Are Person- and Areal-Level Socio-Economic Status Asoosciated With the Onset and Progression of Diabetic Retinopathy?

Abstract

Purpose : Diabetic retinopathy (DR), a leading cause of vision loss, has been equivocally linked with poor socio-economic status (SES) but longitudinal data are rare. We investigated the impact of person- and areal-level SES on the onset and progression of DR in a multiethnic Asian cohort.

Methods : From the Singapore Epidemiology of Eye Diseases Study, a population-based cohort study, we included 1189 Malays and Indians (mean age [SD]:58.9 [9.1] years, 50.6% female) with diabetes, SES data, and gradable fundus photos at baseline and follow-up visits (mean follow-up duration: 6.2 [0.9] yr). Person-level SES (PLSES) parameters were assessed using education (≤6yr/>6), monthly income (<SG$2000/≥2000) and residence (£2 rooms/>2) data. Areal-level SES (ALSES) was assessed using a published socio-economic disadvantage index, created with 12 variables from the 2010 Singapore population census. A high score indicated poor ALSES. DR was categorized as none, minimal, mild, moderate and severe. Incident DR was defined as none at baseline but at least minimal at follow-up; and DR progression as at least a 2-step increase in severity at follow-up from any DR initially. Modified poisson regression models (MPRMs) estimated the longitudinal relationship between PLSES and ALSES and the onset and progression of DR, adjusted for traditional DR risk factors.

Results : Of the 1189 participants, 769 (64.7%), 942 (80.4%), and 91 (7.7%) had low education; monthly income; and lived in small public flats, respectively. The median ALSES score was 103.7 [range 79-120]. At follow-up, 157/834 (18.8%) and 94/355 (26.5%) participants developed DR incidence and progression, respectively. In MPRMs, every ALSES SD increase was independently associated with a 32% higher risk of DR incidence (RR [95% CI]: 1.32 [1.14-1.53]). Similar risks were found in Malays (1.43[ 1.02-2.00]) and Indians (1.27 [1.01-1.60]). For PLSES, low monthly income (1.73 [1.06-2.83]) was significantly associated with incident DR but not low education (0.74 [0.51-1.08]) or living in small public flats (1.63 [0.88-3.04). No associations were found for SES measures with DR progression.

Conclusions : Low income and overall areal SES significant increase the risks of incident DR in two large ethnic groups in Singapore. Strategies targeting those with diabetes with low income and living in socio economically challenged areas may prevent the onset of DR.

This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.