Abstract
Purpose::
To compare the agreement in the assessment for glaucomatous change in serial optic disc stereo photos between experts and non-experts and to re-evaluate the agreement between non-experts after a training session.
Methods::
Three experts (senior readers for the European Glaucoma Prevention Study) evaluated for change a set of two serial optic disc color stereo-slides of 40 patients, masked from the temporal sequence of the slides (Ophthalmology 2003;110:340-4). Thirty seven (37) non-expert ophthalmologists independently evaluated twice the same set of optic disc stereo slides. The second evaluation was masked from the previous evaluation and took place the same day after a training session on a separate slide set. Kappa statistics were used to evaluate the inter-observer agreement. A Mann-Whitney U test was used to compare the kappa coefficients between experts and non-experts. A Wilcoxon signed-rank test was used to compare the kappa coefficients of the non-experts before and after the training session.
Results::
The mean inter-observer kappa coefficient of the experts and non-experts was 0.55 (95% CI: 0.47-0.63) and 0.33 (95% CI: 0.27-0.40), respectively (Mann-Whitney U test, p=0.014). After a training session, the mean kappa coefficient of the non-experts increased to 0.38 (95% CI: 0.32-0.43) (signed rank test, p=0.068).
Conclusions::
The agreement between non-expert ophthalmologists in assessing serial optic disc stereo photos was poor but improved to almost fair after a training session.
Keywords: optic disc • imaging/image analysis: clinical • imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound)