Evaluation of a Protocol for Integrated Speech Audiometry.

Purpose This project was aimed at evaluating the reliability, validity, and clinical utility of a protocol for integrated measurements of the most comfortable level (MCL) and uncomfortable level (UCL) for speech, in combination with the speech recognition threshold (SRT). We also evaluated the validity of using spondee words when measuring speech MCL and UCL. Method In a randomized block design, equal numbers of women and men with normal hearing, aged 18-29 years, were assigned to each of 3 experimental stimulus conditions: spondee singlets, spondee triplets, or connected discourse ( n = 12 per group). Following measurement of the SRT, a modified method of limits was employed to establish, on a 7-point loudness rating scale, an ascending MCL, a descending MCL, and an ascending UCL. A single instructional set covered all loudness measurements. Test times were tracked electronically to assess clinical efficiency. All test conditions were repeated during each of 2 separate test sessions. Results Mean SRTs, MCLs, and UCLs across the 3 different experimental groups were found not to differ statistically or clinically (mean differences < 5 dB). Intrasession and intersession reliability for the various measures were excellent, and testing of all listeners was completed in a timely manner. In a follow-up experiment with adults with normal hearing who were only a decade older than participants in our main experiment, the older group was found to have significantly higher MCLs and UCLs. Conclusions Spondee words can be used routinely to obtain reliable, valid, and clinically efficient measures of MCLs and UCLs for speech, in a protocol combined with the SRT. Spondees, presented singly, yielded the greatest level of efficiency overall. Results support a recommendation to obtain an ascending measurement of MCL prior to a descending measurement and to establish the MCL by averaging the 2 values.