Abstract
Purpose :
Teleretinal imaging (TRI) is valuable in screening for diabetic retinopathy (DR) and other conditions, but its utility as a screening tool in large, underserved populations with many barriers to care needs to be established. We performed a retrospective, observational clinical study in the Harris Health System (HHS, Houston, TX) to (1) evaluate the compliance with follow-up of TRI patients referred for in-clinic examination, and (2) determine the accuracy of TRI diagnoses in this population with regards to DR severity and presence of diabetic macular edema (DME).
Methods :
A list of HHS patients who were evaluated by TRI and referred for clinic examination between 2013-2016 was cross-referenced with patient charts from HHS hospitals and clinics. Thresholds for referral were severe non-proliferative DR or moderate DME. Patients were excluded if they were under the age of 18 or were already established patients at one of these eye clinics. Information obtained from the charts included patient follow-up status as well as diagnosis of DR severity/DME on TRI and clinical exam.
Results :
Of 206 TRI patients referred for clinical exam, 102 (49.5%) attended a clinic appointment. Of those requiring a second appointment within 4 months, 75% attended this follow-up appointment. Overall, between TRI and clinical exam diagnoses, there was moderate agreement (kappa = 0.39, 95% confidence interval [CI] 0.27-0.52) in grading DR. There was agreement within one level of DR severity in 82.2% (95% CI 75.7-90.3%) of patients. The positive predictive value for detecting referable-level DR was 71.8% (95% CI 60.3-81.1%), and for detecting center-involving DME was 20.6% (95% CI 9.3-38.4%).
Conclusions :
Less than half of referred TRI patients presented for a clinic appointment, suggesting there may be additional barriers to ophthalmic care in a county population. However, once patients established care in the system, a substantial portion were compliant with continued follow-up. TRI was highly predictive for referable-level DR, which is critical for avoiding overburdening clinics. It was poorly predictive for detecting center-involving DME, which is likely related to the limitation of two-dimensional photos. Further investigations into addressing barriers to care and refining the TRI pathway with the integration of additional diagnostic modalities (e.g., optical coherence tomography) is warranted.
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.