PURPOSE: Test-retest reliability of the CSV-1000 (Vector Vision) has only been reported for one adult sample. We measured the reliability of this instrument in both children and adults and also investigated the effect of changing the examiner on test-retest reliability. METHODS: Test-retest log contrast sensitivity (CS) measurements were obtained for 19 young adults and 15 children by the same examiner. Test-retest log CS data were obtained from 21 young adults with different examiners. Reliability was calculated using the Bland-Altman limits of agreement, the coefficient of repeatability (COR), and the intraclass correlation coefficient. RESULTS: All three estimates of reliability for the CSV-1000 chart are low for both children and adults using the standard recommended testing protocol. If the test-retest log CS data are obtained from the same examiner then the reliability is improved, but not significantly so. CONCLUSIONS: The reliability of the CSV-1000 is low, even if the same examiner obtains test-retest data. The data indicate that this test is unlikely to be sensitive enough to provide useful information for the clinician as is, but we suggest modifications of the procedure that may significantly increase test reliability.
PURPOSE: Test-retest reliability of the CSV-1000 (Vector Vision) has only been reported for one adult sample. We measured the reliability of this instrument in both children and adults and also investigated the effect of changing the examiner on test-retest reliability. METHODS: Test-retest log contrast sensitivity (CS) measurements were obtained for 19 young adults and 15 children by the same examiner. Test-retest log CS data were obtained from 21 young adults with different examiners. Reliability was calculated using the Bland-Altman limits of agreement, the coefficient of repeatability (COR), and the intraclass correlation coefficient. RESULTS: All three estimates of reliability for the CSV-1000 chart are low for both children and adults using the standard recommended testing protocol. If the test-retest log CS data are obtained from the same examiner then the reliability is improved, but not significantly so. CONCLUSIONS: The reliability of the CSV-1000 is low, even if the same examiner obtains test-retest data. The data indicate that this test is unlikely to be sensitive enough to provide useful information for the clinician as is, but we suggest modifications of the procedure that may significantly increase test reliability.
Authors: Merina Thomas; Rebecca F Silverman; Filippos Vingopoulos; Megan Kasetty; Gina Yu; Esther L Kim; Amro A Omari; Katherine A Joltikov; Eun Y Choi; Leo A Kim; David N Zacks; John B Miller Journal: J Vitreoretin Dis Date: 2020-11-05