Purchase this article with an account.
D.C. Hoffman, G. Li, L. Jiang, D.E. Gaasterland, J. Caprioli; Identifying Visual Field Progression with a Linear Mixed Model . Invest. Ophthalmol. Vis. Sci. 2003;44(13):76.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: To identify progressing visual fields with a mathematical model that takes into account the spatial and temporal correlations between test locations. Methods: 319 eyes from 319 patients from the Advanced Glaucoma Intervention Study (AGIS) with at least 7 Humphrey 24-2 perimetric exams (mean 20.9 ± 5.26) over at least 5 years (mean 9.39 ± 1.91 years) of follow up were evaluated. All exams had an initial AGIS score of 17 or less and a reliability rating of 2 or better. The average age at the time of the first exam was 64.3 (± 9.61) years. The standard for progression was the AGIS scoring system as published. A linear mixed model with equicovariance structure was used for this study. The model assumed the average slope of a glaucoma hemifield cluster was a fixed effect and the deviation of locations within that cluster was a random effect. The linear mixed model is a special linear regression that allows variation within hemifield clusters to provide additional and previously unavailable spatial and temporal information to determine progression of visual field series. Results: The AGIS standard for progression identified 89 eyes (27.9%) as progressing. The linear mixed model identified 153 eyes (47.9%) as progressing. Sixty eight of the 89 eyes (76%) identified as progressing by the AGIS standard were also identified by the linear mixed model The kappa between the two methods was 0.32 with a percent agreement of 66%. Conclusion: The determination of visual field progression should take into account both spatial and temporal correlations between visual field test locations without assuming that test locations are independent. The linear mixed model with equicovariance structure provides additional information not otherwise available and shows fair agreement with the AGIS standard to identify visual field progression.
This PDF is available to Subscribers Only