PHANTOM is a novel open-field binocular integrated refraction system that uses device-specific binocular target presentation and lensless refractive simulation to perform objective refraction, subjective refraction, and visual acuity testing in one device. However, evidence is limited regarding how corrective refraction derived from such systems compares with prescriptions obtained using conventional subjective refraction combined with binocular balancing. We aimed to compare refractive correction determined using the conventional method with that obtained using PHANTOM and evaluate its potential role as an adjunctive tool for spectacle prescription. We compared PHANTOM with a conventional clinical refraction workflow. Twenty healthy young adults (20.3 ± 0.9 years) were subjected to both conditions in a randomized order, with examinations performed by one experienced orthoptist; right eye data were analyzed. Conventional refraction comprised monocular autorefraction plus subjective refraction with prism-dissociated duochrome binocular balancing. Agreement (PHANTOM–conventional) was assessed using the Bland–Altman analysis. Objective spherical equivalent (SE) was − 3.09 ± 2.25 D vs. −3.10 ± 2.24 D (bias 0.02 D; 95% limits of agreement LoA − 0.55 to 0.58). Subjective SE was − 2.78 ± 2.41 D vs. −2.79 ± 2.39 D (bias 0.01 D; LoA − 0.56 to 0.58) with no proportional bias. Proportional bias was detected in objective J45 (p = 0.01) and subjective J0 (p = 0.006). Visual acuity was − 0.17 ± 0.02 vs. −0.16 ± 0.04 logMAR (bias − 0.01; LoA − 0.10 to 0.08). Examination time was 199.7 ± 33.9 vs. 527.6 ± 70.0 s (61.6% shorter; p < 0.0001). In healthy young adults under controlled study conditions, participants completed the integrated PHANTOM protocol and achieved visual-acuity endpoints comparable to those obtained with conventional refraction. Under these conditions, SE agreement was generally good and the examination time was substantially reduced, whereas astigmatism components warrant further investigation. Further validation in broader clinical populations is required before these findings can be generalized.
Iwata et al. (Tue,) studied this question.