This is an edited transcript of the live expert e-seminar presented on 1/24/11. Register here to view the recorded course.
Selecting the ideal hearing aid for a patient is a blend of art and science. While a basic pure-tone configuration may have once steered our decision making, real-world speech tests have now taken center stage as an objective measuring tool with which to select amplification. Speech-in-noise (SIN) testing for some may be a very routine practice, and for others may seem a bit daunting because it is an unknown. This article will attempt to familiarize either audiologist with the underlying fundamentals of conducting these tests clinically and how to use that information to make better hearing-aid-selection decisions. It is well understood that to use and continue use of a test, we have to understand how it fits into our clinical practice and how we can get the most bang for our buck when we invest both our time and resources.
Why Conduct SIN Tests?
There are many good reasons as to why SIN testing can be very beneficial when used routinely in the clinic. One thing we know is that SIN tests can directly address the most common complaint that patients have, which is an inability to hear well in background noise. Because it is a common complaint in all age ranges, the results we get on SIN tests can provide some very valuable insights into what might be the most appropriate amplification strategy. The results of these tests may indicate quite clearly if someone needs directional microphones, stronger noise reduction programming, extra signal processing to try to manage the background noise, or that they are in fact doing so well that we do not need to emphasize these things at all. Perhaps most importantly, it gives us more precision in the way we counsel patients about realistic expectations.
There is logical as well as emotional reasoning for conducting SIN tests in the clinic. Let's look at some of the logical reasons first. MarkeTrak VIII (Kochkin, 2010) shows some ubiquitous, yet very valuable, information. The survey indicates that the number of patients who are extremely dissatisfied with their hearing aid in noisy situations is 14%, and those that love it in noise are only at 11%, with the other 75% falling somewhere in between. These are all patients with technology that is less than four years old. The breakdown of some of the "noisy situations" was as follows: Restaurants- 14% dissatisfied, 18% very satisfied; Sports events- 11% dissatisfied, 18% very satisfied; School and classroom where noise can be a huge problem- 10% are dissatisfied and only 18% are very satisfied. The take home point is that there is not a big difference between those that love and those that hate their amplification in common noisy environments. Because patients continue to be tremendously dissatisfied in some of the noisier background situations, as clinicians we need to think of ways to more carefully identify these problems during the pre-fit and manage them with amplification strategies. The same MarkeTrak survey (Kochkin, 2010) also looked at reasons why good hearing aid candidates ended up putting their hearing aids in the drawer. The overwhelming number-one reason was the inability of the hearing aids to perform well in noise. From a logical perspective, when patients are struggling in noise and not adopting hearing aids because of background noise, it makes sense to have a reliable test to measure someone's ability or inability to hear in those challenging situations that lead to dissatisfaction and non use.
There is also an emotional aspect to conducting SIN tests routinely in the clinic. There is actually a little bit of data around this, as well. Carole Rogin (2009) discussed a study conducted by the Better Hearing Institute that examined the consumer's journey through the hearing aid testing and purchasing process, as well as the reasons people reported for being delighted with their hearing aids. From this study, Rogin concluded that in order to delight patients, professionals need to provide high tech, high-touch service delivery. High tech would include SIN testing as it actually tests something that patients are encountering on a daily basis, and that contributes to a better, more thorough evaluation. The whole concept of high tech high-touch is a great way to drive what Rogin (2009) calls hearing aid delight. Additionally, hearing aid delight might simply be that patients that are so enthralled with the way you are delivering services that they will spread the message through word of mouth to generate even more business. So from purely a business standpoint it makes sense to use SIN testing because it does contribute to that high-touch high tech service delivery.
The Fundamentals Behind Hearing in Noise
Hearing loss can be generally categorized into two types: loss of audibility and loss of clarity. We know from basic hearing science that the loss of audibility, or volume, can be attributed to damage of the outer hair cells. We also know that there is a fairly predictable relationship between the thresholds and the amount of gain a patient needs to restore audibility. Loss of clarity, on the other hand, is attributed to damage of the inner hair cells or central auditory nervous system. We also know that, for the most part, there is a pretty unpredictable relationship between the audiometric thresholds and that loss of clarity. Loss of clarity is distortion-based and is not remedied by additional gain or volume. It can, however, be quantified with SIN testing that directly measures something called signal to noise ratio loss (SNR loss). Because you cannot completely get insight into a patient's difficulty based on pure tones or word recognition scores alone, SIN tests were developed to create more real-world listening scenarios and evaluate a person's aided performance against a normal performance-intensity function. SIN tests help you decipher and quantify how much distortional loss there might be.
Another key point to consider when we are testing is the fact that language is redundant. There are actually two types of redundancy: extrinsic and intrinsic. Extrinsic redundancy is being able to use rules of language such as syntax and grammar to fill in missing blanks. For, example, because you have grown up and innately know English, if you miss a couple of words you can often times fill in the gaps because of that innate knowledge of your language. This kind of redundancy is a key compensatory strategy as we age. Intrinsic, or internal, redundancy is the auditory system's ability to carry the message to the language centers of the brain. This is also something highly affected in an aging system. This leads to some very important clinical questions: how does aging affect speech perception in noise? How does it affect internal redundancies? And how does central auditory processing affect speech perception in noise? An interesting study by Wong and colleagues at Northwestern University (2010) compared QuickSIN scores and the MRI of the prefrontal cortex on a group of 15 older adults to a group of 14 younger adults. The results suggested a decline of volume in thickness of the prefrontal cortex as a result of aging, which contributed to a declining ability to perceive speech in noise. What they also found is that the larger frontal cortex volume actually compensated for declining peripheral hearing. It is interesting to think that somebody who has a larger prefrontal cortex volume may, in the face of aging, actually perform better in noise.
Another review by Akeroyd (2008) in the U.K. compared 20 studies over a 20-year period and examined the relationship between speech perception in noise and central auditory processing factors. Nineteen of the twenty studies indicated a fairly strong relationship between speech perception and central auditory processing factors. One of the main conclusions was that even though hearing loss may occur peripherally, hearing loss remains the primary predictor of speech perception in noise, although there is a strong secondary affect from cognition and central auditory processing. In light of these findings, we should strongly consider the relationship between central auditory processing, aging, and speech perception in noise, and make sure our SIN testing reflects those challenges accordingly.
What Tests Should We Be Using In the Clinic?
Luckily, there is a choice when it comes to a variety of testing materials. There are two different methodologies when it comes to evaluating the auditory processing system: top down or bottom up. Bottom-up processing simply means that you are evaluating more of those lower-level processes such as audibility and perhaps intelligibility. You are going to use words to measure that. Top-down processing would be a measurement of the entire system including memory, comprehension and intelligibility. Likely, you would want to use sentences for this. Of course, there are pros and cons associated with each.
If you are a top-down kind of a person using sentences, you must take into account that memory and other central auditory processing mechanisms might influence your results. This is an acceptable approach when a patient scores poorly on a SIN test, because it does not matter necessarily if it is a central problem or peripheral problem. You know that person is going to struggle in background noise with amplification based on your SIN results. On the other side of things, you might want to get a more precise idea of what is going on with the entire system. Therefore, you might be more of a bottom-up person where you measure words in noise and then maybe some other central auditory processing test to measure that system. There is no "right" methodology to use. The answer really depends on what you are trying to accomplish in your clinic and how much time you have to spend with each patient. The good news is there are tests available for both bottom-up and top-down people.
What Are We Trying to Evaluate?
When we ask that question to ourselves, there are a few things to keep in mind. Most of us have very limited time in our busy clinic, and the time should be valuable to both you and the patient. Whatever test we use on the patient has to have high face validity. The patient has to understand what it is you are trying to measure, because that will probably result in a solid recommendation or treatment plan that the patient understands, which ends up being more cost efficient for both parties. There are four things that we know contribute to a more successful outcome, or maybe on the flip side, derail a fitting and lead to an unsuccessful outcome. Those would be speech intelligibility in everyday listening, annoyance from background noise, comfort in noise, and tolerance of loud sounds. For the purposes of our discussion, we will focus on speech intelligibility in everyday listening (or noise) and annoyance from noise.
Speech perception testing is always performed suprathreshold in many different arrangements to evaluate different variables. We can look at intelligibility versus comfort, annoyance or tolerance, or actual performance in quiet versus noise, fixed versus adaptive, and sentences versus words. Let's look at the whole idea of speech in quiet versus speech in noise. Some tremendous work has been completed out of the VA by Richard Wilson and Rachel McArdle. In a 2005 study, Wilson and McArdle were comparing word recognition percent-correct scores in quiet versus word recognition scores in noise as SNR loss. They concluded that a majority of the participants did, in fact, score well in quiet. Of the group who heard particularly well in quiet, only some of the subject group scored below 5 dB SNR loss (meaning better performance) on the QuickSIN test (Etymotic Research, 2001; Killion et al., 2004). A larger majority of the subject group scored higher (poorer) on the QuickSIN with SNR losses of greater than 10 to 12, indicating tremendous difficulty in background noise. So both groups had at least an 80%-correct word recognition score in quiet with a wide range of scores in noise. This easily reminds us that good word recognition in quiet does not automatically indicate good word recognition in noise. The obvious conclusion would be that you really need to be measuring speech recognition in quiet and speech recognition in noise separately.
So what are some of the options that are available to you? Some of the objective intelligibility tests which feature adaptive signal-to-noise ratios (SNR) are:
- Hearing in Noise Test (HINT) (Nilsson, Soli, and Sullivan, 1994);
- Words in Noise (WIN) (Wilson, 2003; Wilson and Burks, 2005);
- QuickSIN (Etymotic Research, 2001; Killion et al., 2004),
- Bamford-Kowal-Bench SIN (BKB-SIN) (Etymotic Research, 2005; Bench, Kowal, and Bamford, 1979; Niquette et al., 2003).
Other tests that feature fixed-level noise are:
- Connected Speech Test (CST) (Cox, Alexander, and Gilmore, 1987);
- Speech Perception in Noise Test (SPIN) (Kalikow, Stevens, and Elliott, 1977).
From a clinical perspective, the priority should be on measuring speech intelligibility in common listening situations. The QuickSIN is a reasonable choice for clinicians because it is easy to administer and score and can be done in a very short period of time. It is actually measuring a recording at six different SNRs all the way from +25 SNR down to 0 SNR. Each list consists of six sentences, with one sentence presented at each SNR level. This test uses a female talker with four competing talkers, which is classified as informational masking. It is a very challenging, realistic environment. The scoring sheet tallies key words correct subtracted from 25.5. This gives you the SNR loss, which is then compared to a normative table telling you if the loss is normal, mild, or so forth. Ideally, this test should be performed in the unaided condition with earphones to get ear-specific results and compare left to right. The QuickSIN is presented at a "loud but ok" level which is typically around 70 to 75 dB HL. Once the audiometer and external settings are in place, there is no other manual manipulation; it is all done on the CD. You begin with two practice lists and then continue with, ideally, at least two list pairs per ear to get your test score. It takes ten minutes to perform the practice tests and each ear individually, a little more if you choose to do more lists. Once completed, calculate your SNR loss as by adding up the correct number of key words and subtracting that number from 25.5. For example, if your patient scored 19 words correctly (out of 30 total), you would take 25.5 - 19 = 6.5 dB SNR loss. If you perform two lists, you would average both SNR loss scores to get your total (i.e. 6.5 from list 1 and 7.5 from list 2 averaged together would total 7 dB SNR loss.)
To use this information for counseling and hearing aid selection, SNR loss is divided into four levels of "difficulty" or severity of difficulty in noise. An SNR loss of 0-2 dB indicates that once audibility is restored, the individual should do pretty well in most background noise situations. A mild score of 3-6 dB SNR loss estimates that the patient will likely do okay as long as good directional microphones are in use. When a patient scores a moderate 7-12 dB SNR loss, they will experience problems in noise even with good directional microphone and noise reduction technology. And lastly, an SNR loss of greater than 12 dB indicates that the patient will struggle significantly with hearing aids alone and should consider the use of a personal FM system if listening in noise is important to the patient. These are the kinds of patients that require extra counseling and examination of realistic expectations from amplification, even if it is top-of-the-line amplification. This kind of scoring breakdown makes it easy to see that you would tailor your counseling very differently for each group. It helps you to be more precise in setting and achieving expectations.
Background noise is obviously very annoying from a normal-hearing standpoint, but how much greater is it for the person with an impaired auditory system? Audiologists who have been fitting hearing aids for some time are acutely aware of the group of patients who, no matter what you do, continue to struggle in background noise. But what happens if they struggle more with annoyance and not necessarily intelligibility? The Acceptable Noise Level test or the ANL (Nabelek et al., 1991; 2004; 2006) quantifies what level of noise patients deem as "acceptable."
Here are some things we know about the ANL (Nabelek et al., 2006). The ANL score is around 10 dB for both normal and hearing-impaired individuals. Although the ANL is similar to an SNR loss, it is not correlated with SNR intelligibility measures. ANL scores greater than about 13 are associated with hearing aid failure, meaning failure of not using the hearing aids. We also know that scores less than about 8 are associated with longer hearing aid use. Additionally, the scores are not related to age, gender, or degree of hearing loss. The ANL is measured by establishing the patient's most comfortable listening level (MCL) and increasing the background noise over the MCL over a period of trials. A listener with a low ANL score can accept greater levels of background noise without being annoyed. A listener with a high ANL has a tendency to be an unsuccessful hearing aid user. "Unsuccessful" may be defined as a measure of time, such as, "I only wear my hearing aids two hours a day or less." Nabelek and her University of Tennessee colleagues (2006) also found that if you define hearing aid success as the number of hours per day that you wear hearing aids, the ANL can actually predict that with about 83% accuracy. The ANL, therefore, is a great way to identify red flags for non use.
Calculating ANL is very simple. First, establish the patient's MCL using recorded running speech. Then, introduce background noise and increase in 2dB steps until the patient tells you they can no longer tolerate the running speech at MCL with the amount of background noise with which they are competing. Now subtract the background noise level (BNL) from the MCL to calculate ANL (e.g. MCL-BNL=ANL). This is not an intelligibility test. It is strictly a test of annoyance. Nabelek et al (2006) tested different groups in the unaided condition to identify an average ANL based on hours per day of hearing aid use. The groups were divided into full-time, part-time, and non-users. Full-time users had an ANL score of 7; part time users 13.5 and non users 14. Keeping in mind that a lower score means higher noise tolerance, there was a noted difference between the full-time user group and the part-time user group. This suggests that people who wear their hearing aids only an average of four hours a day or less (part-time) have a significantly higher annoyance with noise.
A study published in The Hearing Journal (Taylor, 2008) looked at the relationship between the ANL and the International Outcome Inventory for Hearing Aids (IOI HA) (Cox et al., 2000; Cox, Stephens, and Kramer, 2002) which is a fairly common self report outcome measure that is used in research. We separated patients into one of three possible groups when tested in the unaided condition: low (0 to 6), medium (7 to 12) and high (≥13) ANL scores. After they were categorized based on their unaided ANL score, they were fit with hearing aids the way you typically would in a clinic, and then 30 to 45 days after the fit we measured their outcome on the IOI HA. The results were almost too good to be true. Patients who had a very low outcome on the IOI HA had a tendency to be the ones with highest ANL scores as well. This study concluded that the ANL might be a good predictor of final outcome.
Objective Intelligibility Testing
The last test to mention is the Performance Perceptual Test (PPT) which was created by Dr. Gabrielle Saunders (2002). The PPT compares two things: how the patient thinks they understand speech and how they actually understand speech when measured objectively. Dr. Saunders added an additional measure to the standard PPT, the performance perceptual discrepancy (PPDIS), which is a measure of how a patient might misjudge his or her individual hearing abilities. To find out the PPDIS, you begin with the actual performance component and have the patient repeat HINT sentences and calculate their performance in noise. Secondly, you gather the perceptual component by asking the patient how well they can understand sentences in noise. You are looking at the disconnect between actual performance and the patient's perception of their performance. Patients may over- or underestimate their ability in noise, and some may have an accurate reflection of their performance. The PPT can be used as a counseling tool when you the category in which your patient falls after subtracting perceived SNR loss from measured SNR loss. Underestimators score a -3 or worse, meaning that they actually do fairly well in noise, but they judge their performance to be much worse. Accurate patients have a PPDIS from -3 to 1, meaning that their performance and idea of their performance are not far off. The overestimators score 2 or higher, and are often in denial about how well they perform in noise. They often have more difficulty than they acknowledge. By using the PPT, you can identify potential hearing-aid related problems at the outset before the patient is even fit. Saunders and Forsline (2006) developed a five-step PPT-based counseling approach that can be used with results from the PPT and PPDIS.
Which Test is Superior?
Mueller, Johnson, and Weber (2010) published an article on AudiologyOnline comparing the QuickSIN, ANL, and PPT tests. The study looked at 20 satisfied hearing aid users and their results on each of the three pre-fitting speech tests. They found that the QuickSIN, ANL, and PPT scores are not correlated. In short summary, all three of these tests have a place in the clinic during the prefitting assessment because all three of these tests measure something different. Therefore, each has the potential to contribute to a more successful fitting. They all bring a unique score to the table to help you make a better clinical decision. It is up to you individually to determine the value of the testing to you and your patients.
Additional research that is being completed at the date of this article will be presented at the AudiologyNOW 2011 conference in Chicago regarding a matrix Unitron has cooperatively developed with Jill Bernstein to evaluate potential patient benefit and satisfaction at prefitting using the ANL and QuickSIN measures.
Because background noise is a real, everyday problem for hearing aid users, the implementation of test measures to estimate the degree of difficulty is beneficial to clinicians in selecting appropriate amplification. As studies show that background noise and a patient's perception of background annoyance and tolerance can affect hearing aid use, we can use speech-in-noise tests as a positive counseling tool to help patients evaluate their expectations and reach their listening potential.
Akeroyd, M.A. (2008). Are individual differences in speech reception related to individual differences in cognitive ability? A survey of twenty experimental studies with normal and hearing-impaired adults. International Journal of Audiology, 47(s2), S53-S71.
Bench, J., Kowal, A., & Bamford, J. (1979). The BKB (Bamford-Kowal-Bench) Sentence Lists for partially-hearing children. British Journal of Audiology, 13, 108-112.
Cox, R.M., Alexander, G.C. & Gilmore, C.A. (1987). Development of the connected speech test (CST). Ear and Hearing, 8(S): 119S-126S.
Cox, R.M., Hyde, M., Gatehouse, S., Noble, W., Dillon, H., Bentler, R., Stephens, D., Arlinger, S., Beck, L., Wilderson, D., Kramer, S., Kricos, P., Gagne, J.P., Bess, F., Hallberg, & Lillemor. (2000). Optimal outcome measures, research priorities, and international cooperation. Ear and Hearing, 21(4), 106S-115S.
Cox, R.M., Stephens, D., & Kramer, S.E. (2002). Translations of the International Outcome Inventory for Hearing Aids (IOI-HA). International Journal of Audiology, 41, 3-26.
Etymotic Research. (2005). Bamford-Kowal-Bench Speech-in-Noise Test (Version 1.03) [Audio CD]. Elk Grove Village, IL: Author.
Etymotic Research. (2001). Quick Speech-in-Noise Test [Audio CD]. Elk Grove Village, IL: Author.
Kalikow, D.N., Stevens, K.N., & Elliott, L.L. (1977). Development of a test of speech intelligibility in noise using sentence materials with controlled word predictability. Journal of the Acoustical Society of America, 61, 1337-1351.
Killion, M. C., Niquette, P. A., Gudmundsen, G. I., Revit, L. J., & Banerjee, S. (2004). Development of a quick speech-in-noise test for measuring signal-to-noise ratio loss in normal-hearing and hearing-impaired listeners. The Journal of the Acoustical Society of America, 116, 2395-2405.
Kochkin, S. (2010). MarkeTrak VIII: Consumer satisfaction with hearing aids is slowly increasing. The Hearing Journal, 63(1), 19-32.
Mueller, H.G., Johnson, E.E., & Weber, J. (2010, March 29). Fitting hearing aids: A comparison of three pre-fitting speech tests. AudiologyOnline, Article 2332. Direct URL: www.audiologyonline.com/articles/article_detail.asp?article_id=2332 Retrieved February 9, 2011, from the Articles Archive on www.audiologyonline.com
Nabelek, A.K., Freyaldenhoven, M.C., Tampas, J.W., & Burchfield, S.B. (2006). Acceptable noise level as a predictor of hearing aid use. Journal of the American Academy of Audiology, 17(9), 626-639.
Nabelek, A.K., Tampas, J.W., & Burchfield, S.B. (2004). Comparison of speech perception in background noise with acceptance of background noise in aided and unaided conditions. Journal of Speech, Language, and Hearing Research, 47, 1001-1011.
Nabelek, A., Tucker, F., & Letowski, T. (1991). Toleration of background noises: Relationship with patterns of hearing aid use by elderly persons. Journal of Speech and Hearing Research, 34, 679-685.
Nilsson, M., Soli, S.D., & Sullivan, J.A. (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.
Niquette, P., Arcaroli, J., Revit, L., Parkinson, A., Staller, S., Skinner, M., & Killion, M. (2003, March). Development of the BKB-SIN Test. Paper presented at the annual meeting of the American Auditory Society, Scottsdale, AZ.
Rogin, C. (2009, October 13). Top ten reasons for hearing aid delight. AudiologyOnline, Recorded Course 14615. Retrieved February 5, 2011, from the e-Learning section on www.audiologyonline.com
Saunders, G., & Cienkowski, K. (2002). A test to measure subjective and objective speech intelligibility. The Journal of the American Academy of Audiology, 13, 38-49.
Saunders, G.H, & Forsline, A. (2006). The Performance-Perceptual Test (PPT) and its application to hearing aid counseling. The Hearing Review, 13(13), 18-25.
Taylor, B. (2008). The Acceptable Noise Level Test as a predictor of real-world hearing aid benefit. The Hearing Journal, 61(9), 39-40, 42.
Wilson, R. H. (2003). Development of a speech in multitalker babble paradigm to assess word-recognition performance. Journal of the American Academy of Audiology, 14, 453-470.
Wilson, R. H., & Burks, C. A. (2005). The use of 35 words to evaluate hearing loss in terms of signal-to-babble ratio: A clinic protocol. Journal of Rehabilitation Research and Development, 42, 839-852.
Wilson, R.H., & McArdle, R. (2005). Speech signals used to evaluate functional status of the auditory system. Journal of Rehabilitation Research & Development, 42(4), 79-94.
Wong, P.C., Ettlinger, M., Sheppard, J.P., Gunasekera, G.M., & Dhar, S. (2010). Neuroanatomical characteristics and speech perception in noise in older adults. Ear and Hearing, 31(4), 471-479.