Abstract
Purpose :
To describe the findings, to assess diagnostic accuracy and to analyze the factors influencing the detection process.
Methods :
A telemedicine web-based platform (DYSEO) was designed and developed to enable specialists to remotely evaluate images. Four thousand subjects over 55 years were randomly selected by age and gender groups from the sanitary census and invited to participate. Measures of visual acuity and ocular pressure were obtained, and macular and optic disks images were acquired with a non-mydriatic retinograph (Topcon) and a portable SD-OCT (iVue, Optóvue). Eight ophthalmologists evaluated the cases on a deferred basis and the report was automatically sent to the primary care center (PCC) with an ocular health advice. Definite diagnosis was established in a second complete ophthalmic examination (including gonioscopy and visual field) at the Hospital. All patients screened were invited to be examined at Hospital. Descriptive statistics were applied. Sensitivity, specificity and predictive values were calculated.
Results :
After a phone call, 1655 subjects initially accepted to participate in the study and, of these, 1006 (60%) did attend and were fully screened at PCC. The quality of the images was good in 69.8 % for the retinographies and 92.5% for optic disk’s OCTs. The screening program identified 195 cases (19.4%) with suspicion of glaucoma. The number of cases detected increased with increasing age (p< 0.005). Optic disc photographs and OCT were classified as suspicious of glaucoma in 128 (12.7%) and 127 (12.6%), respectively with a Kappa agreement index of 0.37. A total of 481 subjects (48%) were fully examined at the hospital. Global sensitivity, specificity, negative predictive value and positive predictive value were 70.1%, 97.4%, 84% and 94.4%, respectively. Diagnostic accuracy of screening increased with increasing age, greater glaucoma severity and with expert evaluators (p< 0.05).
Conclusions :
The screening program identified glaucoma suspects in 19.4% of cases examined and offered high specificity and predictive values to detect glaucoma. Agreement between photographs’ and OCT images’ evaluation was moderate but they offered additive information.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.