PRecision Integrated Saturation Monitor (PRISM)

Conventional pulse oximeters are prone to motion artifacts due to their placement on high-activity limbs. Traditionally, inpatient pulse oximetry relies on clinical staff to verify the accuracy of readings and distinguish between real and false desaturation events. However, in outpatient environments, caregivers are left to manage and interpret signals that may not provide reliable information about their child's health. PRISM aims to improve pulse oximetry monitoring for children. We will test an innovative pulse oximeter prototype that mitigates motion artifacts by shifting the monitoring location from a high-activity limb to the abdomen while incorporating motion detection capabilities. Previously, we have successfully developed an abdomen-worn prototype for collection of ECG signals for infant monitoring (Figure 1). We obtained high-quality signals with a robust correlation to the gold-standard ECG. Our preliminary results demonstrate the feasibility of collecting physiological signals from the abdomen, supporting its use as a novel, baby-friendly monitoring site. Figure 1. Comparison of ECG signals between gold-standard chest ECG and abdomen ECG. Protocol Version 2: 11 June 2026 IRB NUMBER: STUDY00000178 IRB APPROVAL DATE: 6/29/2026 IRB EXPIRATION DATE: 6/2/2027 We will incorporate a commercially available pulse oximetry sensor into a biocompatible, flexible, 3D printed housing. The child-focused design will be specific for application to the abdomen and will clip onto the diaper or pants of the baby; the tension from the diaper or pants will provide ample force to ensure sufficient contact between the device and the skin to achieve robust signals (Figure 2). The device will be removed with each diaper change. Figure 2. Schematic demonstrating the proposed pulse oximeter and its abdominal application As pulse oximetry uses optical sensors, external factors such as ambient light may interfere with data acquisition. Additionally, maintaining reliable contact with the skin is crucial to ensure the emitted light penetrates the underlying tissue, rather than reflecting off the skin's surface. Here, we will design the sensor housing to shield the pulse oximetry sensors from ambient light and provide a contact surface flexible enough for direct skin-to-sensor contact. The PRISM prototype is intended to minimize the impact of motion artifacts, a common phenomenon that leads to inaccurate readings and unnecessary alarms. By shifting the monitoring location from high-activity limbs to the more stable abdomen, we aim to enhance the accuracy and reliability of the device's measurements compared to traditional pulse oximeters. This design choice is particularly relevant for infants and young children, who often experience restlessness and movement during routine monitoring.

Successful accomplishment of this will result in a working prototype with motion artifact-resistant SpO2 reading error of less than 3.0% RMSE from SaO2. This development will allow the caregiver to continuously monitor accurate oxygen status Hypoxemic conditions, or medical problems stemming from low oxygen saturation, can result in half a million deaths annually. Oxygen saturation levels, which determine hypoxic conditions, can be measured directly from arterial blood. This is the gold standard of oxygen saturation measurement and is often called SaO2. Oxygen saturation measured from pulse oximetry, referred to as SpO2, uses light transmission and can allow for early detection of hypoxic conditions. It's a popular belief that melanin can absorb light in pulse oximetry, leading to the overestimation of SpO2 in darker skin. Although individuals with darker skin are observed to have a SpO2 overestimation, clinicians have yet to devise a method to accurately predict oxygen saturation given this overestimation[1]. Various clinical data has been curated to specifically address the issue of biases and disparities of pulse oximeter measurements in darker skin tones including the MIT Critical Datathon 2023 utilizing MIMIC IV and the blood-gas and oximetry linked dataset (BOLD)[2]. One issue with these datasets is race is often used as a proxy for skin tone, but there is great variability in skin color, even within a single race. Without an algorithm to account for how variation in skin tone biases in SpO2 measurement, a disproportionate amount of people of color may continue to receive incorrect diagnoses, leading to health inequities.

These incorrect diagnoses can lead to hidden hypoxemia, which is a condition where the gold standard oxygen saturation of a patient is less than 88% while the pulse oximeter measurement reads over 88%5. Studies show a significant disparity with black patients experiencing nearly three times the frequency of undetected hypoxemia compared to white patients6. They also emphasize the need to integrate pulse oximetry data with other clinical and patient-reported information to address this bias in medical technology.

These results are expected to generate evidence for proof-of-concept for this advanced next-generation pulse oximetry to be further tested and commercialized.