Abstract
Purpose :
Many reading tests use very short passages, read aloud, without a test of comprehension. We created a silent, naturalistic reading test with longer passages that simulate common daily reading tasks and features questions to test comprehension. The purpose of this study was to evaluate the reading test in people with low vision and compare the results to those from a previous study of normally-sighted subjects.
Methods :
The test consists of six passages based on scientific articles for children. Each contains 300 words and only includes the 5,000 most commonly used words in the English language. Flesch-Kincaid reading ease for the passages ranged from 69.4 to 78.9 and grade level for the passages ranged from 5.3 to 6.8. Three comprehension questions accompany each passage. Each subject read six passages and answered the questions associated with those passages. Subjects who were able read half of the passages with a +3.00D add over distance correction and half with a handheld video magnifier. Those unable to read the passages with a near addition used the video magnifier for all passages. Reading speed was calculated and the percentage of correct responses to comprehension questions was assessed.
Results :
Sixteen subjects with low vision participated, 63% of whom reported reading difficulty. Mean±SD age was 60±17 with a range of 20 to 84 years. Mean better eye logMAR BCVA was 0.63±0.15. The average passage reading speed was 140±40 words per minute, compared to 269±65 for 30 normally-sighted subjects (mean age = 37±15). Passage reading speed was significantly correlated with contrast sensitivity (p<0.001) and MNRead critical print size (p=0.022) in subjects with low vision. Eighty-five percent of subjects with low vision answered at least one comprehension question incorrectly, however most subjects missed relatively few (mean of 0.36 incorrect responses per passage), indicating generally good comprehension of the passage content.
Conclusions :
This naturalistic reading test performed well in people with low vision and may be useful in research and clinical settings. Its advantages include similarity to reading material commonly encountered in daily life and the ability to assess comprehension in addition to speed. It showed an association with the established MNRead test and contrast sensitivity, and subjects were able to comprehend the passage content as verified by the comprehension questions.
This is a 2020 ARVO Annual Meeting abstract.