Guidelines of the French Society of Otorhinolaryngology-Head and Neck Surgery (SFORL) and the French Society of Audiology (SFA) for Speech-in-Noise Testing in Adults
European Annals of Otorhinolaryngology, Head and Neck Diseases
Joly, C.-A., Reynard, P., et al. (2022).
European Annals of Otorhinolaryngology, Head and Neck Diseases, 139(1), 21-27.
This is a clinical practice guideline providing recommendations for speech-in-noise testing in adults.
French Society of Otorhinolaryngology-Head and Neck Surgery; French Society of Audiology
Recommendations for the assessment of gain with hearing rehabilitation are as follows:
<ul>
<li>"Prosthetic gain is assessed exclusively in free field" (Expert opinion; p. 26).</li>
<li>"For quick estimation, adaptive tests to determine [the 50% speech reception in noise threshold] SRT-50n are to be preferred" (Expert opinion; p. 26).</li>
<li>"Procedures using several fixed [speech-to-noise ratios] SNRs give a fuller picture of performance in noise" (Expert opinion; p. 26).</li>
<li>"If the patient is discouraged by the difficulty of the test and to avoid the floor and ceiling effects inherent to fixed SNR tests, the speech signals should be delivered at a favorable SNR, so as to assess, for example, SRT-70n (SNR allowing 70% intelligibility)" (Grade B; p. 6).</li>
<li>"As learning improves SRT-50n over successive trials, gain will be overestimated if testing begins under the least favorable conditions (bare ear); if, on the other hand, SRT-50n is measured first with the hearing aid, such overestimation can be avoided" (Expert opinion; p. 26).</li>
</ul>
Recommendations for the initial diagnosis of hearing loss in adults are as follows:
<ul>
<li>"When used for initial diagnosis, audiometry in noise should be part of a complete hearing work-up, with pure-tone audiometry, speech audiometry in silence and objective measurements" (Grade A; p. 26).</li>
<li>"Adaptive procedures to determine [the 50% speech reception in noise threshold] SRT-50n are to be preferred for initial diagnosis if a rapid standardized assessment of intelligibility in noise is sought for all subjects" (Grade A; p. 26).</li>
<li>"In case of discordance between pure-tone and speech audiometry and in the absence of retro-cochlear abnormality on [magnetic resonance imaging] MRI, a specific disorder should be screened for: auditory processing disorder, auditory neuropathy spectrum, hidden hearing loss" (Grade A; p. 26).</li>
</ul>
Recommendations for speech audiometry in noise tests in adults are as follows:
<ul>
<li>"To take account of the signal processing algorithms in present-day devices, noise should be delivered via at least 4 loudspeakers ahead of the signal and continuously" (Expert opinion; p. 26).</li>
<li>"Assessing the efficacy of directional microphones, [contralateral routing of signals] CROS devices or bimodal adaptation requires at least 5 loudspeakers, to reproduce real life. The assessment method should allow the speech signal azimuth to be varied so as to test different spatial configurations" (Expert opinion; p. 26).</li>
<li>"The main measurements of binaural gain (demasking, head shadow, summation) are made in free field with 2 loudspeakers positioned on a 1-[meter] radius semicircle" (Grade A; p. 27).</li>
<li>"Before measuring binaural gain with signal and noise sources separated, a test familiarization step should be performed with both sources located frontally" (Grade A; p. 27).</li>
</ul>
Recommendations for choosing masking noise for speech-in-noise testing in adults are as follows:
<ul>
<li>"Stationary noise should be used if identical noise is required for test-retest and is useful for comparison between groups and conditions" (Expert opinion; p. 24).</li>
<li>"Cocktail-party noise is close to real life, enabling ecologic assessment useful for assessing hearing difficulty or loss" (Expert opinion: p. 24)</li>
</ul>
"Only tests with normal values defined in [speech-to-noise ratio in decibels] SNR dB can be used for indicating auditory rehabilitation [e.g., hearing aids]" (Expert opinion; p. 26).
"To ensure validity, speech audiometry in noise should begin with at least 2 blank trials, allowing the subject to become familiar with the test procedure" (Grade A; p. 25).
"Air conduction with transducers [inserts or headphones] is recommended for screening, initial diagnosis, indicating rehabilitation and whenever separate assessment of the two bare ears is required" (Expert Opinion; p. 26).
Recommendations for choosing speech material for speech-in-noise testing in adults are as follows:
<ul>
<li>"To measure intelligibility in a situation optimally close to real life, tests using sentences are to be preferred" (Grade A; p. 24).</li>
<li>"Pseudo-words, logatomes or numbers can be used to circumvent language-level issues and mental effects, especially in case of limitations such as imperfect mastery of the language, cognitive disorder, etc." (Expert Opinion; p. 4).</li>
</ul>
Recommendations for free field delivery for speech-in-noise testing in adults are as follows:
<ul>
<li>"Only free-field delivery allows assessment or comparison of gain with hearing aids, the contribution of binaurality or perception in noise assessment close to real-life situations" (Expert opinion; p. 26).</li>
<li>"Use of 2 loudspeakers enables signal and noise to be delivered separately and frontally or assessment of the contribution of binaurality, reproducing physiological interaural differences in intensity and time" (Grade A; p. 26).</li>
<li>"Using at least 5 loudspeakers (4 for noise and 1 for signal) is recommended to avoid acoustic interference" (Grade C; p. 26).</li>
</ul>
