Prospective Validation of an Artificial Intelligent System for Eye Emergency Triage and Primary Diagnosis

Ophthalmic emergencies are acute vision-threatening disorders, for which a delay in prompt emergency response could result in catastrophic vision loss. Triage is an effective process for ensuring that timely emergency care is provided despite limited resource by prioritizing patients to appropriate orders for visits. Historically, registered nurses classify emergency patients based on personal experiences with high variation. Additionally, primary healthcare providers have been conventionally at the forefront of providing first aid care. However, most of ocular emergencies are wrongly diagnosed or referred due to non-eye specialists' limited knowledge and training in the ophthalmology.

Here, the investigators established and validated an artificial intelligence system, EE-Explorer, to triage eye emergencies and assist in primary diagnosis using metadata and ocular images. This system has been integrated into a website to be prospectively validated in the real world.