4

Characteristics of Hearing and Speech Tests

This chapter focuses on item 2 in the committee’s Statement of Task:

Describe the characteristics of hearing tests, administered in the sound field, either binaurally or monaurally, in either quiet or noise, that are in use for those with cochlear implants, and describe to the degree possible:

1. The availability of the selected tests with respect to the instruments themselves, trained administrators of the tests, and insurance coverage or costs incurred with testing;
2. The patient burden of undergoing these tests;
3. Whether testing procedures or parameters, or the appropriateness of the test itself, vary based on the age of the person being tested;
4. Whether the test outcomes are expected to vary based on demographic or other patient characteristic factors, including repeated testing with the same instrument; and
5. The validity, specificity, sensitivity, reliability, and generalizability of the tests.

The committee selected several sentence tests and word tests to describe in this chapter. The tests selected are those that are commonly used to evaluate hearing loss in adults and children with cochlear implants. The chapter begins by addressing a few of the issues in the state of task that are crosscutting, such as test administrators and costs associated with testing. The

chapter then discusses the characteristics of each of the following sentence tests and word tests in order of their chronologic development:

Sentence Tests

• Central Institute for the Deaf (CID) Sentences
• City University of New York (CUNY) Sentences
• Hearing in Noise Test (HINT)
• HINT-Children (HINT-C)
• Quick Speech in Noise Test (QuickSIN)
• Bamford-Kowal-Bench Speech in Noise (BKB-SIN)
• Arizona Biomedical (AzBio) Sentences Test
• Pediatric AzBio Sentences Test

Word Tests

• Phonetically Balanced Kindergarten (PBK) Words
• Northwestern University Test No. 6 (NU-6) Words
• Maryland Consonant–Nucleus–Consonant (CNC) Words
• Lexical Neighborhood Test and Multisyllabic Neighborhood Test (LNT and MLNT)
• Words in Noise Test (WIN)
• Digit Triplet Test

The characterizations of individual tests are followed by a section on considerations beyond auditory testing, given that hearing loss affects multiple functional abilities. The information requested in items “a” through “e” above is provided to the degree that is available in the literature. Several of the test characteristics could not be addressed because the information required did not appear to be available in the published literature. Specifically, the committee conducted a literature search on the validity, specificity, sensitivity, reliability, and generalizability of each of the tests described in the chapter. Although committee members found many published papers regarding the reliability of the tests that are summarized in this chapter, they did not find published literature about test validity, specificity, sensitivity, and generalizability. The committee also did not find published literature on outcome variation by patient demographic for all of the tests described. When this information was not available, the committee suggests some patient characteristics that might affect outcomes, based on the committee’s professional judgment.

Finally, the chapter provides conclusions and a recommendation based on material covered in Chapters 1 through 4.

CROSS-CUTTING ISSUES

Making decisions about adult cochlear implant candidacy has historically relied heavily on the use of sentence-level speech perception tests. As described in previous chapters, advancements in cochlear implant technology and the resulting improvements in cochlear implant outcomes have necessitated the development and use of increasingly difficult sentence tests over the past several decades.

In the United States, audiologists are the most likely administrators for tests of unaided and aided speech recognition, because those tasks fall within the scope of audiology clinical practice and audiologists are familiar with associated current procedural terminology (CPT) codes for unaided speech audiometry and aided speech recognition. Unaided speech audiometry can be administered and billed solely as a speech reception threshold or SRT (92555), SRT combined with unaided word recognition testing (92556), or both combined with conventional pure-tone audiometry (92557). Assessing aided speech recognition is included under the evaluation of aural rehabilitation status and billed as 92626. Note that CPT 92626 is a time-based code and should not be billed for aided speech recognition assessments lasting less than 31 minutes. For shorter assessments it is recommended that the audiologist bill 92700 (unlisted otorhinolaryngologic service or procedure) and be sure to include appropriate documentation in the clinical report (ASHA, 2020). While there are a number of audiology and otolaryngology practices that are employing audiology technicians, at present technicians are not able to bill for speech audiometry, speech recognition, or the evaluation of aural rehabilitation status. Thus, to ensure compliance with Medicare billing practices, a licensed audiologist should complete the assessments of speech recognition.

In terms of patient burden, the costs of the test are covered by most private and public insurance programs, although patients who are uninsured may face out-of-pocket expenses. From the patient perspective, the burden of completing a sentence or word test is low in terms of time and complexity; however, patients might experience frustration if they realize they cannot accurately understand the test items.

To ensure that test results are not influenced by factors unrelated to perceptual ability, it is important to consider a patient’s chronologic age and language level when selecting a particular sentence test. The vocabulary and grammatical structure of test items may be of particular concern; for example, if a child is given a sentence test developed for adults, it might contain many words not in the child’s vocabulary so that the score is likely to be lower than if all of the words in the test were known to the child. The vocabulary factor also affects list equivalence if a child has different levels of familiarity with the words within various lists. Test selection must

be approached carefully to ensure that the testing is measuring the desired factors (i.e., assessing speech perception abilities) rather than language level. Those same concerns hold for word tests as well.

SENTENCE TESTS

The following sections characterize sentence tests that are commonly used to evaluate hearing loss in adults and children with cochlear implants. The tests are presented in the order of their chronologic development.

Central Institute for the Deaf Sentences

In 1955, Silverman and Hirsh from CID constructed and recorded a set of 100 sentences for use in speech testing (i.e., the CID Sentences). The content and form of the sentences met a set of criteria related to vocabulary and sentence structure laid out by a working group spanning the disciplines of experimental phonetics, linguistics, psychology, and statistics that was hoping to improve the “face validity” of speech testing. Early evaluation of these sentences revealed a close relationship between performance on them and performance on a sample of continuous discourse (Giolas, 1966). The original CID Sentences were revised by Harris and colleagues in 1961 in an attempt to provide greater homogeneity of sentence length and were afterwards known as the Revised CID Sentence Lists.

Availability

The CID Everyday Speech Sentences are available for purchase from Oaktree Products.

Reliability

Giolas and Duffy (1973) investigated the equivalency of scores obtained with the original and revised lists. As part of this work, high-quality master tapes were made of all sentences (10 lists of 10 sentences for each version of the test) being spoken by one male talker with a general American accent. The recordings were distorted using low-pass filtering at 420 Hz to produce a sufficient number of error responses to allow for list comparisons. The results strongly suggested that under the study’s test conditions none of the lists within the original or revised CID sentences could be considered to be equivalent for most testing purposes.

City University of New York Sentences

As originally developed, the CUNY Sentences consist of 48 sets of 12 topic-related sentences where all sentences within a set relate to the same topic. Each set consists of four statements, four questions, and four commands (Boothroyd et al., 1985). Sentence length varies from 3 to 14 words.

Availability

The CUNY Sentences are available for purchase from Auditec, Inc.

Reliability

To determine the reliability and equivalency of the original 48 sentence sets, nine adults with normal hearing were provided visual-only information from recordings of a woman producing the sentences, supplemental audio information containing only fundamental frequency information, and the topic of the sentence (Boothroyd et al., 1985). Listeners were inexperienced lipreaders. Their task was to repeat as much of each sentence as possible. For the given recording and test conditions, the CUNY sentence sets were found to be of equivalent difficulty. However, short-term leaning effects were present as revealed by improving scores over the first eight or nine sets of sentences. The 95% confidence intervals of a single score of approximately 85 percent was assumed to be +/– 16 percentage points, excluding the short-term learning effect described above. It is important to note that the apparent learning effects may have been due to improvements in lipreading performance, improvements in the ability to integrate limited acoustic information with visual information, or increased familiarity of the task.

Hearing in Noise Test

The HINT, described in detail in Chapter 3, was developed by Nilsson et al. (1994) as a measure of sentence speech reception thresholds. The HINT makes use of modified BKB sentences (Bench et al., 1979), which were originally constructed for use with British children. These sentences contain common nouns and verbs found in transcriptions of children’s speech and were designed to be scored based on the recognition of key words. Modifications included removing British idioms but maintaining the same sentence length. Great care was taken to match the phonemic content of the 25 10-sentence-long lists.

Availability

To verify the availability of the HINT, the committee corresponded with three individuals at Natus, a company that produces, among other products, medical equipment used in the diagnosis of hearing loss. Natus previously distributed the HINT Pro System, an audiometer used for assessing hearing, which contained a library of all of the various versions of the HINT. The committee learned that this system is no longer produced and that the HINT is not otherwise available for stand-alone purchase. In addition, it is not accessible for many patients (see Chapter 3). While it is true that a number of long-established cochlear implant clinics may still possess a copy of the HINT, those clinics are not necessarily geographically dispersed. This means that individuals might need to travel hundreds of miles in order to find an audiologist who is capable of offering the HINT, greatly increasing patient burden in terms of time and cost.

Reliability

Test validation studies confirmed that reliable sentence-level speech reception thresholds can be obtained in quiet and in noise with adaptive procedures using the HINT’s short lists of brief sentences. Confidence intervals (CIs) suggest that the use of one 10-sentence list per condition makes it possible to detect differences of 2.98 dB (decibels) in quiet and 2.41 dB in noise. CIs improve as the number of sentence lists increases. The reliability of sentence speech reception thresholds was also found to vary as a function of bandwidth, but was only substantially degraded when the bandwidth dropped below about 2,000 Hz,¹ suggesting that the test is fairly robust to the variations in audibility associated with hearing loss and personal sensory devices. While variability is greater in listeners with hearing loss, the reliability of the test in these listeners is still quite close to that measured in listeners with normal hearing (Gilbert et al., 2013; Robson, 2001; Schafer, 2010).

The HINT sentence lists are intended for adaptive measures of speech reception thresholds. The popular use of these sentences in the evaluation of cochlear implant patients disregards this intent in that the sentences are presented at fixed levels in quiet or in the presence of fixed-level noise. Because of that, the reliability demonstrated during initial test construction does not necessarily apply to the common use of these sentences with cochlear implant users.

___________________

¹ The hertz (Hz) is the derived unit of frequency in the International System of Units and is defined as one cycle per second.

Hearing in Noise Test-Children

The HINT-C is composed of a subset of sentences from the HINT that could be accurately identified by a group of 5- and 6-year-old children. In total there are 10 lists of 10 sentences each for use in adaptive threshold testing in quiet or in noise.

Availability

The HINT-C is no longer available for purchase by cochlear implant audiologists or others (see HINT availability).

Reliability

Given that the HINT-C is composed of a subset of sentences from the HINT, reliability is expected to be similar between the two tests. Performance on the HINT-C is age-dependent. Specifically, 6- to 12-year-old children have been shown to perform significantly more poorly than older children and adults (Nilsson et al., 1996; Schafer, 2010).

Quick Speech in Noise Test

In 2004, Killion et al. (2004) developed a shorted and improved version of the SIN test (Etymōtic Research, 1993), which they named the QuickSIN test. The original SIN test uses 360 Institute of Electrical and Electronics Engineers (IEEE) sentences (Rothauser, 1969) as target speech and four-talker babble (Auditec of St. Louis, 1971) as “noise.” Each sentence is scored on five key words, with partial credit available for partially correct answers; however, inter-observer reliability of partial credit scoring has been found to be low (Bentler, 2000). While the original IEEE sentences lists were phonetically balanced, these balanced groupings were not maintained in the SIN test. Another concern about the SIN test is that Bentler (2000) demonstrated that it suffered from both ceiling and floor effects. The updated QuickSIN test uses a significantly smaller number of the original SIN’s IEEE sentences (a total of 72), which were carefully selected to ensure a balancing of lists, and it does not allow for partial credit scoring. The QuickSIN is composed of 12 equivalent lists, each containing 6 different IEEE sentences. A single sentence (as opposed to five sentences in the SIN test) is presented at each of six decreasing signal-to-noise ratios (SNRs) between successive sentences (from 25 to 0 dB SNR, in 5 dB steps). The listener’s score is calculated as the SNR the listener requires to understand 50 percent of the sentences’ key words. As its name implies, the QuickSIN

is fast to administer, with each list taking approximately 1 minute to complete, compared with 6 minutes for a single SIN test block.

Availability

The QuickSIN Test is available for purchase from Etymōtic Research.

Reliability

The QuickSIN’s 12 equivalent lists have a standard deviation of 1.4 dB for SRN estimation when a single list is administered, which is slightly better than what would have been expected based on the SIN test, an improvement that is likely due to the more careful selection of sentences for the QuickSIN. In terms of reliability, a single QuickSIN list is accurate to ± 2.2 dB at the 80% CI (2.7 dB for a 95% CI), which the authors suggest is adequate for clinical testing. Reliability improves when more than one list is administered. For example, averaging two lists increases accuracy to from ± 2.2 dB to ± 1.6 dB for an 80% CI (and from ± 2.7 to ± 1.9 dB for a 95% CI). When comparing two conditions, the use of multiple QuickSIN lists is important. For example, when using only one list per condition, results would have to be greater than 3.2 dB to be considered statistically different, given an 80% CI; that difference score decreases to 2.2 dB for two lists, to 1.8 dB for three lists, with further, diminishing decreases as the number of test lists continues to increase.

Bamford-Kowal-Bench Speech in Noise

The BKB-SIN test is a sentence test that was developed primarily for use with both children and patients who are candidates for or already have cochlear implants (Etymōtic Research, 2005). Like the HINT, it uses Americanized BKB sentences that are short, highly redundant, rich with semantic and syntax context cues, and at a first-grade reading level. The main differences between the HINT and the BKB-SIN are the methods used to determine the 50 percent point and the type of background noise. While the HINT uses a modified adaptive procedure that brackets the 50 percent point, the BKB-SIN, like QuickSIN and the WIN (described below), relies on a descending level of noise and the Spearman-Karber equation to calculate the 50 percent point.

The BKB-SIN, when presented in quiet at 70 dB SPL, is the standard test for evaluating candidacy for adult cochlear implantation in the United Kingdom (NICE, 2019). BKB-SIN sentences are also used after implantation to demonstrate improvement (Cullington and Aidi, 2017). The BKB sentences used in the BKB-SIN and the HINT provide more semantic

context than the IEEE sentences used in the QuickSIN (Wilson et al., 2007). Thus, better recognition scores should be expected on the BKB-SIN and the HINT than on the QuickSIN and the WIN.

Availability

The BKB-SIN test can be purchased from Auditec, Inc., and Etymōtic Research.

Reliability

The reliability of the BKB-SIN is related to the number of test items, age, and cochlear implant use. The largest increase in reliability by incremental increases in the number of lists is observed between the use of one versus two lists (0.5 dB for adults with normal hearing, 0.9 dB for adult cochlear implant users, 1.2 dB for 5–6-year-olds, 0.7 dB for 7–10-year-olds, and 0.7 for 11–14-year-olds), which adds just 3 minutes to testing. The test manual provides information regarding the magnitude of difference that can be reliably measured when comparing two conditions as a function of number of lists for adults with normal hearing, adult cochlear implant users, and children.

Arizona Biomedical Sentences Test

The AzBio Sentences Test was devised by Spahr and Dorman at Arizona State University specifically as an experiment to compare the speech understanding abilities of high-performing patients implanted with different cochlear implant systems, in everyday listening environments (Spahr and Dorman, 2005; Spahr et al., 2007). The relatively complex sentences within each list are spoken by four different talkers using a conversational speaking style and rate, which allows for an estimation of real-life performance.

In 2008, Gifford and colleagues applied the tests more generally to evaluate the speech perception abilities of hearing impaired people with cochlear implants. They reported that the AzBio test produced results that were highly correlated with monosyllabic word scores and did not suffer the same ceiling effects in quiet as other sentence tests (ceiling effects are observed when subjects consistently score at or near 100 percent, such that any further improvement is unable to be detected). Ceiling effects are minimized in quiet and in noise because the materials are more difficult than the HINT sentences (Gifford et al., 2008; Spahr et al., 2012). Fabry et al. (2009) agreed with that assessment and suggested that AzBio sentences could be of value in evaluating hearing in cochlear implant users, both pre- and post-implant. Cochlear implant manufacturers have thus moved

to include AzBio sentence lists in a new battery of tests that will serve as the standard for assessing pre- and post-implant hearing.

Availability

The AzBio Sentences Test can be purchased from Auditory Potential, LLC.

Reliability

In 2012, Spahr et al. (2012) set out to create and validate a new set of sentence tests to add to the AzBio sentence corpus. Four talkers, two male and two female, were selected to each record 250 sentences. The Spahr et al. (2012) validation study found that 29 of the 33 lists of sentences yielded scores that, when averaged over the 15 cochlear implant users tested, were not statistically different from one another. However, individual listeners demonstrated considerable variability in performance across lists. The researchers chose 15 of the 29 lists to include in the AzBio Sentences Test. They noted that the test had not been validated for children and that further research would need to be conducted to assess the reliability of the sentences for specific populations.

Pediatric Arizona Biomedical Sentences Test

The popularity of the original AzBio test resulted in clinicians using these materials with their pediatric patients. Concerns regarding the appropriateness of the content of some of these sentences, paired with the fact that the test was too difficult for some poorer-performing adult cochlear implant users, led to the creation of a new set of materials. The validated Pediatric AzBio Sentences Test consists of 16 equivalent lists of 20 sentences spoken by a single female talker (Spahr et al., 2014).

Availability

The Pediatric AzBio Sentences Test can be purchased from Auditory Potential, LLC.

Reliability

Spahr et al. (2014) created recordings of 450 sentences (3–12 words in length) that had been generated by 5- to 12-year-old children, with and without hearing loss, during everyday conversations. All of the sentences

were recorded by one female talker and screened for inclusion by presenting them to 30 kindergarten and first-grade children with normal hearing. Following a sentence intelligibility estimation procedure in which adults listened to versions of the sentences that had processed through a 15-channel noiseband vocoder, a type of filtering that produces a signal that contains the same kind of speech information available to a cochlear implant user, a total of 320 sentences were chosen and divided into 16 lists of 20 sentences each. The mean list intelligibility per list was 78.6 percent correct. The equivalency of these pre-assigned lists was validated by a group of experienced adult cochlear implant users and pediatric hearing aid and cochlear implant users. The performance achieved by adult and pediatric cochlear implant users averaged 74 percent correct and 77 percent correct, respectively. Statistical analysis revealed no significant differences across lists. The authors provide CIs for the administration of one list per condition and two lists per condition that are based on methods originally described by Thornton and Raffin (1978). The availability of 16 equivalent lists provides the opportunity to assess changes in performance across conditions over time.

WORD TESTS

The section above highlighted the sentence tests that have been developed and can be used for individuals with cochlear implants. This section will focus on word tests.

Phonetically Balanced Kindergarten Words

The PBK test (Haskins, 1949) was the product of the unpublished master’s thesis work of Harriet Haskins. While four test lists were generated originally, only three of the lists have been used clinically because the fourth list (List 2 in the original work) did not produce results equivalent to the other three. At the time they were chosen for the PBK test, the test words were all found within the 2,500 words of highest frequency spoken by preschool children (IKU, 1928). However, it is likely that today’s very young children with hearing loss may be unfamiliar with a number of the words contained within the PBK lists (Kirk et al., 1995; Meyer and Pisoni, 1999). In this open-set test of monosyllabic word understanding, children are asked to repeat back words presented to them without any visual cues.

Availability

The PBK test can be purchased from Auditec, Inc.

Reliability

Haskins collected speech perception data from adults with normal hearing listening to the PBK lists. It was during this testing that List 2 was identified as being more audible than the other lists. For this reason, only Lists 1, 3, and 4 from the original thesis work are used in clinical practice. Meyer and Pisoni (1999) were unable to find evidence that either children or listeners with hearing loss were tested when the reliability of these word lists was evaluated. In their own work, Meyer and Pisoni (1999) found that the mean frequency of words from List 2 was higher than the mean frequency of words from the other lists, which likely accounts for them as having been described as “more audible” in Haskins’s thesis. These authors also found that the ratio of the frequency of the stimulus to the density of the lexical neighborhood was more favorable for the words in List 2 than for the words in the other lists. Finally, findings of Kluck et al. (1997) suggest there will be age effects on the PBK test that may be related to word familiarity. Specifically, they found that 4-year-old children with normal hearing performed at near ceiling levels on the PBK test, while 3-year-old children performed slightly lower. The investigators also asked the parents of those 3- and 4-year-olds to judge whether their children were familiar with each of the PBK words. Results suggested that the words were slightly less familiar to the 3-year-olds but that, overall, most words on the PBK test were familiar to the general population of preschool age children.

Northwestern University Test No. 6 Words

The NU-6 test was first described in 1966 by Tillman and Carhart. This test contains four lists of 50 CNC monosyllabic words. The authors took great care to preserve the phonemic distribution of the pool of 1,263 words initially selected by Lehiste and Peterson (1959), who recommended that each initial, medial, and final phoneme appears in a single list of 50 words with the same relative frequency as found in the total pool of words.

Availability

The NU-6 test can be purchased from Auditec, Inc.

Reliability

Experiments conducted with 40 listeners with normal hearing and 12 listeners with hearing loss suggest both good inter-list equivalence and high test–retest reliability. For both groups of subjects, the maximum test–retest difference was always less than 10 percent (five words) and almost always (i.e., 92 percent of the time) within 6 percent (three words).

Maryland Consonant–Nucleus–Consonant Words

The Maryland CNC Words test is routinely used in U.S. Department of Veterans Affairs medical centers as part of the audiological Compensation and Pension exam, which determines hearing disability related to military service (Mendel et al., 2014). Causey et al. (1984) used Peterson and Lehiste’s revised CNC words lists (Peterson and Lehiste, 1962) to develop this open-set, monosyllabic CNC test. Recordings were made by one male talker speaking the test words within the phrase “Say the [test word] again.”

Availability

The Maryland CNC Words test is available for order from Richard Wilson at Arizona State University.

Reliability

Causey et al. (1984) examined list equivalency by presenting all lists to a total of 40 male veterans with sensorineural hearing loss, 10 listeners per group, with each group assigned to listen at 20, 30, 40, or 44 dB sound level (SL) in relation to their speech reception thresholds. Results suggested that 6 of the 10 available lists produced equivalent scores (i.e., Lists 1, 3, 6, 7, 9, and 10). Performance across those lists varied by a range of only 3.6 percent.

Test–retest reliability was examined in 10 listeners who were tested twice with an average span of 4 weeks between test sessions. They listened at a 44 dB sensation level. The differences in average scores between the two test sessions were never greater than 5.6 percent. The correlations between test sessions were significant for all 10 lists, with the strongest correlation between List 1 and List 2. When evaluated using Thornton and Raffin’s (1978) 95 percent critical different values for a 50-item test, only one listener showed a significant difference across the two test sessions and, then, only for List 2. Based on both of these experiences, the developers suggested using the six equivalent and reliable lists when assessing speech perception abilities.

Lexical Neighborhood Test and Multisyllabic Neighborhood Test

Given that both word frequency and lexical similarity affect speech understanding, Kirk et al. (1995) set out to develop new word recognition tests, one monosyllabic (the LNT) and one multisyllabic (the MLNT), in which the lexical properties of the test items were carefully controlled. They used a measure of lexical similarity that considered the number of

“neighbors,” or words that differed by just one phoneme from a target word. The frequency of a word combined with its “density” within a lexical neighborhood can be used to sort words into dense neighborhoods (i.e., words that have many lexical neighbors) and sparse neighborhoods (i.e., words with few lexical neighbors). Words can be further classified from a scale of “easy” (with the easiest being high-frequency words from sparse neighborhoods) to “hard” (with the hardest being low-frequency words from dense neighborhoods).

Test stimulus words were drawn from Logan’s (1992) analysis of a sample of child language obtained from the Child Language Data Exchange System database (MacWhinney and Snow, 1985). Specifically, words produced by 3- to 5-year-old children were selected and sorted into four lists, each composed of 25 monosyllables for the LNT and 25 two- or three-syllable words for the MLNT. Two of the word lists for each of these tests were lexically easy, and two were lexically hard. While now available as recorded lists, during the initial development and testing, the words were presented via monitored live voice to a group of pediatric cochlear implant users. The analysis of this testing revealed that young cochlear implant users were sensitive to the acoustic–phonetic similarity among the LNT’s words in that they performed better with the easy lists than with the hard lists. Furthermore, word recognition scores were higher for multisyllabic than monosyllabic words for both easy and hard lists. In addition to information about speech perception ability, the LNT and the MLNT allow for an examination of the perceptual processes underlying spoken word recognition and may also be used to better understand the organization of sound patterns of words in young children’s lexicons and the processes used to access these patterns.

Availability

The LNT and the MLNT can be purchased from Auditec, Inc.

Reliability

As a follow-up to the initial development and testing of the LNT and the MLNT lists, Kirk et al. (1999) investigated the reliability and inter-list equivalency of recordings of these lists. Their results showed that even with recorded presentations, performance was better on easy lists than hard ones, there were no differences between matched lists, and test–retest reliability was high. High reliability was demonstrated in two ways: (1) no difference in LNT and MLNT scores over separate test sessions and (2) strong correlations (r ≥ 0.83) between test sessions.

Words in Noise Test

The WIN was developed by Wilson et al. (2003) to provide a test that would qualify speech understanding in multi-talker babble in terms of hearing loss for speech, expressed in terms of signal-to-babble ratio (S/B) loss. The resultant instrument presents 10 NU-6 words (described above) spoken by a female talker (Tillman and Carhart, 1966) at each of seven S/Bs, in 4 dB steps from 0 to 24 dB, with the babble being comprised of three male and three female talkers speaking about various topics. The use of monosyllabic words in the WIN provides a good measure of basic auditory function because the effect of memory and linguistic context on recognition are minimized. Additionally, monosyllabic words recorded in isolation decrease the chance of blurring phoneme boundaries with co-articulation (Wilson et al., 2007).

Availability

The WIN is available for order from Richard Wilson at Arizona State University.

Reliability

The WIN is sensitive to the presence of even high-frequency hearing loss. The 90th percentile for listeners with normal hearing was observed to be 6 dB S/B. Based on this observation, any test outcome suggesting a S/B ratio greater than 6 dB is considered an abnormal finding. Even though all listeners with hearing loss who participated in experiments to develop and evaluate the WIN test had monosyllabic word understanding abilities equal to or greater than 80 percent correct in quiet (mean 89 percent), more than 95 percent of them had abnormal WIN scores. Mean data revealed a 6–9 dB difference in the 50 percent points for listeners with normal hearing and hearing loss, which is consistent with findings obtained by others (Dubno et al., 1984; Tillman et al., 1970; Wilson and Strouse, 2002) who have examined the speech perception in noise abilities of these two groups of listeners.

Digit Triplet Test

Smits et al. (2004) developed a speech-in-noise test for use as an objective, self-test for hearing loss in Dutch listeners. They made the fully automated test controllable by a computer and available for taking over the telephone. In approximately 3 minutes, the test can produce a speech reception threshold using an adaptive approach based on the perception of 20 strings of 3 single digits (a total of 23 strings are chosen randomly from

a body of 80 triplets), spoken by a female talker and presented in a background of speech-shaped noise (matched in long-term average spectrum to the digits), which is not affected by telephone type or listening environment.

It is becoming increasingly common to measure speech understanding with numerical digits. This type of testing has an advantage compared with sentence tests in that it does not rely on the listener’s cognitive ability to recognize contextual cues and limited language ability does not impede testing as much with other tests (Cullington and Aidi, 2017). Digit triplet testing is also easier than sentence tests to conduct over the phone or the Internet or to self-administer (Leensen et al., 2011; Smits and Houtgast, 2005; Smits et al., 2006; Watson et al., 2012). The digit triplet test also minimizes lexical knowledge, which can be useful for non-native speakers of English (Ramkissoon et al., 2002). Testing digit triplets in noise has been used to screen hearing in hundreds of thousands of people worldwide (Stam et al., 2015; Williams-Sanchez et al., 2014). It has also been used successfully with both adults and children with cochlear implants (Kaandorp et al., 2015; Mishra et al., 2015), its results correlate well with those of a sentence test in noise, and it has been shown in at least one Dutch study to have adequate repeatability (Kaandorp et al., 2015).

Availability

Availability in the United States is unknown.

Reliability

The digit triplet test is highly reliable, as evidenced by its measurement error of less than 1 dB (i.e., standard deviation of repeated measurements within subjects). This high level of reliability is observed in the controlled conditions of audiology clinics as well as in listeners’ homes. Using a standard of a sentence speech reception threshold test (Plomp and Mimpen, 1979) administered via headphone, the sensitivity and specificity of the digit triplet are 0.91 and 0.93, respectively, which the authors cite as evidence of its utility for screening purposes. A digit triplet SNR of –4.1 dB serves as the cut-off for normal and impaired hearing.

Table 4-1 presents a summary table of the information provided in this chapter on the individual tests.

CONSIDERATIONS BEYOND AUDITORY TESTING

Auditory testing, including both audiometric thresholds and speech testing, is considered the gold standard of assessing auditory function in individuals with hearing loss. Cochlear implant candidacy and outcomes

TABLE 4-1 Reliability and Other Notable Characteristics of Selected Sentence and Word Tests

are often determined based on a defined performance level obtained from auditory testing. As described in this chapter, there are many different speech tests available that can be used in specific patient populations. However, hearing loss can have impacts beyond what is captured during auditory testing. Many studies have shown that significant hearing loss can affect emotional, social, academic, and occupational elements of life. Additionally, there is evidence to suggest that measurements of speech recognition tests do not always correlate with subjective perception of auditory performance (e.g., Bentler, 2005; Wackym et al., 2007).

The consequence of hearing loss on quality of life in cochlear implant users can be measured using self-report questionnaires, such as two versions of the Cochlear Implant Quality of Life (McRackan et al., 2019), the Cochlear Implant Function Index (Coelho et al., 2009), and the Nijmegen Cochlear Implant Questionnaire (Hinderink et al., 2000). These are questionnaires that have been specifically designed for use in the cochlear implant population. There are also many questionnaires that exist that can provide an assessment of hearing handicap, perceived disability, or listening effort. Some examples include the Hearing Handicap Inventory for Adults/Elderly (Newman et al., 1990; Ventry and Weinstein, 1982), the Communication Profile for the Hearing Impaired (Demorest and Erdman, 1987), the Abbreviated Profile of Hearing Aid Benefit (Cox and Alexander, 1995), and the Speech, Spatial and Qualities of Hearing Scale (Gatehouse and Noble, 2004).

Questionnaires are available that can be completed by parents of young children or by clinicians to supplement audiological test results and provide important information regarding a child’s auditory skills. Examples of questionnaires available for use with young children with hearing loss include the LittlEars (Weichbold et al., 2005), the Infant–Toddler version of the Meaningful Auditory Integration Scale (Zimmerman-Phillips et al., 2000), and the Auditory Skills Checklist (Meinzen-Derr et al., 2007). Examples of questionnaires designed for use with children with hearing loss over the age of 3 years include the Meaningful Auditory Integration Scale (Robbins et al., 1991), the Parents’ Evaluation of Aural/Oral Performance of Children (Ching and Hill, 2007), and the Auditory Skills Checklist (Meinzen-Derr et al., 2007).

To fully evaluate auditory function, it can be helpful to include the subjective perspective from the patient or from their parent. Self-report or parent-report measurements, when used as a supplement to auditory threshold and speech testing, can help capture the complete picture of the impact of hearing loss in a given individual.

CONCLUSIONS AND RECOMMENDATION

As discussed in Chapter 3, the HINT has several limitations in its characteristics and deviation from its intended use (i.e., the use of a single male speaker and speech-spectrum background noise, and administration in quiet or fixed noise presentation rather than its intended adaptive design). Furthermore, the Minimum Speech Test Battery (MSTB) recommendations note that “advances in technology, improvements in outcomes, and changes in candidacy criteria have resulted in ceiling effects on the HINT sentences when presented in quiet” (Gifford et al., 2008, 2010; MSTB, 2011). U.S. Food and Drug Administration usage in effectiveness studies and unclear candidacy criteria from insurance providers (e.g., Centers for Medicare & Medicaid Services) add to the limitations of the test. Finally, due to its exclusion from the most recent MSTB and the fact that it is no longer available for purchase, the HINT is difficult for clinics across the United States to obtain.

Overall, the HINT has been widely used to measure cochlear implant candidacy and post-operative outcomes since its inception in 1994. However, the test characteristics, the state of cochlear implant technology, and the environment that made the HINT a common choice of assessment in 1994 are different in 2021.

Word recognition testing as employed in most audiology clinics includes administration of a phonemically balanced word list such as the NU-6 (Tillman and Carhart, 1966), CID W-22 (Silverman and Hirsh, 1955), or the Maryland CNC (Causey et al., 1984) word lists. Monosyllabic word recognition is currently an accepted measure for determining U.S. Social Security Administration (SSA) disability as related to hearing loss for individuals without cochlear implants. Monosyllabic word recognition is currently the standard for pediatric cochlear implant candidacy, and the field is moving toward use of a monosyllabic word recognition criterion for determining adult candidacy in the United States (e.g., Buchman et al., 2020). Additionally, monosyllabic word recognition has been used to characterize postoperative outcomes for both adult and pediatric cochlear implant recipients for more than two decades (MSTB, 2011). Furthermore, SSA has been using monosyllabic words to determine initial and continued eligibility for SSA benefits for individuals with hearing loss not treated with cochlear implantation. For those reasons, the committee recommends monosyllabic word tests, rather than sentence tests.

The committee believes that is the direction that SSA should be adopting for assessment of cochlear implant recipients. Chapter 6 provides more specific recommendations for characteristics of hearing tests that should be used to assess hearing loss in individuals with cochlear implants.

REFERENCES

Gatehouse, S., and I. Noble. 2004. The Speech, Spatial and Qualities of Hearing Scale (SSQ). International Journal of Audiology 43(2):85–99.

Gifford, R. H., J. K. Shallop, and A. M. Peterson. 2008. Speech recognition materials and ceiling effects: Considerations for cochlear implant programs. Audiology and Neurotology 13(3):193–205.

Gilbert, J. L., T. N. Tamati, and D. B. Pisoni. 2013. Development, reliability, and validity of PRESTO: A new high-variability sentence recognition test. Journal of the American Academy of Audiology 24(1):26–36.

Giolas, T. 1966. Comparative intelligibility scores of sentence lists and continuous discourse. Journal of Auditory Research 6:31–38.

Giolas, T. G., and J. R. Duffy. 1973. Equivalency of CID and revised CID sentence lists. Journal of Speech and Hearing Research 16(4):549–555.

Haskins, H. 1949. A phonetically balanced test of speech discrimination for children. Evanston, IL: Northwestern University.

Hinderink, J. B., P. F. M. Krabbe, and P. Van Den Broek. 2000. Development and application of a health-related quality-of-life instrument for adults with cochlear implants: The Nijmegen Cochlear Implant Questionnaire. Otolaryngology—Head and Neck Surgery 123(6):756–765.

IKU (International Kindergarten Union Child Study Committee). 1928. A study of the vocabulary of children before entering the first grade. Washington, DC: International Kindergarten Union.

Kaandorp, M. W., C. Smits, P. Merkus, S. T. Goverts, and J. M. Festen. 2015. Assessing speech recognition abilities with digits in noise in cochlear implant and hearing aid users. International Journal of Audiology 54(1):48–57.

Killion, M. C., P. A. Niquette, G. I. Gudmundsen, L. J. Revit, and S. Banerjee. 2004. Development of a Quick Speech-in-Noise test for measuring signal-to-noise ratio loss in normal-hearing and hearing-impaired listeners. Journal of the Acoustical Society of America 116(4 I):2395–2405.

Kirk, K. I., D. B. Pisoni, and M. J. Osberger. 1995. Lexical effects on spoken word recognition by pediatric cochlear implant users. Ear and Hearing 16(5):470–481.

Kirk, K. I., L. S. Eisenberg, A. S. Martinez, and M. Hay-McCutcheon. 1999. Lexical neighborhood test: Test–retest reliability and interlist equivalency. Journal of the American Academy of Audiology 10(3):113–123.

Kluck, M., D. Pisoni, and K. Kirk. 1997. Performance of normal-hearing children on open-set speech perception tests. In H. C. Nusbaum, D. B. Pisoni, and C. K. Davis (eds.), Research on spoken language processing, report no. 21. Bloomington, IN: Speech Research Laboratory. Pp. 349–366.

Leensen, M. C. J., J. A. P. M. De Laat, and W. A. Dreschler. 2011. Speech-in-noise screening tests by internet, part 1: Test evaluation for noise-induced hearing loss identification. International Journal of Audiology 50(11):823–834.

Lehiste, I., and G. E. Peterson. 1959. Linguistic considerations in the study of speech intelligibility. Journal of the Acoustical Society of America 31(3):280–286.

Logan, J. S. 1992. A computational analysis of young children’s lexicons. Bloomington, IN: Indiana University Press.

MacWhinney, B., and C. Snow. 1985. The child language data exchange system. Journal of Child Language 12(2):271–295.

McRackan, T. R., B. N. Hand, C. A. Velozo, J. R. Dubno, J. S. Golub, E. P. Wilkinson, D. Mills, J. P. Carey, N. Vorasubin, V. Brunk, M. L. Carlson, C. L. Driscoll, D. P. Sladen, E. L. Camposeo, M. A. Holcomb, P. R. Lambert, T. A. Meyer, C. Thomas, A. C. Moberly, N. H. Blevins, J. B. Larky, R. P. Herzano, M. E. Hoffer, S. M. Prentiss, R. N. Samy, S. P. Gubbels, J. Brant, J. B. Hunter, B. Isaacson, J. Walter Kutz, R. K. Gurgel, D. M. Zeitler, C. A. Buchman, J. B. Firszt, R. H. Gifford, D. S. Haynes, R. F. Labadie, and Cochlear Implant Quality of Life Development Consortium. 2019. Cochlear Implant Quality of Life (CIQOL): Development of a profile instrument (CIQOL-35 Profile) and a global measure (CIQOL-10 Global). Journal of Speech, Language, and Hearing Research 62(9):3554–3563.

Meinzen-Derr, J., S. Wiley, J. Creighton, and D. Choo. 2007. Auditory skills checklist: Clinical tool for monitoring functional auditory skill development in young children with cochlear implants. Annals of Otology, Rhinology and Laryngology 116(11):812–818.

Mendel, L. L., W. D. Mustain, and J. Magro. 2014. Normative data for the Maryland CNC test. Journal of the American Academy of Audiology 25(8):775–781.

Meyer, T. A., and D. B. Pisoni. 1999. Some computational analyses of the PBK test: Effects of frequency and lexical density on spoken word recognition. Ear and Hearing 20(4):363–371.

Mishra, S. K., S. P. Boddupally, and D. Rayapati. 2015. Auditory learning in children with cochlear implants. Journal of Speech, Language, and Hearing Research 58(3):1052–1060.

MSTB (Minimum Speech Test Battery). 2011. Minimum speech test battery for adult cochlear implant users. http://www.auditorypotential.com/MSTBfiles/MSTBManual2011-06-20%20.pdf (accessed October 17, 2020).

Newman, C. W., B. E. Weinstein, G. P. Jacobson, and G. A. Hug. 1990. The Hearing Handicap Inventory for Adults: Psychometric adequacy and audiometric correlates. Ear and Hearing 11(6):430–433.

NICE (National Institute for Health and Clinical Excellence). 2019. Cochlear implants for children and adults with severe to profound deafness. https://www.nice.org.uk/guidance/ta566/resources/cochlear-implants-for-children-and-adults-with-severe-to-profound-deafness-pdf-82607085698245 (accessed October 15, 2020).

Nilsson, M., S. D. Soli, and J. A. Sullivan. 1994. Development of the Hearing in Noise Test for the measurement of speech reception thresholds in quiet and in noise. Journal of the Acoustical Society of America 95(2):1085–1099.

Nilsson, M. J., S. D. Soli, and D. J. Gelnett. 1996. Development of the Hearing in Noise Test for Children (HINT-C). Los Angeles, CA: House Ear Institute.

Peterson, G. E., and I. Lehiste. 1962. Revised CNC lists for auditory tests. The Journal of Speech and Hearing Disorders 27:62–70.

Plomp, R., and A. M. Mimpen. 1979. Improving the reliability of testing the speech reception threshold for sentences. International Journal of Audiology 18(1):43–52.

Ramkissoon, I., A. Proctor, C. R. Lansing, and R. C. Bilger. 2002. Digit speech recognition thresholds (SRT) for non-native speakers of english. American Journal of Audiology 11(1):23–28.

Robbins, A. M., J. J. Renshaw, and S. W. Berry. 1991. Evaluating meaningful auditory integration in profoundly hearing-impaired children. American Journal of Otology 12(Suppl):144–150.

Robson, H. 2001. Using a variety of signals to improve real-ear testing. Trends in Amplification 5(2):77–79.

Rothauser, E. H. 1969. IEEE recommended practice for speech quality measurements. IEEE Transactions on Audio and Electroacoustics 17(3):225–246.

Schafer, E. C. 2010. Speech perception in noise measures for children: A critical review and case studies. Journal of Educational, Pediatric, & (Re)Habilitative Audiology 16:4–15.

Silverman, S. R., and I. J. Hirsh. 1955. CX problems related to the use of speech in clinical audiometry. Annals of Otology, Rhinology and Laryngology 64(4):1234–1244.

Smits, C., and T. Houtgast. 2005. Results from the Dutch speech-in-noise screening test by telephone. Ear and Hearing 26(1):89–95.

Smits, C., T. S. Kapteyn, and T. Houtgast. 2004. Development and validation of an automatic speech-in-noise screening test by telephone. International Journal of Audiology 43(1):15–28.

Smits, C., P. Merkus, and T. Houtgast. 2006. How we do it: The Dutch functional hearing-screening tests by telephone and internet. Clinical Otolaryngology 31(5):436–440.

Spahr, A. J., and M. F. Dorman. 2005. Effects of minimum stimulation settings for the MED EL Tempo+ speech processor on speech understanding. Ear and Hearing 26(4 Suppl):2S–6S.

Spahr, A. J., M. F. Dorman, and L. H. Loiselle. 2007. Performance of patients using different cochlear implant systems: Effects of input dynamic range. Ear and Hearing 28(2):260–275.

Spahr, A. J., M. F. Dorman, L. M. Litvak, S. Van Wie, R. H. Gifford, P. C. Loizou, L. M. Loiselle, T. Oakes, and S. Cook. 2012. Development and validation of the AzBio sentence lists. Ear and Hearing 33(1):112–117.

Spahr, A. J., M. F. Dorman, L. M. Litvak, S. J. Cook, L. M. Loiselle, M. D. DeJong, A. HedleyWilliams, L. S. Sunderhaus, C. A. Hayes, and R. H. Gifford. 2014. Development and validation of the Pediatric AzBio sentence lists. Ear and Hearing 35(4):418–422.

Stam, M., C. Smits, J. W. Twisk, U. Lemke, J. M. Festen, and S. E. Kramer. 2015. Deterioration of speech recognition ability over a period of 5 years in adults ages 18 to 70 years: Results of the Dutch online speech-in-noise test. Ear and Hearing 36(3):e129–e137.

Thornton, A. R., and M. J. M. Raffin. 1978. Speech-discrimination scores modeled as a binomial variable. Journal of Speech and Hearing Research 21(3):507–518.

Tillman, T. W., and R. Carhart. 1966. An expanded test for speech discrimination utilizing CNC monosyllabic words. Northwestern University Auditory Test no. 6. Technical report SAM-TR-66-55. Brooks Air Force Base, Texas: USAF School of Aerospace Medicine. https://pdfs.semanticscholar.org/008e/54f6708d34231a350af7e7b585b53a8c048c.pdf (accessed January 15, 2021).

Tillman, T. W., R. Cahart, and W. O. Olsen. 1970. Hearing aid efficiency in a competing speech situation. Journal of Speech and Hearing Research 13(4):789–811.

Ventry, I. M., and B. E. Weinstein. 1982. The Hearing Handicap Inventory for the Elderly: A new tool. Ear and Hearing 3(3):128–134.

Wackym, P. A., C. L. Runge-Samuelson, J. B. Firszt, F. M. Alkaf, and L. S. Burg. 2007. More challenging speech-perception tasks demonstrate binaural benefit in bilateral cochlear implant users. Ear and Hearing 28(Suppl 2):80S–85S.

Watson, C. S., G. R. Kidd, J. D. Miller, C. Smits, and L. E. Humes. 2012. Telephone screening tests for functionally impaired hearing: Current use in seven countries and development of a U.S. version. Journal of the American Academy of Audiology 23(10):757–767.

Weichbold, V., L. Tsiakpini, F. Coninx, and P. D’Haese. 2005. Development of a parent questionnaire for assessment of auditory behaviour of infants up to two years of age. Laryngo-Rhino-Otologie 84(5):328–334.

Williams-Sanchez, V., R. A. McArdle, R. H. Wilson, G. R. Kidd, C. S. Watson, and A. L. Bourne. 2014. Validation of a screening test of auditory function using the telephone. Journal of the American Academy of Audiology 25(10):937–951.

Wilson, R. H., and A. Strouse. 2002. Northwestern University Auditory Test no. 6 in multitalker babble: A preliminary report. Journal of Rehabilitation Research and Development 39(1):105–114.

Wilson, R. H., H. B. Abrams, and A. L. Pillion. 2003. A word-recognition task in multitalker babble using a descending presentation mode from 24 db to 0 db signal to babble. Journal of Rehabilitation Research and Development 40(4):321–327.

Wilson, R. H., R. A. McArdle, and S. L. Smith. 2007. An evaluation of the BKB–SIN, HINT, QuickSIN, and WIN materials on listeners with normal hearing and listeners with hearing loss. Journal of Speech, Language, and Hearing Research 50(4):844–856.

Zimmerman-Phillips, S., A. McConkey Robbins, and M. J. Osberger. 2000. Assessing cochlear implant benefit in very young children. Annals of Otology, Rhinology and Laryngology 109(12 II Suppl):42–43.

This page intentionally left blank.