Clinical Evaluation Of Automated Binocular Refraction Using Eye Refract® Compared With Conventional Refraction Techniques

Published 2026 - 44th Congress of the ESCRS

Reference: POREF204 | Type: Free paper | DOI: 10.82333/1mfj-x965

Authors: Dr Mariana Pires Garcia* ¹ , Dr Pedro Cardoso Teixeira ¹ , Dr João Chibante Pedro ¹ , Dr Lígia Figueiredo ¹

¹Ophthalmology Department,Unidade Local de Saúde de Entre Douro e Vouga, E.P.E.,Santa Maria da Feira,Portugal

Purpose

Before new automated refraction systems are integrated into routine clinical practice, validation under real-world conditions is essential. This study evaluates the agreement between automated binocular refraction using Eye Refract® (Visionix®, Luneau Technology, France) and conventional objective refraction with subjective refinement, as well as patient preference and experience with the automated method.

Setting

Ophthalmology Department, Unidade Local de Saúde Entre Douro e Vouga, Portugal.

Methods

This cross-sectional study included 79 patients undergoing routine ophthalmology consultation. Each participant underwent two refraction protocols: conventional refraction, consisting of autorefraction with Tono-Ref II® (Nidek Co., Ltd., Japan), followed by subjective refinement; and automated binocular refraction following Eye Refract algorithm. Sphere, cylinder, axis and spherical equivalent (SE) were analysed. Patient experience and preference between refraction methods were assessed using a structured questionnaire. A masked comparison of the two prescriptions was performed to evaluate subjective visual preference. Agreement was analysed using Wilcoxon signed-rank tests and Bland–Altman analysis.

Results

Mean age was 61.7±18.0 years, 59.5% were female. With conventional refraction, median sphere was +0.25D(1.25), cylinder −0.75D(0.81), axis 90°(37.5), and SE 0.00D(1.38), while with Eye Refract median sphere was +0.50D(1.69), cylinder −0.88D(0.94), axis 92°(39.5), and SE 0.00D(1.34). SE difference was statistically significant (p = 0.018), showing a slight hyperopic shift with Eye Refract. Bland–Altman analysis showed a mean bias of +0.17D(SD 0.63) with 95% limits of agreement −1.06 to +1.40 D. Regarding patient experience, 51.9% preferred Eye Refract, 49.4% considered it faster, and comprehension was equivalent in 41.8% and clearer with Eye Refract in 46.8%. In masked comparison, 68.6% reported better vision with Eye Refract prescription.

Conclusions

Automated binocular refraction using Eye Refract demonstrated good agreement with conventional subjective refraction, with a small hyperopic bias of limited clinical relevance. Patients showed a preference for the automated examination and frequently reported superior visual performance in masked prescription comparison, supporting its integration as a reliable alternative in routine clinical practice.