Pediatric Amblyopia Risk Investigational Study (PARIS)

Abstract

Abstract: : Purpose: We sought to compare the accuracy of pediatric screening (for amblyopia and its risk factors) performed by physician extenders to that performed by optometrists and ophthalmologists. We determined test re-test reliability for LEA visual acuity (VA) and non-cycloplegic autorefraction by physician extenders, and compared these findings to those obtained by an optometrist or ophthalmologist. Methods: We prospectively evaluated 200 consecutive children, aged 3-6 years. Screening exams by physician extenders consisted of VA, Random Dot E Stereotest (RDE), and two noncycloplegic autorefractions employing a SureSight^tm autorefractor. More formal exams by an ophthalmologist or optometrist included cycloplegic refraction (CRNS), cycloplegic and noncycloplegic SureSight^tm autorefraction (AR), and classical pediatric ophthalmologic evaluations. Results from each group of examiners were compared and analyzed using Pearson correlation coefficients and univariate regression analysis performed with ANOVA statistical programs. Results: Our patient cohort was 46% female, and 60% African-American. Inter-test reliability of VA was moderate (R=0.59, p=0.0001), with a mean difference of 0.1 (SD 0.1) LogMAR units. Intra-test and inter-test reliability for noncycloplegic AR was very high for astigmatism (R=0.89, p=0.0001), moderate for spherical equivalent (SE) (R=0.56, p=0.0001), and weak for anisometropia (R=0.26, p=0.005). Agreement between noncycloplegic AR and CRNS was high for astigmatism (R=0.83, p=0.0001), moderate for SE (R=0.55, p=0.0001), yet not correlated for anisometropia (R=0.12, p=0.18). Correlation between cycloplegic AR and CRNS was good for astigmatism (R=0.77, p=0.0001), high for SE (R=0.85, p=0.0001), and moderate for anisometropia (R=0.48, p=0.0001). Conclusions: Noncycloplegic AR using a SureSight^tm autorefractor provided a fast and reliable assessment of astigmatism whether performed by physician extenders or more trained professionals. Noncycloplegic AR provided a less valid assessment for SE and an unreliable assessment of anisometropia. Accuracy of the autorefractor for SE and anisometropia increased with cycloplegia. Noncycloplegic AR appears to be valuable and amenable to use by office personnel in the recognition of astigmatism. Cycloplegic AR improves detection of ametropia and anisometropia. Detection of such amblyopic risk factors makes SureSight^tm autorefraction a valuable device in screening for amblyopia, even when performed by physician extenders.

Keywords: screening for ambylopia and strabismus • astigmatism