Barriers to eye care among participants of a mobile eye clinic

Abstract

Purpose: Barriers to health care present complex challenges to improving eye health in the U.S., yet few studies have quantitatively measured the impact of these barriers. This study investigated the time elapsed since a person’s most recent eye examination (TLEE) as a measure of barriers to eye care. Methods: Participants (N = 1699) from mobile clinic vision health screenings completed demographic and subjective visual function questionnaires, and then underwent comprehensive eye exams. Associations between demographics, subjective visual function, and visual acuity were analyzed with respect to TLEE. Results: Many Hispanic (34.9%) and uninsured (28.6%) participants had no previous eye exam. Although most Caucasians had a previous eye exam, 40.2% did not have an exam in the previous four years. The majority of participants with diabetes were non-compliant with recommendations of annual eye examinations. Conclusion: These results describe barriers that are specific to subpopulations, providing useful information for efforts to improve eye care access.

Keywords:

• ophthalmology
• visual acuity
• access
• hispanic
• caucasian
• diabetic retinopathy

PUBLIC INTEREST STATEMENT

Visual impairment has a large impact on health and quality of life, and timely preventative eye care is important. Existing studies on barriers to eye care access focus on established patients in clinics, and do not account for screening guidelines. This study instead aims to explore barriers among the general population in a quantitative manner using data from the Casey Eye Institute Outreach (CEIO) Program. This mobile screening program provides no cost vision screening and on site comprehensive eye exams for un- and under-insured adults throughout Oregon. It was found that 34.9% Hispanic participants had no previous eye exam. Furthermore, 40.2% of Caucasians and the majority of participants with diabetes had not had eye exams within the recommended time period. These findings will help the CEIO and other agencies better understand the populations that need to be targeted for improved eye care access.

1. Introduction

Visual impairment exerts a substantial impact on health and quality of life, and can lead to life dissatisfaction, disability, and decreased physical activity (Crews et al., 2016). Between 10.7% to 32.1% of visually impaired individuals are depressed (Casten, Rovner, & Tasman, 2004; Skalicky & Goldberg, 2008; Zhang et al., 2013). Furthermore, chronic visual conditions present as an economic burden to the United States economy, and accounted for approximately $139 billion of national costs in 2013 according to Prevent Blindness America (Wittenborn & Rein, 2013).

Several sources of health care barriers can delay or prevent eye examinations that can reduce vision loss. Economic barriers include affordability of health services, lack of health insurance, decreased work flexibility to address health care needs (Goins, Williams, Carter, Spencer, & Solovieva, 2005; Kullgren, McLaughlin, Mitra, & Armstrong, 2012; Macinko, Shi L, B, & Wulu, 2003; Owsley et al., 2006; Syed, Gerber, & Sharp, 2013) and transportation issues among the elderly (Goins et al., 2005; Lee et al., 2009; Owsley et al., 2006; Syed et al., 2013). Underinsured and uninsured patients utilize vision care services less frequently than insured patients (Lee et al., 2009), potentially due to the high out of pocket cost, lack of vision care coverage under public health insurance plans (Quigley, Park, Tracey, & Pollack, 2002), and not understanding the difference between vision care coverage and eye health coverage (Li, Xirasagar, Pumkam, Krishnaswamy, & Bennett, 2013). Geographic barriers include distance to appropriate health care providers and permanent or seasonal travel obstacles, such as mountains and winter road conditions (Tsui, Siedlecki, & Deng et al., 2015). These barriers are particularly prevalent in rural communities, where individuals may have to travel long distances to see an eye care provider (Tsui et al., 2015). Cultural barriers include language, rural/urban cultural differences, ethnic differences in health behavior, and attitudes toward traditional care versus the modern health care system in the U.S. (Chen, Vargas-Bustamante, Mortensen, & Ortega, 2016; Rhodes et al., 2016). Low levels of health literacy may also impede patient education efforts by providers within the clinical care setting (Baker et al., 1996; Friedman et al., 2013; Schillinger et al., 2002).

Barriers to health care access present complex challenges to improving eye health, and need to be better understood if effective and sustainable solutions are to be identified (Skalicky & Goldberg, 2008; Wittenborn & Rein, 2013). These barriers often lead to long delays in accessing eye care for underserved and high-risk populations. The American Academy of Ophthalmology (AAO) recommends preventive eye exams every 2–4 years between 40 to 54 years of age, every 1–3 years between 55 to 64 years of age, and every 1—2 years over 65 years of age (AAO Hoskins Center for Quality Eye Care, 2015). People with diabetes should have a preventive eye exam at least once a year. Yet, data from vision-screening programs are indicating that barriers to preventive eye exams persist. In the rural northeastern United States data collected between 2011 and 2014 indicated that participants last sought eye care an average of 7.1 years prior to the screening (Tsui et al., 2015). In addition, an Ohio-based study revealed that the largest portion of their screening participants (39%) between 2009 and 2011 last received a dilated eye exam 10 or more years prior (Friedman et al., 2013).

Much of the existing research on eye care barriers is based on whether participants have received an eye exam previously, but does not include quantitative data on eye exam history despite AAO screening guidelines. The variable of time elapsed since last eye exam might provide better insight into certain populations receiving inadequate eye care. The Casey Eye Institute Outreach (CEIO) Program is a mobile screening program intended to serve un-insured and under-insured participants by providing no cost vision screening and on site comprehensive eye exams for adults throughout Oregon during planned service trips typically lasting one to seven days. The clinic is staffed by volunteer eye care providers and trained community members. CEIO data collected between October 2014 and July 2016 contain several metrics that may serve as a proxy for barriers to preventive eye care including race/ethnicity, health insurance status, geographic location, as well as motivators for seeking recommended eye care (i.e., perceived vision decencies and history of diabetes). Using the AAO recommendations for preventive eye exams, this paper will present data on associations between possible indicators of access barriers based on TLEE. This study will allow CEIO to better address the unique needs of underserved areas in Oregon.

2. Methods

2.1. Ethics

The Oregon Health & Science University Institutional Review Board granted approval in 2013 to perform a retrospective chart review on screening participants, and this study was conducted in accordance with the principles in the Declaration of Helinski. Participants were provided informed consent. Those who did not consent were not included in this study, however, they were still provided the service. This occurred less than one time per event on average.

2.2. Study population

The CEIO program partnered with 60 local health and social service partner agencies throughout Oregon to provide free vision health screenings and on-site comprehensive exams for individuals identified as un—or under- insured. Inclusion criteria for at-risk participants was defined as un- or under- insured persons who have not had access to preventive eye exams. Under-insured was defined as insurance that does not cover eye exams. The mobile clinic made an average of 28 agency visits, and screened an average of 1,200 participants, per year. Partner agencies included federally qualified health centers, homeless shelters, community centers, senior centers, advocacy groups, and American Indian tribal health and wellness centers. All partner agencies signed a contract assuming responsibility for referrals to clinical eye care as well as to dispense eyeglass prescriptions.

Data from screenings held between October 2014 and June 2016 were analyzed. Only data from participants’ first screening were analyzed. The outreach program is designed to serve adults only, however occasionally someone may ask to have their child screened due to symptoms of eye problems. Therefore, participants less than 18 years old were excluded. Participants pregnant at the time of the screenings were excluded due to the temporary effect of hormones that could alter their visual acuity temporally. Participants who did not complete the majority of the demographic survey or the complete exam process were also excluded.

2.3. Screening process

Participants completed a demographic and subjective visual function questionnaire inquiring about race/ethnicity, gender, health insurance coverage, zip code of residence, history of diabetes, and TLEE. Subjective measures of vision were collected. Participants indicated whether they had no, little, moderate, or severe difficulty reading a newspaper and reading street signs according to the National Eye Institute Visual Function Questionnaire (Crews et al., 2012; Mangione et al., 2001), a widely used tool in existing research. Near visual acuity was then measured with a hand held Snellen Eye chart and distance visual acuity with a Snellen Eye chart at 20 feet. Autorefraction (NIDEK ARK-530A/ARK-510A), pupil exam, tonometry (Reicher Tono-Pen AVIA), and lensometry (NIDEK LM-600P) were performed and dilation drops (Bausch + Lomb 1% tropicamide, Tampa, FL) were administered if anatomic narrow angles or risk factors for narrow angle glaucoma were not identified. An eye doctor performed manifest refractions using a phoropter (Recihert Phoroptor model 11635/11635B), and a comprehensive eye health exam with a slit lamp and indirect ophthalmoscope. Participants requiring prescriptive spectacle correction were provided a prescription and subsidized or free spectacles. If participants required further evaluation or medication based on abnormal exam findings, they were referred to a clinic-based eye doctor for further evaluation. Access to clinical eye care was managed by partner agencies. Interpreters were provided at each screening as needed through the partner agency. Paper charts were transcribed into OnBase® (version 15.0.1.84; Westlake, Ohio). Analysis was conducted on de-identified data to ensure confidentiality.

2.4. Statistical analysis

Participant data was analyzed using SAS Version 9.4 (SAS Institute, Cary, NC). Participants were classified as rural or urban based on the Metropolitan Statistical Area map by the Oregon Office of Rural Health (Oregon Office of Rural Health, 2016). Additionally, participants who identified with more than one racial category were considered of mixed race. World Health Organization visual acuity criteria were used to classify Snellen visual acuity ranges into visual impairment categories of mild [20/30–20/70], moderate [20/80–20/200], and severe (20/200 -) (World Health Organization, 1997).

All predictor variables, including demographics, subjective symptoms, or objective visual acuity, were analyzed as categorical measures. The outcome variables included: 1) having eye exam history (yes/no) as a binary measure; 2) TLEE as a continuous measure. For the binary outcome variable, Pearson chi-square tests were run as the initial bivariate analysis. Logistic regression models were further fit with individual predictor variable, and age variable. Age-adjusted Odds Ratios (ORs) and 95% confidence interval were calculated from these models. For the continuous outcome variable of TLEE, its distribution appeared to be right-skewed. It was categorized into <1, 1-<2, 2-<3, 3-<4, 4-<10 and 10+ years for summary tables of descriptive statistics. The data were logarithmically transformed to achieve approximately normal distribution when conducting statistical testing. Two-sample student t-test (for dichotomous predictor variables) or one-way ANOVA (for predictor variable with ≥ 3 levels) was used as the initial bivariate testing. General linear models were further fit with individual predictor variable and age variable, and age-adjusted comparisons between different levels were calculated. All possible pairwise comparisons were conducted with Tukey’s adjustment. A two-sided α level of significance of 0.05 was used in statistical tests.

3. Results

A total of 1,699 participants underwent vision health screenings. The majority were female (56.3%), over the age of 40y (76.1%), and lived in an urban setting (61.1%) (Table 1). The racial and ethnic makeup of the population was primarily Hispanic (34.3%), Caucasian (34.1%), and American Indian/Alaska Native (AI/AN; 14.7%). Just over half of the population had some form of health insurance (54.8%), and 29.5% had diabetes.

Table 1. Demographic breakdown of participants who have and have not had an eye exam prior to the screening

	All (N = 1699)	History of Eye Exam (N = 1344)	No History of Eye Exam (N = 261)	p	Age-adjusted Odds Ratio (OR) ^†
Demographics	N (%)	n (%)	n (%)		Estimate (95% CI)	p**
Gender				0.73		0.68
F	955 (56.3)	771 (84.0)	147 (16.0)		1 (ref)
M	740 (43.7)	570 (83.3)	114 (16.7)		0.94 (0.72, 1.24)	0.68
Unknown/Unreported*	4 (n/a)	3 (n/a)	0 (n/a)
Race/Ethnicity				<.0001		<.0001
Caucasian	556 (34.1)	505 (96.4)	19 (3.6)		1 (ref)
Hispanic	559 (34.3)	349 (65.1)	187 (34.6)		0.08 (0.05, 0.14)	<.0001
AI/AN	239 (14.7)	231 (98.3)	4 (1.7)		2.14 (0.72, 6.39)	0.17
Black	53 (3.3)	39 (86.7)	6 (13.3)		0.27 (0.10, 0.71)	0.0085
Asian	91 (5.6)	70 (86.4)	11 (13.6)		0.19 (0.09, 0.43)	<.0001
Multiethnic/Other	131 (8.0)	97 (77.0)	29 (23.0)		0.14 (0.07, 0.26)	<.0001
Unknown/Unreported*	70 (n/a)	53 (n/a)	5 (n/a)
Health Insurance				<.0001		<.0001
None	639 (45.2)	434 (71.4)	174 (28.6)		1 (ref)
Public	678 (48.0)	593 (94.1)	37 (5.9)		5.14 (3.50, 7.53)	<.0001
Private	96 (6.8)	88 (91.7)	8 (8.3)		4.12 (1.94, 8.75)	0.0002
Unknown/Unreported*	286 (n/a)	229 (n/a)	42 (n/a)
Location				0.0024		0.090
Rural	632 (38.9)	543 (87.4)	78 (12.6)		1 (ref)
Urban	994 (61.1)	747 (81.6)	168 18.4)		0.77 (0.57, 1.04)	0.090
Unreported*	73 (n/a)	54 (n/a)	15 (n/a)
Age (y)				<.0001		<.0001
18–39	405 (23.9)	298 (78.2)	83 (21.8)		1 (ref)
40–54	611 (36.1)	441 (76.6)	135 (23.4)		0.91 (0.67, 1.24)	0.55
55–64	390 (23.0)	339 (91.4)	32 (8.6)		2.95 (1.91, 4.57)	<.0001
65+	288 (17.0)	263 (96.0)	11 (4.0)		6.66 (3.48, 12.76)	<.0001
Unknown/Unreported*	5 (n/a)	3 (n/a)	0 (n/a)
History of Diabetes (y)				0.0007		0.026
None	1197 (70.5)	981 (82.0)	216 (18.1)		1 (ref)
Yes [1–5]	168 (9.9)	143 (85.0)	25 (14.9)		1.17 (0.74, 1.86)	0.50
Yes 5+	235 (13.8)	216 (91.9)	19 (8.1)		1.99 (1.20, 3.31)	0.0075
Unreported/Unknown *	5 (0.3)	4 (n/a)	1 (n/a)

*Unknown, unreported, or missing values have not been considered in calculations; ^†OR for age is unadjusted.

Table 1 describes participant history of having an eye exam based on demographics, potential barriers to access, and motivators for seeking eye care. There was little variation of access indicators among males and females, and rural and urban participants, across all time categories. Within each ethnic group except Caucasians, most participants had an eye exam within the past year compared to other time categories. Participants received an eye exam within the past year more often than for any other TLEE category (e.g. 1-<2 years, 2-<3 years, etc.) for each ethnic group including 28.1% of Hispanics, 45% of AI/AN, 48.7% of Blacks, 45.7% of Asians, and 43.3% of multiethnic participants. Notably, the most common TLEE for Caucasians was 4–10 years (26.7%), compared to 1 year (19.6%). Among insured participants, 32.0% of publically insured and 36.4% of privately insured participants had an eye exam within the last year, compared to 26.5% of uninsured participants. However, AAO recommendations for eye examinations were not met for 28.4% of uninsured and 31.7% of those on public insurance, compared to 14.8% of those with private insurance. Across age categories, 30% of 40-54yo, 36.7% of 55-64yo, and 39.9% of 65 and older had not met the AAO recommendations. Less than half of the diabetic participants received AAO recommended annual screenings.

Age showed significant association with TLEE (Table 2). After adjusting for age, ethnicity, health insurance, and diabetes history were also significant variables. Compared to Caucasians, all ethnicities had a significantly shorter mean in (TLEE), especially AI/AN (p < 0.0001). Privately insured (p = 0.037) sought eye care sooner than uninsured participants. As participant age increased, the (TLEE) decreased, especially for 65+ yo (p = 0.02). Participants with longer history of diabetes had eye exams more recently than those diagnosed more recently (p < 0.0001).

Table 2. Demographic breakdown of participants with previous history of eye exam based on TLEE

	Years since Last Eye Exam						Ln(Years Since Last Eye Exam)
Demographics	(0-<1] n %	(1-<2] n %	(2-<3] n %	(3-<4] n %	(4-<10] n %	10+ n %	Mean ± Std Dev	P Value (Age adjusted)^†
Gender								0.20
F	231	172	102	61	153	52	0.85 ± 1.13
	30.0	22.3	13.2	7.9	19.8	6.7
M	177	100	62	39	132	52	0.93 ± 1.24
	31.1	17.5	10.9	6.8	23.2	6.7
Unknown/Unreported*							n/a
Race/Ethnicity								<0.0001
Caucasian	99	100	60	43	135	68	1.21 ± 1.09	Ref
	19.6	19.8	11.9	8.5	26.7	13.5
Hispanic	98	73	61	28	69	20	0.87 ± 1.09	<.0001
	28.1	20.9	17.5	8.0	19.8	5.7
AI/AN	104	52	24	13	29	9	0.45 ± 1.16	<.0001
	45.0	22.5	10.4	5.6	12.6	3.9
Black	19	7	3	2	5	3	0.57 ± 1.21	0.0056
	48.7	18.0	7.7	5.1	12.8	7.7
Asian	32	12	5	5	15	1	0.43 ± 1.29	<.0001
	45.7	17.1	7.1	7.1	21.4	1.4
Multiethnic/Other	42	17	10	6	13	9	0.65 ± 1.36	<.0001
	43.3	17.5	10.3	6.2	13.4	9.3
Unknown/Unreported*							n/a
Health Insurance								0.042
None	115	88	67	41	95	28	0.95 ± 1.06	Ref
	26.5	20.3	15.4	9.5	21.9	6.5
Public	190	117	61	37	128	60	0.89 ± 1.23	0.97
	32.0	19.7	10.3	6.2	21.6	10.1
Private	32	24	12	7	10	3	0.61 ± 0.97	0.037
	36.4	27.3	13.6	8.0	11.4	3.4
Unknown/Unreported*							n/a
Location								0.16
Rural	164	116	70	27	110	56	0.92 ± 1.17
	30.20	21.36	12.89	4.97	20.26	10.31
Urban	224	148	86	73	164	52	0.87 ± 1.19
	30.0	19.8	11.5	9.8	22.0	7.0
Unreported*							n/a
Age (y)								0.0026
18–39	78	65	35	27	72	21	0.93 ± 1.08	Ref
	26.2	21.8	11.7	9.	24.2	7.1
40–54	119	93	60	37	85	47	0.98 ± 1.21	0.95
	27.0	21.1	13.6	8.4	19.3	10.7
55–64	104	63	48	28	67	29	0.90 ± 1.15	0.98
	30.7	18.6	14.2	8.3	19.8	8.6
65+	107	51	21	9	60	15	0.64 ± 1.27	0.02
	40.7	19.4	8.0	3.4	22.8	5.7
Unknown/Unreported*							n/a
History of Diabetes (y)								<.0001
None	245	200	122	76	241	97	1.01 ± 1.18	Ref
	25.0	20.4	12.4	7.8	24.6	9.9
Yes [1–5]	67	28	12	9	20	7	0.63 ± 0.99	0.0005
	46.9	19.6	8.4	6.3	14.0	4.9
Yes 5+	94	45	29	16	24	8	0.47 ± 1.18	<.0001
	43.5	20.8	13.4	7.4	11.1	3.7
Unreported/Unknown*							n/a

*Unknown, unreported, or missing values have not been considered in calculations; ^†P-value for age is unadjusted; P values were calculated using the Wald test in unadjusted logistic regression models. For predictor variables with ≥3 levels that were statistically significant, all pairwise comparisons were further performed with Turkey adjustment. Only the P values from the tests that compare with reference level are shown here.

In addition to demographic factors, data on the relationship between subjective and objective visual acuity versus TLEE were gathered (Table 3). Participants who reported moderate difficulty reading street signs had the highest rate of compliance with recommended eye exams, (87.5%; OR = 2.16, CI 95% [1.52, 3.07]), compared to those with no difficulty. Those with moderate objective distance visual impairment in the better (93.0%; OR = 2.64, CI 95% [1.20, 5.82]) and worse eyes (90.6%, OR = 1.87, CI 95% [1.15, 3.06]) had more recent TLEE than those with mild and extreme visual impairment. Average (TLEE) was not significantly associated with subjective distance visual acuity ranges (adj. p = 0.53), or objective distance visual acuity ranges in the better and worse eye (adj. p = 0.69 and 0.43, respectively). Average (TLEE) significantly increased as difficulty reading newspapers increased (adj. p = 0.0023). However, objective near visual acuity (adj. p = 0.40) was not significantly associated with (TLEE).

Table 3. Visual acuity (VA) breakdown of participants with previous history of eye exam based on TLEE

			Years since Last Eye Exam						Ln(Years Since Last Eye Exam)
Visual acuity	Total Participants	Participants with a history of eye exams (N = 1344) n (%)	(<1] n %	(1-<2] n %	(2-<3] n %	(3-<4] n %	(4-<10] n %	10+ n %	Mean ± Std Dev	P Value (Age adjusted)
Distance VA
Difficulty (subjective)**										0.53
none/little	403	312 (77.4)	96	59	40	24	67	26	0.91 ± 1.17
			30.8	18.9	12.8	7.7	21.5	8.3
moderate	566	495 (87.5)	140	92	61	40	110	52	0.96 ± 1.20
			28.3	18.6	12.3	8.1	22.2	10.5
extreme	72	62 (86.1)	13	25	10	2	8	4	0.79 ± 1.00
			21.0	40.3	16.1	3.2	12.9	6.5
unreported*	564	475 (n/a)							n/a
Objective—Better Eye										0.69
[20/10–20/20]	597	466 (78.1)	136	99	67	39	92	33	0.90 ± 1.07
			29.18	21.24	14.38	8.37	19.74	7.08
[20/30–20/70]	883	764 (86.5)	232	148	87	56	171	70	0.89 ± 1.26
			30.37	19.37	11.39	7.33	22.38	9.16
[20/80–20/200]	100	93 (93.0)	33	23	10	5	14	8	0.74 ± 1.11
			35.5	24.7	10.8	5.4	15.1	8.6
(20/200 -)	18	15 (83.3)	6	2	0	1	5	1	0.91 ± 0.90
			40.0	13.3	0.0	6.7	33.3	6.7
Unknown*	7	6 (n/a)							n/a
Objective- Worse Eye										0.43
[20/10–20/20]	295	217 (73.6)	66	39	29	22	47	14	0.91 ± 1.05
			30.41	17.97	13.36	10.14	21.66	6.45
[20/30–20/70]	1010	859 (85.1)	257	174	110	62	180	76	0.91 ± 1.19
			29.92	20.26	12.81	7.22	20.95	8.85
[20/80–20/200]	212	192 (90.6)	56	43	23	11	42	17	0.82 ± 1.30
			29.2	22.4	12.0	5.7	21.9	8.9
(20/200 -)	81	70 (86.4)	28	16	2	6	13	5	0.65 ± 1.20
			40.0	22.9	2.9	8.6	18.6	7.1
Unknown*	7	6 (n/a)	323	213	139	84	227	90	n/a
			30.0	19.8	12.9	7.8	21.1	8.4
Near VA
Difficulty (subjective)**										0.0023
none/little	374	290 (77.5)	106	56	37	16	57	18	0.71 ± 1.20	Ref
			36.5	19.3	12.8	5.5	19.7	6.2
moderate	619	515 (83.2)	143	96	70	42	114	50	0.99 ± 1.16	0.002
			27.8	18.6	13.6	8.2	22.5	9.7
extreme	123	103 (83.7)	25	26	13	6	20	13	1.01 ± 1.18	0.056
			24.3	25.2	12.6	5.8	19.4	12.6
unreported*	489	436 (n/a)							n/a
Objective—Both eyes										0.40
[20/10–20/20]	345	285 (82.6)	80	56	34	26	62	27	0.96 ± 1.09
			28.07	19.65	11.93	9.12	21.75	9.47
[20/30–20/70]	752	626 (83.2)	190	133	80	38	136	49	0.87 ± 1.18
			30.35	21.25	12.78	6.07	21.73	7.83
[20/80–20/200]	129	115 (89.2)	39	20	14	9	23	10	0.83 ± 1.25
			33.9	17.4	12.2	7.8	20.0	8.7
(20/200 -)	38	34 (89.5)	16	6	5	1	5	1	0.49 ± 0.98
			47.1	17.7	14.7	2.9	14.7	2.9
Unknown*	341	284 (n/a)							n/a

*Unknown, unreported, or missing values have not been considered in calculations.

**Difficulty (subjective) = Participants self-reported difficulty w/street signs for Distance VA or reading the newspaper for Near VA

4. Discussion

Answering the challenge of the rising prevalence of vision loss will require a better understanding of vision health access among races and ethnicities, for a range of other population groups (the elderly, poor, and isolated), and for those seeking care for particular eye diseases (National Academies of Sciences, Engineering, and Medicine, 2016). To investigate vision health barriers in Oregon, we analyzed data from a vision health outreach program with a broad geographic reach across the state to identify barriers in association with ethnicity and race, insurance status, age, and a diagnosis of diabetes.

The greatest vision health care access disparity was identified among Hispanics, who were less than one tenth as likely to have had a previous eye exam compared to Caucasians (Unzueta et al., 2004; Zhang et al., 2012). Less than half of Hispanics (49.0%) had an eye exam previously, compared to 60–68% of Black, Asian, and multiethnic participants. Studies have reported that Hispanics seek health care in a manner that is particularly based on the severity of their symptoms and their trust in a provider, which is increased with providers of the same gender/ethnicity (Larkey, Hecht, Miller, & Alatorre, 2001). Future efforts to increase access for Hispanics should emphasize prevention instead of seeking care for symptoms, and encourage outreach by and access to Hispanic providers (Juckett, 2013). Hispanics account for over 35% of patients at community clinics and health centers (Ortega, Rodriguez, & Vargas Bustamante, 2015), and 41% of non-citizen, non-legal permanent resident Hispanics report using community clinics or health centers for their health care (Pew Research Center, 2008). Ensuring that eye care is encouraged and integrated into these clinics and health centers might improve access among Hispanics. In contrast, AI/AN participants of the CEIO program were twice as likely as Caucasians to have had a previous eye exam. While statistically insignificant, this finding begs further exploration.

Previous studies have shown that Hispanics face limited access to eye care. However, these access barriers/issues vary relative to Caucasians depending on the setting or state investigated (Morales et al., 2010; Zhang et al., 2007). Differences in state-offered health insurance and other financial barriers to health care, as well as rural isolation in some settings, may be key influences among these unique populations, according to an analysis of 21 states between 2006–2009 (Chou, Barker, & Crews et al., 2012). While Oregon was not examined in the aforementioned study, our data showed that vision health care access utilization is low among Hispanics in Oregon. This finding might also be explained by health insurance disparities, since Hispanic Oregonians (88.9%) have the lowest rate of health insurance coverage among all races (95.7%) (Oregon Health Authority, 2015).

Caucasian participants were distinguished by their high likelihood of having had a previous eye exam and for having the longest average duration since their previous eye exam. Over 25% of Caucasians had their most recent eye exam over 4 years prior to their screening exam with the CEIO program. Higher rates of minority participants, such as 49% of Hispanics and 67.5% of AI/AN, had eye exams in the previous two years compared to 39.4% of Caucasians. The cause is unknown, but perhaps, the lack of access to or participation with focused community-oriented resources among Caucasians may partially explain this finding. Recently reported increases in premature deaths among Whites, largely due to poor behavioral health, risk taking behavior, and chronic disease suggest that these findings may fit into a larger picture of inconsistent health maintenance among Caucasians (Stein, Gennuso, Ugboaja, & Remington, 2017). This relatively unexplored finding may deserve further study.

The rate and overall odds of having had an eye exam in the past were higher among those with public and private insurance compared to the uninsured, a finding consistent with previous studies (Lee et al., 2009). However, nearly 22% of both publically insured and uninsured participants had their last eye exam 4-10y ago. Although health insurance coverage increased after implementation of the Affordable Care Act (ACA), the associated marketplace health plans that usually do not include vision care coverage for adults (Li et al., 2013; Oregon Health Authority, 2017; Zhang et al., 2008). Furthermore, even with the ACA, stand-alone vision plans are only available through private insurance companies, which may be unaffordable to certain populations. As well, publically insured individuals are more likely to be of a lower socioeconomic status (Zhang et al., 2013), a group commonly associated with other factors like income and employment status that limit vision health excess (Shavers, 2007).

Our study adds to previous findings that older populations are more vulnerable to eye care disparities (Prevent Blindness America, National Eye Institute, 2008). While 10% more of those in the 65+ yo age group had eye exams within the last year compared to all other age categories, 28.5% of 65+ yo did not meet the AAO recommendations for preventive eye exams, in spite of Medicare insurance eligibility. This finding is concerning since the AAO recommends annual eye exams for persons in this age category who are much more likely to experience eye disease. Medicare Part B only provides coverage for annual eye exams for individuals with diabetes or at high risk for glaucoma; routine annual eye exams for refraction are not covered (Centers for Medicare and Medicaid Services, 2014). This might explain low eye care utilization among the elderly who do not have diabetes.

Diabetes has profound effects on health, and missing more than one clinic visit has been shown to have a positive correlation with increased weight, A1C, and lipids (Mallow, Theeke, Barnes, Whetsel, & Mallow, 2014). Diabetes also exerts a strong effect on vision health that increases with duration of the disease. Our findings indicated that participants with a longer history of diabetes had a better history of eye exams. Individuals with 5+ years of diabetes were twice as likely to have had an eye exam compared to non-diabetics and those with <5 years of diabetes history. This is possibly because longer-term diabetes is associated with worsening vision and more awareness of disease complications, prompting more careful monitoring of eye exam adherence by primary care providers. Less than half of the participants with diabetes received an eye exam within the previous year, failing to meet the AAO annual screening guidelines for diabetics. While some participants with 1-5y of diabetes history might be newly diagnosed Type I diabetics, not requiring annual exams until five years after diagnosis (American Academy of Ophthalmology Retina/Vitreous Panel, 2016), new diagnoses of Type II diabetes is more common among older adults and requires annual eye exams starting at the time of diagnosis.

Subjective or objective distance or near visual acuity were thought to be potential motivators for accessing eye exams, yet showed inconsistent association with eye exam history and frequency. Decreased subjective and objective distance acuity were significant predictors for having eye exam history. Subjective problems with distance visual acuity might motivate individuals to seek care more than near visual acuity concerns, since distance vision is required for a great variety of daily tasks like driving and work responsibilities (Qiu, Wang, Singh, & Lin, 2014). Furthermore, individuals might be motivated to meet the distance visual acuity requirements set by the Department of Motor Vehicles (DMV) in order to obtain or renew their license.

Other studies have suggested that rural residents might travel to urban areas for health care, including eye health services, yet rural/urban location was not significantly associated with TLEE. Previous studies have also found that there is great variability in frequency of eye care utilization between major cities around the country (Gower, Friedman, & Haller et al., 2014). Furthermore, Medicare beneficiaries from rural and urban areas had similar amount of care and satisfaction with care access, despite lower provider density and slightly longer travel times in rural areas (Stensland, Akamigbo, Glass, & Zabinski, 2013). The impact of location as a barrier to eye care seems to vary in different parts of the United States.

Overall, the barriers to eye care identified in our study can be addressed in multiple ways. Mobile eye clinics can bring eye care to individuals, thereby potentially overcoming transportation, location, and financial barriers. Such clinics can also target community clinics and health centers to reach specific populations, such as Hispanics. Telemedicine can also be used to overcome these barriers and provide long-term monitoring for eye pathology, as it has demonstrated an increase in the percentage of diabetic retinopathy screening exams (Mansberger et al., 2013).

These findings should be interpreted in light of several limitations. In this non-randomized study, biases were introduced from self-selection and partner agency selection of screening participants, as well as the potential for inducement by offering free exams and spectacles. However, this was addressed with logarithmic data transformation to normal distribution, reducing the effect of skew. Furthermore, this study might have benefited from more refined survey questions, such as whether participant insurance covered eye care, specifying whether subjective visual acuity was reported with or without corrective lenses, and distinguishing between Type I or II diabetes. For the purposes of this study, adherence to AAO guidelines of annual diabetic screenings was determined based on a hard cutoff of TLEE ≤ 1y, while recognizing that a number of participants may have had eye exams within several months of that time frame. Finally, there were relatively few Black, Asian, and privately insured participants. Although the results were significant, a higher sample size would yield more definitive findings.

5. Conclusion

Our study yields several key findings, which include a paucity of eye exams among Hispanics in Oregon, long duration between eye exams among Caucasians, and inadequate frequency of eye exams among diabetic individuals per AAO recommendations. A positive history of eye exam and frequency of eye exams were higher among individuals with a longer history of diabetes, and those with subjective and objective distance visual acuity problems. While older, insured, and diabetic participants received more frequent eye exams, more effort needs to be made to increase the number of individuals meeting current screening recommendations. The variable of TLEE provided unique insight into the complexity of eye care access disparities. Current screening guidelines and standard of care recommendations by the AAO for a variety of eye disorders specify time intervals between eye exams. Investigating barriers to care therefore could involve TLEE to identify populations in need of more frequent eye exams.

6. Abbreviations

Time since last eye exam (TLEE), American Academy of Ophthalmology (AAO), Casey Eye Institute Outreach (CEIO), odds ratio (OR), American Indian/Alaska Native (AI/AN), confidence interval (CI), Affordable Care Act (ACA), Department of Motor Vehicles (DMV)

Declaration of Interest

None of the authors have any proprietary interests or conflicts of interest related to this submission.

This submission has not been published previously and it is not being considered for any other publication.

Cover image

Source: Author.

Additional information

Funding

This work was supported by the National Institutes of Health [P30 EY010572]; Research to Prevent Blindness [CEI/OHSU].

Notes on contributors

Mitchell V. Brinks

The Casey Eye Institute domestic outreach team, led by Dr. Mitchell V Brinks, partners with communities, academic, and professional organizations to advance the understanding of eye disease and to develop effective programs to reduce avoidable blindness in Oregon and nationally. The Casey Eye Institute Adult Outreach Program provides vision screening to under and un-insured individuals and is active in creating policy at the national level to address vision health. Current research and advocacy efforts include studying the cause, and the solution, for avoidable blindness in Oregon.