Speech Audiometry
Speech Audiometry
Speech Audiometry
2
Speech recognition threshold (SRT) patient, and is then lowered in predeter-
mined step sizes to a level at which 50%
The SRT is the most frequently used recognition is attained.
speech threshold test. It is a measure of the
intensity level at which the listener is able 1. Examine the pure tone thresholds ob-
to correctly repeat 50% of words presen- tained for 500, 1000 and 2000 Hz in
ted. This level should correspond roughly the test ear
to the average of the pure tone audiometry 2. Use the lowest two of the three values
thresholds at 500, 1000 and 2000 Hz 2. to calculate a two-frequency PTA
When the pure tone audiogram is steeply 3. Start testing at an intensity level 25dB
sloping (high frequency hearing loss), the above the two-frequency PTA by
SRT corresponds better with either the best presenting one spondee word. If using
two of three pure tone average (PTA) monitored live-voice, take care to
frequencies (500, 1000 & 2000 Hz) or the present the two syllables of the word
best one of these three frequencies 3. The with equal intensity by carefully
SRT is generally con-ducted using spon- watching the VU-meter and ensuring
dee words i.e. single words that comprise that the meter indicates the same level
two syllables with equal emphasis placed for both syllables
on each syllable. 4. If the listener is unable to repeat the
word at the initial presentation level,
SRT test procedure increase the intensity by 10dB and
present another word
Before testing, the listener is made 5. Continue this process until the listener
familiar with the test materials to ensure is able to correctly repeat the presen-
that all the words are known to him/her. ted word
This is done by providing a written list of 6. Once the listener correctly repeats a
the words to the patient to read aloud, or word, lower the intensity by 5dB and
by presenting the words at a comfortable present another word
intensity (e.g. 30-40 dB above the PTA of 7. Continue lowering the intensity in
500, 1000 and 2000 Hz) and asking the steps of 5dB until the listener makes
patient to repeat the words (if the patient an error
finds this intensity too low, words can also 8. Once this occurs, the main part of the
be presented at a slightly higher intensity test is commenced
for the purposes of familiarisation). Any 9. Start the main threshold search 10dB
words that the listener cannot repeat cor- above the level where the listener
rectly or that is unfamiliar are excluded made the first error. Present three
from the test. Before starting the test, in- words at this level
struct the patient to repeat the words he/she 10. If all three words are repeated
hears, even if the words become very soft, correctly, lower the intensity by 5dB
and he/she is allowed to guess if uncertain. and present three more words
11. Once the listener makes an error with
Chaiklin-Ventry SRT test procedure any of the three words, present three
more words at the same level to get a
The SRT may be measured in many ways. score out of six
A simple technique is that described by 12. Continue to descend in 5dB steps,
Chaiklin and Ventry 4; it is a descending each time presenting six words, until a
method, meaning that the test commences level is reached where the patient
at an intensity level clearly audible to the
3
correctly repeats three of six (50%) of
the words
13. This level is then recorded on the test
form as the SRT. If a patient repeats
more than three of six words correctly
at one level (e.g. 4/6 at 25dB) and less
than three of six at the next presenta-
tion level (e.g. 2/6 at 20 dB), record
the SRT as 20-25dB on the test form
14. Conduct the test in the opposite ear, so Figure 3: Insert earphones
as to determine the SRT in each ear
independently Supra-aural earphones (Figure 1)
4
(three steps of 5dB) from the original indicate the site-of-lesion if pathology is
masking level. present. Patients with cochlear or retro-
7. If the patient still attains 50%, then the cochlear hearing loss or pathology are
masked threshold is the same as the generally unable to attain 100% speech
unmasked threshold. Record this on the recognition.
audiogram, along with the masking
level that was used (e.g. 20-35dB). Speech discrimination can be measured
8. If the patient is unable to attain 50% using a variety of test materials. Mono-
recognition at the previously measured syllabic single words preceded by a carrier
SRT level once masking is introduced, phrase are most frequently used. Because
keep the masking level constant and measuring speech discrimination involves
increase the speech level in the test ear testing at three or more intensity levels,
by 5dB, presenting six words at that test materials should include a number of
level. lists so that a complete list of words can be
9. Continue this process until the patient used at each presentation level. A number
attains 50% recognition. of commercially available tests in English
10. Increase the masking level by 5dB and are available (see Lawson and Peterson for
present six more words, repeating this a list of tests 5). Two frequently used
process until the masking level is 15dB English word lists are the Northwestern
higher than the initial level. Auditory Test No. 6 (NU-6) 6 and the CID
Auditory Test W-22 7. Lists that are phone-
tically balanced (contain the same number
Speech discrimination score of each particular phoneme in each list) are
particularly popular as they ensure that
The speech discrimination score (also listeners with a steeply sloping audiogram
called word recognition score) is the most who have difficulty with high frequency
commonly used test of supra-threshold phonemes will not find some of the lists
speech perception. easier than others. Ideally patients should
be tested in their native language (mother
This is an important test in the audiological tongue). If test materials are not available
test battery, as it indicates the patient’s in the patient’s first language and they are
ability to hear and understand speech at proficient in English, it should be possible
typical conversational levels. It also indi- to conduct the test using English materials,
cates how well a patient can perceive provided that the patient is familiar with all
speech if the presentation level is increase- the words used. If the patient has limited
ed; this helps the clinician to predict the English vocabulary it might be necessary
potential benefits from amplification. to use a closed set test where the listener
Some patients, especially those with sen- chooses a response from a limited number
sorineural hearing loss and hearing losses of options e.g. the Modified Rhyme
that do not affect all frequencies equally, Hearing Test 9.
are unable to achieve 100% speech recog-
nition, regardless of how much the presen- Standard word discrimination/recognition
tation level is increased. These patients score measurement
will benefit from strategies to improve the
signal-to-noise ratio towards greater clarity 1. Instruct the patient to listen carefully to
of speech signals (e.g. advanced digital the presented words and to repeat them
hearing aids, FM systems etc.). Speech aloud.
audiometry results can also be used to
5
2. Presentation of each test item may be this. Calculate the percentage of words
preceded with a “carrier phrase” e.g. repeated correctly by doubling the
“say the word”; although results of score out of 50.
research into the differences between 7. If the patient attains 100% recognition
this method and where no carrier at the initial presentation level, present
phrase is used are inconclusive 2. What a 2nd list at a level 20dB below this
is most important is that, whichever level, or at a conversational level (i.e.
method is selected (i.e. with/ without 60dB HL) if the patient’s thresholds
carrier phrase), it should be used were elevated, to obtain an impression
consistently throughout the test and of how well the patient would hear at
across lists. normal conversational levels.
3. If audiometric results (air and bone 8. If the patient is unable to attain 100%
conduction pure tone audiometry) recognition (or very close to 100%) at
indicate normal hearing or a the initial presentation level, present
conductive loss, start testing at 30dB another list at an intensity level 20dB
above the measured SRT, or if the SRT above the initial level. If time allows, a
is not available, 30dB above the 3rd list is presented at a level 20dB
average pure tone thresholds measured higher than that used for the 2nd list, or
at 500, 1000 and 2000 Hz. Most 10dB higher if the presentation level is
patients with normal hearing should uncomfortable for the patient.
attain 100% speech recognition at 25- 9. With cochlear hearing loss, speech
40dB above their SRT 10. Patients with recognition is expected to increase up
conductive loss are also usually able to to a certain point with increasing
attain 100% recognition at intensity intensity (maximum recognition point),
levels 30-40 dB above their own SRT. and then to stabilise.
4. If a patient’s audiogram indicates 10. With retrocochlear lesions speech
sensorineural hearing loss (both air recognition improves as intensity in-
and bone conduction thresholds creases up to a point (maximum recog-
elevated), start testing at 40dB above nition point) after which recognition
the patient’s SRT 8. Patients with deteriorates at higher intensities. This
cochlear or retro-cochlear hearing loss phenomenon is called “roll-over” and
or pathology are generally unable to is considered to be an indication of
attain 100% speech recognition, retrocochlear pathology. However, the
regardless of presentation level. The degree of deterioration in speech
objective with such patients is to recognition at increased intensities that
determine what the maximum percen- can be considered a clinically signifi-
tage of speech recognition is, and at cant roll-over may differ according to
what intensity level this occurs. the speech materials used 3 and should
5. Start the test by presenting a complete be interpreted with caution and in
list of words (typical word lists usually conjunction with other site-of-lesion
comprise 50 words) at the selected tests such as acoustic reflexes, reflex
starting intensity level. decay test or Auditory Brainstem
6. Record how many of the words the Response measurements. The formula
patient repeats correctly. It is useful to used to calculate the degree of roll-
mark words in the list that the patient over is as follows: (maximum score –
finds difficult or repeats incorrectly; minimum score)/maximum score; the
when a patient mistakes a particular maximum score indicates the maxi-
phoneme for another, make a note of mum % of recognition the patient
6
attained, and the minimum score indi- bone gap) + 20dB (safety factor) =
cates the minimum % recognition 65dB of masking required in the non-
attained at an intensity level higher test ear.
than the level where the maximum
score was obtained. Reporting speech audiometry results
7
intensity indicated on the x-axis of the rollover should therefore be considered in
graph. conjunction with other audiological find-
ings to determine the likelihood of a retro-
cochlear lesion being present.
8
5. Lawson G. Peterson M. Speech Pretoria, South Africa
Audiometry, (2011) Plural Publishing, [email protected]
Inc., San Diego
6. Tillman TW, Carhart R. An expanded Claude Laurent, MD, PhD
test for speech discrimination utilizing Professor in ENT
CNC monosyllabic words, North- ENT Unit
western University Auditory Test No. 6 Department of Clinical Science
Tech Report SAM-TR-66-55, (1966) University of Umeå
USAF School of Aerospace Medicine, Umeå, Sweden
Brooks Air Force Base, Texas. [email protected]
7. Hirsh IJ, Davis H, Silverman SR, Rey-
nolds EG, Eldert E, Benson, RW. De- Johan Fagan MBChB, FCS(ORL), MMed
velopment of materials for speech au- Professor and Chairman
diometry. Journal of Speech and Hear- Division of Otolaryngology
ing Disorders. 1952; 17(3): 321-37 University of Cape Town
8. Kramer SJ. Audiology: science to Cape Town
practice, (2008) Plural Publishing, Inc., South Africa
San Diego [email protected]
9. Kruel EJ, Nixon JC, Kryter KD, Bell
DW, Lang JS, Schubert ED. A
proposed clinical test of speech OPEN ACCESS GUIDE TO
discrimination, Journal of Speech and AUDIOLOGY & HEARING AIDS
Hearing Research. 1968; 11(3): 536-52
10. Schoepflin JR (2012). Back to basics: FOR OTOLARYNGOLOGISTS
speech audiometry. http://www.entdev.uct.ac.za
11. http://www.audiologyonline.com/articl
es/back-to-basics-speech-audiometry-
6828. Last accessed on 15 December
2012
The Open Access Atlas of Otolaryngology, Head &
Neck Operative Surgery by Johan Fagan (Editor)
[email protected] is licensed under a Creative
Author Commons Attribution - Non-Commercial 3.0 Unported
License
Editors