Aims Clinically, picture acuity tests are thought to overestimate visual acuity (VA) compared with letter tests, but this has not been systematically investigated in children with amblyopia. This study compared VA measurements with the LogMAR Crowded Kay Picture test to the LogMAR Crowded Keeler Letter acuity test in a group of young children with amblyopia.

Methods 58 children (34 male) with amblyopia (22 anisometropic, 18 strabismic and 18 with both strabismic/anisometropic amblyopia) aged 4–6 years (mean=68.7, range=48–83 months) underwent VA measurements. VA chart testing order was randomised, but the amblyopic eye was tested before the fellow eye. All participants wore up-to-date refractive correction.

Results The Kay Picture test significantly overestimated VA by 0.098 logMAR (95% limits of agreement (LOA), 0.13) in the amblyopic eye and 0.088 logMAR (95% LOA, 0.13) in the fellow eye, respectively (p<0.001). No interactions were found from occlusion therapy, refractive correction or type of amblyopia on VA results (p>0.23). For both the amblyopic and fellow eyes, Bland-Altman plots demonstrated a systematic and predictable difference between Kay Picture and Keeler Letter charts across the range of acuities tested (Keeler acuity: amblyopic eye 0.75 to −0.05 logMAR; fellow eye 0.45 to −0.15 logMAR). Linear regression analysis (p<0.00001) and also slope values close to one (amblyopic 0.98, fellow 0.86) demonstrate that there is no proportional bias.

Conclusions The Kay Picture test consistently overestimated VA by approximately 0.10 logMAR when compared with the Keeler Letter test in young children with amblyopia. Due to the predictable difference found between both crowded logMAR acuity tests, it is reasonable to adjust Kay Picture acuity thresholds by +0.10 logMAR to compute expected Keeler Letter acuity scores.