2
Photoscreening for Refractive Error and Strabismus With a Smartphone App

Saturday, October 26, 2013: 8:43 AM
W304 A (Orange County Convention Center - West Building)
Joannah Vaughan, M.B.A., Talitha Dale, B.S. and David Huang, MD, PhD, Casey Eye Institute, Oregon Health & Science University, Portland, OR

Objectives

To detect risk factors for amblyopia using a smartphone app.

Methods

The iCheckKids photoscreening app (iCheck Health Connection, Inc., Portland, OR) was used with an iPhone 5.  The app takes standardized flash photographs at a target working distance of 28 inches in 2 modes (portrait and landscape).  Refraction was measured using crescent width, pupil diameter, and corneal diameter from the 2 photographs, according to the eccentric photorefraction principle.  Strabismus was measured by the position of the corneal reflex relative to the pupil and cornea.  Children between the ages of 3 and 5 years referred for eye examination at the Elks Preschool Vision Screening Program were examined on 1 day.  Photoscreening were performed in a dimly lit room prior to cycloplegic eye drops.  Clinical examination including cover test and cycloplegic retinoscopic refraction were then performed.  AAPOS referral criteria for refractive error and strabismus were used.  

Results

There were 23 subjects were examined.  4 subjects (17%) were excluded from analysis due to poor photographic quality: 3 were gazing off axis and 1 had small pupils.  Of the remaining 19 subjects, 7 were referred by clinical examination: 5 for astigmatism, 1 for myopia, and 1 for exotropia.  iCheckKids correctly referred 6/7 of these subjects (86% sensitivity), 1 subject with astigmatism was missed.  Of the 6 subjects referred by both clinical and iCheckKids criteria, the diagnosis agreed in 5 of the 6, but in one case the clinical examination identified astigmatism while iCheckKids identified exotropia.  There were no false positive referrals by iCheckKids (100% specificity). 

Conclusions

Smartphone photoscreening could be an effective method of detecting risk factors for amblyopia.  Diagnostic accuracy appeared adequate but better quality control is needed through training and automation.  Automated quality analysis software is needed for real-time feed back to the operators, particular with regard to patient fixation and pupil size.