HOTV letter-matching testing may offer advantages over tumbling E testing for vision screening in younger children, particularly those younger than 8 years.
Investigators conducted the cross-sectional observational study in Shanghai, China, as part of a national multicenter cohort study. The analysis included 995 children aged 3 to 15 years who completed both HOTV and tumbling E uncorrected visual acuity tests in randomized order, separated by a 30-minute rest period. Participants underwent comprehensive ophthalmic examinations, although cycloplegic refraction data were unavailable for 341 preschool children.
The primary outcome was mean uncorrected visual acuity (UCVA), measured using logarithm of the minimum angle of resolution (logMAR). Secondary outcomes included referral rates for reduced UCVA, screening performance for amblyopia and refractive errors, testability, and testing time.
Because performance on both tests was poorer when the test was administered second, investigators used only the first test administered to compare mean UCVA. This analysis included 980 children. Overall, mean UCVA was 0.15 logMAR, approximately 20/25, with HOTV vs 0.20 logMAR, approximately 20/32, with tumbling E. The difference was greatest among children aged 3 to younger than 4 years, reaching 1 logMAR line. Differences remained greater than 3 letters among children younger than 8 years but narrowed to 1 letter or less among those aged 8 years or older.
Among 183 preschool children, 19% met referral criteria for reduced UCVA with HOTV vs 31% with tumbling E, a difference of approximately 13 percentage points. The difference narrowed with age: referral rates were 27% vs 29%, respectively, among 244 lower primary school children and 64% vs 61% among 227 upper primary and secondary school children.
Screening performance showed a similar age-related pattern. Among preschool children, the true-positive rate was 0.68 with HOTV vs 0.44 with tumbling E, while specificity was 0.93 vs 0.79. Among lower primary school children, true-positive rates were 0.83 vs 0.76, respectively. Screening performance was similar between the tests among older children.
Of 21 children with suspected amblyopia, would not have met the suspected amblyopia criteria based on tumbling E results alone. For refractive-error screening, the tests generally performed comparably, although HOTV had modestly higher performance for detecting myopia and astigmatism. The authors noted that the small numbers of refractive errors in the younger age groups may have increased variability in these estimates.
Testability was high with both approaches. Among 1,053 children assessed, 95% completed HOTV testing and 97% completed tumbling E testing. Among 69 preschool children without previous vision-testing experience, testability was 80% with HOTV vs 83% with tumbling E. Among 53 inexperienced preschool children who completed both tests, HOTV required approximately 6 additional seconds for instruction and 8 additional seconds for testing, or approximately 14 additional seconds overall.
The study was limited by the administration of both tests on the same day. Despite the 30-minute rest period, performance was poorer when either test was administered second, suggesting a fatigue effect and limiting direct within-child comparisons. Testing on separate days could provide a more valid intraindividual comparison. Both approaches also produced false-negative results, particularly for hyperopia among preschool children, with no differences in sensitivity or false-negative rates observed across educational stages.
The findings supported consideration of HOTV testing for vision screening in children younger than 8 years, with the largest differences between the tests observed at younger ages.
Disclosures: Morgan reported personal fees from Eyerising International outside the submitted work. Xiangui He reported grants from the National Natural Science Foundation of China during the conduct of the study.
Source: JAMA Ophthalmology
