Accuracy of the Spot and Plusoptix photoscreeners for detection of astigmatism

Purpose To evaluate the accuracy of the Spot (V2.0.16) and Plusoptix S12 (ROC4, V6.1.4.0) photoscreeners in detecting astigmatism meeting AAPOS referral criteria in students from a population with high prevalence of astigmatism. Methods Students attending grades 3-8 on the Tohono O'odham reservation were examined. Screening was attempted with both the Spot and Plusoptix photoscreeners. Results were compared to cycloplegic refraction. Screening attempts providing no estimate of refractive error were considered fail/refer. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for detection of refractive errors were determined using AAPOS referral criteria and receiver operating characteristic area under the curve (ROC AUC) analysis was conducted for measures of astigmatism. Agreement between screening and cycloplegic refraction measurements of astigmatism, spherical equivalent, and anisometropia were assessed using t tests and correlation analyses. Results A total of 209 students were included. Of the total, 116 (55%) met examination-positive criteria based on cycloplegic refraction, with 105 of those (90%) meeting the criterion for astigmatism. Measurements success rates were 97% for Spot and 54% for Plusoptix. Comparing the Spot and the Plusoptix, sensitivity was 96% versus 100%, specificity was 87% versus 61%, PPV was 90% versus 76%, and NPV was 94% versus 100% for detection of refractive error. Both screeners overestimated astigmatism by 1/3 D to 2/3 D. AUC for astigmatism was 0.97 for Spot and 0.83 for Plusoptix. Conclusions In this highly astigmatic population, the Spot and the Plusoptix had similar sensitivity, but the Spot had better specificity and measurement success rates. Compared with results from study samples with lower rates of astigmatism, our results highlight the need to assess the ability of screening instruments to detect individual types of refractive errors.