Determining the Consistency of Scores Across Two Measurements of the Visual System: Test-retest Reliability
Determining the Consistency of Scores Across Two Measurements of the Visual System: Test-retest Reliability
Binocular vision tests are guided by the principle that uniform motion and focus of the eyes may be trained, but exercising the simultaneous movement of the pupils is necessary and important in achieving optimal and lasting results (Horwood & Toor, 2014). These tests have been in use for over 70 years and represent the foundation of therapy for visual deficiencies such as the inability of the eyes to move together. Despite their frequent use, no one has ever formally evaluated how good these tests are.
One such evaluation method is reliability. A reliable test is one that will give the same result when repeated on multiple occasions (i.e. test-retest reliability) or by different people (inter-rater reliability). An unreliable test gives widely varying results when repeated, which would make changes in a patient's test results difficult to interpret. Therefore, this study aims to determine the test-retest reliability of seven binocular vision tests across two independent measurements. The investigators will measure 20 healthy people ages 18 to 35 years for each of these tests on two separate occasions, one week apart. Since these tests have been in use for many years, we are expecting the two measurements to be within acceptable range of each other.
The tests being investigated measure: 3D vision (i.e. gross stereoscopic acuity), uniform movement of the eyes (i.e. "motor punctum proximum"), ability of the eyes to adapt to a target moving towards and/or away from the eyes (i.e. binocular fusion with convergence and divergence), quick movements of the eyes (i.e. saccadic movements), natural alignment of the eyes (i.e. anatomic oculomotor deviation), and double vision as a target is moved towards the eyes (i.e. convergence fusional proximum).
The results of each test will be analyzed individually. If the tests are perfectly reliable, we would expect the results of the first test to equal the results of the second test for each participant. To examine the test-retest reliability the difference between the first and second test for each individual and across all individuals' scores summed together will be assessed.
Binocular vision tests are guided by the principle that uniform motion and focus of the eyes may be trained, but exercising the simultaneous movement of the pupils is necessary and important in achieving optimal and lasting results (Horwood & Toor, 2014). These tests have been in use for over 70 years and represent the foundation of treatment for convergence insufficiency and other visual dysfunctions. Despite their frequent use, these tests have yet to be formally evaluated for their validity or reliability in healthy adults.
One such method of evaluation is reliability. A reliable test is one that will provide the same results when repeated on multiple occasions (i.e. test-retest reliability), by different people (i.e. inter-rater reliability), or the same person (i.e. intra-rater reliability). An unreliable test yields widely varying results when repeated, which would make changes in a patient's test results difficult to interpret. In this study, test-retest reliability will be estimated using a single clinician (equivalent to intra-rater reliability) as it examines the degree to which a test is able to produce consistent results or whether it is dependent on the situation or the state of the individual being tested (Rousson, Gasser, & Seifert, 2002). In future studies, inter-rater reliability will be examined which is largely dependent on and concerned with adequate training of raters as opposed to the properties of the test itself.
Furthermore, for health measures - such as the binocular vision tests - to be useful in clinically guiding patient management, they must be valid and sensitive to the system of interest (i.e. visual system). We are currently unable to quantify the validity of these binocular vision tests as this domain lacks a "gold-standard" (Hachana et al., 2013), thus, test-retest reliability must be determined first (Broglio et al., 2007). Therefore, this study aims to determine the test-retest reliability of seven binocular vision tests across two independent measurements. Twenty healthy participants aged 18 to 35 years will be measured on two separate occasions, one week apart. Since these tests have been in use for many decades, it is hypothesized that the two measurements will be within acceptable range of each other.
The binocular vision tests undergoing examination differ from optometry tests in that they use more advanced equipment which can measure very small deviations in several domains of the visual system. These seven binocular vision tests measure various elements of the visual system and will be described in detail below:
Gross stereoscopic acuity: (range 0-15 arc seconds) Our binocular vision allows us to see in three dimensions (3D), or more simply, to see depth. In this test, seated participants wearing 3D glasses are shown images. Inability to see depth or 3D will cause images to appear as points instead of objects. The objects are presented in different stages, with each stage requiring them to discriminate different levels of depth perception. The test is scored in optical units, with a range of 0 to 15 arc-seconds. The maximum score corresponds to the level where the last object was identified.
Near point of convergence and near point of convergence - break: (cm) When an object is moving towards our eyes, they symmetrically converge in order to maintain focus. However, there is a point at which our eyes no longer symmetrically converge (point of convergence). This test measures the distance (cm) between the bridge of the nose and point of convergence in seated participants as an object is moved closer to the head.
Positive fusional vergence: (diopters, prism convergence units) This test measures how well someone can adapt to challenges in focusing light on the retina. There are two almost identical tests. One test occurs with an object placed at 3m from the seated participant, and the other with an object at 30cm from the seated participant. Light from an image is passed through a prism. This is analogous to moving the image further away from the body. In response, the eyes must diverge (separate) to focus on the object, just as they would if the image actually moved away from the body. Different prisms are used to create increasing challenges. The score for these tests is simply the maximum amount of prism convergence (dioptres, noted on the prism, as one would note diopters on eye glasses) that the seated participant can accommodate at 3m and at 30cm.
Negative fusional vergence: (diopters, prism convergence units) This is the same test as (3), except that the prisms diverge the light and the participant has to converge their eyes to maintain focus. The score for these tests is simply the maximum amount of prism divergence (diopters, noted on the prism, as one would note diopters on glasses) that the seated participant can accommodate at 3m, and at 30cm.
Saccadic movements or oculomotor capacity: (Score = bad, medium, good) A light appears on the screen and the participant move their eyes to fix on the object. While they eyes adjust, they will temporarily cover small distances until they achieve a fixed focus. These are called saccadic movements. Lights appear and disappear, in different locations on the screen, at a rate of 100 per minute, lasting 2 minutes. The test result is scored by the evaluator based on a global impression over the entire 2 minutes, with 3 separate sub scores on an ordinal scale for quality (bad, medium, good), for synchronization (bad, medium, good) and saccadic correction (many corrections, few corrections, no corrections). The three sub scores are combined into an overall score according to our industry partner's (Apexk) proprietary algorithm.
Anatomic oculomotor deviation: (diopters, prism convergence units) This test measures the eyes' natural deviation (heterophoria) and also allows the detection of strabismus. In strabismus, anatomic deviation is evident and the person's dominant eye is looking at you, but the "lazy/deviated" eye is not. In heterophoria anatomic deviation is not visible to the naked eye and the deviation has to be triggered by covering in sequence, one eye at a time, to trigger the deviation. There are two identical tests: one occurs with an object placed at 3m from the seated participant (far vision), and the other with an object at 30cm from the seated participant (near vision). In this test, seated participants focus on an object. These movements can be seen by the clinician. The clinician covers/uncovers eyes to trigger movements and uses a prism to cancel the movement. The prism that achieves this cancellation is the measure of anatomic deviation. The rating of the prism that achieves this cancellation is considered the score for this test, with one score for the object placed at 3m and another score for the object placed at 30cm. Participants with strabismus are excluded in our study because strabismus is a contra-indication to post-concussion visual training, which is part of our larger study and thus, patients with strabismus do not represent our target population.
Convergence fusional proximum: (diopters, prism convergence units) This test is similar to (2) above. When an object is moving closer to our head, the eyes symmetrically converge. When the object is moved beyond the participant's ability to converge, the participant will start to see two images (double vision). This test measures the distance between the bridge of the nose and point where double vision (cm) occurs in seated participants as an object is moved closer.
In addition, demographic information relevant to our study will be collected in order to appropriately describe the population and explore if these factors modify the test-retest correlations. Variables such as age, sex, highest level of education achieved (i.e. secondary school, CEGEP, university), the use of corrective lenses for vision problems, occupation, and any relevant past medical history (i.e. migraines, vision problems, medication, etc) will be included.
The main objective of this study is to evaluate the agreement and consistency between measurements of the visual system taken at two different times to determine the test-retest reliability of the seven binocular vision tests. To do so, the results of each test will be analyzed individually. If the tests are perfectly reliable, it is expected that the value of the first test is equal to the value of the second test for each participant across the range of values for all individuals. Reliability will be assessed using the intra-class coefficient (ICC) which measures the between-group variance divided by the total variance (sum of between- and within-group variance). The difference between the first and second test for each participant is called the within-group (each participant represents a group) variance. The difference across individuals is the between-group variance. If the within-group variance is 0 (first and second measures are identical), then the ICC = 1 (between-group variance / (between-group variance + 0)). In addition to the ICC, limits of agreement will be estimated using the paired data and illustrated via Bland-Altman plot (Bland & Altman, 1986).
As the binocular vision tests we are examining have applicability in many health conditions (e.g. concussion, Alzheimer's, etc.), this study has the potential to contribute to the improved management of symptoms related to the visual system.
Inclusion Criteria:
Exclusion Criteria: