New AI-powered wearable cameras help detect medication errors

Preventable patient harm remains a pressing issue in healthcare, with medication errors posing a significant threat. These errors contribute to an estimated 140,000 to 440,000 deaths annually in the U.S.

Adverse events related to drug administration account for a substantial portion of this figure, particularly in high-risk settings like operating rooms and intensive care units. Injectable medications alone impact 1.2 million hospitalizations each year, incurring costs of $5.1 billion.

Errors in drug administration occur across various medical environments, including surgery, emergency departments, and primary care. Studies suggest that 5% to 10% of all administered drugs involve some form of error.

In anesthesiology, these mistakes are the most frequently reported critical incidents, with 41% of adverse events occurring in operating rooms. Syringe and vial swaps represent particularly concerning forms of error, with substitution errors accounting for 20% of these mistakes. Another 20% occur when the correct drug is labeled but administered improperly.

Vial swap errors typically happen during the preparation of intravenous injections when clinicians transfer medication from a vial to a syringe. In these cases, incorrect labeling or selecting the wrong vial can result in significant harm or even death.

Such errors have been documented across various specialties, including pediatrics, emergency medicine, and rehabilitation units. Alarmingly, only 2% of these errors are intercepted before they reach the patient in regular hospital wards.

Healthcare providers have implemented several measures to reduce medication errors. These include color-coded labels, tall-man lettering to emphasize differences in drug names, and standardized safety protocols. Prefilled syringes and barcode scanning have also been employed to minimize risks. However, these methods often require active participation by clinicians, who may resort to workarounds to save time.

Studies reveal that up to 62% of clinicians bypass established safety mechanisms during medication administration, tripling the likelihood of errors. This underscores the urgent need for a system that integrates seamlessly into clinical workflows while providing real-time checks.

A groundbreaking solution has emerged in the form of a wearable camera system designed to detect and prevent vial swap errors. This system, developed by researchers using deep learning algorithms, identifies syringe and vial labels in real-time.

By analyzing video footage captured by head-mounted cameras, the system can compare syringe and vial labels to detect mismatches before medication is administered.

The wearable system, tested on a large dataset of drug preparation events, achieved remarkable accuracy. It demonstrated a sensitivity of 99.6% and a specificity of 98.8% in identifying vial swap errors. These results suggest that the technology can serve as a reliable secondary check, providing clinicians with timely alerts to prevent potentially catastrophic mistakes.

Researchers collected extensive data to train the AI model, capturing 4K video of drug preparation by anesthesiology providers in real-world operating room environments. Over 55 days, 418 drug draws were recorded across two hospitals and 17 operating rooms. The video dataset included variations in lighting, clinician techniques, and medication types, ensuring robust training for the model.

Unlike barcode scanning, which requires manual interaction, the wearable system passively analyzes visual cues such as vial cap color, label print size, and syringe shape. This capability is particularly valuable in dynamic clinical settings where clinicians work quickly and may not fully expose labels.

The AI model also distinguishes between medications in the foreground and unrelated items in the background, focusing solely on the drugs being prepared.

Dr. Kelly Michaelsen, an anesthesiology professor and co-lead author of the study, emphasized the importance of preventing errors before they occur.

“The thought of being able to help patients in real time or to prevent a medication error before it happens is very powerful,” she said. Michaelsen added that while perfect accuracy is an aspirational goal, the system’s performance exceeds the 95% threshold desired by most anesthesia providers.

Shyam Gollakota, another coauthor and a professor of computer science, highlighted the challenges of training the AI model. “It was particularly challenging because the person in the OR is holding a syringe and a vial, and you don’t see either of those objects completely,” he explained.

Despite these obstacles, the system’s ability to analyze partial views and rapid hand movements demonstrates its sophistication and reliability.

The wearable camera system represents a significant advancement in patient safety. By providing real-time feedback, it has the potential to transform drug administration practices in operating rooms, intensive care units, and emergency settings.

The researchers envision broader applications for AI and deep learning in healthcare, improving both safety and efficiency across various clinical workflows.

This study, published in npj Digital Medicine, underscores the transformative potential of AI-driven solutions in addressing critical healthcare challenges. The research team, which included experts from Carnegie Mellon University and Makerere University, hopes their work will inspire further innovation in the field.

Funding for the study came from the Washington Research Foundation, the Foundation for Anesthesia Education and Research, and a National Institutes of Health grant.

As healthcare systems continue to face increasing demands, technologies like the wearable camera system offer a promising path forward. By integrating advanced AI capabilities into everyday clinical practices, the medical community can reduce preventable harm and enhance patient outcomes.

While challenges remain, the potential benefits of these innovations make them a worthy investment in the future of healthcare.

Note: Materials provided above by The Brighter Side of News. Content may be edited for style and length.

Like these kind of feel good stories? Get The Brighter Side of News’ newsletter.

The post New AI-powered wearable cameras help detect medication errors appeared first on The Brighter Side of News.