Effect of letter separation on computerised visual acuity measurements: comparison with the gold standard Early Treatment Diabetic Retinopathy Study (ETDRS) chart


  • Competing interests: DAHL and Guys and St Thomas’ Foundation NHS Trust have proprietory interests in the COMPlog visual acuity measurement system. The other authors do not.

Correspondence and reprint requests to: D. Alistair H. Laidlaw.
Tel.: +44 (0) 20 7188 4320; Fax: +44 (0) 20 7188 9193.
E-mail address: alistair.laidlaw@gstt.nhs.uk


Contour interaction plays an important role in letter recognition and has long been known to influence clinical visual acuity measurements. LogMAR charts originally employed between – letter and line spacing of one letter width/height; however, in more recent designs, half letter width spacing has become accepted without validation. COMPlog is a computerised visual acuity measurement device being developed and validated for routine and research clinical use and is available commercially. This study was performed to investigate the effect of letter separation on computerised visual acuity measurements. An iterative series of experiments was performed in which test and retest fully interpolated five letters-per-line logMAR visual acuity measurements were taken in three groups of a total 100 adult subjects. Each group had visual acuity measured using ETDRS charts 1 and 2 and on COMPlog using various combinations of letter and line spacing. The first group consisted of 31 subjects tested using 3.75 and 2.50 stroke width spacing, the second 45 subjects using 2.50 and 1.25 stroke width spacing and the last group, 24 subjects tested using 1.90 stroke width spacing. The methods of Bland and Altman were used to analyse the data. No significant bias was noted between the gold standard ETDRS measurements and those taken using COMPlog with 3.75 stroke width spacing. Significant bias of half a logMAR line was found with crowding of 1.90 and 1.25 stroke widths. A small systematic bias with a reduction of 1 ETDRS letter was found with spacing of 2.50 stroke widths which has not been demonstrated in previous validation studies. Crowding of 2.50 stroke widths is the smallest spacing which can be used in the design of COMPlog.