Authors:James Im¹, Graham E. Trope¹, Yvonne M. Buys¹, Peng Yan², Michael H. Brent¹, Sherif El-Defrawy¹, Ya-Ping Jin¹. ¹University of Toronto, ²Kensington Vision and Research Center.

Author Disclosure Block:J. Im: None. G.E. Trope: None. Y.M. Buys: Membership on advisory boards or speakers’ bureaus; Name of for-profit or not-for-profit organization(s); Allergan. Membership on advisory boards or speakers’ bureaus; Description of relationship(s); Was in advisory board in the past year. P. Yan: None. M.H. Brent: Membership on advisory boards or speakers’ bureaus; Name of for-profit or not-for-profit organization(s); Novartis Canada, Bayer Canada, Roche Canada. Membership on advisory boards or speakers’ bureaus; Description of relationship(s); Sitting on advisory boards. Funded grants or clinical trials; Name of for-profit or not-for-profit organization(s); Novartis Canada, Bayer Canada. Funded grants or clinical trials; Description of relationship(s); Unrestricted research grant support for diabetic retinopathy research and unrestricted educational grant support for the Medical Retina Fellowship. S. El-Defrawy: None. Y. Jin: None.

Abstract Body:

Purpose: To assess the prevalence of and changes in visual impairment (VI) among Canadians with and without diabetes from 1994 to 2014.
Study Design: Population-based cross-sectional survey.
Methods: Data from respondents aged 40+ participating in seven nationwide surveys were analyzed: the National Population Health Survey in 1994/1995 (n=17,626), 1996/1997 (n=81,804) and 1998/1999 (n=17,244) and the Canadian Community Health Survey in 2000/2001 (n=130,880), 2008/2009 Healthy Aging (n=30,865), 2009/2010 (n=124,188) and 2013/2014 (n=127,462). VI was self-reported as unable to see close or distance when wearing glasses/contact lenses. Diabetes was self-reported. Individuals with post-secondary education or higher were grouped into high levels of education. Classification of high- and low-income levels was based on household income and the cut-off varied by survey years. Using the 2016 Canadian population as the standard, the age- and sex-standardised prevalence of VI was calculated to remove influences of temporal changes in age and sex distributions. For analyses stratified by levels of education and income, only sex-standardised prevalence was calculated due to sparse data.
Results: In any survey year, the age- and sex-standardised prevalence of VI was about 2 times higher in diabetics than non-diabetics (e.g. 2.96% vs 1.57% in 2013/2014). Among diabetics, the standardised VI prevalence was 7.11% (95% confidence interval (CI): 6.99%-7.24%) in 1994/1997 and gradually decreased to 2.96% (95% CI: 2.91%-3.01%) in 2013/2014, a 58% reduction over 20 years. Among non-diabetics, the standardised prevalence of VI decreased from 3.38% (95% CI: 3.35%-3.40%) in 1994/1997 to 1.57% (95% CI: 1.56%-1.58%) in 2013/2014, a 54% reduction. Among diabetics with low levels of education, the sex-standardised prevalence of VI reduced by 72% (12.4% in 1994/1997 to 3.43% in 2013/2014). In those with high levels of education, the reduction was 65% (8.31% in 1994/1997 to 2.94% in 2013/2014). Among Canadians without diabetes, the standardised VI prevalence reduced by 65% (5.86% in 1994/1997 to 2.03% in 2013/2014) in those with low levels of education and 79% (5.26% in 1994/1997 to 1.08% in 2013/2014) in those with high levels of education. Similar trends were observed after stratification by income.
Conclusions: From 1994 to 2014, VI was roughly 2 times higher in diabetics versus non-diabetics. For both diabetics and non-diabetics, the standardised VI prevalence decreased in those with high and low levels of education and income. These results likely demonstrate the effectiveness of the collective efforts by eye care providers, researchers, the public and government.