Test-retest reliability of high frequency hearing aid output using two real ear probe microphone placement methods

Abstract

Background: The benefits of extended high frequency (HF) amplification are well established. In particular, research has shown improved speech recognition performance in hearingimpaired children with an increase in hearing aid (HA) bandwidth (BW) from 6 kHz to 10 kHz (Stelmachowicz, Pittman, Hoover, & Lewis, 2001a; Stelmachowicz et al., 2008; Stelmachowicz, Lewis, Choi, & Hoover, 2007). For optimal speech and language development, children require access to as many sounds as possible (Yoshinaga-Itano, 2003a). In addition, many adults with HF hearing losses (< 55 dB HL) benefit from additional HF information in speech recognition tasks (Hogan & Turner, 1998) and prefer a 9 kHz BW over a 5.5 kHz BW (Ricketts, Dittberner, & Johnson, 2008). Although the concept of extended BW HAs has existed for over thirty years (Killion & Tillman, 1982) it is only fairly recently that manufacturers have begun to offer devices with upper frequency limits of above 6 kHz (Kuk & Baekgaard, 2008). The amount of effective amplification provided by these devices has been questioned due to the presence of cochlear dead regions, acoustic feedback and coupling limitations. Real ear probe microphone measures (REM) are considered best practice in the verification of HA performance (Dillon & Keidser, 2003; G. H. Mueller, 2005) and the most important step in the HA fitting process (Kochkin et al., 2010). Probe microphone measures provide objective assessment of HA performance (Jespersen & Moller, 2013) and verify that the gain-frequency response shown in the manufacturer fitting software is representative of HA gain and output in an individual ear canal (Aazh & Moore, 2007). In the clinic, assessing the ability of HAs to amplify above 6 kHz is difficult, as the standard verification method, REM is limited by the measurement range of equipment and methodological concerns such as HF standing waves that could misrepresent ear canal sound pressure. A reliable, robust verification approach is required as part of the fitting process. Despite the advantages of extended BW amplification, there have been few investigations into the verification of HA performance at HFs. Aims: This study investigated the test-retest reliability of a 6 kHz acoustically-assisted and a 25 mm constant insertion depth probe placement method to verify HA performance up to 10 kHz. Both methods are currently used clinically (Audioscan, 2012). The primary aim of this study was to identify which method was more reliable across frequency, gender and trial. In order to assess real-ear sound pressure above 6 kHz a probe microphone with extended measurement BW was coupled to a digital sound level meter (SLM). The second aim was to determine if high frequency measures of HA performance could be obtained using real-ear measurement. Methods: Twenty normal-hearing adult participants (13 female, 7 male) aged 21 to 61 years were recruited. Otoscopy, tympanometry and pure-tone audiometry between 250 Hz to 8 kHz was performed on all 40 otologically healthy ears. A pair of Widex Mind 440 behind-the-ear (BTE) HAs (Appendix A) with slim tubing and open domes was used. Widex Compass (version 5.6) was used to programme HA output using default settings. Measurement equipment consisted of an Etymotics ER-7C (Appendix B) probe microphone system coupled to a Bruel & Kjaer Type 2250 SLM. ICRA noise was played at 70 dB SPL through a desk-level speaker for 30 seconds per measurement and HA output was recorded and saved on the SLM. Both the 6 kHz acoustically-assisted and 25 mm constant insertion depth probe microphone placement methods were undertaken twice on each ear (4 measurements per ear; 8 per participant) to ascertain test-retest reliability. The order of measurements for each method was counterbalanced to remove any ordering effect on results. In the acoustically assisted method a 6 kHz tone was played and the SPL in the ear canal was monitored as the probe was advanced along the earcanal, at a point approximately 10 mm from the eardrum a drop then increase in sound pressure can be observed due to a standing wave. Once the standing wave minima was obtained the probe tube was advanced a further 5 mm. In the constant insertion depth method the probe tube was marked at 25mm from its end and placed along the earcanal so the mark was at the intertragal notch at the earcanal entrance. Results: The standard deviation (SD) of test-retest differences for both methods was at or below + 2.4 dB, within the 6 dB margin considered acceptable (Jespersen & Moller, 2013). A three-way interaction between frequency, method and time was run as part of a repeatedmeasures ANOVA. There was no interaction between frequency, method and time (F(6,114)=0.248, p=0.778), indicating both methods were equally reliable across time and frequency (Figure 6A;B). Whilst both methods had similar test-retest reliability for females, for males the acoustically-assisted method showed greater reliability than the constant insertion depth method across frequencies. Based on the higher reliability for the acousticallyassisted method for males, analysis of the test-retest difference as a function of ear canal volume was investigated. No correlation was found for either method or gender. Conclusions: A safe, accurate and quick method is required for probe microphone verification of extended BW HAs to be clinically feasible. Both methods in this study were reliable and are iv currently used clinically (Audioscan, 2012). They can be performed in less than a minute per ear and do not require any additional equipment, outside that which is required for a HA fitting appointment. Further research is recommended to clarify the extent to which the test-retest reliability reported here extends to inter-tester reliability across clinicians in other practices. The gender difference is consistent with other studies (Aarts & Caffee, 2005), though the reason for this is unclear, as no correlation was found between error and ear canal volume.