Purchase this article with an account.
J.L. Keltner, C.A. Johnson, K.E. Cello, J. Fan, S.E. Bandermann, M.A. Kass, M.O. Gordon, R.W. Beck, Ocular Hypertension Treatment Study Group; Visual Field Quality Control in the Ocular Hypertension Treatment Study (OHTS): and in the 10–Year Optic Neuritis Treatment Trial (ONTT) Follow–Up Study . Invest. Ophthalmol. Vis. Sci. 2005;46(13):3738.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: To compare visual field (VF) quality control (QC) data in the Ocular Hypertension Treatment Study (OHTS) and in the 10–year Optic Neuritis Treatment Trial (ONTT) follow–up study. Methods: We evaluated 46,777 VFs from 1,618 OHTS participants (22 clinics with 1–103 participants/site) and 814 VFs from 315 ONTT participants (14 clinics with 13– 40 participants/site). Technicians were certified for perimetry, but OHTS technicians were also required to submit 2 sets of qualifying VFs. The OHTS protocol required one set of central 30–2 full–threshold VFs every 6 months; the ONTT protocol required one set of 30–2 VFs and an additional set of 24–2 SITA Fast VFs at a one day visit. Results: For OHTS, 2.6% (1,217 of 46,777) of the VFs were unreliable and 0.23% (107 of 46,777) were unusable due to incorrect test parameters. Mean patient and shipment errors decreased by 0.13 each year, exhibiting correlation coefficients of –0.79 (p=0.0063) and –0.67 (p=0.0333), respectively. Test errors decreased over time, except in 1999 when a new perimeter was introduced. Unreliability increased with time but can be explained by increasing patient age and abnormal VFs. While adjusting for age, VF abnormality, use of HFA II, and correlation among VFS, a Generalized Estimating Equations model with unreliability as outcome found no significant association between test year and unreliability. In the ONTT, 4% (31 of 814) of the VFs were unreliable, 95% (528 of 554) of the 30–2 VFs were usable, and only 31% (80 of 260) of the SITA VFs were usable. Conclusions: With increased OHTS study duration, mean patient and shipment errors decreased. In both studies, reliability was high. A greater proportion of ONTT VFs were unusable probably due to lack of technician experience/feedback and ongoing QC monitoring scores.
This PDF is available to Subscribers Only