Li Y, Xirasagar S, Pumkam C, Krishnaswamy M, Bennett CL. Vision Insurance, Eye Care Visits, and Vision Impairment Among Working-Age Adults in the United States. JAMA Ophthalmol. 2013;131(4):499-506. doi:10.1001/jamaophthalmol.2013.1165
SECTION EDITOR: PAUL P. LEE, MD
Author Affiliations: Department of Health Services Policy and Management, Arnold School of Public Health, University of South Carolina (Ms Li and Drs Xirasagar and Pumkam), and South Carolina College of Pharmacy at the University of South Carolina (Dr Bennett), Columbia; and Department of Ophthalmology, MS Ramaiah Medical College, Bangalore, India (Dr Krishnaswamy).
Objectives To compare rates of eye care visits and vision impairment among working-age adults with vision insurance vs without, among the total sample of Behavioral Risk Factor Surveillance Survey respondents and among a subsample of respondents who had diagnoses of glaucoma, age-related macular degeneration (ARMD), and/or cataract.
Design Using the Behavioral Risk Factor Surveillance Survey 2008 vision module data, we examined the likelihood of an eye care visit within the past year and of self-reported visual impairment among 27 152 adults aged 40 to 65 years and among a subset of 3158 persons (11.6%) with glaucoma, ARMD, and/or cataract. Multivariate logistic regression models were used.
Results About 40% of both the study population and the subsample with glaucoma, ARMD, and/or cataract had no vision insurance. Respondents with vision insurance were more likely than those without to have had eye care visits (general population adjusted odds ratio [AOR], 1.90 [95% CI, 1.89-1.90]; glaucoma-ARMD-cataract subsample AOR, 2.15 [95% CI, 2.13-2.17]), to have no difficulty recognizing friends across the street (general population AOR, 1.24 [95% CI, 1.22-1.26]; eye-disease subsample AOR, 1.45 [95% CI, 1.42-1.49]), and to have no difficulty reading printed matter (general population AOR, 1.34 [95% CI, 1.33-1.35]; eye-disease subsample AOR, 1.37 [95% CI, 1.34-1.39]). Respondents from the total sample who had an eye care visit were better able to recognize friends across the street (AOR, 1.07) and had no difficulty reading printed matter (AOR, 1.70), and respondents from the eye-disease subsample who had an eye care visit also were better able to recognize friends across the street (AOR, 1.71) and had no difficulty reading printed matter (AOR, 1.45).
Conclusions Lack of vision insurance impedes eye care utilization, which, in turn, may irrevocably affect vision. Vision insurance for preventive eye care should cease to be a separate insurance benefit and should be mandatory in all health plans.
Millions of Americans experience vision loss or serious impairment due to age-related eye diseases,1,2 the 3 leading causes being glaucoma, age-related macular degeneration (ARMD), and cataract.3 The prevalence of these conditions is expected to escalate to 5.6 million by 2020 with the rapid aging of the US population.1,4 Age-related macular degeneration progressively ruins central vision, seriously affecting the activities of daily living, particularly among men.3,5,6 Primary open-angle glaucoma progressively destroys the optic nerve head and other visual sensory structures,3,7,8 often unnoticed by the individual,9,10 causing permanent vision impairment (particularly visual field loss if untreated).8,10,11 To prevent irreversible vision loss, periodic eye checkups are recommended for the timely detection and treatment of these and many other eye conditions.12 Early detection through screening is critical because physiologic compensation by the contralateral eye delays patient recognition of acuity and visual field impairments, often until the stage of irreversible global vision disability is reached.7,8,13 Early treatment to reduce intraocular pressure in patients with glaucoma can slow or halt vision impairment.8,9,14,15 Cataract, which is also related to aging, appears prematurely among diabetic and other predisposed individuals, and, left untreated, it can lead to irreversible vision impairment. The American Academy of Ophthalmology recommends preventive eye checkups with pupil dilatation every 2 to 4 years for persons 40 to 64 years of age and every 1 to 2 years for persons 65 years of age or older.16
Having vision insurance is a potentially important determinant of use of preventive eye examinations. Our study addresses an important evidence gap regarding the connection between avoidable vision impairment and an addressable deficiency in health services (namely, vision insurance benefit). Vision coverage is currently an add-on benefit, which is often not covered by employment-based health insurance plans, and one of the first benefits to be eliminated to lower health insurance costs.17 We present a population-based study of working-age individuals' use of eye examinations and vision impairment levels, among those with vision insurance vs those without.
Using secondary, population-based survey data, we test the hypotheses that (1) eye care coverage is positively associated with an eye care visit within the past year and (2) persons who had a prior eye care visit have nil or less severe visual impairment, both among the general population and among those with a diagnosis of glaucoma, ARMD, and/or cataract. We used the 2008 Behavioral Risk Factor Surveillance System (BRFSS) vision module data gathered from adults aged between 40 and 64 years residing in 8 states. Respondents 65 years of age or older were excluded because Medicare covers vision care. The key variables of interest were having vision insurance, an eye care visit in the past 12 months, and vision impairment. Vision impairment is measured as difficulty level in (1) recognizing friends across the street and (2) reading a newspaper or magazine or numbers on the phone (using a 4-point Likert scale, with 1 being no difficulty, 2 a little difficulty, 3 moderate difficulty, and 4 extreme difficulty). To create 3 distinct categories representing progressive levels of vision loss with separation of 2 extreme groups by a middle group, we combined “a little difficulty” and “moderate difficulty” into “some difficulty.” Combining the 2 groups minimized confounding by misclassification bias due to variable interpretation of “mild” and “moderate” difficulty across respondents. We used “extreme” difficulty as the reference group. The eye-disease subsample consisted of respondents who answered “yes” to the question whether they had ever been diagnosed with glaucoma, ARMD, or cataract. Weighted multinomial logistic regression analysis was used, controlling for the following variables: having a regular physician, having general health insurance, age, sex, race, education level, and income level.
A total of 27 152 BRFSS vision module respondents 40 to 64 years of age qualified for inclusion in our study. Of these respondents, 3158 (11.6%) indicated having received a diagnosis of glaucoma, ARMD, and/or cataract (Table 1). The prevalence of “some” and “extreme” vision impairment in the general population was 13.3% and 0.8%, respectively, for recognizing friends across the street and 30.4% and 3.4%, respectively, for reading printed matter. Corresponding rates among the eye-disease subsample were 23.6%, 2.8%, 37.1%, and 5.9%, respectively. Among both the total sample and the eye-disease subsample, about 40% had no vision insurance, although 86% and 88.6%, respectively, reported having general health insurance, and similar percentages had a regular physician. Table 2 presents the prevalence of general health insurance and vision insurance. Among the total sample of respondents having general health insurance, 32.3% had no vision insurance, and among the eye-disease subsample of respondents with general health insurance, 31.4% had no vision insurance. About 10% of respondents who lacked general health insurance reported having vision insurance, both in the total sample and in the eye-disease subsample. Overall, about 40% of both the total sample and the eye-disease subsample had no vision coverage. Persons with vision insurance were more likely than those without vision insurance to have had an eye care visit in the past 12 months (64.3% vs 45.3%), and respondents from the eye-disease subsample were more likely than respondents from the total sample to have had an eye care visit in the past 12 months (80.0% vs 61.7%) (Table 3).
Table 4 shows the adjusted associations between vision insurance and eye care visit in the past 12 months. Respondents with vision insurance were twice as likely as those without vision insurance to have had an eye care visit, both in the total sample (adjusted odds ratio [AOR], 1.90 [95% CI, 1.89-1.90]) and in the eye-disease subsample (AOR, 2.15 [95% CI, 2.13-2.17]). Other factors associated with a higher likelihood of an eye care visit among the total sample and the eye-disease subsample were patient demographics (older age, higher education level, higher income level, being unemployed, and being black vs white) and one health behavior characteristic (having a regular physician). Notably, having general health insurance was not a significant predictor of an eye care visit once vision insurance was included in the model.
After adjusting for patient- and provider-level factors, having vision insurance was associated with significantly higher odds of reporting good vision (the AORs in the general sample for “no” vs “extreme” difficulty were 1.20 and 1.30, respectively, for recognizing friends across the street and reading printed matter, respectively, among the total population) (Table 5). The corresponding AORs are higher among the eye-disease subsample (1.50 and 1.40, respectively). Both the direction of association and an increased strength of association among the eye-disease sample relative to the general population are consistent for some vs extreme difficulty (the AORs were 1.10 and 1.30 for the general population, respectively, and 1.30 and 1.60 for the eye-disease subsample, respectively [all statistically significant]). Notably again, having general health insurance was not a significant predictor of vision status once vision insurance is included in the model.
Table 6 shows that having an eye care visit was associated with higher odds of good vision both in the total population and in the eye-disease subsample (with AORs of 1.07 and 1.71, respectively, for no difficulty vs extreme difficulty in recognizing friends on street, and with AORs of 1.70 and 1.45 for reading printed matter). As with earlier models, having general health insurance was not associated with vision status once eye care visit was included in the model.
Our study used BRFSS survey data to examine associations between vision insurance and eye care utilization among the older working-age adult US population aged 40 to 64 years. This is the age group at high risk for eye diseases that cause gradual vision loss that is preventable. This is also the age group generally without Medicare coverage. We also studied a subsample of respondents with a diagnosis of glaucoma, ARMD, and/or cataract, the leading causes of avoidable vision loss. Our study further evaluated the potential downstream consequence of lack of eye care (namely, vision impairment). Contrary to the National Health Interview Survey and disability-focused surveys that elicit information on total blindness or severe vision loss, our study data source (ie, the BRFSS) has questions that enable the assessment of gradations of vision loss (consistent with the progressive nature of vision loss experienced in these conditions in the absence of preventive/early treatment). An additional advantage of the BRFSS questions is that they capture data on the visual impairment hallmarks of the eye diseases studied: global visual field impairment (recognizing friends across the street) and central vision impairment (reading printed matter).
The study finds that having vision insurance increases the likelihood of an eye care visit (consistent with previous studies18- 20) and that a prior-year eye care visit is associated with better vision status, both in the total sample and in the eye-disease subsample. Notably, in the multiple regression models, general health insurance lost significance once vision insurance was included in the model. A new contribution to the literature is our study's finding of the asymmetry between general health insurance and vision insurance. Fully 40% of our study population age group has no vision insurance. Notably, the 85% prevalence of general health insurance masks the 40% lacking vision insurance in this population. A previous study17 using National Health Interview Survey data found that 38.9% of the total adult population with general health insurance reported an eye care visit in the prior year compared with 15.2% among the uninsured. That study,17 covering all adults, included young adults (at low risk for eye disease) and the Medicare population (with almost universal insurance coverage). By comparison, our study focused on older working-age adults and found that 64.3% with vision insurance reported having had an eye care visit in the prior year vs 45.3% of those without vision coverage. It clarifies that vision insurance, rather than general health insurance, is the key determinant of eye care visits. A significant policy implication is that policy makers and researchers should revisit the widespread assumption that health insurance implies coverage for essential health care services needed for preserving general physical health and the capabilities critical for daily living and a basic quality of life. It may be beneficial to closely examine mainstream health plans for other potentially important coverage gaps and specific benefit exclusions faced by the “insured,” and to study their impacts in order to identify the critical benefits that should be required in all comprehensive health plans.
Another major new contribution of our study is the hitherto unexplored subset of the US population: persons with glaucoma, ARMD, and/or cataract. Fully 11.6% of the US working-age population aged 40 to 64 years reported having received a diagnosis of glaucoma, ARMD, and/or cataract, a group at high risk of progressive vision loss. This group was found to have a similar prevalence of lack of vision insurance to that of the general population, and a similarly reduced eye care visit rate in the absence of vision insurance, despite risking permanent visual impairment owing to lack of timely treatment.
Our study empirically tracks an important consequential, process-outcome link both among the total sample of respondents and within the eye-disease subsample. In both groups, respondents who reported having had an eye examination in the prior year, on average, had better vision. These associations highlight the long-term benefits of vision insurance for preventing eye impairment. Fifty percent of all cases of vision impairment in the United States are either preventable or treatable following early diagnosis.1 Our study provides key empirical evidence that addresses the recent debate over the cost-effectiveness of the additional premiums incurred for vision insurance vs the actual preventive eye care utilization achieved.21 Mandatory vision coverage in health plans is estimated to increase health plan premiums by about 3%. Our study's finding that 64.3% of working-age adults with vision insurance completed an eye care visit in the prior year, as did 80.1% of the eye-disease subsample of respondents, suggests that a vision coverage mandate may be good value for the 3% premium increase.22
Our study has major public health significance. First, vision loss is often insidious, with the affected individuals being unaware of their condition until irreversible and disabling levels of impairment are reached. Second, persons who are blind are often homebound and, therefore, “invisible” to society, failing to draw the attention of the public and of policy makers to this hidden scourge. With the increasing graying of America, the looming impacts of avoidable vision impairment are ballooning. Visual disability has serious consequences far beyond the immediate human cost. Most evident are the economic costs in supplemental social security payments, Medicaid costs, and loss of productive employment. Disability-related direct spending on working-age adults in the United States is $357 billion, or 12% of total federal expenditures in 2008.23 Vision disability accounted for 17.2% of the total disabled in 2009.24 Only 38.3% of vision-disabled working-age adults were employed.24 The annual total direct cost to the government (and taxpayers) of blindness and visual disorders among Americans 40 years of age or older is estimated at $35.4 billion, of which Medicaid accounts for $13.7 billion.25,26 The additional financial burden on individuals, caregivers, and nongovernment health care payers is estimated at $16 billion annually.27
Less evident, however, is the personal human impact of severe and intermediate levels of vision impairment (personal injury due to falls, impacts on body functions, serious deterioration in quality of life, depression, and social isolation). Also less evident is the public health impact of road accidents that cause serious injury or death to those with vision impairment and to others. Persons with significant eye conditions, particularly the elderly, are more likely to have road accidents or vehicle collisions, and they are at much higher risk of being at fault.21,22,28,29 In the United States, 4139 people 70 years of age or older died in motor vehicle crashes in 2010, which is 4% higher than in 2009. The financial and human costs of vision impairment could potentially be mitigated by requiring universal vision coverage among working-age adults because the leading causes of vision impairment typically commence in the fifth to seventh decade. Because our study empirically establishes the consequential link between lack of vision insurance and vision damage mediated by its impact on eye care visits, it provides the needed evidence for policy interventions to mandate vision coverage in all standard health plans. Alternatively, federal and state governments may find it beneficial for their own budgets to initiate publicly sponsored eye-screening programs for the uninsured that are similar to those provided under the Best Chance Network for breast and cervical cancer screening.
The finding that fully 11.6% of the working-age population older than 40 years of age reports having glaucoma, ARMD, and/or cataract is concerning. A major implication for medical practice is that any nonacute or wellness visit by patients older than 40 years of age should trigger an emphatic recommendation for an ophthalmic checkup, which should be incorporated into the routine primary care visit workup (similar to the routine orders for cardiovascular risk factor screening and cancer screenings), consistent with the American Academy of Ophthalmology recommendation. Currently, a comprehensive eye examination by an ophthalmologist is not routinely recommended during primary care visits. Our study's findings are encouraging for this development: more than 85% of respondents reported having a regular physician.
Our study had some limitations. The BRFSS is a cross-sectional survey; therefore, causal relationships between vision insurance, eye care visits, and vision impairment cannot be conclusively inferred. However, the time frame of eye care visits and having eye insurance, in the prior year, and current vision impairment strongly suggests a causal association, particularly in light of documented studies showing the vision preservation effect of timely diagnosis and treatment of the conditions studied. Second, our data are obtained from the 8 surveyed states and may not generalize nationally. Third, the BRFSS question on eye care visits has no details of the care received, whether it was a routine preventive checkup or for other eye conditions. Future research should explore the longitudinal and specific impact of vision coverage on preventive eye checkups and the consequential vision preservation impact at a population level. Lastly, BRFSS data are subject to recall bias because they are self-reported, with no verification obtained from medical records.
Correspondence: Sudha Xirasagar, MBBS, PhD, Department of Health Services Policy and Management, Arnold School of Public Health, University of South Carolina, 800 Sumter St, Room 116, Columbia, SC 29208 (firstname.lastname@example.org).
Submitted for Publication: July 9, 2012; final revision received September 4, 2012; accepted September 4, 2012.
Published Online: December 10, 2012. doi:10.1001/jamaophthalmol.2013.1165
Author Contributions: Ms Li and Dr Xirasagar had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Conflict of Interest Disclosures: None reported.