Urban diabetic retinopathy teleophthalmology screening: results and impact at the service corridor

To assess real-world results and the impact on a hospital service corridor for screening for DR through an urban community teleophthalmology service.

Retrospective analysis at the hospital service corridor of 148 diabetics referred to it following DR teleophthalmology screening of 1185 type II diabetics.

Of the screened diabetics, 87.4% (n = 1036) were exempted from face-to-face clinical examination (FFCE) in a traditional hospital eye care pathway and continued monitoring through teleophthalmology under a watch-and-wait attitude, while 12.5% (n = 148) were recommended for an FFCE. The FFCEs revealed that significant DR was present in 48.2% or in 5.6% of this screened diabetic population. Reasons for referral were findings of significant DR in 40.5%, of which diabetic macular edema (DME) represented 86.6%, other incidental significant sight-threatening findings represented 32.4% (4% of the screened diabetics), and insufficient image quality was obtained for the other 27.0%. Optical coherence tomography (OCT) imaging at FFCE confirmed DME in 26.4% and led to treatment. Patients referred for insufficient image quality showed significant pathology in 90.2%, of whom 63.4% underwent further monitoring or treatment. The readers requested the FFCEs for 148 patients within 1 month of the reading in 19.6%, 3 months in 26.3%, 4-12 months in 47.3%, and 12 months in 6.7% over the 34 months of the study. Compliance with FFCEs was 91.9%, absolute in 78.4% and relative in 21.5%. The availability of OCT at the imaging site would have impacted 4.5% of the screened diabetics by enabling teleophthalmology monitoring of 91.6%, identifying just-in-time interventions for DME treatments in 26.4% and reducing by 25% the need for referral of OCT-negative reader-identified DME.

The FFCEs generated at the hospital service corridor by an urban community DR screening teleophthalmology project did not impact negatively on its services; moreover, the service corridor was exempted from providing FFCEs to 87.4% of the diabetic population it serves. This study may help provide cost-efficiency indications for a screening protocol that would include OCT availability at the imaging site and measure its positive effects. While DR of which DME was the main cause of referral for FFCE, incidental significant sight-threatening findings were significant and approached DR as a cause of referral; this supports the recommendation of continued human intervention in DR teleophthalmology screening at this time and for this population, until automatic computer-aided diagnosis systems can recognise biomarkers associated with other significant fundus diseases. As a secondary gain this project benefited individuals in need of care who were lost to the traditional eye care pathway. Good compliance with the follow-up FFCE further supports teleophthalmology in its effort to provide better access to DR screening.