Purchase this article with an account.
D. Molinaro, L. Cerulli, F. Ricci, M. Cesareo; Moorfields Regression Analysis (MRA) and Glaucoma Probability Score (GPS) Agreement in Daily Clinical Practice. Invest. Ophthalmol. Vis. Sci. 2008;49(13):4220. doi: https://doi.org/.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To evaluate the agreement of HRT III MRA and GPS classification without respect to operator experience and ocular diagnosis and to investigate factors affecting disagreement.
409 eyes of 217 Caucasian patients. Females 57%, Males 43%. Mean age: 50.5 ± 16.3. All subjects underwent a full eye examination. The HRT III examination was performed in part by different operators: experienced and HRT trained ohpthalmologists, HRT trained postdoctoral students and HRT trained technicians. We considered only the Global classification of both MRA and GPS. INCLUSION CRITERIA: Clear Lens. Refractive error within ≤ 6D equivalent sphere and ≤3D astigmatism. Image quality > Good (MPHSD < 40 µm). EXCLUSION CRITERIA: Lens opacities. History of previous intraocular surgery. Optic disk abnormality (drusen or a tilted disk). Institutional Review Board approval was obtained and all subjects provided signed informed consent.
(preliminary) MRA and GPS agreement was equal to 68% (277 eyes); MRA and GPS disagreement was equal to 31% (127 eyes); GPS failed classification in 5 eyes (1%) that were excluded from the study.
The agreement (or disagreement) between two different diagnostic decision-support system implemented in the same instrument has obvious implications. MRA and GPS showed similar diagnostic performances in different studies. Nevertheless, in most of the studies that deal with this issue, HRT examinations have been carried out by experienced operators on the basis of more stringent inclusion criteria in the contest of scientific protocols. We aimed to evaluate the MRA and GPS agreement in a group of patients referred to our Ophthalmology Clinic to perform HRT III examination. Since our Clinic is an University center, different operators with various degree of experience can perform this test on different patiens referred to the clinic from any public or private ophthalmologist. This clinical ophthalmology setting seems to be a more realistic situation to evaluate MRA and GPS agreement, with obvious clinical and therapeutic consequences. The percentage of disagreement between MRA and GPS found in this study was expected since MRA is operator dependent while GPS analysis is, at least in part, an operator-independent system of classification. Logistic regression analyses to determine the effect of different covariates on the classifications are in progress.
This PDF is available to Subscribers Only