[The Freiburg Vision Test. A computer-assisted procedure with sequential strategy].

Single Landolt Cs were presented in one of four positions on a monitor. Subjects responded by pressing one of four buttons (forced choice). A computer selected the size of the Landolt Cs on a logarithmic scale using the "Best PEST" algorithm (Best Parameter Estimation by Sequential Testing), a modern procedure to measure psychophysical thresholds. For comparison, conventional measurements according to DIN 58220 (Deutsche Industrie Norm) were performed with Landolt Cs projected in eight positions, using three out of five correct responses as the threshold criterion. Examination of 23 patients (most of them with senile cataract) did not show any significant difference between the two tests in either visual acuity or reproducibility on two consecutive days. The coefficients of variation between sessions were 22% for the Freiburg test and 26% for the DIN test and 18% between the two tests, pooled over two sessions. The Freiburg acuity test thus appears to be numerically equivalent to the DIN 58220 procedure. In addition, it has a number of advantages: (1) examiner-dependent variability is reduced; therefore, the test can be used by inexperienced examiners; (2) the forced choice technique is used rigorously; (3) mistakes in confounding oblique directions are largely avoided; (4) the procedure is about twice as fast.