Historically, audiologists and others who fit hearing aids often have relied on the results of pre-fitting tests to predict patient benefit from amplification in everyday listening situations. Traditional word-recognition testing or speech-intelligibility-in-noise tests have frequently been used as part of this pre-fitting test battery, although there is little evidence of their effectiveness in predicting real-world hearing aid outcomes.1 This article examines the use of a different speech test, the Acceptable Noise Level Test, to predict real-world hearing aid outcome.
There are a few reasons why audiologists and hearing instrument specialists might be interested in predicting how patients will perform in the real world with hearing aids. One, most clinicians pride themselves on providing patients with a detailed explanation of test results, along with recommendations based on these results. Predicting benefit allows patients to leave the pre-fitting appointment with a more precise and individualized idea as to how much benefit they will receive from hearing aids before they start using them.
Secondly, being able to predict benefit allows the practitioner to position herself or himself as the expert. Experts don't have all the answers, but they are able to gather meaningful information, organize it, and accurately inform patients about what they should expect from a treatment. This concept is similar to what surgeons do prior to an operation. Using test results, they give the patient a prognosis of the expected outcome in everyday living situations,
Finally, all of us feel a little more at ease with a major purchase when we are offered some reassurance that it will indeed work where and when we need it most. If hearing healthcare providers can rely on certain pre-fitting tests to make more accurate predictions of real-world outcomes with hearing aids, patients will be more likely to feel confident about the decision to purchase hearing aids.
USE OF THE ANL TEST
The Acceptable Noise Level (ANL) test has recently emerged as a pre-fitting test procedure. The test, developed some 15 years ago by Anna Nabelek, PhD, and colleagues at the University of Tennessee, measures the listener's reaction to background noise while listening to speech.2 Research has shown the unaided ANL to be a useful predictor of hearing aid use.3 Nabelek et al. suggest that full-time hearing aid users have significantly lower unaided ANL scores than do part-time users or non-users.4 This test also has been used in hearing aid research to demonstrate the advantage of directional microphones in noise.3
In a recent study, Mueller et al. were able to show a significant improvement in aided ANL scores when digital noise reduction (DNR) was activated compared with when it was not.5 Although an improvement in the measured ANL with DNR activated does not equate to real-world hearing aid use or benefit, it is a positive result supporting the use of DNR for combating annoyance with noise.
While we know there is at least a weak correlation between amount of hearing aid use and amount of hearing aid benefit, no previous study has shown the unaided ANL to be a good predictor of other dimensions of hearing aid outcome, such as satisfaction or benefit. The goal of this study was to investigate the pre-fitting predictive value of the ANL for self-reports of hearing aid benefit and satisfaction.
As mentioned, Nabelek et al. reported that unaided ANL scores were predictive of hearing aid usage.4 In their study, they divided the 191 participants, all experienced hearing aid wearers, into three categories based on how much they used amplification. Group 1 (n = 69) consisted of full-time hearing aid users, subjects in Group 2 (n = 69) were classified as part-time users, and Group 3 (n = 53) was made up of non-users.
The researchers found that full-time users could accept more background noise than the part-time users and non-users. The amount of hearing aid use could be predicted by comparing the ANL scores of part-time and non-users with those of full-time users. Full-time/successful users had consistently lower ANL scores than part-time and non-users, meaning they could tolerate greater background noise.
Nabelek et al. demonstrated that patients with low (<7) or very high (>13) ANL scores could be classified as successful or unsuccessful users, respectively.4 Their study, however, used subjects who had been wearing hearing aids, and asked them to classify their use time by completing a questionnaire during their first year of hearing aid use. Thus, they did not use the unaided ANL as part of the pre-fitting assessment test battery.
Two clinics collected data for the study reported here. Twenty-seven people participated in the study over a 2-month period. They were drawn at random from a larger group of people, all of whom had met the following criteria: They were first-time users with no worse than a severe high-frequency hearing loss who had been fitted bilaterally with digital hearing aids using multi-channel wide dynamic range compression. The 27 subjects ranged from 56 to 82 years of age, with a mean of 68 years. All had sensorineural hearing loss without any medical complications. Their pure-tone thresholds had to be no worse than 50 dB at 500 Hz or 75 dB at 4000 Hz.
Pre-fitting testing consisted of the pure-tone audiogram, the NU-6 word-recognition test (in quiet), and the unaided QuickSIN.6 All speech audiometry testing was conducted bilaterally under earphones. The presentation level for speech audiometry was conducted at the same intensity level at a “loud MCL” by using a bracketing procedure and asking the patient to let the audiologist know when a “loud, but okay MCL” had been achieved.
The ANL test was conducted as part of the pre-fitting test battery. The original ANL test,4 which uses a male talker as the primary stimulus and 12-talker babble as the competing background noise, was conducted bilaterally under earphones.7
In the ANL test procedure, the listener is asked first to make a most comfortable loudness (MCL) judgment with the male talker as the primary stimulus. Once the MCL is determined, background noise is introduced and the listener is asked to determine the background noise level (BNL). The BNL is defined as the maximum level of background noise the listener can “put up with” without becoming tense or tired while still listening to the primary talker.4 The BNL was introduced in 2-dB steps and conducted twice. The ANL is calculated by taking the difference between the MCL (dB HL) and BNL (dB HL).
All hearing aids were fitted using standard clinical protocols, including the matching of a prescriptive gain target (NAL-NL1) verified with probe-microphone measures. Participants were able to return for hearing aid adjustments as needed during the initial 30 days of use. Approximately 30 days post-fitting, all participants were asked to complete the International Outcome Inventory for Hearing Aids (IOI-HA).6 Participants were administered the IOI-HA while waiting in the reception area during a routine post-fitting appointment.
The IOI-HA was selected for this study because it captures several domains of real-world outcome and is easy to administer in a busy clinical setting. Each question on the IOI-HA has five possible answers on a 1 to 5 scale. It can be scored in the following manner: The answer for each question corresponding to the best outcome is given a 5, and the answer corresponding to lowest outcome is given a 1.
The creators of the IOI-HA determined that the seven questions could be grouped into two separate factors. Factor 1, which comprises questions 1, 2, 4, and 7, is interpreted as encompassing introspection about the hearing aids (“me and my hearing aids”). Factor 2, comprising questions 3, 5, and 6, is interpreted as reflecting the influence of the hearing aids on the individual's interactions with the outside world (“me and the rest of the world”).6
Figure 1 shows the dispersion of overall IOI-HA scores for the 27 participants. An analysis of variance shows a significant correlation between the unaided ANL and the overall IOI-HA score (F = 6.872, p<0.05). That is, as the overall IOI-HA score became smaller (poorer), there was a tendency for the individual ANL score to be higher. i.e., poorer.
Participants in this study were assigned to one of three groups based on their unaided ANL test score. The classification of these three groups was based on the Nabelek et al. study mentioned previously.1 Table 1 shows the number of participants in each of the three unaided ANL groups.
Figure 2 shows the mean overall scores on the IOI-HA for the three ANL groups. Although group 3 with the highest un-aided ANL scores has significantly lower overall scores on the IOI-HA, results indicate that all three groups, on average, are receiving a favorable outcome from amplification when compared with published IOI-HA norms.7
To investigate how much of the variance in overall IOI-HA scores could be accounted for from the unaided ANL, a regression analysis on the overall IOI-HA was employed. Results showed that the ANL test accounted for 16.8% (R-value = 0.41) of the variance of the overall IOI-HA score, a significant finding (F = 5.054, p = 0.033).
For the Factor 1 and Factor 2 sub-questions of the IOI-HA, independent analyses of variance were performed to determine if either factor was correlated with the unaided ANL. For Factor 1, which is considered to comprise intrinsic dimensions of outcome, the unaided ANL score is a strong predictor of responses to questions 1, 2, 4, and 7 collectively (F = 12.253, p = 0.002), accounting for 32% of their variance. Figure 3 shows the relationship between the unaided ANL scores and responses to the Factor 1 sub-questions of the IOI-HA. Observe that the four participants with the highest unaided ANLs had the lowest scores on Factor 1 questions.
For Factor 2, which consists of extrinsic aspects of hearing aid outcome, the unaided ANL score had a non-predictive relationship to responses to questions 3, 5, and 6 collectively (F = 3.487, p>0.05) and did not reach a level of statistical significance.
These results suggest that for intrinsic dimensions of hearing aid outcome (“me and my hearing aids”) the unaid-ed ANL score is a very strong predictor of final outcome as measured by Factor 1 of the IOI-HA. This finding suggests that the unaided ANL score can predict intrinsic factors related to hearing aid outcome, such as how patients feel about their hearing aids.
The purpose of this study was to examine the predictive relationship between the unaided ANL and self-reports of hearing aid outcome as measured on the IOI-HA. Unlike the Nabelek et al. study, the ANL scores in this study were collected before participants wore hearing aids. Therefore, this study directly reflects how the unaided ANL might be used in daily clinical practice to predict hearing aid benefit.
Although the sample size was small for two of the three ANL groups, there is a statistically significant trend suggesting that the unaided ANL score contributes to the final outcome as measured by the IOI-HA, with 16% of the overall IOI-HA score accounted for in the unaided ANL.
More specifically, there is a significant finding that Factor 1 variables on the IOI-HA can be predicted from the unaided ANL score. Given the strong predictive relationship between Factor 1 (“me and my hearing aid”) IOI-HA scores and the unaided ANL, knowing the unaided ANL scores before the fitting provides quantifiable insights into how patients perceive their handicap prior to using hearing aids. In other words, the results of this study suggest that taking the time to measure the patient's unaided ANL can contribute to the pre-fitting counseling process in a meaningful way.
Past research on the ANL has shown it to be a good predictor of hearing aid usage.4,5 Although this is a positive finding, hearing aid use does not necessarily equate to overall hearing aid satisfaction and benefit. Results from this study are among the first data to show that it may be possible to predict hearing aid benefit from the unaided ANL score. Given this finding, the ANL would appear to be a valid predictor of self-reports of hearing aid benefit and could be used as a standard part of the pre-fitting clinical test battery.