Modification of Plusoptix referral criteria to enhance sensitivity and specificity during pediatric vision screening

Abstract

To determine the impact of using several different proposed sets of referral criteria on the specificity and sensitivity of the plusoptiX S08 photoscreener for detecting amblyopia risk factors. During a 2-month period, 144 children ages 9 months to 14 years were screened at the Tennessee Lions Eye Center before receiving a comprehensive eye examination and cycloplegic refraction. Three previously published sets of referral criteria were used for screening, including the manufacturer's criteria and the criteria proposed by Arthur and colleagues, which are nearly identical to the gold standard examination failure thresholds proposed by the Vision Screening Committee of the American Association of Pediatric Ophthalmology and Strabismus (AAPOS). Modifications of these criteria also were evaluated. The screening results obtained by the plusoptiX S08 were compared with the results from the gold standard pediatric ophthalmologic examination, and the respective sensitivities and specificities of each set of referral criteria in detecting amblyopia risk factors identified by the AAPOS Vision Screening Committee were calculated. The manufacturer's criteria yielded high sensitivity (100%) but very low specificity (37%). The Arthur criteria, which used the values for the AAPOS-defined amblyopia risk factors as referral criteria, maintained sensitivity (89%) and greatly improved specificity (76%). Two modifications of the Arthur criteria further increased specificity with minimal loss of sensitivity. The manufacturer's criteria have excellent sensitivity but low specificity, warranting modification; other criteria increase specificity with minimal effect on sensitivity.