A Novel Automated Visual Acuity Test Using a Portable Head-mounted Display

• Published in: Optometry and vision science Vol. 97; no. 8; pp. 591 - 597
• Main Authors: Ong, Sze Chuan, Pek, Li Cheng (Ivy), Chiang, Tsuey Ling (Carol), Soon, Hock Wei, Chua, Kuang Chua, Sassmann, Chanakarn, Razali, Muhammad Azri Bin, Koh, Teck Chang (Victor)
• Format: Journal Article
• Language: English
• Published: United States Lippincott Williams & Wilkins 01.08.2020; American Academy of Optometry
• Subjects: Adult; Aged; Equipment Design; Female; Humans; Male; Middle Aged; Prospective Studies; Reproducibility of Results; Vision Disorders - diagnosis; Vision Disorders - physiopathology; Vision Tests - instrumentation; Visual Acuity - physiology
• ISSN: 1040-5488; 1538-9235; 1538-9235
• DOI: 10.1097/OPX.0000000000001551

SIGNIFICANCEWe developed a head-mounted display (HMD) as an automated way of testing visual acuity (VA) to increase workplace efficiency. This study raises its potential utility and advantages, analyzes reasons for its current limitations, and discusses areas of improvement in the development of this device. PURPOSEManual VA testing is important but labor-intensive in ophthalmology and optometry clinics. The purpose of this exploratory study is to assess the performance and identify potential limitations of an automated HMD for VA testing. METHODSSixty patients from National University Hospital, Singapore, were enrolled in a prospective observational study. The HMD was constructed based on the Snellen chart, with single optotypes displayed at a time. Each subject underwent VA testing of both eyes with the manual Snellen chart tested at 6 m from the subject and the HMD. RESULTSFifty-three subjects were included in the final analysis, with an incompletion rate of 11.7% (n = 7). The mean difference in estimated acuity between the HMD and Snellen chart was 0.05 logMAR. However, 95% limits of agreement were large at ±0.33 logMAR. The HMD overestimated vision in patients with poorer visual acuities. In detecting VA worse than 0.30 logMAR (6/12), sensitivity was 63.6% (95% confidence interval, 0.31 to 0.89%), and specificity was 81.0% (95% confidence interval, 0.66 to 0.91%). No significant correlation existed between mean difference and age (r = −0.15, P = .27) or education level (r = 0.04, P = .76). CONCLUSIONSAdvantages of our novel HMD technology include its fully automated nature and its portability. However, the device in its current form is not ready for widespread clinical use primarily because of its low accuracy, which is limited by both technical and user factors. Future studies are needed to improve its accuracy and completion rate and to evaluate for test-retest reliability in a larger population.