The current study provides an evaluation of self-reported eye care utilization and eye disease among older adults by examining response agreement with information obtained from medical records. Agreement was substantial for the question regarding time of last eye examination; however, the results suggest that eye care utilization estimates based on this question will overestimate the number of respondents who received an eye exam within the previous year and underestimate the number receiving an eye exam greater than two years ago. Qualitatively, for the four eye disease questions investigated, agreement was substantial for one, moderate for two and slight for the other.
The study’s prevalence estimates for cataracts are much greater than previous estimates for a similarly aged population.8,11,12,31,32
This is primarily due to the present study combining responses for current cataract and past cataract removal. Nonetheless, the concordance for cataracts in the current study was comparable to previous research.11,12
A large proportion (85%) of the current study’s subjects who reported no cataracts, in fact had medical record information indicating cataracts; as might be expected, most (92.9%) of these participants had not had an IOL. Poor agreement may be due to eye care providers noting a diagnosis of cataract in the early stages of opacity, yet because the vision impairment is minor or the patient cites no vision complaints, the provider does not believe it necessary to inform patients of the early condition.
The self-reported prevalence of glaucoma in the current study was similar to those reported in recent analyses of similarly aged adults, one that that relied on 2005 through 2008 BRFSS data,8
and the other, 2002 NHIS data.32
Two conditions related to glaucoma, but diagnostically distinct from it, are ocular hypertension and glaucoma suspect. In caring for patients diagnosed with ocular hypertension or glaucoma suspect and in discussing their disorder with them, it would not be surprising if eye care providers may mention the term “glaucoma”. Thus it would not be surprising if some patients got the impression that they had glaucoma. Broadening the classification of glaucoma to include glaucoma suspect (N=104) and ocular hypertension (N=26) increased the number of subjects diagnosed with glaucoma-related conditions from 200 (10.8%) to 330 (17.7%). Agreement for this broad diagnostic category with self-report of glaucoma was reduced from 0.73 to 0.60 suggesting that those with ocular hypertension or glaucoma suspect diagnoses tended to correctly understand that they did not have glaucoma. Of the 200 subjects with medical record information indicating glaucoma, 165 (77.0%) correctly self-reported their glaucoma diagnosis. For the 130 subjects with medical record information indicating glaucoma suspect or ocular hypertension, 109 (83.9%) correctly self-reported no glaucoma diagnosis. Thus, a large proportion of participants with a medical record diagnosis of glaucoma, glaucoma suspect and ocular hypertension were able to accurately report, suggesting that self-report of glaucoma is valid.
Self-reported prevalence of macular degeneration was slightly lower than those of a recent analysis of similarly aged adults that relied on 2005 through 2008 BRFSS data,8
but similar to self-reported prevalence among adults 75 and older participating in the 2002 NHIS.32
Differences may be due to the current study’s inclusion of only current drivers. Since the ability to obtain a license to drive a motor vehicle and one’s decision to actually be a driver is influenced by vision,33
it is likely that participants had better vision health on average compared to similarly aged individuals regardless of licensure status from the general population.
The results indicated moderate agreement for the question regarding diabetic retinopathy. A recent study which investigated the prevalence of diabetic retinopathy using 1997–2010 BRFSS data found that although the prevalence of diabetes had increased over the study period, the prevalence of age-adjusted vision impairment and diabetic retinopathy among those reporting diabetes had decreased from 23.7% to 16.7%.34
It is encouraging that the anticipated increase in diabetic retinopathy has not occurred, but the authors note that the unclear validity of the self-report of vision impairment was a study limitation. Results of the current study indicate under-reporting for self-reported diabetic retinopathy but do not provide information as to whether reporting is differential with regard to diabetes diagnosis, subject age, or time period of report.
The current study found a significant difference in agreement between African Americans and Whites for self-reported macular degeneration with medical record (κ=0.14 versus κ=0.43, p<0.01). Previous research is ambiguous regarding the role of race in the accuracy of self-reported medical conditions.13,16,26,27
Racial differences in health literacy35
might contribute to African Americans being less familiar with eye disease such as glaucoma.36
An analysis of focus groups of ophthalmologists and optometrists from a similar geographic region as the current study found that a majority (74%) of their comments were negative about older African Americans’ attitudes about vision and eye care. In general, the eye care professionals were concerned that older African Americans did not make eye care a priority and did not fully understand the importance of preventive strategies and available treatments.37
Recent research has supported the proposition that physician-patient communications play an important role in race based health disparities,38–41
which may have contributed to the agreement patterns reported in the current study for macular degeneration and glaucoma.
Early detection and treatment of eye diseases is important in maintaining eye health and decreasing disability.42–44
Mobility and quality of life are negatively impacted by vision impairment resulting in heavy personal burdens and high costs to society; thus, early identification and treatment are the keys to decreasing eye disease, vision impairment and blindness.45–48
Consequently, population based information on eye care utilization is valuable in assessing the quality and determinants of eye care. In the present study agreement between self-report and medical record for this variable was substantial and the proportion of subjects who reported having been to an eye care provider in the previous year was comparable to previous research.8
Nonetheless, subjects’ medical records suggested that participants were overestimating eye care visits that occurred within one year and underestimating visits that occurred after two years. Because age-related eye diseases are often asymptomatic, current recommendations for older adults, i.e., those >65 years of age, from the American Academy of Ophthalmology and the American Optometric Association are to receive a comprehensive dilated eye examination every one to two years.49,50
Strengths of this study are as follows. The current study was a large, population-based examination of agreement between self-report and medical record on eye care utilization and major age-related chronic eye conditions among adults aged 70 years and older. Medical records were available for a large proportion of subjects and the person abstracting medical charts was masked to subject responses. In addition, all medical records were obtained from ophthalmologists and optometrists.
One important methodological consideration is how to handle self-reported responses of “not sure.” Few subjects self-reported “not sure” for time of last eye examination (N=7) and slightly more responded “not sure” for questions regarding cataracts (N=15), glaucoma (N=11), macular degeneration (N=17), and diabetic retinopathy (N=7). The current analysis omitted subjects who responded in this manner from the respective prevalence calculation. Based on information obtained from subjects’ medical records, 14/15 (93.3%) had cataracts, 7/11 (63.6%) had glaucoma, 5/17 (29.4%) had macular degeneration and 2/7 (28.6%) had diabetic retinopathy. To include those responding “not sure” in the analysis would require assuming that either none or all had the condition in question, thereby driving prevalence estimates down or up, respectively. However, given the small number of subjects (< 1%) in the current study who self-reported “not sure” to any of the eye disease questions, the impact on results was minimal.
The current study did not consider education level or cognition. Increased educational attainment has been reported to increase the validity of self-reported diagnoses that result in hospitalizations,11
and for some chronic conditions, e.g., diabetes, hypertension, myocardial infarction, and stroke.21
Some studies have reported that educational attainment does not play a significant role in self-report validity among elderly subjects,10,25,16
and the relationship between level of cognition and the validity of self-reported medical information is not well established.20,25,28
Nonetheless, relatively few (6.4%) participants did not complete high school, and only 2.4% performed less than normal on cognition assessment.
Only licensed drivers were included in the sample but the National Highway Traffic Safety Administration estimates that approximately 90% of older adults are licensed drivers.51
Thus, the prevalences of eye diseases and vision impairment are likely lower in the study population than in the general population which would result in conservative estimates. And finally, the study population did not include a meaningful number of participants (< 1%) who were not White or African American and did not include any participants less than 70 years of age.
In summary, the current study suggests that national estimates based on self-report underestimate eye disease and overestimate eye care utilization in the 70 years of age and older population. Future research should investigate these associations in other populations, e.g., < 70 years of age, Hispanic and Asian American populations.