HSC Section 8_April 2017

J. FINBOW ET AL.

as expected, that performance is better (lower SNR) in all device conditions with noise to the poorer ear, i.e., in the condition in which the noise is attenuated by the head shadow before it reaches the better ear to mask speech. The addition of an SSD aid should lower the SNR with noise to the better ear, as it would route more of the speech signal to the better ear. We were not able to show a significant improvement with either SSD aid in this condition, but importantly, nor was there a sig- nificant difference between devices, unlike older studies (11,12,15,24). Subjective results for both questionnaires, which are focused on the particular handicaps of SSD, seem to be similar for both CROS and BAHD (Figs. 4 and 5) as well as conditions and duration of use. This finding is in con- trast to older studies, which admittedly used a different questionnaire, the APHAB, but often favored the BAHD (10,18). Perhaps the most telling aspect of the results is that out of the eight participants, four expressed a pref- erence for the CROS device, despite at least 6 months of experience with a BAHD. This finding was not de- pendent on the device as two of four Cochlear BP100 users preferred the BAHD, with one undecided, and with the Oticon Medical Ponto Pro, three of four users pre- ferred the CROS. The main reasons for preferring the CROS was the sound quality, and reasons for preferring the BAHD were comfort; those participants who pre- ferred the BAHD also expressed difficulty with retention of the dome tips with the CROS and the annoyance of wearing two aids. It should be noted that for the CROS device, disposable dome tips were used. Custom-fitted ear tips may have further improved comfort for the CROS. This study is the first study to compare BAHD on a percutaneous implant with newer models of CROS de- vices and to randomize the exposure. Although previous studies have randomized exposure (e.g., Hol et al. [14]), the BAHD was on a headband. The main limitation of this study is the small number of participants. Although the number of participants is comparable to some previous studies (3,14,17), clearly this work should be considered a pilot study to explore if one device provides greater benefit than the other in objective or subjective testing. Differences between the devices, if they exist in objective or subjective measurements, are likely to be relatively small if this study can be considered a point estimate. Obviously, future work will need to include a larger number of participants to determine if differences be- tween devices emerge. Other factors to explore include whether the Unitron CROS aid used in this study is rep- resentative of all modern wireless CROS aids from other manufacturers, such as Phonak. It is not known whether there are differences among different CROS hearing aid models and different BAHD models. We did not have a large enough sample to compare Oticon Medical and Cochlear Corporation BAHD products. Moreover, the CROS device experience may have been enhanced by giving participants more than a 2-week trial and with custom tips rather than domes; these issues should be considered in future research.

Older published studies comparing CROS and BAHD (11,12,15,24) (including a systematic review of these studies [10]) have generally concluded that BAHD were preferred for SSD, particularly because of subjective scores rather than objective improvements in speech in noise performance. Two newer studies (3,13) used a BAHD on a headband rather than a percutaneously im- planted device and found better speech-in-noise results with noise directed to the better ear with the CROS rather than the BAHD. Because the skin can attenuate the signal in the high frequencies substantially (32), the headband is not a fair comparison of device function in actual usage. All these studies used wired CROS aids. Patients have expressed dissatisfaction with the older wired CROS devices, citing poor cosmetics, discomfort with occlusion of the better ear, poor sound quality and distortion, social stigma, ineffective reduction of high ambient noise, elec- tromagnetic interference by other devices, and interference with sounds heard in the better ear as reasons for being dissatisfied with the CROS (2,23,24,33). Since these studies, evolution in both technologies (CROS and BAHD) has occurred, particularly in the CROS wearing experience. Modern CROS aids avoid a physically wired connection passing behind the head, are smaller, and do not occlude the functioning ear as much as older devices. However, both devices have changed in ways that should improve the sound quality, such as the sophistication of their signal processing of noise and speech and adaptive directionality of the microphones. Hence, the relative rankings of these devices may well differ from previous studies. Contralateral routing of sound, whether by CROS or BAHD, can be both deleterious and helpful in the hearing experience. When speech is directed to the poorer ear, with noise at the better ear (e.g., SpeN0), then routing speech to the better ear should increase performance (12,24). However, in the S0Nbe condition, contralateral routing actually should decrease perfor- mance (12,24). In everyday life, the overall benefit will depend on the relative abundance of these conditions, which may vary from person to person. Across devices, the ability of the processor to distinguish noise from speech and to suppress it, as well as the adaptive direc- tionality of the microphone system, will determine how deleterious the S0Npe situation will be. It is important to note that the degree of head shadow alleviation is programmable in the CROS device, whereas it is much more limited by the physics of head impedance for bone conduction in the BAHD, particularly in the high frequencies (34). Our results overall show the trends that we would ex- pect, i.e., an increase in word recognition with SSD aids compared to the unaided state in the SpeN0 condition and a decrease in the S0Npe condition (Fig. 2), an effect also seen with the QuickSin (Fig. 3). With our relatively low numbers, neither the benefit or decrement is significantly different from unaided in Figure 2. Figure 3 again shows the difference between the con- dition in which the SSD aid is useful (noise to better ear) and harmful (noise to poor ear). Our results confirmed,

Otology & Neurotology, Vol. 36, No. 5, 2015

153