The precision of wavefront refraction compared to subjective refraction and autorefraction.

PURPOSE To determine the precision (repeatability) of several methods of calculating refraction from higher-order wavefront aberration data and to compare these wavefront refractions with lower-order (LO) wavefront refraction, subjective refraction, and autorefraction. METHODS Four clinicians refracted 16 normal participants aged 23.6 +/- 1.2 years, 69% female with an average spherical equivalent refractive error of -3.03 +/- 2.55 D, median sphere -2.50 D (minimum -7.50, maximum +4.75), and median cylinder -0.50 D (minimum -3.00, maximum 0). Participants were cyclopleged and underwent subjective refraction, autorefraction on two machines (Nidek AR-800, Topcon KR-8000), and wavefront sensing using the Wavefront Sciences Complete Ophthalmic Analysis System. Wavefront error was used to calculate: LO refraction, refractions that incorporated higher-order spherical and astigmatism terms from up to the 4th, 6th, and 10th orders (PCM4, PCM6, and PCM10), and a method based on optimizing image quality metrics [wavefront analysis technology (WAT) refraction]. Within and between examiner agreements for total dioptric difference were determined using Bland-Altman limits of agreement (LOA). RESULTS The interexaminer LOA for individual measurements for M, J0, J45 were: Topcon (+/-0.18, +/-0.10, +/-0.06), Nidek (+/-0.28, +/-0.16, +/-0.09), LO (+/-0.17, +/-0.10, +/-0.06), PCM4 (+/-0.26, +/-0.09, +/-0.06), PCM6 (+/-0.37, +/-0.17, +/-0.34), PCM10 (+/-0.54, +/-0.32, +/-0.40), WAT (+/-0.28, +/-0.20, +/-0.15), and subjective refraction (+/-0.48, +/-0.20, +/-0.13) and averaging across three measures LOA: Topcon (+/-0.15, +/-0.08, +/-0.05), Nidek (+/-0.21, +/-0.13, +/-0.07), LO (+/-0.12, +/-0.06, +/-0.04), PCM4 (+/-0.16, +/-0.05, +/-0.04), PCM6 (+/-0.23, +/-0.09, +/-0.19), PCM10 (+/-0.29, +/-0.19, +/-0.24), and WAT (+/-0.18, +/-0.12, +/-0.10). The within-examiner LOA for M, J0, J45 were: Topcon (+/-0.08, +/-0.04, +/-0.02), Nidek (+/-0.13, +/-0.07, +/-0.05 D), LO (+/-0.11, +/-0.07, +/-0.04), PCM4 (+/-0.17, +/-0.07, +/-0.04 D), PCM6 (+/-0.28, +/-0.12, +/-0.24 D), PCM10 (+/-0.42, +/-0.24, +/-0.32 D), and WAT (+/-0.19, +/-0.14, +/-0.09 D). CONCLUSIONS All objective refractions except for PCM10 were more repeatable across clinicians than subjective refraction. The precision of all refractions were improved by an expected amount through averaging over multiple measurements. Wavefront refractions were not as precise as standard autorefractions, although not clinically significantly worse.