Near Visual Acuity Measurements by Community Screeners Using Digital (Peek) Testing Versus Conventional Charts in India

The Peek digital near vision test has been previously validated in a trial setting; here it is assessed in clinic (stage 1) and community (stage 2) settings. The study was carried out in the catchment area of Dr. Shroff's Charity Eye Hospital, Mohammadi, Uttar Pradesh, India, with a total of 768 participants. Stage 1 assessed the interobserver variability of Peek near vision impairment (NVI) screening in 168 clinic participants, with three trained community screeners. Stage 2 compared Peek to conventional chart testing for NVI screening using Cohen's kappa coefficient, sensitivity, and specificity and, for quantitative near visual acuity (NVA) measurement, using Bland-Altman limits of agreement (LoA) in 600 participants with two screeners. In stage 1, interobserver variability using Peek ranged from 96.43% to 98.21% (kappa = 0.92-0.96). In stage 2, there was overall agreement in 95.8% of cases (kappa = 0.91). Peek testing had a sensitivity and specificity of 91.25% (95% confidence interval [CI], 87.22-94.1) and 99.41% (95% CI, 97.86-99.84), respectively. For NVA testing, the 95% LoA between Peek and chart testing were within -0.11 and +0.07 logMAR. Mean test time was 40.3 seconds (95% CI, 38.8-41.7) for Peek versus 46.6 seconds (95% CI, 45.5-47.7) for a conventional chart. The previously demonstrated validity of Peek testing was maintained when used by trained community screeners. The Peek near vision test can be used as a validated method of NVA/NVI measurement in research, clinical and community settings.