Recognition acuity in children measured using The Auckland Optotypes

Abstract

Purpose: Sloan letters displayed by the Electronic Visual Acuity (EVA) system are the gold standard for recognition acuity measurement in research settings. However, letters are not always appropriate for children. The Auckland Optotypes (TAO) are a new, open-access set of 10 pictograms available in regular and vanishing formats. We sought to assess feasibility of using both formats of TAO for measuring visual acuity (VA) in children using a Bayesian adaptive staircase, in a community setting. Methods: We tested 121 children (5–12 years old) with both formats of TAO, a handheld flipchart vision screener (Parr vision test), as well as the gold standard EVA. We measured feasibility of the three comparison tests in three ways. First, using limits of agreement (LoA) with EVA, second, calculating area under the receiver operating characteristic curve (AUC), and finally, investigating trial-by-trial responses. Results: Agreement between tests was within test-retest reliability of EVA measures (LoA TAOregular  = ±0.14, LoA TAOvanishing  = ±0.15, LoA Parr  = ±0.16 logMAR). TAO tests were highly effective at identifying children with vision impairment (AUC TAOregular  = 0.96, AUC TAOvanishing  = 0.95), whereas Parr was less effective (AUC Parr  = 0.82). In 5–6 year old children there was an enhanced advantage of TAO (AUC TAOregular  = 0.97, AUC TAOvanishing  = 0.98) over Parr (AUC Parr  = 0.75). Although each child completed 16 trials, approximately 10 trials were sufficient to achieve excellent LoA, and six trials sufficient for accurate screening. Conclusion: Threshold VA assessment and vision screening are feasible using both vanishing and regular formats of TAO.