Tech Topic | October 2017 Hearing Review

This is a summary of a full study report submitted to an international peer reviewed journal. Parts of this work have been presented at the IHCON 2016 conference (Granlibakken, Tahoe City, California, USA).

A study that assesses both the test-retest repeatability and the precision (matching accuracy) of the new Phonak TargetMatch system.

Real-ear measurements (REMs) have been accepted as best practice in the field of audiology.1,2 They provide a way of measuring hearing aid performance in the ear which can lead to more accurate and individualized hearing aid fittings.2,4 A more individualized fitting has several advantages for both the client and the hearing care professional: less follow-up appointments, overall higher client satisfaction, increased patient loyalty, reduction of emotional stress, and an overall positive perception of clinical services delivered. The overall effect of having a more accurate hearing instrument fitting—customized for the individual client—is a better listening and fitting experience.4,5

Open-fit hearing aids are one of the most commonly-used devices today. They have many benefits, such as reduced occlusion, improved own-voice perception and sound quality, and increased localization performance.6 Performing REMs with open-fit hearing aids requires special considerations due to the effect of the output of the hearing aid on the reference microphone, which potentially leads to misleading results.7,8 However, several studies have proven REMs to be reliable with a range of different coupling types, including open fits.9

Although the benefits of REMs have been clearly proven (see the article by Amlani et al in this edition of Hearing Review), the number of hearing care professionals performing REMs is relatively low.5 This may be due to non-integrated REM systems not meeting the needs of hearing care professionals.

One solution that can address these needs is TargetMatch from Phonak. TargetMatch is a verification software that is fully integrated in Phonak Target fitting software. It provides the hearing care professional with a step-by-step guided workflow on performing REMs. The software then automatically makes fine-tune adjustments to the hearing aids to match as closely as possible to both proprietary and standard fitting formulae prescription targets (the system, rather than the audiologist, adjusts the gain). One advantage of TargetMatch is that it reduces the complexity, which some hearing care professionals perceive, and may increase the likelihood of hearing care professionals performing REMs.

TargetMatch is designed to efficiently measure all necessary transformations, and therefore the amount of actions at the patient’s ear is kept to a minimum. Also, the system has been developed to assist with probe tube insertion with a guided probe tube placement feature, providing visual feedback on insertion depth by means of a “parking sensor.” When the probe tube is at the ideal location, a clear indication via a green check is provided, and the system subsequently moves onto obtaining real-ear unaided gain (REUG) measurements.

As part of the implementation of TargetMatch, it was important to confirm the reliability and accuracy of the system. The British Society of Audiology1 recommends tolerances for mismatch to targets for prescription rationales of no more than ±5 dB at 250, 500, 1000, and 2000 Hz, and of ±8 dB at 3000 and 4000 Hz. The goal of TargetMatch is for an even closer match to prescribed targets. To verify the results of the system, a study was carried out at the University of Leuven’s  ExpORL Research Group in Belgium.

Study Objectives

The aim of this study was to assess both the test-retest repeatability and the precision (matching accuracy) of TargetMatch. Therefore, the following two questions were addressed:

1) Precision and repeatability. How repeatable and precise is TargetMatch for different acoustic coupling types (eg, open dome, power dome, and custom earmold)?

2) Accuracy. How accurate is the match to prescriptive targets from standard fitting formulas when using TargetMatch versus Manual Fit to Target (manually performing REMs, followed by manually adjusting gain curves to attempt to match to prescriptive targets)?

Methodology

Two different groups of participants were recruited in this investigation: 1) a normal-hearing group to address the research Objective #1 above, and 2) a hearing-impaired group to address research Objective #2 above.

The normal-hearing group was recruited to test the repeatability (test-retest) of the TargetMatch system. This group consisted of 7 subjects (6 female, 1 male) with hearing thresholds of ?20 dB HL across the frequencies of 250-8000 Hz. Subjects’ ages ranged between 21 to 24 years. All had normal middle-ear function. These study participants were fit with Phonak Bolero V 70P devices and different acoustic couplings (these listeners do not have specific acoustic coupling or amplification needs).

Three different acoustic couplings were tested with each subject: open domes, power domes, and custom made earmolds. The custom earmolds were produced with a 1 mm vent to allow for comfort. Subjects were fit using TargetMatch according to the NAL-NL2 fitting rule10 for a simulated mild-to-moderate N2 hearing loss11 (Figure 1). This hearing loss was chosen in order to create a prescription with enough gain to run a realistic verification, while keeping the gain level low enough to be comfortable for the normal-hearing participants. Acoustic coupling conditions were randomized for each participant. Each coupling was tested twice (test-retest).

Figure 1. The black line shows an N2 hearing loss11 which was the simulated hearing loss for the normal hearing group. Blue and red lines reflect the mean audiogram (±1 SD) for left and right ears, respectively, for the hearing-impaired group.

Figure 1. The black line shows an N2 hearing loss11 which was the simulated hearing loss for the normal hearing group. Blue and red lines reflect the mean audiogram (±1 SD) for left and right ears, respectively, for the hearing-impaired group.

The group of hearing-impaired participants was recruited to investigate the accuracy of matching prescriptive targets using TargetMatch compared to manually matching targets using a non-integrated REM solution. The hearing-impaired group consisted of 10 hearing impaired listeners (5 male, 5 female) ages 22 to 76 years (mean age = 41 years, SD = 22 years). The mean audiogram (±1 SD) for left and right ears is shown in Figure 1. There was a relatively wide variance in the hearing thresholds of the participants. This was done in order to test the flexibility of the TargetMatch system with regard to the accuracy for different degrees of hearing loss. All participants had a sensorineural hearing loss and normal middle-ear function. They were all experienced hearing aid users with hearing aid experience ranging from 6 to 22 years.

All hearing-impaired subjects were fit with Phonak Bolero V 70P devices, with the acoustic coupling most appropriate for their hearing loss and according to the DSLv5 prescriptive fitting formula.12 This formula was selected to evaluate how well targets are matched with an algorithm that provides the highest amount of gain for the respective hearing loss, especially at high frequencies.

Transfer functions produced via TargetMatch and via Manual Fit to Target were compared to prescriptive targets in order to compare the mismatch to target for both fitting methods. For both groups of participants, TargetMatch, with its guided probe-tube placement available in Phonak Target 4.3, was used in combination with an Otometrics AURICAL FreeFit to perform integrated REM verification. Prior to each fitting, the acoustic parameters were set (venting diameter, tubing, etc), the prescriptive formula was chosen, and a feedback test was performed. Regarding the settings of the hearing aids, all adaptive features were set in default mode; self-learning features and SoundRecover were deactivated, and feedback cancellation was activated.

Transfer functions consisting of the real-ear unaided gain (REUG), the real-ear occluded gain (REOG), the real-ear to coupler difference (RECD), and aided measures (real-ear aided response or REAR) were measured in both the normal-hearing and hearing-impaired participants using the TargetMatch workflow as they were sitting 1 meter from the Aurical loudspeaker. The same transfer functions were measured using Manual Fit to Target. For the aided measurements, the standard ISTS speech signal13 was presented at three input levels: 50 dB, 65 dB, and 80 dB SPL.

Test-retest measures were performed for all transfer functions for both the normal-hearing and hearing-impaired subjects, and for each fitting strategy (TargetMatch and Manual Fit to Target). For transfer functions that require accurate placement of the probe tube, such as REUG, the probe tube was removed and re-inserted each time. With TargetMatch, the guided probe-tube placement functionality was used. With the Manual Fit to Target, a constant insertion depth approach was used. Probe tubes were marked for a consistent distance of 27 mm and placed at the intertragal notch with the aim of reaching an optimal distance of 5-6 mm from the tympanic membrane.

RESULTS

Precision and Repeatability

How repeatable and precise is TargetMatch for different acoustic coupling types (eg, open dome, power dome, and earmold)? The test-retest precision for the REUG was found to be very high for all acoustic couplings. Up to 4000 Hz, unaided measures were observed to have a test-retest precision of better than 1.5 dB (Figure 2). Frequencies higher than 4000 Hz have a test-retest precision of ?2 dB, although in this frequency region the reliability of REMs is lower due to the low pass filtering of the tube.14 These small test-retest differences verify the accuracy of the probe-tube positioning system of TargetMatch.

Figure 2. Test-retest precision for the REUG for the normal hearing group fit via TargetMatch. Graph shows the root mean square of the standard deviation of the repeated measurements over frequency.

Figure 2. Test-retest precision for the REUG for the normal hearing group fit via TargetMatch. Graph shows the root mean square of the standard deviation of the repeated measurements over frequency.

Absolute test-retest differences for the REAR, using different acoustic couplings are shown in Figure 3. The REAR is the response measured in the ear canal with the hearing aid switched on. Overall, differences are smaller than or equal to 2 dB, except sometimes at higher frequencies. This shows that across coupling types, even for open fittings, TargetMatch is reliable. The larger test-retest differences for higher frequencies across all couplings, confirms that REMs are in general not as reliable for high frequencies, due to the tube used in REM which generates a low band pass filter.14,15

Figure 3. Box plots of absolute test-retest differences for the REAR per acoustic coupling type, for the normal-hearing participant group. Data are pooled across ears and levels. Crosses represent means.

Figure 3. Box plots of absolute test-retest differences for the REAR per acoustic coupling type, for the normal-hearing participant group. Data are pooled across ears and levels. Crosses represent means.

Accuracy

How accurate is the match to prescriptive targets from standard fitting formulas when using TargetMatch versus Manual Fit to Target? Figure 4 shows the target mismatch for the hearing-impaired listeners when fit using Manual Fit to Target (red) and using TargetMatch (blue). With TargetMatch, targets are well matched within 3 dB in the frequency range of 400-4000 Hz. When compared to the Manual Fit to Target, target matching across most frequencies between 800-5000 Hz is significantly better with TargetMatch. This outperforms the acceptable tolerances (±5 dB at frequencies 250, 500, 1000, and 2000 Hz, and of +/- 8 dB at 3000 and 4000 Hz) recommended by the BSA.1 Additionally, Figure 4 indicates that TargetMatch outperforms Manual Fit to Target in the difficult-to-fit high frequency range (above 4000 Hz).

Figure 4. Mean target mismatch for hearing-impaired listeners fit via Manual Fit to Target (in red) and TargetMatch (in blue). Data are pooled across ears and levels.

Figure 4. Mean target mismatch for hearing-impaired listeners fit via Manual Fit to Target (in red) and TargetMatch (in blue). Data are pooled across ears and levels.

Conclusion

This investigation of the test-retest reliability of TargetMatch for the normal-hearing group proves its reliability across different acoustic couplings. The results demonstrate that the system is very accurate and can accommodate a wide variety of acoustic couplings.

The results from the hearing-impaired group, using an algorithm such as DSLv5, demonstrates the matching accuracy of the TargetMatch system, as real ear targets are well matched within 3 dB. This outperforms both the BSA guidelines for accuracy and precision, as well as a fitting which has been manually matched to target using a standard non-integrated REM system. Given that TargetMatch also reduces complexity, the system should enable more hearing care professionals and end-users to access the proven benefits of verification during the fitting process.

Acknowledgements

This is a summary of a full study report submitted to an international peer reviewed journal. Parts of this work have been presented at the IHCON 2016 conference (Granlibakken, Tahoe City, California, USA).

Screen Shot 2017-09-22 at 4.33.21 PM

Correspondence can be addressed to HR or Dr Latzel at: [email protected]

Citation for this article: Latzel M, Denys S, Anderson S, Francart T, Wouters J, Appleton-Huber J. An integrated REM system with proven accuracy and reliability. Hearing Review. 2017;24(10):36-39.

References

  1. British Society of Audiology. Guidance on the use of real ear measurement to verify the fitting of digital signal processing aids. July 2007. Available at: http://www.thebsa.org.uk/wp-content/uploads/2014/04/REM.pdf

  2. Valente M, Abrams H, Benson D, et al. Guidelines for the Audiological Management of Adult Hearing Impairment. Audiology Today. October 30, 2006. http://www.audiologyonline.com/articles/guideline-for-audiologic-management-adult-966

  3. Moore BCJ. Speech mapping is a valuable tool for fitting and counseling patients. Hear Jour. August, 2006;59(8):26-30. doi: 10.1097/01.HJ.0000286371.07550.5b

  4. Amlani A. Impact of probe-microphone measurements and NOAH quick-fit on patient satisfaction and loyalty. March 22, 2016. Available at: http://hearinghealthmatters.org/hearingeconomics/2016/amyn-post-consumer-psychology-of-real-ear-services/

  5. Kochkin S. MarkeTrak VIII: Reducing patient visits through verification and validation. Hearing Review. 2011;18(6):10-12. Available at: https://hearingreview.com/2011/06/marketrak-viii-reducing-patient-visits-through-verification-amp-validation/

  6. Winkler A, Latzel M., Holube I. Open versus closed hearing-aid fittings: A literature review of both fitting approaches. Trends Hearing. February, 2016;15:20. doi: 10.1177/2331216516631741

  7. Hallenbeck, S. Considerations in performing real ear measures with open fit hearing instruments. March 24, 2008. Available at: http://www.audiologyonline.com/ask-the-experts/considerations-in-performing-real-ear-279

  8. Switalski W. Three probe microphone measurement techniques to enhance open fittings. Hearing Review. 2011;18(11):40-49. Available at: https://hearingreview.com/2011/10/three-probe-microphone-measurement-techniques-to-enhance-open-fittings/

  9. Jespersen CT, Møller KN. Reliability of real ear insertion gain in behind-the-ear hearing aids with different coupling systems to the ear canal. Int J Audiol. March, 2013;52(3):169-176. doi: 10.3109/14992027.2012.744105

  10. Dillon H, Keidser G, Ching TYC, Flax MR, Brewer S. The NAL-NL2 Prescription Procedure. Stafa, Switzerland: Phonak Focus 2011. Available at: https://www.phonakpro.com/content/dam/phonak/b2b/C_M_tools/Library/focus/en/Focus_40_NAL-NL2_Prescription_GB.pdf

  11. Bisgaard N, Vlaming MS, Dahlquist M. Standard audiograms for the IEC 60118-15 measurement procedure. Trends Amplif. June, 2010;14(2):113-120.

  12. Scollie S, Seewald R, Cornelisse L, et al. The desired sensation level multistage input/output algorithm. Trends Hearing. Fall, 2005;9(4):159-197.

  13. Holube I, Fredelake S, Vlaming M, Kollmeier B. Development and analysis of an International Speech Test Signal (ISTS). Int J Audiol. December, 2010;49(12):891-903. doi: 10.3109/14992027.2010.506889

  14. Dillon H. Hearing Aids. 2nd Ed. Sydney, Australia: Boomerang Press:608

  15. Ulrich J, Hoffmann E. Hörakustik–Theorie und Praxis 3.0. Heidelberg: DOZ Verlag;2017.