A Telemedicine Program for Diabetic Retinopathy in a Veterans Affairs Medical Center—The Joslin Vision Network Eye Health Care Model

Cavallerano AA, Cavallerano JD, Katalinc P, et al. Am J Ophthalmol. 2005;139:597-604.

Although early screening and treatment of diabetic retinopathy (DR) can reduce the risk of vision loss, a substantial proportion of people with diabetes fail to receive either an annual eye exam or vision sparing surgery. In order to increase access to diabetic eye care, the Joslin Vision Network (JVN) digital-video retinal imaging system was created. This innovative telemedicine program allows healthcare providers to obtain retinal images in a nonophthalmic setting and then transmit them to experts for evaluation and treatment recommendations. The goal of this retrospective observational study was to characterize the extent of DR in an outpatient setting using the JVN.

In this study, the JVN was tested at the outpatient clinic of the Veteran’s Affairs Medical Center in Togus, Maine. The study enrolled 1219 consecutive patients who were diagnosed with either diabetes, impaired fasting glucose (IFG), or impaired glucose tolerance (IGT). No patients were excluded on the basis of preexisting ocular or systemic conditions. Retinal images were taken by a certified JVN image acquisition specialist using a Topcon TRC-NW6S digital retinal camera. Images were obtained according to the JVN protocol, which requires nonmydriatic, nonsimultaneous stereoscopic retinal images of three 45º fields and an external image of each eye, with additional images taken as needed. Images were then transferred to the JVN Reading Center at the Beetham Eye Institute of the Joslin Diabetes Center in Boston, where they were read by expert readers. Readers evaluated the images to determine the level of DR and diabetic macular edema (DME; calculated based on a modified Early Treatment Diabetic Retinopathy Study classification), as well as the presence of any ocular abnormality unrelated to diabetes. Readers then generated reports that included DR and DME diagnosis and a treatment plan based on ocular findings, patient history, and interval since last eye exam. Reports were electronically transmitted back to each patient’s treating physician. In addition to the imaging, patients also received basic education on the prevention of DR and other diabetes-related complications of the eye.

Because the study was conducted in a VA setting, participants were primarily older (mean age 63.2 yrs) men with type 2 diabetes. Results indicated that 63% of eyes screened showed no evidence of DR, 16% had mild nonproliferative diabetic retinopathy (NPDR), 4.3% had moderate NPDR, 1.4% had severe NPDR, 0.8% had very severe NPDR, and 0.9% had proliferative diabetic retinopathy (PDR). In terms of DME, 78.3% had no evidence of DME, 1.4% had early DME, and 0.7% had clinically significant macular edema (CSME). A number of images were ungradable (13.1% for both levels of DR and 6.6% for DME). More than half of the ungradable eyes (44.7%) also showed evidence of cataract. A wide range of nondiabetic ocular disorders were also observed in this population. There are numerous benefits associated with the use of the JVN digital-video retinal imaging system. First, no pupil dilation is required, making the procedure safer, more convenient, and shorter in duration. Patients can receive DR screening in a nonophthalmic setting, and the results are immediately available. The JVN protocol also has a patient education component, which may improve patient awareness of diabetic complications and actions that can be taken to prevent them. Further, although not done in this study, previous work has demonstrated a high level of agreement between management decisions made using JVN and decisions made following clinical examination and fundus photography. Thus, the JVN is a viable method of identifying the severity of DR, DME, and nondiabetic ocular conditions and for recommending the appropriate follow-up treatment plan.