Diagnostic accuracy of non-mydriatic fundus photography as a triage and telemedicine tool for patients with vision loss.

Abstract

Objectives: Fundoscopy can be challenging for non-ophthalmologists. For emergency physicians, non-mydriatic fundus photography is superior to other forms of ophthalmoscopy in sensitivity, specificity, and inter-examination agreement. We report on a prospective cross-sectional study evaluating the use of non-mydriatic photography as a triage and telemedicine tool for patients with vision loss in a Canadian emergency setting.

Methods: Images of both eyes were obtained by a non-ophthalmologist using a handheld, non-mydriatic fundus camera and shared with a fellowship-trained ophthalmologist without patient context. The reviewer was asked to (1) select the best photo obtained for each eye and rate image quality on a Likert scale, (2) comment on the presence or absence of fundus abnormalities and (3) provide an opinion on whether findings would have changed patient disposition if available at the time of the initial Emergency Department (ED) exam.

Results: Of 57 patients evaluated in the ED for vision loss, 22.8 % had a documented fundus examination. 86.8 % of images were deemed to have acceptable quality (Likert scale≥2). Factors limiting image quality included media opacity, pupillary miosis, photosensitivity, and eyelid/periorbital abnormalities. Of patients with relevant abnormalities, 0 % were identified by emergency physicians. In contrast, 37.5 % of patients with relevant findings were identified on review of images alone (specificity=100 %).

Conclusions: Fundoscopy is infrequently performed in the emergency setting in patients presenting with vision loss. Non-mydriatic fundus photography is a cost-effective method of fundus examination for non-expert examiners and can be reliably used as a telemedicine tool for remote ophthalmology consultation.