PURPOSE: To determine whether the diagnostic accuracy of judging serial optic disc color stereo photographs for glaucomatous change by non-expert ophthalmologists changed after a training session. METHODS: 24 ophthalmologists in training at the University Hospitals Leuven classified 50 eyes with varying severity of glaucoma as stable or progressing based on the appraisal of serial optic disc stereo photographs. A comparison between the diagnostic accuracy of residents (n=18) and trainees (n=6) was made before and after a training session. RESULTS: The mean agreement (kappa) with the reference standard before training was lower for the trainees than for the residents. The mean kappa before training was 0.37 for the residents and 0.29 for the trainees (p = 0.18). The mean agreement with the reference standard improved significantly after a training (from 0.29 to 0.56 [p = 0.031 for the trainees, and from 0.37 to 0.48 for the residents [p = 0.005]). The overall mean kappa was 0.35 pre-training and 0.50 post-training (p < 0.001). CONCLUSIONS: The agreement and diagnostic accuracy of residents in training in appraising serial optic disc photos improved significantly after a training session.
PURPOSE: To determine whether the diagnostic accuracy of judging serial optic disc color stereo photographs for glaucomatous change by non-expert ophthalmologists changed after a training session. METHODS: 24 ophthalmologists in training at the University Hospitals Leuven classified 50 eyes with varying severity of glaucoma as stable or progressing based on the appraisal of serial optic disc stereo photographs. A comparison between the diagnostic accuracy of residents (n=18) and trainees (n=6) was made before and after a training session. RESULTS: The mean agreement (kappa) with the reference standard before training was lower for the trainees than for the residents. The mean kappa before training was 0.37 for the residents and 0.29 for the trainees (p = 0.18). The mean agreement with the reference standard improved significantly after a training (from 0.29 to 0.56 [p = 0.031 for the trainees, and from 0.37 to 0.48 for the residents [p = 0.005]). The overall mean kappa was 0.35 pre-training and 0.50 post-training (p < 0.001). CONCLUSIONS: The agreement and diagnostic accuracy of residents in training in appraising serial optic disc photos improved significantly after a training session.