Purpose
To determine if the quality of photographs taken of eye findings with an iPhone camera are of sufficient quality to triage patients presenting to ophthalmologists on-call.
Methods
Patients presenting to ophthalmology residents at Stanford Hospitals and Clinics and Lucille Packard Children’s Hospital consented to participate in this Stanford IRB approved-study. All subjects received routine clinical examination and treatment and had relevant ocular findings photographed with the iPhone 4s or 5 using our group's proprietary anterior segment adaptor. The photographs were uploaded to a secure, HIPAA compliant database for later review by two independent graders who were also provided with very brief clinical information and the initial consult question. They rated the quality of the photographs on a scale of 1-5 with 1 being “inadequate for any diagnostic purpose”, 3 “only able to exclude emergent findings” and 5 “ideal quality” for evaluation of each part of the eye. They also selected a triage decision for the patient - “see immediately”, “give next-day clinic follow-up”, and “see later same day." These were analyzed by an independent clinician for appropriateness of response compared to actual diagnosis and treatment found in the chart. Finally, readers were asked if they would be comfortable defending the triage decision in a malpractice suit situation.
Results
19 patient examinations with photographs were collected and fully evaluated by the readers. Of these, the triage decision made by the readers was appropriate for the diagnosis made by bedside examination 97.3% of the time. The two readers made the same triage decision 63.2% of the time and felt that they were comfortable defending the decision in a malpractice suit 73.7% of the time. Table 1 shows the percentage of images when the readers felt that the quality was at least a 3 - emergent findings could be excluded. This was over 50% for the external examination, eyelids, conjunctiva/sclera, cornea, iris, optic disc, macula, and vessels. However, the rate that emergent findings were excludable was <50% for the periphery, vitreous, lens, and anterior chamber.
Conclusions
Photographs of the anterior segment taken from iPhone4s or better may be potentially useful in the remote triage of patients. However, there are limitations to viewing some parts of the eye.
Keywords: 550 imaging/image analysis: clinical •
552 imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound)