BRABYN, JOHN PhD; SCHNECK, MARILYN PhD; HAEGERSTROM-PORTNOY, GUNILLA OD, PhD, FAAO; LOTT, and LORI PhD
The impetus for the SKI study was twofold; namely the need for more information about vision function and its impact among the older population and the need for better vision tests that address the everyday viewing conditions under which elders (even those with good acuity) frequently complain of difficulties.
Relatively few data exist on the vision function of older persons—especially the rapidly growing group who are >75 years old. We wished to examine the vision of this population in more detail and assess the interaction between vision function, visual task performance, health and functioning, and self-report. To examine impact and risk factors, we began a longitudinal study. The resulting project has many similarities to the contemporaneous SEE study, 1 but differs in several respects including extension to an older age group and use of a wider variety of vision measures.
Even elders with good standard acuity often report visual difficulties under everyday conditions (driving at night or at dusk, adjusting to dim indoor lighting after coming in on a sunny day, reading menus in dimly lit restaurants, etc.). With a few recent exceptions, most studies on vision have been restricted to standard vision measures that do not reflect these nonideal conditions. 2–4 We therefore wished to characterize vision using tests that more accurately reflect “real-world” impairments in the older population to provide a better understanding by both patient and clinician of an individual’s visual capabilities for everyday tasks and help guide rehabilitation efforts.
We tested a random subset of a population previously used for epidemiological studies by our collaborators, the Buck Center for Research in Aging. The Buck study group was a stratified random sample of the older population living in Marin County, California, with approximately equal numbers in the 55 to 64, 65 to 74, 75 to 84, and ≥85 years age groups. Of 1284 subjects invited into our study, 900 agreed to participate. (Those who refused were generally older and had worse vision, based on Buck Center interview data.) Participants ranged from 58 to 102 years of age, including 449 persons >75 years and 297 >80 years; the median age of the whole sample was 74.9 years. (Additional details of the subject sample appear in Haegerstrom-Portnoy et al. 5) At the time of the second round of testing, an average of 4 years later, 139 of the original 900 subjects had died, 42 had moved away, 11 could not be contacted, and 112 declined retesting; therefore, 596 were tested. Once again, subjects who refused were older than those who accepted. A third round of testing is currently underway.
Tests and Questionnaires
Vision tests were chosen using two criteria: first, to represent the conditions under which older persons commonly report problems in everyday life and, second, to be as inexpensive, simple, and rapid to administer as possible to ensure clinical practicality. The test battery has been described in detail elsewhere. 5 Standard high-contrast distance letter acuity was measured using the Bailey-Lovie chart 6; the general format was replicated in several of the other tests for direct comparison. For example, a low-contrast (gray on white) versions of the letter chart were used at distance for measuring low-contrast acuity. A similar low-contrast chart for use at near was also used without and with a surrounding diffuse glare source to measure performance in the presence of veiling glare. 7 Performance under low light and contrast was assessed using the SKILL (Smith Kettlewell Institute Low Luminance) card—a letter chart of similar format but with black letters on a dark gray background. 8 This provides an effective reduction in light level by a factor of 10 while conducting the test under normal room lighting. A simple glare recovery test was used, which measured the time taken to recover sufficient vision to read a line on the SKILL card two lines above the person’s previously established threshold after looking at a bright (3300 cd/m2) glare source for 1 min. Visual fields were measured on a standard perimeter using a standard screening test with bright targets, which was then repeated with the addition of a central distracting task (to count the number of times the central fixation point blinked off during the test cycle). The Frisby stereo test was used to assess the ability to detect depth at three different coarse disparities. 9 Contrast sensitivity was assessed with the Pelli-Robson contrast sensitivity test consisting of a sequence of large (1°, requiring 20/240 acuity) letters of steadily diminishing contrast. 10 No test took longer than about 3 min. Because we were interested in vision under realistic, everyday conditions, all tests were performed with both eyes viewing and with the person’s usual glasses for distance or near as appropriate.
Visual task performances included reading as measured by the Pepper Visual Skills for Reading Test. 11,12 This is an oral reading test of words and letters and is scored in terms of accuracy and speed. A test of dark adaptation was used in which the subject was timed while walking parallel to the wall from one marker (a night light on the wall) to another, first in normal room lighting and then immediately after the room lights were turned off to examine the effects of changing light levels. Other visual tasks including timed grocery item search and a test of recognition of faces and facial expressions 13 are not discussed here, nor are findings of examination of driving records. However, driving restriction was considered.
Also not further discussed are findings of a vision questionnaire based on the Visual Activities Questionnaire. 14 Additional questionnaire data were available from the Buck Center Health and Function Interviews 15 which had been carried out on the same population. These interviews included questions on living circumstances, medical conditions, daily activities, driving, and many other aspects of life. They also included tests to evaluate depression (Center for Epidemiologic Studies Depression Scale) and cognitive functioning (Short Portable Mental Status Questionnaire).
A number of physical tests were performed as part of the Buck data gathering on the cohort. These included hearing and balance tests, a manual dexterity task, a walking task (number of 10-ft lengths walked in 1 min), tandem and semitandem stands, and a chair stand task (number of times the subject could arise from a chair in 1 min). Blood pressure was also measured.
Which Vision Functions Decline Most with Age?
Many vision results from the first round of testing have been reported in Haegerstrom-Portnoy et al., 5 Brabyn et al., 16 and Schneck et al. 17 We summarize the findings in terms of the relative magnitudes of the deficits among the various dimensions of vision function. We determined how many times worse than young normals a particular measure (median value) is at any age, thereby giving a direct comparison of relative decline in different aspects of visual function. Fig. 1 plots the factor by which the elders’ median performance is worse than the young normal values as a function of age. The Figure clearly shows a wide range in decline among vision functions with age. For example, critical flicker fusion frequency changed very little with age; even the oldest age group is less than a factor of two worse than the values for young normal observers (42 vs. 50 Hz). The eldest individuals’ median high contrast acuity was 20/50, only 2.4 times worse than young normals. On the other hand, vision in the presence of glare, glare recovery time, and attentional visual field size all showed large declines across age; the eldest group was 12 to 18 times worse than normals on these tasks.
Vision and Visual Task Performance
Fig. 2 plots (open squares) the reading performance (measured with the Pepper Test) of the sample across age. Reading performance (corrected reading rate) declined with age. Compared with the 55- to 64-year-old age group, reading speed in the ≥85 years group averaged a 50% decline. This is not surprising given the large declines seen in many aspects of vision performance. Even visual acuity declines to about 40% of the value in youth. The solid squares in Fig. 2 show the decline in reading performance in the subsample of individuals who retain good visual acuity (20/30 or better). Even among this group, there is substantial decline in reading performance with age, particularly among the very old.
There are many implications to this finding. First, it reminds us that an individual at age 85 with visual acuity of 20/30 is unlike an individual of age 60 with the same acuity, in terms of both vision under nonoptimal conditions and performance on visual tasks. Second, it points out that many other visual functions are critical to the reading task.
Indeed, multivariate analysis showed that among older subjects (≥75 years) with good acuity, vision function, cognitive factors, and motor factors all contribute to reading performance, accounting for 52% of the variance. Among the individual vision tests, SKILL Dark performance was most highly correlated with corrected reading speed (r = −0.49). Size of the attentional visual field, which has a cognitive component, is also closely related to reading performance, whereas standard visual field size is not. 18 These associations are illustrated in Fig. 3. Standard field size is similar among younger and older observers with good or poor reading performance. The individuals who maintain very good reading performance (≥100 words per minute) into old age (≥75 years) are those who maintain their attentional visual fields. In contrast, attentional field size is significantly reduced in subjects from the older group with poor reading performance (compared with good and poor readers aged ≤75 years or older good readers).
This study also concerns itself with the association between vision (and other health variables) and driving among the elderly. Driving records showed that this aged sample had very few driving incidents (accidents or violations) during the risk analysis period. This is attributable to a number of factors, not least of which is the tendency of elders to restrict their driving to relatively low-risk situations 19,20 (e.g., daylight, nonfreeway, and good weather in familiar places) and to impose self-restriction in the presence of vision problems, known eye conditions, and prior incident history. 21,22 We explored the relation between performance on each of our large battery of vision tests and driving self-restriction. 23 Taking age and sex into account, failing tests of high-contrast visual acuity, low-contrast acuity in glare, stereopsis, attentional field size, and color vision were all associated with self-restriction. 23
Some Associations between Vision and Physical Function
A complete discussion of our findings regarding vision and physical function are forthcoming (West CG, et al., submitted manuscript). Three aspects of physical performance (timed walking, chair stand, and tandem stand) previously assessed in older adults 24 are considered as well as self-reported mobility limitations. Failure on the walking test was nine or fewer 10-ft lengths completed in 1 min. Failure on the chair stand was inability to rise from a straight-back chair at least four times in 1 min with arms crossed over the chest. Failure on the tandem stand was inability to maintain a full tandem stand (heel of one foot to toe of the other) for at least 10 s. For self-reported mobility, subjects were asked two questions: “Are you able to walk up and down stairs without help?” and “Are you able to walk a half-mile without help?” Subjects answering “no” to both questions were classified as having mobility limitations. To explore the relationships between vision and physical function, odds ratio analyses were used. Table 1 summarizes the results, giving the odds ratios for associations that were statistically significant (p < 0.05). The odds ratio method of analysis, which is common in epidemiological studies, compares the likelihood of a particular outcome (e.g., mobility limitation) across levels on some independent measure (e.g., visual acuity). More specifically, the Table shows the (significant) increase in probability of failing a physical performance measure or having a mobility limitation for a unit change in each vision measure. Units were 0.1 log unit for all measures except stereopsis (0.3 log unit), field diameters (10%), and impact of light on walking (1 s). It is worth noting here that the odds ratio values are dependent on the size of the unit, and that the apparently small values of the significant ORs presented here reflect our choice of rather fine units, such as a line (0.1 log unit) of acuity loss.
Both timed walking and chair stand were significantly associated with all measures of spatial vision and with both standard and attentional fields. Chair stand ability was also related to stereopsis and the impact of (sudden reductions in level of) light on walking, whereas walking performance was associated with both critical flicker fusion frequency and sensitivity to low-frequency flicker. Tandem stand performance showed fewer associations, limited to contrast sensitivity, stereopsis, and attentional field diameter. Only mobility limitation was found to be associated with retinal photostress recovery, and it was also associated with distance high- and low-contrast acuity, acuity in glare, contrast sensitivity, fields, and the impact of light on walking.
A Brief Mention of Longitudinal Findings
Longitudinal studies of vision in the aged are relatively rare because early studies of this type measured only visual acuity, which is relatively insensitive to age- and disease-related vision changes. Thus, the changes seen were small. For example, the Beaver Dam Study 25 found very little decline (averaging only 3.5 letters per decade) in standard high-contrast acuity in the group tested, and we found a similarly small decline in our group. However, we find, as we did in our cross-sectional comparisons, that virtually all other measures of vision (except flicker) declined much more dramatically and that there was considerable variation among measures with respect to the rate of longitudinal decline. For most measures, there was good agreement between the rate of decline observed across age groups and longitudinally, however this was not uniformly so. For example, the longitudinal decline in SKILL Dark chart acuity was much greater than expected based on cross-sectional findings, and, particularly among the very old, the decline in acuity in glare was somewhat less.
Our cross-sectional data 5 indicated that the rate of decline in all aspects of vision function increases with age and does so in such a way that a single exponential function fits all spatial vision measures equally well. But the slope at any given age differed among functions so that the common “aging function” lies at different locations along the age axis for different functions. For example, high- and low-contrast acuity “aged” later than, for example, acuity in glare or SKILL Dark chart acuity, which were displaced toward younger ages. Fig. 4 illustrates a similar pattern among spatial vision measures in the longitudinal data. Fig. 4 plots the rate of decline (log units per decade age change) for five measures of spatial vision for two age groups (<75 years, N = 347; ≥75 years, N = 249). The Figure illustrates that, in both age groups, high- and low-contrast acuity declined less rapidly than other measures, with SKILL Dark chart acuity showing the largest decline. Furthermore, the rate of decline was significantly larger in the older age group than the younger for acuity in glare (p < 0.05), contrast sensitivity (p < 0.01), and SKILL dark chart acuity (p < 0.00001), but not for high- and low-contrast acuity.
A goal of any study of aging is to be able to identify those individuals who are likely to suffer unusually drastic losses in the future (and to identify factors contributing to the decline). We found indeed that some measures do show promise for improving prediction of future severe vision loss. For example, median high-contrast acuity loss for the sample over the period between testing (4 years) was about two letters. Individuals whose SKILL Dark Chart acuity at the first test was 20/100 or poorer (38% of the group of 596 people who were re-tested) were 13 times more likely to lose nearly three lines of acuity (15 letters) over the 4-year inter-test interval than those with better SKILL dark acuity.
Independence vs. Redundancy of Vision Measures
A major theme of the present study is to establish a practical test protocol for vision in the elderly. We considered the following as necessary parts of that practicality: simplicity of tests for testers and subjects; low cost; each test is rapid; relating meaningfully to daily function; and predictive of future status. Efficiency of the test battery is also crucial, because even if each test takes just a few minutes, a large battery may take considerable time and tax the subjects/testers. Therefore, it is important to avoid including unnecessary tests that provide little or no additional information. The large range of visual function among the sample on each of the measures used here provides us with an opportunity to evaluate the “redundancy” of each measure with visual acuity.
The argument has been made in the past that tests that are highly correlated with one another, or with visual acuity, ought not to be included in the same battery because the performance on one is well-predicted by performance on the other. We examined the strength of prediction of performance on other spatial vision measures that are highly correlated with acuity. 26 Our results showed that numerically high correlation coefficients between the different vision measures masked wide individual differences and that an individual’s performance on other measures cannot be well predicted from acuity. For example, contrast sensitivity (measured with the Pelli-Robson Chart) was highly correlated with high-contrast distance visual acuity (r = 0.86). Yet, the 95% confidence limits of prediction of contrast sensitivity based on high-contrast acuity spans 0.72 log units, nearly one-third of the range covered on the chart itself! Fifty four percent of the subjects fell outside the range of ± 0.1 log unit; 23% were outside a 0.4 log unit (8 letter) range (a factor of 2.5) away from the value predicted from acuity. 26 The strength of association (r, r2) between spatial vision measures and high-contrast acuity and the strength of prediction (95% confidence limits) are presented in Table 2. The results indicate that despite high correlations, spatial vision of individuals cannot be well predicted from acuity measurements alone. This highlights the importance of incorporating additional vision tests, and particularly those that more closely resemble everyday viewing conditions.
This paper summarizes findings of a study of health, function, and vision of elders living in Marin County, California, a suburb of San Francisco. This sample was better educated and more affluent than those of some other large vision studies 1–4,25 and is largely white. Marin County residents have lower mortality rates than the U.S. population in general, but “while death has been postponed the prevalence of reported diseases and disability has not.”15 There were few significant differences in measures of health and function compared with other elderly populations with significantly lower socioeconomic status. 15 Furthermore, our visual acuity findings agreed well with other large aging studies, 5 with the exception of the SEE study, which had unusually “good” acuities.
Earlier longitudinal studies in elders, such as the Beaver Dam Study 25 and Blue Mountains Study 27 measured only visual acuity and found very small changes. More recently, our and a few other studies (e.g., Rubin et al. 1) have incorporated other more sensitive tests. The present study is unique because it includes a large proportion of individuals at much older ages. A general finding of this study and the SEE study 1 was that large changes across age are observed for most vision measures other than standard acuity. Visual acuity was well maintained in most individuals into very old age. In this study, the largest changes were seen for vision in the presence of glare, time to recover vision after exposure to glare, stereopsis, and the size of the visual field in the presence of a central attentional load. 5
Thus, despite maintained acuity, many elderly people are effectively visually impaired under conditions of everyday life—changing light level, in glare, or in attentionally demanding situations such as driving. Despite the fact that many measures are highly correlated with visual acuity, it is not possible to predict, from the standard acuity measure alone, what an older individual’s vision will actually be under the more difficult conditions he or she is likely to encounter in daily life. Two individuals with the same standard acuity may have very different performance levels under conditions such as low contrast or glare. Many older individuals, particularly the very old, have greatly impaired performance on tasks such as reading, despite continued good acuity.
Testing vision under nonideal, real-world conditions would enable elders and clinicians to better identify those circumstances under which the individual needs to exercise caution as well as how to modify the environment to optimize vision (e.g., adding lights, increasing contrast, and reducing glare). In addition, use of these more subtle vision measures may help in monitoring disease progression and predicting individual risks for future serious vision loss.
The observed changes with age in acuity and contrast sensitivity, in addition to loss of stereopsis and greatly reduced visual fields under conditions of divided attention combine in individually idiosyncratic patterns with expected significant impact on everyday life. Naturally, vision is not the only important variable impacting task performance; for example, the decline in reading speed with age regardless of visual acuity indicates the importance of cognitive and motor abilities in visual tasks. Visual impairment is clearly associated with other health and functioning variables such as depression, cognitive status, chronic disease, and physical abilities. Although many of the results have not yet been analyzed, the availability of longitudinal data should clarify the nature of these and other relationships in future analyses, allowing determination of individual risk factors for deficits in vision, health, and daily functioning.
We thank Dr. C. West, Dr. G. Gildengorin, and the staff and volunteers at the Buck Center for Research in Aging.
Supported by National Eye Institute Grant EY09588 to JAB and by The Smith-Kettlewell Eye Research Institute.
Received July 14, 2001; revision received February 23, 2001.
1. Rubin GS, West SK, Munoz B, Bandeen-Roche K, Zeger S, Schein O, Fried LP. A comprehensive assessment of visual impairment in a population of older Americans: the SEE Study: Salisbury Eye Evaluation Project. Invest Ophthalmol Vis Sci 1997; 38:557–68.
2. Kahn HA, Leibowitz HM, Ganley JP, Kini MM, Colton T, Nickerson RS, Dawber TR. The Framingham Eye Study: I. Outline and major prevalence findings. Am J Epidemiol 1977; 106:17–32.
3. Ganley JP, Roberts J. Eye Conditions and Related Need for Medical Care among Persons 1–74 Years of Age, United States, 1971–72. Hyattsville, MD: U.S. National Center for Health Statistics, 1983.
4. Tielsch JM, Sommer A, Witt K, Katz J, Royall RM. Blindness and visual impairment in an American urban population: the Baltimore Eye Survey. Arch Ophthalmol 1990; 108:286–90.
5. Haegerstrom-Portnoy G, Schneck ME, Brabyn JA. Seeing into old age: vision function beyond acuity. Optom Vis Sci 1999; 76:141–58.
6. Ferris FL, Kassoff A, Bresnick GH, Bailey I. New visual acuity charts for clinical research. Am J Ophthalmol 1982; 94:91–6.
7. Bailey IL, Bullimore MA. A new test for the evaluation of disability glare. Optom Vis Sci 1991; 68:911–7.
8. Haegerstrom-Portnoy G, Brabyn J, Schneck ME, Jampolsky A. The SKILL Card. An acuity test of reduced luminance and contrast: Smith-Kettlewell Institute Low Luminance. Invest Ophthalmol Vis Sci 1997; 38:207–18.
9. Frisby JP. The Frisby stereotest: amended instructions. Br Orthopt J 1980; 37:108.
10. Pelli DG, Robson JG, Wilkins AJ. The design of a new letter chart for measuring contrast sensitivity. Clin Vis Sci 1988; 2:187–99.
11. Whittaker S, Watson G, Steciw M. Low Vision Research and Training Center Final Report, Part 2. Washington, DC: U.S. Department of Education Office of Special Education and Rehabilitative Services, National Institute of Handicapped Research, 1983: 19–38.
12. Baldasare J, Watson GR, Whittaker SG, Miller-Shaffer H. The development and evaluation of a reading test for low vision individuals with macular loss. J Vis Impairment Blind 1986; 80:785–9.
13. Ekman P, Friesen WV, Ellsworth P. Emotion in the Human Face: Guidelines for Research and an Integration of Findings. New York: Pergamon Press, 1972.
14. Sloane ME, Ball K, Owsley C, Bruni JR, Roenker DR. The visual activities questionnaire: developing an instrument for assessing problems in everyday visual tasks. In: Noninvasive Assessment of the Visual System. OSA Technical Digest Series, Vol 1. Washington, DC: Optical Society of America, 1992: 26–9.
15. Reed D, Satariano WA, Gildengorin G, McMahon K, Fleshman R, Schneider E. Health and functioning among the elderly of Marin County, California: a glimpse of the future. J Gerontol A Biol Sci Med Sci 1995; 50:M61–9.
16. Brabyn JA, Haegerstrom-Portnoy G, Schneck ME, Lott LA. Visual impairment in elderly people under everyday viewing conditions. J Vis Impairment Blind 2000; 94:741–55.
17. Schneck ME, Haegerstrom-Portnoy G, Lott LA, Brabyn JA. Ocular contributions to age-related loss in coarse stereopsis. Optom Vis Sci 2000; 77:531–6.
18. Lott LA, Schneck ME, Haegerstrom-Portnoy G, Brabyn JA, Gildengorin GL, West CG. Reading performance in older adults with good acuity. Optom Vis Sci 2001; 78:316–24.
19. Janke M. Age-related disabilities that may impair driving and their assessment: literature review (Report to National Highway Traffic Safety Administration). Sacramento, CA: California Department of Motor Vehicles, 1994.
20. Planek TW, Condon ME, Fowler RC. An investigation of problems and opinions of aged drivers. Chicago, IL: National Safety Council, 1968.
21. Owsley C, Ball K, Sloane ME, Roenker DL, Bruni JR. Visual/cognitive correlates of vehicle accidents in older drivers. Psychol Aging 1991; 6:403–15.
22. Ball K, Owsley C, Stalvey B, Roenker DL, Sloane ME, Graves M. Driving avoidance and functional impairment in older drivers. Accid Anal Prev 1998; 30:313–22.
23. West CG, Schneck ME, Oman D, Gildengorin G, Reed D. The safe older driver: sensory and medical characteristics. (Final report to the American Association of Retired People Andrus Foundation). Washington DC: AARP Andrus Foundation, 1996.
24. Salive ME, Guralnik J, Glynn RJ, Christen W, Wallace RB, Ostfeld AM. Association of visual impairment with mobility and physical function. J Am Geriatr Soc 1994; 42:287–92.
25. Klein R, Klein BE, Lee KE. Changes in visual acuity in a population: the Beaver Dam Eye Study. Ophthalmology 1996; 103:1169–78.
26. Haegerstrom-Portnoy G, Schneck ME, Lott LA, Brabyn JA. The relation between visual acuity and other spatial vision measures. Optom Vis Sci 2000; 77:653–62.
27. Attebo K, Mitchell P, Smith W. Visual acuity and the causes of visual loss in Australia: the Blue Mountains Eye Study. Ophthalmology 1996; 103:357–64.