Peek acuity as an alternative method for visual acuity measurement for vision screening in children

• Main Authors: Mohd Noordin, Amirul Asraf, Che Azemin, Mohd Zulfaezal, Musa, Arni, Abdul Rahim, Muhammad Afzam Shah, Shafee, Noor Wafirah, Ahmad, Norsham, Md Mustafa, Md-Muziman-Syah, Yusof @ Alias, Firdaus
• Format: Article
• Language: English
• Published: IIUM Press, International Islamic University Malaysia 2023
• Subjects: RE Ophthalmology
• Online Access: http://irep.iium.edu.my/108745/6/108745_Peek%20acuity%20as%20an%20alternative%20method.pdf; http://irep.iium.edu.my/108745/; https://journals.iium.edu.my/ijahs/index.php/IJAHS/article/view/865
• Institution: Universiti Islam Antarabangsa Malaysia

Introduction: Accurate and reliable clinical measurement of visual acuity (VA) is necessary to establish the needs for further clinical investigations. A simpler option for VA measurement is the use of smartphones, in which the measurement method is programmed in an application that enables automated and self-administered VA testing. Aims: This study aimed to investigate the validity of PEEK Acuity (PA), a smartphone’s application for VA measurement, in testing children’s vision, especially for vision screening programs. Methodology: Data were obtained from a vision screening program involving children aged 3–13 years. One hundred thirteen participants were included in the study (mean age 7.47±2.33years old, 59 boys). The participants’ VA was measured using the PA and Snellen chart in random order. Both eyes were tested, with 226 data points used in the statistical analysis. The Kolmogorov-Smirnov test revealed non-normal data. The VA between the PA and Snellen tests was compared using the Wilcoxon signed rank test. Test-retest variability was calculated for both tests using intraclass correlation (ICC) analysis. Spearman’s rank correlation and the Bland-Altmann test were plotted to evaluate the agreement between tests. Results: The mean VA recorded using PA was 0.066±0.007 logMAR, while that recorded using Snellen was 0.103±0.009 logMAR (z=-4.119, p<0.0005, Wilcoxon signed rank test; ≈ two-letter differences). The ICC between the PA and Snellen was 0.641.Spearman’s rank correlation revealed a significant correlation between PA and Snellen (r=0.508, p<0.001). Conclusion: There was a moderate correlation between the PA and Snellen scores. The PA application showed good agreement with conventional Snellen. Thus, smartphone applications such as PA may serve as an alternative to measure VA in children, especially in vision screening programs.