Autorefractors are used for objective refraction. They are widely used in eye care, as they provide fast measurements and do not require a trained operator. The accuracy and repeatability of autorefractors have been investigated since their launch on the market. They are proven to be not accurate enough to replace subjective refraction. It was shown that conventional instruments with built-in artificial targets overestimate myopia. Open-field autorefractors, which measure under natural conditions, were proven to be more accurate as they relax accommodation more efficiently. Factors influencing the accuracy of autorefraction have also been studied. Autorefractors are more reliable in adults compared to children and when cycloplegia is applied.
This dissertation consists of 3 parts. The first one aimed to evaluate the accuracy of the conventional autorefractor Nidek ARK 510A, the open-field autorefractor Shin-Nippon NVision-K 5001, and the aberrometer Visionix L80 Wave+. The second part analyzed short-term refractive state variation of the eye using the mathematical formalism of h-vectors. The purpose was to investigate the impact of viewing conditions (artificial target vs. natural view at distance), accommodative dysfunctions, binocular vision anomalies, and ocular surface diseases on the accuracy of autorefraction and the type and range of patterns of the short-term refractive state variation. The aim of the third part of the dissertation was to compare the accuracy of the fotorefractor Plusoptix A12C, autorefractors ARK 510A and NVision-K 5001 among children and adolescents. This study also investigated the effects of visual anomalies on autorefraction accuracy and short-term refractive variation.
In the first study, the autorefraction measurements, obtained by ARK 510A, L80 Wave+, and NVision-K 5001, were compared with subjective refraction. The study was conducted among 51 young adults aged 19–23 years. Conventional spherocylindrical notation was transformed to spherical equivalent (SE) and Jackson cross-cylinder components J0 and J45. Mean differences in SE between subjective refraction and autorefraction were 0.25 ± 0.53 D (p < 0.001), 0.34 ± 0.55 D (p < 0.001), and −0.18 ± 0.42 D (p < 0.001) for ARK 510A, L80 Wave+, and NVision-K 5001, respectively. They were statistically significant. Aberrometer gave the most accurate astigmatism measurements.
In the second study, 64 adults aged 23–60 were examined. A full optometric examination, autorefraction (20 readings for each eye) using ARK 510A and NVision-K 5001, and the OSDI eye surface disorders diagnostic questionnaire were performed. The spherocylindrical notation was converted into vector h and plotted in an orthogonal, three-dimensional coordinate system. The impact of visual anomalies on the accuracy of autorefraction was statistically insignificant. The graphical analysis allowed to observe a large variety of types and ranges of the short-term refractive variation. For the most cases, the type of pattern was dependent on the viewing conditions (artificial vs. natural). The impact of accommodative dysfunctions on the results from the conventional autorefractor, and the impact of binocular vision anomalies on the results from the open-field instrument were observed in analysis of individual plots.
The third part of the study evaluated the accuracy of autorefraction without cycloplegia in school-aged children and adolescents. The study compared results from A12C, ARK 510A, and NVision-K 5001 with subjective refraction. The smallest difference was observed for NVision-K 5001, and the biggest – for ARK 510A. The short-term refractive state variation from ARK 510A and A12C changed mainly in the spherical value, which indicates the accommodative state changes. The distributions from NVision-K 5001 were more complex and characterized by high variability of the axis and cylinder power, which is caused by unstable fixation during observations in natural conditions.