A comparison of two automated devices that measure refractive error

The present study compared the reliability, validity, and screening effectiveness of the Welch Allyn SureSight Autorefractor (WASS) and the PlusoptiX S09 Vision Screener (PS09). Eighty-nine children attending appointments at a pediatric ophthalmology clinic were tested twice each with the PS09 and the WASS. Each child then completed the gold standard examination of refractive error, cycloplegic retinoscopy, with one of two pediatric ophthalmologists. Refractive error scores from the two devices were compared to cycloplegic retinoscopy. Results indicated that the PS09 yielded better reliability than the WASS on both spherical refractive error (Coefficients of Repeatability [CORs] = 1.21D v. 1.63D) and cylindrical refractive error (CORs = 0.50D v. 0.58D). Although the PS09 yielded better agreement with cycloplegic retinoscopy than did the WASS on spherical refractive error (CORs = 3.53D v. 4.19D), the validity of both devices was quite poor. Furthermore, both devices significantly underestimated hyperopia. Compared to the PS09, the WASS yielded slightly better agreement with cycloplegic retinoscopy on cylindrical refractive error (CORs = 0.87D v. 1.06D). In terms of screening effectiveness, the WASS yielded superior sensitivity (WASS = 69%; PS09 = 46%), but the PS09 yielded superior specificity (PS09 = 90% v. 54%). These results demonstrate the impact of the selection of pass/fail criteria, and therefore, the screening accuracy of each device was also calculated. The PS09 was the more accurate device (PS09 = 66%; WASS = 62%). In all, the analyses suggest that the PS09 is the superior device, but only by a very small margin.