.  Empirically, error of +/-15% is deemed adequate for clinical use.  For the Wearable Pressure Sensor, this level of accuracy occurs at >20 mm Hg compression.  Most compression garments and dressings exert >20 mm Hg.  Accuracy improves as pressure increases.  At the low end, pressure of