EXECUTIVE SUMMARY

Disability insurance has been a part of the Social Security insurance system in the United States since 1956. The Social Security Administration (SSA) has an obligation to establish criteria for eligibility that will ensure that people who are truly disabled are provided benefits. SSA recently initiated several studies of the processes and criteria that are used to determine disability benefit eligibility. The SSA asked the National Research Council to study its methods of determining disability for people with visual impairments, to recommend changes that could be made now to improve the process and outcomes, and to identify research needed to develop improved visual disability determination methods over the long term.

Two major concerns motivated this study. SSA's first concern is about the reliability of the tests as they are now usedthe tests used for many years to determine visual disability, as now performed in optometrists' and ophthalmologists' offices, may not be as reliable as they could be. The second concern is about the predictive validity of current testsat present, SSA measures the impairment of visual functions, but the relationships between such impairments and disability in the performance of real-world visual tasks in the workplace have not been clearly established. Deficiencies in these vital areas would diminish the fairness and credibility of the disability determination process.

TESTS OF VISUAL FUNCTIONS

Measurements of visual acuity and visual fields are fundamental and should continue to be the primary tests for disability determination. The reliability and value of both tests would be greatly improved by the adoption of standardized testing procedures using modern instruments.

The measurement of contrast sensitivity can detect aspects of visual impairment that are not expressed in measurements of acuity or visual fields. In certain circumstances, this measure adds important information to the assessment of claimants who do not show severe impairments of visual acuity or visual fields.

Impairments of other aspects of visual functiondisorders of binocularity, glare sensitivity or recovery, color vision, visual searchare on their own not generally disabling, the literature provides little evidence that they are major contributors to disability, and these functions do not warrant primary assessment. However, impairments in these aspects of vision can contribute to disability under some circumstances.

The current procedure for computing "visual efficiency" does not permit adequate characterization of the visual performance of persons with severe low vision. The current procedure also makes distinctions, no longer necessary, between the performance of aphakic (lacking natural lenses) and phakic eyes.

The relationships between visual functions and real-world functional capacity do not suggest a natural cutoff point for disability. There are no sharp inflections in performance scores or self-reports of performance abilities corresponding to specific acuity, field, or contrast sensitivity scores. The setting of criterion scores for disability is a policy decision to be made by SSA.

VISUAL TASK PERFORMANCE

The committee examined whether disability due to visual impairments could be measured directly, rather than by estimating it from tests of visual function. The committee explored tasks in four domains: reading, mobility, social participation, and tool use, having concluded that these adequately capture important visual requirements of everyday life and jobs. Acceptable tests of performance are not yet available in most of the domains examined. Tests of reading are the closest to being ready for use.

• Reading tests are available that would be usable for disability determination after modest additional research and development (mainly standardization and norming). Tests of reading ability could provide important information for the assessment of functional capacity in the "vocational factors" steps of the disability determination process.
• For safe and efficient orientation and mobility, the most important aspects of visual function are contrast sensitivity, visual fields, and acuity. Our recommended tests of visual function assess these.
• For driving, at present there are no standard tests of driving ability available for determining driving fitness in those who are visually impaired.
• For tool use and manipulation, the data are insufficient for us to recommend any battery of performance-based tests that would determine visual disability in this domain.
• Social participation should not be a high priority for testing at this time, but it may be worthy of reconsideration in the future, as candidate tests emerge.

OTHER MEANS OF ASSESSING DISABILITY

The committee evaluated the usefulness of two other means of assessing disability: job analysis databases, which include information on the importance of vision to job tasks or skills, and measures of health-related quality of life, which take a person-centered approach to assessing visual functioning.

• Data from the Position Analysis Questionnaire were helpful in determining the importance of some visual functions in the workplace, but the proprietary nature of the dataset permitted only limited inferences about today's employed population.
• Health-related quality of life instruments could be useful in efforts to identify important everyday and work tasks affected by visual impairment and possibly to quantify relationships between visual functions and task performance. However, such tests would not be useful in the disability determination process for individual claimants.

High-quality data on visual function and employment for a random sample of the working-age population could help efforts to identify the relationship between visual function and employment outcomes. Because serious vision limitations are relatively rare in the working-age population, no current nationally representative datasets contain adequate samples of people with significant vision limitations. Without information on a random sample of working-age people with serious visual limitations that includes information on the social environment in which they work, it is not possible to establish the relationship between visual function and employment.

TESTING OF INFANTS AND CHILDREN

Most school-age children (i.e., at least 6 years of age), can be tested with standard adult tests of visual acuity and contrast sensitivity, as well as with short versions of adult procedures for testing visual fields. Although vision improves slightly between the early school years and adulthood, adult standards for disability determination are appropriate to apply to children whose visual acuity, visual fields, and contrast sensitivity can be tested with adult methods. A child whose vision subsequently improves beyond the level of disability status will be identified in the review that is required by law to be conducted by SSA at least every three years for beneficiaries under age 18 years.

Methods for assessing visual acuity in infants and children who cannot be assessed with adult tests have been developed, validated, and normed. These methods can be used for disability determination. There are no standardized, widely available methods for assessment of visual fields in children who cannot perform adult perimetry procedures. For these children, disability determination must, of necessity, be based on the clinician's judgment about the child's peripheral vision based on clinical tools. The use of contrast sensitivity information in determining disability in children who cannot be assessed with adult techniques is not appropriate, due to the absence of standardized, normed tests for young children and to the absence of data indicating the effects of poor contrast sensitivity on activities of everyday life in this population.

Sound statistical principles have been used in establishing SSA's recently published guidelines for disability determination in children, which set criteria of two standard deviations below the same-age norm performance for "marked" impairment and three standard deviations below it for "severe" impairment of function.

RECOMMENDATIONS

Tests of Visual Functions

For testing visual acuity, our recommendations are similar to those of the Committee on Vision in its 1980 and 1994 reports (National Research Council, 1980; 1994). We recommend that visual acuity charts should contain the same number of optotypes in each row; the space between optotypes in a row should be at least as wide as the optotypes in that row; and the size of the optotypes should decrease in 0.1 log unit steps from row to row. Chart luminance should be at least 80 cd/m², with 160 cd/m² optimal, free from glare, with a level of contrast between optotypes and background that is above 80 percent. The person being tested should be encouraged to read as many optotypes on the chart as possible and to guess at an optotype if he or she is unsure. Acuity results should be scored on an optotype-by-optotype basis, since this scoring procedure produces lower test-retest variability than does row-by-row scoring.

For disability determination, visual acuity should be tested under binocular conditions, since this provides the most representative measure of an individual's everyday vision. However, if acuity must be tested monocularly rather than binocularly, the acuity of the better eye should be used for disability determination.

Given the history and legislation behind the current SSA standard of "20/200 or worse distance acuity" as the principal criterion for visual disability, the committee recommends continuation of the 20/200 cutoff criterion. Since we recommend a visual acuity chart design that would include optotypes at the 20/160 level, applying the "20/200 or worse" criterion literally to scores obtained with such a chart would set the effective criterion to "worse than 20/160 distance acuity." The scoring of the charts currently used in disability determination sets the effective criterion at "worse than 20/100." The recommended charts have a 20/100 line that would allow SSA to maintain the effective criterion at its current position, but SSA must make the decision on whether this should be done.

For testing visual fields, the committee recommends that the current SSA standard should be revised so that disability determinations are based on the results of automated static projection perimetry rather than Goldmann (kinetic, nonautomated) visual fields.

We propose the following criteria for any perimeter to be used by SSA for disability determination:

• The automated static perimeter should be capable of performing threshold testing using a white size III Goldmann target and a 31.5 apostilb (10 cd/m²) white background.
• The perimeter should be capable of measuring sensitivity for the central 30° radius of the visual field with equal numbers of target locations in each quadrant of the field, and target locations no more than 6° apart.
• The perimeter should be a projection perimeter or should produce measures that are equal to those obtained on a projection perimeter.
• The perimeter should have an internal normative database for automatically comparing an individual's performance with that of the general population.
• The perimeter should have a statistical analysis package that is able to calculate visual field indices, particularly mean deviation or mean defect (MD), which is the average deviation of visual field sensitivity in comparison to normal values for the central 30° radius of the visual field.
• The perimeter should demonstrate high sensitivity (ability to correctly detect visual field loss) and specificity (ability to correctly identify normal visual fields).
• The perimeter should demonstrate good test-retest reliability.
• The perimeter should have undergone clinical validation studies by three or more independent laboratories with results published in peer-reviewed ophthalmic journals.

We recommend that SSA use the MD (mean deviation or mean defect) score to characterize impairment.

The committee recommends that contrast sensitivity be assessed as a supplementary basis for disability determination for claimants with visual acuity between 20/50 and 20/200 and other evidence or self-report of visual impairment greater than would be expected from the acuity score. The following criteria should be met:

The test used should

• be simple to administer;
• require no sophisticated equipment;
• be well-standardized, reliable, valid, and sensitive to visual loss;
• be relatively insensitive to changes in focus, viewing distance, and illumination;
• provide a single score that can be compared with normative data.

Currently, one available testthe Pelli-Robson Contrast Sensitivity Testis known to meet these criteria.

SSA should use log contrast sensitivity score in determining disability.

Severe impairments in other visual functionsdisorders of binocularity, glare sensitivity or recovery, color vision, visual searchcould be taken into account as "adjustments" in the disability determination process.

In the matter of combining scores on multiple visual impairments to arrive at an aggregate impairment score and setting cutoff scores for disability determination, the committee recommends that:

• SSA no longer calculate central visual efficiency and visual field efficiency. Because the recommended indices of visual acuity, visual fields, and contrast sensitivity use logarithmic scales, appropriately weighted addition of scores will provide a simple, direct aggregate measure of impairment.
• Research be undertaken to examine directly how different kinds of impairments interact in determining overall visual performance. In the absence of such research, SSA should continue to treat impairments of different aspects of visual function as though they operate independently in determining overall visual performance.
• SSA support research to inform policy decisions about the levels of impairment of visual functions that should qualify as disability.

The committee also makes the following recommendations for all testing of visual functions.

• All tests should be administered under standardized conditions using modern instruments and with the claimant wearing the best tolerable correction.
• Tests such as acuity should be administered and scored binocularly unless circumstances dictate that this is not appropriate (e.g., in diplopia, in which binocular performance would be worse than that of the better eye). If binocular scores are not obtainable, the better eye score should be used in disability determination.
• To improve the reliability and accuracy of testing, SSA should develop and promote, through both regulations and its professional relations function, clear guidelines or criteria for visual function test materials and procedures (charts, perimetry equipment, test protocols, etc.). SSA should consider establishing a test quality assurance advisory panel with both scientific and clinical expertise in vision testing to evaluate new tests and approve those that meet SSA's needs.
• SSA should consider developing standards for test administration, in consultation with the ophthalmological and optometric communities, and explore ways to ensure that such standards are met by the professionals who test SSA claimants. This could greatly improve the reliability of testing. Possibilities range from initiating an accreditation or certification system for providers and their test facilities to establishing dedicated test centers that would operate under SSA supervision.

The committee recommends the following additional research efforts related to visual function testing:

• SSA should support research relating the outcome of visual assessment with such tools as visual acuity charts to an individual's ability to function in the workplace and in society. This will allow future evaluation of the adequacy of the present cutoff criterion of 20/200.
• SSA should support research on the visual functions for which testing is not now recommended, to explore how they contribute to disability and how they can best be measured. This would provide scientific support for any future decision to include such measures in determination of visual disability.

Visual Task Performance

The committee strongly recommends that SSA invest in the development of tests of visual task performance for future use in disability determination.

• Appropriate tests exist to measure reading disability. Reading speed, reading acuity, and critical print size all are potentially relevant to the evaluation of disability. SSA should support research to develop a normative database for promising reading tests.
• As soon as tests are available that meet SSA needs, a test of reading vision should be included as a component in the assessment of individuals with vision impairment who receive vocational assessment when they fail to meet the medical listings.
• There is a need for tests of driving ability to determine fitness in those who are visually impaired. SSA should support research and development of driving tests for use in vocational assessment.
• Standard, performance-based tests of tool use developed for use in personnel selection and placement should be studied for possible utility in helping to determine disability due to vision.

Testing of Infants and Children

• Use same-age norms, not age equivalents, in evaluating visual acuity performance in children too young to be assessed with adult visual acuity tests.
• Until standardized tests are available for younger children, do not test fields by perimetry until children are old enough to be assessed with static perimetrybetween 6 and 8 years of age.
• Until standardized tests are available for younger children, do not test contrast sensitivity until children are old enough to be assessed with adult tests of contrast sensitivitybetween 6 and 8 years of age.