More than eight million catheters are placed for invasive arterial pressure measurement in the United States with 36% of ICU patients receiving one. Invasive arterial pressure measurements require insertion of a catheter into an artery of a patient (typically radial artery) and connecting it to a pressure transducer. For an arterial pressure measurement to be accurate and clinically relevant, the pressure transducer must be maintained at the level of a hydrostatic reference point on the patient’s body (typically the heart). The transducer is frequently placed in a plastic holder clipped to the IV pole next to the patient and moved up and down when the patient’s heart height changes.
Maintaining the pressure transducer in the correct position is a task requiring considerable vigilance and manual manipulation by healthcare providers as there is no alarm system to detect discrepancies between the transducer height and patient’s heart. Height discrepancies can lead to inaccurate blood pressure measurements and result in erroneous treatment decisions. If bed height or patient position changes or a transducer falls to the floor, the blood pressure displayed will be erroneous. Automating this currently manual task of adjusting the transducer height would reduce errors and improve patient care. Nursing and anesthesia provider task overload can be a serious problem, suggesting that automation of routine tasks like this is valuable and may help to prevent errors and adverse outcomes.
Anesthesiologists at UW Medicine reached out to my group seeking a sensing solution to this problem. At a high level, if we can estimate changes in height of the patient's heart, we can translate that to pressure changes and adjust the arterial blood pressure readings, without the need for manually adjusting the transducers height.
We realized that this problem is exactly what researchers in the augmented and virtual reality (AR/VR) community are also addressing. The goal in the AR/VR community is to track devices in the 3D space using various sensors and signals. In fact, my group had previously designed millimeter-level acoustic tracking algorithms that uses acoustic signals to track AR/VR devices.
Building on our prior work in the augmented reality space, we designed a height tracking system that uses a small, wireless, low-power wearable device and a speaker array that can automatically track the height of the wearable device and make corrections to the pressure measurements without manually adjusting the height of the pressure transducer. The low-power wearable device is equipped with a Bluetooth chip, an accelerometer, and a microphone, and can be affixed to a patient via an electrocardiogram (ECG) electrode, as shown in the figure above. The speaker array (localizer) is comprised of four speakers that emit inaudible acoustic signals that are captured by the wearable device. Using these signals, the height tracking system continuously and accurately determines the three-dimensional position of the wearable device on the patient and uses the z-plane (vertical) information to correct the blood pressure measurements.
We evaluated our system with 26 patients undergoing cardiac surgery who required invasive arterial pressure monitoring at the University of Washington Medical Center. As seen in the figure, the mean arterial pressure measured by the clinical transducer, when compared to the calculated mean arterial pressure using the height tracking system, differed on average by 0.19 ± 2.8 mmHg with 90% of the measurements within 4.3 mmHg.
Given the increased workload demands on nurses and physicians, our proof-of
concept technology may improve accuracy of pressure measurements and reduce the task burden for medical staff by automating a task that previously required manual manipulation and close patient surveillance.
Above all, our experience with this paper underscores the significance of interdisciplinary research that uncovers connections between disparate research domains, such as VR/AR and blood pressure monitoring. Though seemingly unrelated, prior VR/AR research has provided useful technical solutions that effectively tackle crucial issues within the medical domain.
Please sign in or register for FREE
If you are a registered user on Nature Portfolio Bioengineering Community, please sign in