The Prevalence of Visual Impairment and Blindness
Our estimates of prevalence and the prevalent population with visual impairment and blindness are listed in table 3.1 and 3.2. A National Eye Institute (NEI) meta-analysis reports the prevalence of visual impairment and blindness based on eight population-based studies of persons aged 40 or older.[2, 3] We use the NEI prevalence rates for ages 40 and older. The NEI study did not include any persons younger than age 40, to our knowledge no nationally representative data on the prevalence visual impairment and blindness are available for this age group, except a study based on NHANES data that reported prevalence rates for persons with and without diabetes.
Table 3.1. Prevalence of Visual Loss, 2005–2008 NHANES
Table 3.2. Prevalent Population with Visual Loss, 2005–2008 NHANES
For the overall population younger than age 40, we estimated the prevalence of visual impairment and blindness based on autorefractor-corrected visual acuity as measured in the National Health and Nutrition Examination Survey (NHANES) data from 2005 through 2008 (Table 3.1). We used SAS statistical software version 9.2 (Cary, NC) with SAS-Callable SUDAAN Version 10.0.1 (Research Triangle Park, NC) to adjust for the complex design of the NHANES sample. We assigned individuals to normal vision, mild impairment, moderate impairment, or blind based on the corrected visual acuity of the better-seeing eye. Cut-off values for impairment category were based on U.S. thresholds: >20/40 for mild impairment, >20/80–20/200 for moderate impairment, and >20/200 for blindness. Respondents who did not have an acuity test due to self-reported blindness were included in the prevalence of blindness. Acuity tests were not administered to NHANES respondents younger than age 12, and no nationally representative data exist on the prevalence of non-correctable bilateral vision loss in this population. We imputed the prevalence of vision loss among children younger than age 12 by adjusting the age 12 to 17 NHANES prevalence using age-specific incidence of severe impairment and blindness as identified in U.K. surveillance data. We estimated the prevalence and confidence intervals (CIs) for each variable of interest using the Taylor Linearization Method. We estimated the population of each impairment category by multiplying prevalence rates by 2010 Census population estimates.
We assume the use of NHANES data includes certain limitations. Screening is based on acuity testing using near distance charts performed with whatever vision correction devices the participants are currently using, followed by autorefraction testing. Only a subsample of respondents aged 40 or older are assessed for reduced visual field or contrast sensitivity. Given the lack of appropriate validation data for persons younger than age 40, we compared the prevalence of visual impairment and blindness for persons aged 40 or older in NHANES to the NEI meta-analysis values for validation purposes. The NEI meta-analysis reports the prevalence of visual impairment and blindness based on eight population-based studies of persons aged 40 or older.[2, 3] The NEI study found prevalence of visual impairment and blindness among the population aged 40 or older to be 1.98% and 0.78%, respectively. For the population aged 40 or older, we estimated the prevalence of visual impairment and blindness in NHANES at 2.19% and 0.22%, respectively. Thus, the NHANES estimates identify higher overall prevalence of impairment but lower prevalence of blindness. Ordered based on prevalence of any impairment, the NHANES prevalence rates would fall fourth among the eight studies included in the NEI report.
One complication of using NHANES prevalence for younger ages and NEI prevalence for ages 40 and older is that mild and moderate impairment appear to decrease from ages 18-39 to ages 40-64. We assume that much of this difference is due to improved acuity testing and refraction correction in the NEI studies versus that achieved by autorefractor in NHANES. However, when assessing NHANES prevalence by age, we see nearly the same reduction in impairment from the 18-39 group to the 40-64 group. Indeed, this pattern is also apparent in the analysis of NHANES prevalence conducted by the CDC. Figure 11 shows age-specific prevalence of visual impairment and blindness from the NEI and NHANES. The figure shows the NHANES-derived prevalence rates for the median of each age group as blue diamonds, and the NEI-based prevalence rates for persons aged 40 and older as the red line. This analysis uses NHANES prevalence for ages less than age 40 and the NEI prevalence for ages 40 and older, resulting in an apparent drop in mild and moderate impairment at age 40. However, even if we were to use NHANES prevalence for ages 40 and older, we would still see this apparent decrease in prevalence at age 40.