Repeatability and Validity of the Discovery System

Purpose: Evaluating an instrument’s repeatability and agreement are critical when utilizing it in research. The purpose of this thesis was to determine between-visit repeatability of the Discovery System aberrometer and the agreement between measurements made on two different Discovery Systems, a Grand Seiko WAM-5500 autorefractor, and the Complete Ophthalmic Analysis System (COAS-HD) aberrometer.

Methods: Twenty-five healthy, young adults participated in two study visits separated by no more than 14 days. At each visit, cycloplegic measurements were made on the right eye. Experiment 1) Five measurements were made using the same Discovery System at two visits. At Visit 2, ten measurements were also made using the Grand Seiko WAM-5500. Discovery System refractive error values were calculated for a 3-mm and 6-mm pupil diameter. Between-visit repeatability and between-instrument agreement were evaluated using difference versus mean plots and calculating the 95% limits of agreement (LoA). Experiment 2) At Visit 2, five measurements were also made on a second Discovery System and the COAS-HD. Instrument order was randomized at each visit. Zernike coefficients were calculated over a 6-mm pupil through the sixth radial order. Between-instrument agreement was assessed using difference versus mean plots and the 95% LoA.

Results: The mean age (±SD) and spherical equivalent refractive error were 23.4 ± 1.7 years and -2.91 ± 1.85 D, respectively. Experiment 1) There were no between-visit differences for M, J0, J45, higher-order root mean square (RMS), third-, fourth-, fifth-, and sixth-order RMS (all p > 0.25). Repeatability for defocus and higher-order RMS were ±0.31 D and ±0.095 microns, respectively. For a 3-mm pupil, the Discovery System measured more positive values than the Grand Seiko for M, J0, and J45 (p < 0.005). For a 6-mm pupil, there was no difference between instruments for M (p = 0.45). Experiment 2) There were no differences between the two Discovery Systems for fourth- and fifth-order RMS (all p > 0.39). There was a difference in third-order RMS (p < 0.005, 0.055 microns) and sixth-order RMS (p < 0.05, -0.005 microns) with horizontal coma (C 3, 1) accounting for the majority of the difference (p < 0.001, 0.010 microns). Between the Discovery System and COAS-HD, no differences were found for fifth- and sixth-order RMS, and spherical aberration (all p > 0.10). Differences were found for third-order RMS (p < 0.001, 0.154 microns) and fourth-order RMS (p < 0.05, -0.029 microns). Horizontal coma (C 3, 1) accounted for the majority of the third-order RMS difference between the two instruments (p < 0.001, 0.230 microns).

Conclusion: The Discovery System is repeatable and would be an appropriate instrument to measure cycloplegic refractive error and higher-order aberration changes. Although repeatable, this thesis found that Discovery System measurements are not interchangeable with those made on the Grand Seiko or COAS-HD. The designer of the Discovery System is in the process of completing an update to address the between Discovery System differences.