AI enabled wearable camera system detects potential errors in medication delivery

Developed and tested at the University of Washington, the video system has shown that it can recognise and identify medications drawn in busy clinical settings. The AI is said to have achieved 99.6 per cent sensitivity and 98.8 per cent specificity at detecting vial-swap errors.

The team’s findings are reported in npj Digital Medicine.

The system could become a critical safeguard, especially in operating rooms, intensive-care units and emergency-medicine settings, said co-lead author Dr Kelly Michaelsen, an assistant professor of anaesthesiology and pain medicine at the University of Washington School of Medicine.

“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 in a statement. “One can hope for a 100 per cent performance but even humans cannot achieve that. In a survey of more than 100 anaesthesia providers, the majority desired the system to be more than 95 per cent accurate, which is a goal we achieved.”

Drug administration errors are the most frequently reported critical incidents in anaesthesia and the most common cause of serious medical errors in intensive care. Furthermore, an estimated 5 per cent to 10 per cent of all drugs given are associated with errors. Adverse events associated with injectable medications are estimated to affect 1.2 million patients annually at a cost of $5.1bn.

Syringe and vial-swap errors most often occur during intravenous injections in which a clinician must transfer the medication from vial to syringe to the patient. About 20 per cent of mistakes are substitution errors in which the wrong vial is selected, or a syringe is mislabelled. Another 20 per cent of errors occur when the drug is labelled correctly but administered in error.

MORE FROM MEDICAL & HEALTHCARE

Safety measures, such as a barcode system that reads and confirms a vial’s contents, are in place to guard against such accidents, but practitioners can forget this check during high-stress situations because it is an extra step in their workflow.

The researchers’ aim was to build a deep-learning model paired with a GoPro camera that recognises the contents of cylindrical vials and syringes and issues a warning before the medication enters the patient.

To gather their data, the investigators collected 4K video of 418 drug draws by 13 anaesthesiology providers in operating rooms where setups and lighting varied. The video captured clinicians managing vials and syringes of select medications. These video clips were later logged and the contents of the syringes and vials denoted to train the model to recognise the contents and containers. 

The video system does not directly read the wording on each vial, but scans for other visual cues such as vial and syringe size and shape, vial cap colour, and label print size.

“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. Some letters [on the syringe and vial] are covered by the hands. And the hands are moving fast. They are doing the job. They aren’t posing for the camera,” said Shyam Gollakota, a co-author of the paper and professor at the UW's Paul G. Allen School of Computer Science & Engineering. 

Further, the computational model had to be trained to focus on medications in the foreground of the frame and to ignore vials and syringes in the background.

“AI is doing all that: detecting the specific syringe that the healthcare provider is picking up, and not detecting a syringe that is lying on the table,” said Gollakota.