Feasibility and performance of eye movement perimetry for glaucoma triage in Tanzania: A prospective cross-sectional study in a real-world setting.

Triaging access to glaucoma diagnostics could allow patients to receive earlier care to slow disease progression. There is a particular need in lower-resource settings where the cost and lack of portability of standard automated perimetry (SAP) restrict access. Eye movement perimetry (EMP) uses eye-tracking technology to provide a rapid assessment of the visual field using a tablet computer. A cross-sectional study in Tanzania was undertaken to assess the feasibility of EMP as an initial glaucoma triage tool. EMP hit rate correlated with SAP mean deviation in 370 eyes (Spearman's rank correlation coefficient, ρ = 0.441, p < 0.001) and graphical comparisons indicated good localisation of defects. In emulated triage of 232 participants, EMP exhibited very high (100.0%) sensitivity (95%CI: 95.5 to 100.0%) in identifying eyes with mean deviation ≤ -6 dB), meaning that a high EMP hit rate was strong evidence of a healthy visual field. However, specificity was 20.7% (95%CI: 15.0 to 27.8%), meaning that many individuals with no severe field defect would require SAP to confirm this. Specificity was limited by EMP test failure, which occurred in 75 of 480 eyes (15.6%), as well as export or processing failure which occurred in 35 eyes (7.3%). Participants preferred EMP to SAP testing and agreed that it was easy to perform. In a real-world setting, EMP rapidly and accurately identifies patients without a significant visual field defect. However, refinement is required to reduce the high number of unnecessary referrals or false positives for EMP to prove useful as a triage tool.