Purchase this article with an account.
Tomohiko Yamada, David Leske, Pamela Moke, Nick Parrucci, James Ruben, Jonathan Holmes; Comparing Failure Criteria for a New PC-Based Pediatric Vision-Screening Test. Invest. Ophthalmol. Vis. Sci. 2013;54(15):3985.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
We previously reported the development and initial validation of the Jaeb Visual Acuity Screener (JVAS), a standardized PC-based optotype vision-screening test. In our previous study, sensitivity and specificity of the JVAS were found to be comparable to established automated screening methods in 3- to 7-year old children. Nevertheless, different failure criteria can be applied to the JVAS, which may influence its sensitivity and specificity. We originally used a criterion of 3 of 4 letters, but we also collected data on screening up to 5 letters. The purpose of this present study was to evaluate the sensitivity and specificity of 3 failure criteria by comparing against a gold-standard eye examination.
142 children aged 3 to 7 years (mean 5.3 years) seen in a pediatric optometry practice were screened with the JVAS prior to undergoing a complete eye examination including cycloplegic Retinomax autorefraction (gold standard). Testing with the JVAS commences with 2 large single surround optotypes (HOTV for <7 years and e-ETDRS for 7 years), followed by up to 5 optotypes at an age-specific threshold level (3 to 5 years 20/40, 6 years 20/32, 7 years 20/25). Three JVAS failure criteria were evaluated: inability to correctly identify 3 of 4, 3 of 5, or 4 of 5 letters at the age-specific threshold level. Gold-standard examination failure criteria were defined by published AAPOS criteria for amblyogenic factors to be detected (J AAPOS 2003;7:314-316) with updated visual acuity age norms (J AAPOS 2008;12:145-149 and Optom Vis Sci 2009;86:607-612).
66 (46%) of 142 children met gold-standard exam failure criteria. For the JVAS 3 of 4 criteria, sensitivity was 83% and specificity 82%, with a positive predictive value (PPV) of 80% and a negative predictive value (NPV) of 85%. For the JVAS 3 of 5 criteria, sensitivity was 80% and specificity 91%, with a PPV of 88% and a NPV of 84%. For the JVAS 4 of 5 criteria, sensitivity was 85% and specificity 80%, with a PPV of 79% and a NPV of 86%.
The new JVAS provides a universal standardized screening method for 3- to 7-year-old children using any standard PC. Setting the JVAS failure criteria at an inability to correctly identify 4 of 5 age-specific letters yielded the highest sensitivity, whereas setting the criteria at 3 of 5 yielded the highest specificity. Using failure to identify 3 of 4 letters provides a balance of sensitivity and specificity.
This PDF is available to Subscribers Only