Identifying dyslexia risk for sport-related concussion management: Sensitivity and specificity of self-report and...

Dyslexia is the most common type of learning disability and studies have shown that student-athletes with learning disabilities sustain more concussions than their non-affected peers. However, current methods of dyslexia identification in college students are potentially invalid because they rely on students to self-report formal dyslexia diagnoses.

To test the accuracy of self-report against two alternative methods of dyslexia screening, 94 college students completed three dyslexia symptom inventories, two rapid naming (RAN) tasks, and a standard word reading measure.

Reliability was acceptable for screening purposes on the inventories (α =.70, -.72), and excellent for RAN (α = .91, -.94). Specificity was acceptable (82.5%), but sensitivity was low (14.3%) when students self-reported suspected diagnoses of reading impairment. Sensitivity and specificity were higher for the digit RAN task (71 and 98%, respectively) compared to the letter RAN task (57 and 90%). Sensitivity (92.7%) and specificity (92.5%) were optimal when a cut-score of ≥27 seconds was used. A binary logistic regression showed digit RAN alone significantly predicted whether students were classified as typical or inefficient readers, p< .001, whereas the most reliable dyslexia inventory alone did not, p=.284. Including inventories along with RAN provided no additional predictive value.

Self-report inventories missed many cases of inefficient word reading. The digit RAN task classified 93.6% of the cases correctly compared to 72.3% for self-report inventory. Thus, we recommend that neuropsychologists working with college concussion management programs add to their baseline screening protocols the digit RAN task, which can be completed in less than one minute.