Medication administration errors are alarmingly common, with estimates suggesting they account for 5% to 10% of all administered drugs. These errors can lead to severe adverse events, affecting approximately 1.2 million patients annually and costing the healthcare system $5.1 billion. The most frequent errors arise during intravenous injections, where clinicians may mistakenly select the wrong vial or mislabel a syringe.

While existing safety measures, like barcode scanning, are in place, they may be overlooked in high-pressure situations.

A groundbreaking wearable camera system developed by researchers at the University of Washington is set to transform medication delivery in clinical settings. By leveraging artificial intelligence, the system detects potential errors during drug administration, significantly enhancing patient safety.

In recent tests published in npj Digital Medicine, the AI-driven video system demonstrated exceptional accuracy, achieving 99.6% sensitivity and 98.8% specificity in identifying vial-swap errors. This technology could serve as a critical safeguard in high-stakes environments such as operating rooms and intensive care units, where medication errors can have dire consequences.

Dr. Kelly Michaelsen, co-lead author and assistant professor of anesthesiology and pain medicine, emphasized the profound impact of this innovation: “The thought of being able to help patients in real time or to prevent a medication error before it happens is very powerful.

This new camera system, paired with a GoPro, utilizes deep learning to recognize the contents of vials and syringes, providing real-time alerts before medications are administered. The training process involved months of collecting 4K video footage from various clinical settings, enabling the model to learn visual cues, such as the size and color of vials and syringes.

Article written by Technology Network

22/10/2024

Source:

Technology Network

https://www.technologynetworks.com/informatics/news/ai-powered-wearable-camera-system-detects-medication-errors-392386