Automated Detection of Horizontal Strabismus and Home Control Evaluation Using a Mobile Phone App

Abstract

Purpose: To evaluate the reliability, validity, and usability of a mobile phone app that captures facial images to identify strabismus.Methods: We enrolled patients who visited our pediatric strabismus clinic and provided consent for participation in the study. The facial photographs of the patients were automatically captured by the app. The app used the photographs to identify ocular misalignment based on the comparison of the positions of the corneal reflex and limbus of both eyes. The results from the app were compared to the physician’s diagnosis. Intraclass correlation coefficients were used to assess the test-retest and interrater reliability of the app. Additionally, the guardians captured photographs of the patients through this app for 1 month at home to evaluate the frequency of manifest exotropia at home. And the results were compared to those of other strabismus control scales.Results: A total of 103 patients were enrolled and 10 patients were enrolled for the home control test. Based on the physician’s diagnosis, the sensitivity, specificity, positive predictive value, and negative predictive value of the app were 74.6%, 87.5%, 90.4%, and 68.6%, respectively. The interclass correlation coefficients for inter-rater and test-retest reliability were 0.801 (<i>p</i> = 0.011) and 0.828 (<i>p</i> < 0.001), respectively. The frequency of intermittent exotropia recorded by the app at home showed a strong positive correlation with other control scales (rho > 0.770, <i>p</i> = 0.009).Conclusions: This mobile phone app is easily accessible and useful for the rapid determination and recording of strabismus, particularly in uncooperative patients. Also, this app may be used as a reliable indicator of the frequency of manifest strabismus at home.