Years ago, the usual audiometry session consisted of pure-tone threshold testing across frequency (the audiogram) and, for a “comprehensive” examination, it included a test of word intelligibility under quiet conditions. Compared to that period, we now have a plethora of tests to evaluate a person’s auditory status. Diagnostic hearing tests can identify abnormalities anywhere in the auditory system, from the middle ear, through the midbrain, and up to the auditory cortex.  A person’s ability to recognize speech can be examined by a host of standardized tests, from the understanding of individual phonemes to the ability to recognize sentences in the presence of background noise interference. The challenge nowadays is no longer to develop additional tools in the clinical armamentarium, but to ensure that the ones we have are administered when necessary.

Concomitant development has also taken place in our ability to evaluate the performance of hearing aids. At first, audiologists were limited to speech discrimination tests of questionable validity in assessing a person’s performance with hearing aids, coupled with the ubiquitous question, “How does this sound?”  Nothing wrong with this question; I have asked it myself and of myself. But a brief listening experience with a new hearing aid or adjustment in some unrealistic situation (say, a quiet office) is no substitute for an objective evaluation of a hearing aid’s performance.

About twenty-five or thirty years ago, when affordable hearing aid test boxes were developed, it became possible for the average hearing center to objectively assess a hearing aid’s performance. Up to this point, an audiologist would have to depend upon the information provide by the manufacturer, i.e., the hearing aid “specifications” (consisting of a description of the hearing aid’s amplification characteristics, special features, etc). The manufacturers obtained this information using an industry-wide standardized test procedure. While important and, indeed, necessary, these specifications did not directly relate to the specific hearing aid being selected for the particular individual. In this procedure, which is still being used by every hearing aid manufacturer, the hearing aid output is delivered into a small cavity (a 2cc coupler) which is intended to simulate the ear-canal dimensions of an average adult; but these results do not apply to any specific human being.  This is a significant point and will be elaborated on below.

The equipment the manufacturers would employ in performing this measure was much too sophisticated and expensive for the average hearing aid center. When commercially available hearing aid text boxes did become available, audiologists could now check the electroacoustic performance of a specific aid for themselves. They no longer had to depend solely on the specifications provided by the manufacturer. Additionally, audiologists could now supplement the standardized tests run by the manufacturer with additional ones applicable specifically to the particular individual being fit with the hearing aid. This capability was a major step forward in the entire hearing aid selection process and is still very useful as a broad guide in the initial selection of a hearing aid for a client. But, as with the manufacturer’s specifications, the information developed with these new hearing aid test boxes was still based on the dimensions of a coupler that simulated the presumed ear canal cavity of an average ear.

With the advent of digital hearing aids, it became necessary to program a hearing aid before fitting it. (Previously, hearing aids were simply adjusted with a screwdriver!) Programming requires the incorporation of specific amplification target goals. These targets are based on extensive clinical research and can be used “as is” or modified for particular individuals.  In the usual programmer, targets and their variations are displayed in a visual format on a computer screen. Adjustments can be made while the aid is being worn, with the consequent listening changes evaluated by the hearing aid user. Each such adjustment is also visually displayed on the computer screen. However, the response curves that are shown on the screen display only the relative changes in the amplification characteristics of the hearing aid, i.e., in comparison to one another. They do not actually show the amplified sound pattern actually occurring in an individual’s ear canal.

Unfortunately, many hearing aid dispensers appear to view either or both the coupler measures and/or the programming displays as indications of the amplified sound patterns being delivered into a person’s ear canals. However, as already stated, neither the coupler responses nor the programming displays are meant to apply to any specific person. Consider children, for example: We know that – because children’s ear canals are much smaller than adults – a given hearing aid’s output would be much higher for them than it would be for grown ups. (This is because greater sound pressure is produced in a smaller cavity relative to a larger cavity, given the same sound input). Thus, if audiologists depended solely on coupler specifications to fit a hearing aid to a child, they could be overamplifying the child, i.e., providing too much sound. On the other hand, if the ear canal of some adult were larger than the average, the actual hearing aid output would be less than the coupler indications in that person’s ear. This individual could be underamplified, perhaps for some frequencies more than others. A truly accurate hearing aid fitting would take both of these possibilities into consideration.

Programming displays also have limited (though still very useful) applicability. These do not include the possible acoustic variations introduced by microphones, the hearing aid receiver (the “loudspeaker”), or the variability in individual ear canal dimensions. Therefore, while both coupler measures and programming displays are extremely useful, they do not tell us the actual sound energy being delivered by a particular hearing aid in the ear canal of a specific individual. For that, we need real-ear measures.

In real-ear measures, what is displayed is the actual acoustic energy that exists within the ear canal of a particular person. They can be done with and without a hearing aid.  When a person is wearing an aid, a flexible tube is inserted alongside the hearing aid and terminates between the tip of the earmold and the eardrum. This tube leads from the ear canal to a microphone situated outside the ear.  The sounds detected by this “probe-tube” system reveal the real-ear output of the hearing aid for the individual wearing it.  It’s not a guess and not an estimate; it depicts the actual sounds reaching the eardrum of the hearing aid user. No other measure of hearing aid performance can do this. Real-ear measures should be done routinely during every hearing aid selection procedure and repeated routinely during follow-up appointments (more on this below).

Figure 1 shows the basic set up of a real-ear measure for the unaided condition. The probe tube can be seen leading from a small black box (containing the microphone and associated electronics) into the ear canal. Figure 2 shows the set-up with a hearing aid in place. In this figure, the tube has been inserted alongside the hearing aid; its termination point is slightly beyond the tip of the earmold.

Audiologists have many options when selecting the specific type of test stimuli to use. The first generation of real-ear systems employed a tone stimulus that swept from the low to the high frequencies. With the development of multi-band hearing aids, many of which included various types of automatic functions (automatic gain control, noise management, feedback cancellation, etc.), it became necessary to develop stimuli other than pure-tones with which to evaluate a hearing aid’s performance.  Composite noise, which simulates exposure to an average speech signal, was introduced and is still being used. With this stimulus, it is possible to predict how much speech audibility is being provided by a particular hearing aid; however, composite noise is still an artificial stimulus. What is clearly not an artificial stimulus is an actual speech signal. In the newest generation of real-ear measurement systems, real speech is being used as test stimuli. This added capability has been termed “speech mapping.”

In using an authentic speech signal it is possible to test a hearing aid with all its special features operative. The overall impact of these features can then be viewed just as they would normally affect a speech signal in real life. Effectively, with a speech mapping procedure, what you see (on the computer screen) is what you get. Of course, an actual speech signal is ever-changing as words flow and intonation varies. It is, however, possible to freeze a snapshot of the speech signal at any instant in time and to depict an acoustical average of the speech energy. The signal can be generated “live”; that is, a spouse or parent can talk into the hearing aid, or prerecorded stimuli can be used. Certainly, there is no more valid test signal than real speech. What makes the display particularly valuable, however, is that the speech signal is charted as an audibility curve relative to the threshold of hearing, thus displaying just what a person can or cannot hear. This is shown in Figures 3 and 4.

The test results are plotted in reference to the familiar audiogram, as depicted by the blue line with the X’s. All speech sounds that fall below the hearing thresholds are inaudible (i.e., not loud enough for a person to hear). The asterisks at the various frequencies indicate the point at which sound becomes too loud. The major amplification goal is to “package” the speech energy between the threshold of hearing and the threshold of discomfort. In other words, speech should be loud enough to be heard but not so loud as to be uncomfortable. The green striped area in Figure 3 shows the speech signal. As can be seen, the acoustical energy in a speech signal stretches from low to high frequencies, with an intensity range of approximately 30 dB at any frequency. The green line running through the striped area depicts the average level of the speech signal.

Now note, in figure 3, how the amplified speech signal is mainly audible only in the low frequencies. This person (and yes, this represents a real human being) can only hear those amplified speech sounds that fall above (higher than) his auditory thresholds. However, as can be observed, most of the high frequency energy in the speech signal falls below (i.e., is less than) the thresholds. This represents the part of the speech signal that is inaudible to him. This person can “hear” alright, because of the low frequencies, but would have a great deal of difficulty “understanding,” because he cannot hear the high frequencies. Of course, there are a number of other reasons why people with hearing loss complain that they can often “hear but not understand,” but one major one is displayed here, in a visual form that is very easy to understand.

But this situation is relatively easy to correct by reprogramming the hearing aid and, with a real-ear system, directly observing the results of the modification. This can be seen in Figure 4. This is the same person, but his hearing aid has now been adjusted for a more appropriate amplification pattern. Now note where the green area is located (remember, this depicts the energy in a real speech signal); the entire amplified speech signal falls above (is higher than) his auditory thresholds, but is below the threshold of discomfort. The entire speech signal is now audible.  This looks to me like an excellent hearing aid fitting. While other dimensions will often enter into the hearing aid fitting process (i.e. different input sound levels, compression characteristics, etc.) audibility still has to be the major goal in fitting a hearing aid. If a person cannot hear the speech signal, there is no way in the world he or she is going to be able to understand it.

The differences between Figures 3 and 4 represent just one example of the value of doing real-ear measures. Many other examples can be given. The interactions between a particular person’s hearing loss and his or her specific hearing aid requirements are uniquely personal. Each person has to be tested individually. I do not understand how hearing aid dispensers can fit hearing aids without knowing whether the aid is providing their clients with the most appropriate level of audibility across frequency. It’s worth repeating: There is no more important dimension to comprehending a speech signal than audibility. It is impossible to comprehend a sound signal that one doesn’t hear at all.  When heard just partially, some degree of comprehension is often possible, thus obscuring an inadequate hearing aid fit; however this degree of comprehension would come at a cost of much effort, lots of guesses, and many errors. Real-ear measures are not perfect, but they do take the hearing aid fitting process one major step forward.

Clearly, therefore, real-ear tests can provide hearing aid dispensers with valuable hearing aid fitting information. Unfortunately, while about fifty percent of dispensers possess the requisite equipment, less than a quarter “almost always” administer real-ear tests to their clients.  One reason they frequently give is “that it takes too much time.”  In my opinion, considering the stakes involved for people being fit with hearing aids, there is little merit to this argument.  According to some leading experts in this field with whom I have consulted, it takes approximately ten to fifteen minutes to run a complete real-ear test battery. This doesn’t sound like too much time to me. It’s not as if there is any serious professional disagreement about the merits of real-ear testing. According to the “Guidelines for the Audiological Management of Hard of Hearing Adults” recently published by the American Academy of Audiology, the inclusion of these measures is strongly recommended in all hearing aid fittings.

One recent hearing aid development may help undercut the “too much time” argument. Starkey recently developed a line of hearing aids in which one’s personal aid doubles as a real-ear measuring device. In this procedure, the flexible probe-tube leads back to the microphone of the hearing aid itself,  instead of outward to an external microphone (as seen in Figure 5). The hearing aid microphone now becomes the real-ear measuring microphone. The aid is prepared for personal programming, just as in any digital hearing aid, with the aid tethered to the external programming unit by a wire connecter.

However, the programmer, instead of simply controlling and adjusting the aid’s electroacoustic characteristics, is now also designed to both measure the amplified sounds in the ear canal as well as to produce the sound stimuli (via the hearing aid receiver) that is to be measured. It evidently requires only a little more time to set up than would be necessary in fitting any hearing aid, but with accurate real-ear measures as a pay-off.  Using one’s personal hearing aid in this fashion is rather a neat development.

One other interesting feature of this system is that is also compares the real-ear measure with the actual coupler response for that specific hearing aid (completed in the factory and stored in the aid). The difference between these two measures (real ear and coupler) gives what is termed the “real ear to coupler difference” (RECD).  This can be a powerful metric, particularly when fitting hearing aids to young children, since only one probe-tube measure need be done; any further analysis of the hearing aid’s performance can then be accurately predicted via easily obtained coupler measures.

In Summary

Real-ear measures provide the most accurate portrayal of the sound amplification characteristics of a particular hearing aid for a specific individual. While other means of evaluating the performance of a hearing aid are valuable, they should not be used in lieu of real-ear measures. It is this latter measure, and only this latter measure, that shows the final product of the amplification process. While there may be amplification dimensions that cannot be displayed, at least easily, at the present state of the art real-ear measures are our best bet to actually see what it going on where it counts the most – in a person’s real-ear.  How people actually then use this raw acoustic information available to them is another, though equally important, question, one that takes us to another often neglected area, i.e. aural rehabilitation.