His follow-up study assessed whether using GOANNA translated to more accurate classification of progression, or worsening, of visual field defects in glaucoma. To do this, we simulated 156 visual fields that were stable and 156 visual field results that showed progression. This allowed us to determine the sensitivity (the proportion of simulated visual field defects correctly classified by GOANNA as getting worse) and the specificity (the proportion of simulated visual field defects correctly classified as stable by GOANNA). Measuring sensitivity and specificity allowed us to evaluate the performance of GOANNA against other visual field algorithms (in particular ‘ZEST’, which is currently incorporated into several commercial visual field testing devices: for example the Humphrey Matrix perimeter and the Medmont perimeter). Our results suggest that the improvement in precision and accuracy around scotoma borders using GOANNA translates to earlier and more accurate detection of visual field progression.
This study has recently been published in Ophthalmic and Physiologic Optics, the Research Journal of the College of Optometrists in the UK. You can read the abstract of the published article here: http://onlinelibrary.wiley.com/doi/10.1111/opo.12184/abstract