Setting appropriate pass or fail cut-off criteria for tests to reflect real life listening difficulties in children with suspected auditory processing disorder

Abstract

ObjectiveThis paper explores the pass or fail cut-off criteria, the number of test fails, and the nature of tests that are most appropriate in predicting listening difficulties (LiD) in children with suspected APD (SusAPD). MethodsOne hundred and nine English-speaking children (67 males, 42 females) aged between 6 and 11 years with SusAPD were assessed. The Children's Auditory Performance Scale (CHAPS) scores 2 SD below the mean were taken as markers of LiD in different listening conditions. Binary logistic regression analyses were carried out to evaluate the cut-off criterion (2 SD or 1.5 SD or 1 SD below the mean) of failing at least two tests, from the SCAN-C and IMAP test batteries, which significantly predicted LiD. Analyses were also carried out to assess if the group of auditory processing (AP) or cognitive or combination of AP plus cognitive tests were significant in predicting LiD. Receiver Operative Characteristic (ROC) curves were also explored to evaluate how the sensitivity and specificity in confirming LiD varied with the number of test fails. ResultsFiltered Words, Competing Words, Competing Sentences, VCV in ICRA noise, Digit Span, Sight Word Reading and the Cued Auditory Attention tests correlated with one or more of the CHAPS domains. Failing at least two of these tests 1.5 SD below the mean significantly predicted (p<.05) CHAPS Ideal scores 2 SD below the mean, and failing at least two of the tests 1 SD below the mean significantly predicted (p<.05) CHAPS Memory and CHAPS Attention scores 2 SD below the mean. The combination of AP plus cognitive tests had significantly higher ability to predict CHAPS Ideal, Memory and Attention scores, compared to the group of AP or cognitive tests separately. ROC curves showed that failing at least two of the tests was associated with the best sensitivity and specificity in predicting LiD. ConclusionOf the different CHAPS domains only the CHAPS Ideal, Memory and Attention correlated with the APD tests. Failing at least two APD tests from a combination of AP and cognitive tests 1 SD and 1.5 SD below the mean, but not 2 SD, is more appropriate in confirming LiD.