Agreement of visual field interpretation among glaucoma specialists and comprehensive ophthalmologists: comparison of...

To evaluate interobserver agreement and interpretation time for three clinically available formats of visual field presentation: serial Humphrey visual field (HVF), STATPAC2 and PROGRESSOR.

40 field series from the Advanced Glaucoma Intervention Study were presented to eight glaucoma specialists and eight comprehensive ophthalmologists to determine whether each field series was stable or progressive. Interobserver agreement and agreement with Hodapp-Parrish-Anderson criteria were evaluated using κ statistics, and the interpretation time was compared.

For glaucoma specialists, median κ values for interobserver agreement were 0.47, 0.60 and 0.43 for HVF, STATPAC2 and PROGRESSOR, respectively. Respective κ values for comprehensive ophthalmologists were 0.43, 0.43 and 0.35. For glaucoma specialists, median κ values for agreement with Hodapp-Parrish-Anderson criteria were 0.52, 0.67 and 0.52 for HVF, STATPAC2 and PROGRESSOR, respectively. Respective κ values for comprehensive ophthalmologists were 0.41, 0.47 and 0.33. For glaucoma specialists, the mean±SD interpretation time for the series of 40 fields was 63.4±35.9, 57.1±23.1 and 41.1±15.3 min using HVF, STATPAC2 and PROGRESSOR, respectively. Respective interpretation times for comprehensive ophthalmologists were 72.9±38.3, 68.6±30.6 and 51±24.1 min. Interpretation time was decreased when STATPAC2 or PROGRESSOR was used rather than HVF. Time reduction was significant for glaucoma specialists using PROGRESSOR (p=0.02).

For glaucoma specialists, interobserver agreement and agreement with HPA criteria were moderate to substantial. For comprehensive ophthalmologists, interobserver agreement and agreement with HPA criteria were fair to moderate. Field interpretation time may be reduced clinically when using STATPAC2 or PROGRESSOR rather than HVF.