Repeatability and reproducibility of a-scan biometry quantitative findings

To determine repeatability and reproducibility of A-scan biometry quantitative findings in order to uniformly evaluate lenticular cataract changes.

Seventeen eyes from seventeen cataract patients were examined by three (trainee ophthalmologists). Results (ratio of mean internal lens spikes height/lens anterior surface spike height) were compared using coefficient of variation (CV), repeatability coefficient (CR), and intraclass correlation coefficients (ICC).

Intra-observer consistency was CV = 0.36, CR = 0.49, and ICC = 0.778 for the first operator; CV = 0.33, CR = 0.52, and ICC = 0.642 for the second operator; and CV = 0.32, CR = 0.53, and ICC = 0.567 for the third operator. Concerning inter-observer agreement: CV (in 6 of 7 eyes) was > 8%, CR was > 0.21, and intraclass correlation coefficient for all eyes was < 0.81.

Regarding repeatability and reproducibility, although there is a tendency for statistical significance, variability of the results does not allow the use of the method as a complementary clinical tool for comparing results. Standardization of this procedure, in order to achieve higher test reliability, might be the aim of relevant future studies, although there are a number of points that have to be addressed.