Clinical applicability of adaptive speech testing : a comparison of the administration time, accuracy, efficiency and reliability of adaptive speech tests with conventional speech audiometry.

Abstract

Adaptive procedures are a common method of investigating sensory abilities in research settings; however, their use in clinical settings is more limited. Little research has been done investigating the implementation of adaptive procedures into Audiological speech tests, and to date, no studies have compared and evaluated adaptive speech tests with current clinical speech audiometry. This study investigated the advantages of using both closed-set and openset adaptive speech tests in the clinical Audiology setting, with respect to administration time, accuracy, efficiency and reliability. Preliminary testing of the two major adaptive procedures (staircase and maximum-likelihood procedures) was conducted using a selection of different parameters chosen on the basis of the results of previous research (Kaernbach, 1991; García- Pérez, 1998) to determine the optimal procedures and parameters for use in clinical speech tests. Focus was given to the staircase procedures, with comparisons made between tests using variations in step size – constant step sizes and larger step sizes at the beginning – and different termination criteria. It was found that both adaptive closed-set staircase tests (with both step size variations) performed similarly, whereas the adaptive open-set staircase test with larger step sizes at the beginning showed advantages over the equivalent constant step size test in terms of administration time, accuracy and efficiency. The maximum-likelihood QUEST procedure showed advantages over the staircase procedures in terms of administration time; however, the reliability of both this test and conventional speech audiometry was poor, indicating that these tests are not the most suitable tests for a clinical setting. Subsequent clinical testing of the optimal adaptive speech tests using participants with varying degrees of hearing loss found that administration time was similar between conventional speech audiometry and the adaptive closed-set staircase tests when the optimal termination criteria identified in the Preliminary Testing Phase were employed. The adaptive open-set staircase test with larger step sizes at the beginning showed the best accuracy of any of the tests when using the pure-tone average as a reference, while the efficiency of all the adaptive staircase tests was similar. Overall, the results highlight some of the potential advantages of adaptive speech testing in the clinical Audiology setting; however, further studies are required to determine the specific parameters that achieve the best results.