Design: Cross-sectional paired-test study within a national cohort
Population: 995 children aged 3–15 in Shanghai, both charts in random order
Primary outcome: Uncorrected visual acuity (logMAR) and referral accuracy
Effect: Preschool referral 18.6% vs 31.2%; specificity 0.93 vs 0.79

In this cross-sectional study within a Chinese national cohort, 995 children aged 3 to 15 in Shanghai had uncorrected acuity measured with both HOTV and tumbling E logMAR charts, in random order with a rest between, plus cycloplegic refraction and full examination.

Acuity read better on HOTV overall (0.15 vs 0.20 logMAR). The gap was largest in the youngest — about one line at age 3 — and closed to a letter or less by age 8. Among 183 preschool children, HOTV referred fewer (18.6% vs 31.2%), detected more true cases (true positive rate 0.68 vs 0.44) and had better specificity (0.93 vs 0.79).

Tumbling E asks young children to indicate direction, which tests orientation and cognition as well as vision. For school and community screening programmes, which in India commonly use E charts, switching to a matching chart for children under 8 could reduce unnecessary referrals while picking up more true problems.

Consider HOTV letter-matching charts rather than tumbling E for vision screening in children under 8.
Expect tumbling E to underestimate acuity by up to a line in 3-year-olds.
Use age-appropriate referral thresholds for the chart in use.
Use either chart in children 8 and older, where results were similar.

Why it matters

It changes the chart that many screening programmes use for the youngest children.

The statistics, in plain English

Specificity of 0.93 versus 0.79 means HOTV wrongly referred about 7 in 100 children without a problem, compared with about 21 with tumbling E. Each child did both tests, so differences between children do not explain the result.