Famed Nobel Laureate, physicist, Richard Feynman, once said, “The statements of science are not what is true or not true, but statements of what is known with different degrees of certainty.” Two recently published papers, each examining a different angle of the hearing aid fitting process, demonstrate how Feynman’s aphorism accurately describes some of the issues facing hearing aid dispensing practices in the age of disruptive technology and the application of good clinical evidence.
Most would agree the commercial price of hearing aids is directly related to the time and expertise needed in the customization and counseling provided by a licensed professional. A big part of the customization process is tailoring the acoustic parameters of any device to optimize audibility and intelligibility of speech for the patient.
Customization of gain, output and frequency response characteristics of hearing aids are achieved using prescriptive fitting formulas such the NAL-NL2 target. Although scientifically validated prescriptive fitting formulas have been available for more than 40 years, computerized hearing aid software has long sought to make the process of hearing aid verification less time consuming and more convenient by using so-called “first fit” approaches.
Today, “first fit” approaches, which do not include any best practice real ear measurements in their application, are the most popular method for fitting hearing aids. Surveys regularly demonstrate that 70-80% of clinicians do not routinely use real ear measures in the fitting process. However, numerous studies published over the past 15 years shown these “first fit” approaches have many shortcomings.
Score Another One for Real Ear Measurements
For clinicians (and patients) wondering exactly what is lost when real ear measures are not completed during the hearing aid fitting process, a recent article in JAAA, written by audiologists at Washington University in St. Louis, is a must-read. In a blinded, crossover design, lead author Mike Valente and colleagues compared subjective and objective outcomes for a group of 25 hearing aid users who used both a “first fit” and “programmed” hearing aid setting.
All study participants were fitted bilaterally with 16 channel RIC BTEs from the same manufacturer. For the “programmed” fit the audiologists used real ear measures to match the NAL-NL2 fitting target +/- 5 dB for soft, average and loud input levels. Participants were able to wear each of the two fitting approaches (first fit and programmed) for about a few weeks.
A series of aided speech testing was performed, including the Hearing in Noise Test (HINT), as well as subjective outcome measures, including the APHAB. Although there were some interesting individual differences, especially in background noise, there was a significant advantage favoring the “programmed” fit for objective speech tests at 50 dB input levels in quiet. Additionally, results indicated nearly 80% of the patients preferred the “programmed” fit over the “first fit” approach.
In an era in which patients have many hearing aid purchasing options, this study suggests the routine use of prescribed fitting targets verified with real ear measures is much more likely to lead to superior patient outcomes, and thus a patient’s investment in customized hearing aids is worthwhile.
Individualizing the Hearing Aid Fitting with a Self-Adjustment App
Although the Valente et al study provides further evidence that verifying the closeness of a prescribed fitting target with real ear measures is more likely to lead to successful patient outcomes, another recent study, indicates that when given control over key hearing aid features, individuals are likely to take a “programmed” fit and make it even better. In a laboratory study conducted at the University of Minnesota, Peggy Nelson and colleagues, examined the usefulness of a self-adjusted fitting technique to find preferred settings for both quiet and noisy listening situations.
In their study a total of 30 participants, aged 59 to 78 with mild to moderate hearing loss, were provided a mobile application (app) running on an iPod coupled to both ears with insert-type earphones, essentially turning the device into a multiband hearing aid in which the acoustic parameters could be adjusted by the end-user with two controllers on the iPod screen. Prior to the listening trials, the NAL-NL2 prescriptive formula, verified with real ear measures, was used to establish the default starting point. The participants were then asked to use the app to establish their preferred listening levels for a series of quiet and noisy listening conditions. A subset of the 30 participants also completed speech intelligibility in noise testing at various signal to noise ratios for both the “fit to target” and “self-adjusted” settings.
Results of the study are quite thought-provoking. First, participants were quite consistent in their self-adjustments, as the average test-retest gain difference within the group was 5.6 dB for the low frequency bands and 6.9 dB for the high frequency bands. Across all retests, 54.6% were within 5 dB of the first self-adjusted fit and 80.7% were within 10 dB of the first self-adjusted fit, indicating participants were able to make consistent self-adjustments with the app.
Individual differences between the participants, however, were remarkable. Differences between the NAL-prescribed starting point and the self-adjusted setting were as large as 24 dB in the low frequencies and 37 dB in the high frequencies. In general, most participants gave themselves more gain than prescribed in the low frequencies, while the majority chose less gain than prescribed in the highs.
The key finding, though, was the considerable amount of individual variation in preferred gain settings. When given an opportunity to self-adjust, several of the participants deviated quite far from the prescribed NAL settings.
Perhaps the most interesting finding in the study was how self-adjusted deviations from the NAL-prescribed starting point for gain affected speech intelligibility. Although individual self-adjustments away from the NAL-prescribed gain targets were often quite large, they did not alter substantially speech intelligibility performance. More than 80% of the time, according to the authors, speech intelligibility performance was “within 10% points of performance with NAL gain in the same SNR.” For individuals with mild to moderate hearing loss, this study indicates that a user-controlled app has the potential to take an audiologist-driven fit with the primarily goal of optimizing audibility and enable the individual to make it even better.
Does it Matter Who Does the Customization?
Taken together, these two studies show the value of customization of hearing aid fittings. In the case of the audiologist-driven fit, taking the time to verify with real ear measures that a validated target is matched is more likely to result in superior patient outcomes. However, if hearing aid adjustments that are ordinarily conducted by clinicians with fitting software are handed over to patients through a vetted smartphone app – even if the patient self-adjusts quite far from the NAL-derived starting point – it can also result in positive patient outcomes.
At the end of the day, it probably doesn’t matter who does the customization as long it leads to real world patient benefit and satisfaction. Isn’t it encouraging to know that persons with hearing loss could, in the very near future, have two proven approaches to getting optimal hearing aid performance in multiple listening situations? One is the tried and true audiologist-driven fitting approach, relying on matching a scientifically-derived prescriptive target and verifying the real ear measures. On the other hand, for individuals who want to actively participate in the customization of their hearing aids settings, scientifically-derived algorithms that are controlled with an easy-to-use smartphone-enabled app could be a valuable addition to an audiologist’s clinical repertoire.
Neither approach alone provides all the answers, but together they might get us a little closer to more certain patient outcomes.
The Nelson et al study was published in the open access journal, Trends in Hearing and can be found here: http://journals.sagepub.com/doi/full/10.1177/2331216518798264
The Valente et al study, published in the Journal of the American Academy of Audiology (JAAA) is found at this gated site: http://aaa.publisher.ingentaconnect.com/contentone/aaa/jaaa/2018/00000029/00000008/art00005
*Featured image courtesy Wesley Elliott, Army Medicine