Utilization and Feasibility of a Wearable Device in Patients With Sedative Effects of Drugs: Protocol for a Prospective Observational Study for the Advanced Respiratory Monitoring Events in Drug Toxicity (ARM-ED) Study

Comprehensive Summary

The Advanced Respiratory Monitoring in the Emergency Department (ARM-ED) study is a prospective observational protocol designed to evaluate the efficacy of a novel wearable biosensor to detect life-threatening breathing patterns in patients at high risk for drug-induced respiratory depression. The sensor was developed by PneumoWave and can monitor long term data for retrospective review and can alert a healthcare team of current dangerous abnormal breathing patterns. This study seeks to validate the sensor's accuracy in identifying apnea and reduced respiratory effort across various clinical scenarios. This study was 18 months long and involved three groups of patients: an acute toxicity group (patients who present to the ED with actual or potential CNS or respiratory depression secondary to toxicological cause), a PSA group (patients who underwent procedural sedation and analgesia with respiratory and CNS depressant drugs in the ED), and GA group (patients who underwent general anesthesia). As of December 2023, the study successfully completed recruitment with 78 participants across the three cohorts. Preliminary objectives focus on the device's usability, assessing parameters such as battery life, physical comfort, and the frequency of accidental removals, along with its ability to generate high-quality data for future machine learning algorithms. While the final results are anticipated by late 2025, the study establishes a critical framework for moving respiratory monitoring technology out of the hospital and into the community to prevent fatal overdoses among high-risk populations.

Outcomes and Implications

The success of the ARM-ED study in capturing high-quality respiratory data via wearable sensors is a precursor to developing robust machine learning algorithms for overdose detection. By providing a diverse dataset of real-world concerning breathing patterns from clinical settings like procedural sedation and acute toxicity, researchers can train models to distinguish between benign movement and true life-threatening apnea. These algorithms will eventually move beyond simple threshold alerts to predictive models, identifying subtle physiological shifts before a fatal respiratory arrest occurs. Ultimately, this data serves as the foundation for creating autonomous, closed-loop systems capable of alerting emergency services or even triggering automated naloxone delivery in the community.