The Impact of a Diabetes Teleretinal Imaging Program in a Predominantly Hispanic, Urban Patient Population

Abstract

Purpose: : To assess the impact and accuracy of digital fundus photography screening of diabetic patients for diabetic retinopathy (DR) in an urban primary care clinic with a high-risk, predominantly Hispanic patient population through analysis of outcomes in referred patients.

Methods: : Digital fundus photos of type-2 diabetics were taken with a Topcon TRC NW-6S non-mydriatic camera in an urban primary care clinic located in Chelsea, MA. Trained readers documented the severity of DR and made referrals on the basis of these photos. The results of in-office clinical exam of patients referred by this screening method were compared to their fundus photo evaluation.

Results: : Among imaged patients (n=233), 57.1% were female and 42.9% were male with a median age of 52 years (range 22-83). The majority of patients were of Hispanic ethnic background (67.4%), while the remainder of patients reported themselves as either Caucasian (19.3%), African-American (8.6%), Middle Eastern (3.4%) or Asian (1.3%). Mean HbA1c was 7.6%. The majority of photos (89.7%) were of sufficient quality to assess for DR. Forty patients (17.2%) had mild non-proliferative DR (NPDR), three patients (1.3%) had moderate NPDR, two patients (0.9%) had severe NPDR, and one patient (0.4%) had proliferative DR (PDR). Twelve patients (5.1%) had macular edema (ME). Overall, seventy-three patients (31.3%) were referred to ophthalmology. Among referred patients, the indications for referral included DR (38.4%), un-gradable images (24.7%), non-diabetic retinal disease (23.3%), and enlarged optic nerve cup-to-disc ratio (12.3%). Thirty-three (45%) of the referred patients were seen at MEE for clinical evaluation. Average time from image acquisition to clinical exam was 4.2 months (range 17 days - 10 months). There was 75.9% concordance between the image evaluator and clinical examiner regarding the presence of DR. On clinical exam, only one of nine patients (11%) referred for possible ME had clinically significant ME (CSME), which was treated with grid laser. Two patients referred for DR had CSME treated with anti-VEGF injection, one of whom also had PDR and underwent vitrectomy. Four referred patients underwent cataract surgery and/or pterygium excision.

Conclusions: : Evaluation of digital fundus photographs taken in a predominantly Hispanic, urban primary care setting provides a clinically useful screening method to identify high-risk patients with DR as well as non-diabetic eye disease that would benefit from ophthalmologic referral and treatment.

Keywords: diabetic retinopathy • clinical (human) or epidemiologic studies: systems/equipment/techniques • imaging/image analysis: clinical