Abstract
Purpose :
West Virginia has the highest prevalence of diabetes in the United States (as per the CDC), and this has brought diabetic retinopathy (DR) and diabetic macular edema (DME) to the forefront of the state’s ophthalmological concerns. Recent implementation of the Intelligent Retinal Imaging Systems (IRIS) has enabled primary care offices throughout the state to incorporate teleophthalmology into their clinical practice. We utilize data acquired via these systems to determine real-world DR/DME concordance between image findings and subsequent comprehensive eye exams. With previous studies calling into question the efficacy of teleophthalmology in elderly populations (Fatehi et al., 2020), we also use this opportunity to explore the potential influence of patient age on image gradeability.
Methods :
Nonmydriatic cameras (handheld Volk Pictor) were used by trained nurses/staff to acquire 45-degree fundus images from patients at primary care clinics throughout WV. The images were graded by retina specialists at the WVU Eye Institute. Images were noted as gradable or ungradable, and the extent of DR (absent, mild, moderate, severe, or proliferative) and/or DME (absent, mild, moderate, or severe) was described. Data from January 2017 to February 2019 from 2,380 patients who were screened were collected. A Pearson’s chi-squared test was performed to compare the gradeability data found within different age ranges.
Results :
From the 4,760 eyes attempted, we found that 3,686 (77%) were deemed gradable. Patients older than 65 years had significantly fewer gradable eyes (42% versus 79%, p < 0.001). We found that 125 patients (53% with suspected DR/DME via teleophthalmology) followed up with comprehensive eye exams. Conclusions from these exams showed that 86 of these patients (37%) truly had DR/DME and allowed us to determine a positive predictive value of 68.8% for the teleophthalmology screenings.
Conclusions :
This real-world application of teleophthalmology has revealed a greater percentage of ungradable images and less DR/DME concordance compared to previous studies (Haddad et al., 2014). Older age has been shown to correlate with a greater odds of ophthalmology referral for DR (Jani et al., 2017), which may raise concern given our findings that suggest an association between poor image quality and older age.
This is a 2020 Imaging in the Eye Conference abstract.