Envision a hearing aid fitting in which the patient's audiogram is entered into the computer, the hearing aids are inserted into the patient's ears, and, with the push of a button, the hearing aids are adjusted for the patient's hearing loss. You're done! Now you can spend the remainder of the appointment developing a rapport with the patient, counseling on communication techniques, and explaining the use and care of the new devices.
At first glance this might seem appealing, but we must remember that the typical goal of any hearing aid fitting is to achieve acceptable audibility for speech. Despite the sophistication of modern fitting software, initial-fit algorithms, also known as first-fit algorithms, are unable to guarantee appropriate audibility. Thus, it is inappropriate to rely exclusively on the fitting software.1-3
“ ...while not a complete substitute for traditional real-ear measurement, Integrated Real-Ear Measurement offers a significant benefit to patients... ”
Due to several factors, including differences in fitting philosophy, hearing aid size and style, microphone location, vent length and diameter, and variability in the size and shape of the ear canal, the only way to ensure audibility is to complete an in situ measure of the hearing aid response.
There is compelling evidence to support the use of the “gold standard” probe-microphone measurement, and we recommend that clinicians use some form of real-ear measurement for every hearing aid fitting.
However, the fact is, multiple surveys have demonstrated that the majority of audiologists and hearing instrument specialists do not routinely use real-ear verification.4-7 Commonly cited reasons for not making probe-microphone measurements include the cost of the equipment, space limitations, not enough time, and the complexity of determining which hearing aid features should be disabled during the measurement and which input stimulus should be used.8
Finally, many practitioners report discontinuing the use of probe-microphone measures because patients are dissatisfied with the sound quality even after the prescriptive targets have been matched, which then requires the dispenser to do even more fine-tuning. This ultimately forces the clinician to decide if the extra time it takes to complete the real-ear measurement is beneficial.
In response to practitioner suggestions and to the compelling evidence that dispensing professionals are not routinely using probe-microphone measurements, Starkey Laboratories introduced Integrated Real-Ear Measurement as an available feature on its Destiny and Zōn product lines. Although Integrated Real-Ear Measurement (IREM) is not a complete replacement for traditional probe-microphone measurements due to the lack of an external speaker and thus of external stimulus, it does introduce an easy, inexpensive way of integrating real-ear measurements into the fitting process.8 Using the IREM included in the Destiny and Zōn product lines takes less than a minute, requires no external equipment, and enables the fitting software (Inspire OS) to account for the patient's unique ear canal characteristics.
Most clinicians will not accept this new approach and have confidence in it until the reliability and validity of this measurement are established. Furthermore, its benefit for both practitioner and patient must be proven. The purpose of this article is to report findings that demonstrate that IREM does indeed measure what it is intended to measure, is repeatable, and provides a more accurate initial fit leading to appropriate audibility for the patient.
TRADITIONAL REAL-EAR MEASUREMENTS
To establish a foundation for the IREM, a brief review of traditional probe-microphone measurements may be helpful. In this article, we will focus primarily on two measures historically used in a real-ear measurement test battery: the real-ear aided response (REAR) and the real-ear-to-coupler difference (RECD). (See ANSI S3.46-1997 for further details.9)
The REAR is a widely accepted and straightforward measure of hearing aid output in the patient's ear canal in response to a known sound field input.10 Since the REAR is a direct measure of a hearing aid's frequency response, the practitioner can quickly determine what is audible or inaudible to the patient using this single measurement. An alternative to measuring the REAR is to predict it from a coupler measurement using the RECD. The commonly known definition of the RECD is the difference between the hearing aid output measured in the ear canal (REAR) and the hearing aid output measured in a 2-cc coupler in response to the same input stimulus via the same transducer.11 A simple illustration of the RECD appears in Figure 1
The RECD is a tool for verifying hearing aid fittings without requiring the cooperation of the patient, by allowing the practitioner to adjust a hearing aid accurately in a 2-cc coupler rather than in the ear. The RECD also plays an integral role in the calculation of the simulated real-ear response displayed in most manufacturers' fitting software. Without a measured RECD, the fitting software will default to an RECD based on an average ear canal.
FITTING SOFTWARE ACCURACY
As mentioned, recent studies have demonstrated that many manufacturers' fitting software may not accurately display or represent the actual sound pressure levels in the patient's ear.12 According to Fabry, this difference is not surprising, since manufacturers' fitting software must apply or estimate correction factors for the style of earmold or shell type, the microphone location effect (MLE), the venting effects, the characteristics of the receiver tube, and the patient's residual ear canal volume.13
While some of the estimates are fairly uniform and therefore more reliable (for example MLEs), others (differences in individual ear canal volume) represent an average with very large variability and thus have a greater effect on the resulting simulation. If, for example, a patient's ear canal volume is larger or smaller than the Burkhard and Sachs RECD values14 used in the Inspire fitting software, the simulated frequency response displayed in the software will not accurately represent the sound pressure levels in the ear canal. Including the patient-specific RECD is a major step in accurately predicting the actual sound pressure levels in the patient's ear.
IREM: HOW DOES IT WORK?
The IREM in the Starkey products overcomes the limitation of using the average data by measuring individual RECDs, and it does so in seconds without additional cost or any external equipment. The IREM is completed using a specially designed probe tube connected to the microphone of the hearing aid. Once the probe tube is placed in the ear canal and the hearing aid is inserted in the ear, the fitting software sends a command to produce a stimulus generated by the hearing aid. The measurement evaluates the patient's real-ear acoustics and compares them to a digitally stored coupler measurement, which ultimately leads to a patient-specific RECD.
This information is then included in every subsequent hearing aid gain adjustment, without need for additional measurement. For more detailed information, see Yanz and Galster15 and Yanz, Pisa, and Olson.8
RELIABILITY, VALIDITY, AUDIBILITY
Whenever an alternative to a traditional measurement method is introduced, questions such as “Can I trust it?” and “How do I know this measurement is accurate?” arise. The following report on the findings of internal and external clinical trials of Integrated Real-Ear Measurement is designed to answer these questions.
Reliability, in this case, can be defined as whether or not a measurement is consistent and repeatable. During development, reliability measures are routinely completed on all products that incorporate IREM.
In an internal study using the recently released Zōn 7 product, Integrated Real-Ear measures were performed twice by the same tester during the same visit on 29 ears. After the first measurement, each device was removed from the ear and the Integrated Real-Ear assembly was uncoupled from the instrument. After the devices were re-assembled to the Integrated Real-Ear probe tube and the hearing aids were re-inserted, the second real-ear measurement was completed.
As shown in Figure 2, the difference between the two RECD measurements was impressively small—less than 1 dB on average and with a maximum variation of 4 dB at high frequencies. This is well within the 5-dB expected variability between test and re-test for probe-microphone measurements reported by Valente.16 Statistical analysis using an intraclass correlation coefficient revealed a very high correlation of 0.96 when data for the frequencies 250, 500, 1000, 2000, 4000, and 6000 Hz were collapsed.
An external study was conducted at the University of Kansas Medical Center.17 The reliability of IREM was evaluated in 20 adult ears using the Destiny 1600 behind-the-ear hearing aid. The aid was coupled to an E.A.R. disposable foam tip with the probe tube inserted to a depth of 28 mm beyond the tragus. The real-ear measurement was completed three times, twice during the first session and once during a second session. The probe tube and foam tip were removed between measurements. The data were analyzed using an intraclass correlation coefficient across frequencies. Results, collapsed across the frequencies of 250, 500, 1000, 2000, 4000, and 6000 Hz, revealed a reliability coefficient of 0.95. A comparison of the internal and external test/re-test data appears in Table 1. Both sets of data are consistent with the 0.91 reliability coefficients of conventional RECDs reported in Sinclair et al.18
Validity is simply whether a test, or in this case a measurement, measures what it was intended to measure. Criterion-related validity19 is used to demonstrate the accuracy of a measurement by comparing it to a procedure that has already been validated. To establish if the Integrated Real-Ear Measurement is indeed valid, we compared RECD measurements obtained from a commercially available system, the Audioscan Verifit, to those measured by the IREM of the Destiny 1600.
Since the two systems use different transducers to measure the RECD, the main obstacle in minimizing sources of variability was devising a method to obtain coupler measurements and real-ear measurements using both transducers. The solution was to create two identical in-the-ear (ITE) shells, one containing a circuit for a Destiny 1600 ITE hearing aid and the other being an exact replica dummy shell with a receiver tube connecting the faceplate and the end of the canal. This allowed us to measure RECDs on 20 ears using both the Verifit and an ITE Destiny 1600. The steps below represent how this comparison was completed.
STEP 1: The Audioscan Verifit RE770 transducer was coupled to the replica ITE shell via the receiver tube on the faceplate side. The ITE shell was then coupled to the HA-1 coupler provided with the Verifit system. The stimulus was presented and the coupler measurement completed.
STEP 2: The Verifit real-ear probe tube was placed in the patient's ear canal 5 mm beyond the medial end of the hearing aid shell. The ITE shell was then uncoupled from the HA-1 coupler and placed in the patient's ear. The same stimulus was presented using the RE770 transducer, and the REAR was measured.
STEP 3: The coupler measurement was subtracted from the REAR to calculate the RECD.
STEP 4: The Destiny 1600 ITE was coupled to the same Verifit HA-1 coupler. The complex tone used for the IREM was presented via the hearing aid receiver. The coupler measurement was stored to the hearing aid.
STEP 5: The ITE was uncoupled from the HA-1 coupler. The probe tube was then placed in the ear canal 5 mm beyond the medial end of the hearing aid shell and the aid was inserted. The complex tone was again presented and the REAR was measured and stored to the hearing aid.
STEP 6: The coupler measurement was subtracted from the REAR to derive the RECD.
Results from this study are shown in Figure 3. The difference between the two average RECDs was no more than 3.1 dB at any point from 250 to 4000 Hz. Of further note, the standard deviation for both measurements was similar, varying from 2.1 to 3.1 dB for the Destiny and from 1.6 to 3.0 dB for the Verifit.
In a clinical trial involving three external sites, 26 participants were fitted bilaterally with Destiny 1600 CIC hearing aids. Real-ear measurements using the Verifit system were completed in 52 hearing aid fittings both with and without IREM. The NAL-R insertion gain fitting formula was used for comparison.
It is apparent from Figure 4 that the frequency response using the IREM provides a closer match to a prescriptive target than when the individual's unique ear canal characteristics are not applied to the initial fit. This higher gain results in better audibility and thus better outcomes for the patient. Further, the measurements are less variable with the measured RECD than with the average RECD.
Many would agree that an important goal, if not the most important goal, of any hearing aid fitting is to ensure that speech is audible and comfortable for the wearer. Because most fitting software relies on average data to simulate the on-screen frequency response, external real-ear equipment is needed to verify that the goal of speech audibility is reached.
To overcome limitations of software accuracy, coupled with the relatively low usage of external verification equipment, Starkey Laboratories developed a new approach called Integrated Real-Ear Measurement. The clinical testing of IREM demonstrates that a valid measured RECD can be obtained with excellent reliability, resulting in a closer match to a patient's prescriptive targets.
While Integrated Real-Ear Measurement is not a complete replacement for traditional real-ear measurement, it offers a significant benefit to practitioners: the ability to ensure appropriate audibility and comfortable speech to their patients.
The authors thank Katie Plum of the University of Kansas Medical Center for providing data from her Capstone Project. Also, special thanks to those who reviewed this article and offered great suggestions.