Appendix 1
About the author: Ad Snik studied physics at the Eindhoven University of Technology and got the master’s degree in 1976. In 1982, he acquired the doctor’s degree from the same university with a thesis entitled ‘Visco-elastic properties of monomolecular layers’. After two years of lecturing Physics and Electronics, he specialized in medical physics and was registered in 1987 as a medical physicist/ audiologist. In 1988 the author was employed as a clinical audiologist and researcher at the ENT department of the Radboud University Medical Centre and appointed as a professor in 2006 until his retirement in 2017. Since 2010, he worked one day a week as a researcher at the department of Biophysics, Radboud University. Ad Snik participated in more than 250 peer-reviewed papers as an author.
Affiliation: Donders Institute for Brain Cognition and Behavior, Department of Biophysics, Radboud University, Nijmegen, the Netherlands
Some relevant scientific rankings:
Research gate. Score 42,49 (November 2020).
Expertscape objectively ranks people and institutions by their expertise in more than 26,000 biomedical topics. The author is listed in the top 15 of experts in the fields of Hearing Impairment, Hearing Aids, Hearing Loss and Sensory Aids (November 2020).
Disclosure:
The author has and had no financial interests in any of the devices described in this blog, nor in any alternative device. This blog is partly based on research carried out at RadboudUMC, which was supported by Nobelpharma, Entific, Cochlear BAS, Symphonix, Med-El, Otologics, KNO fonds Nijmegen, Heinsius Houbolt fonds, the William Demant Foundation and European Union Grants.
Appendix 2: Validation of the auditory implant fittings; objective outcome measures
A2.1 Introduction
Outcome measures used to validate the fitting of a hearing device can be divided into subjective measures and objective measures. Often, subjective refers to the assessment of patient’s opinion and when referring to objective measures, it concerns hard data, e.g., electrophysiological data. However, when assessing hearing loss or the profit of a hearing aid, objective refers to all data that can be measured with equipment, including hearing thresholds and speech recognition (Humes, 1999). To evaluate a hearing aid fitting, the most popular ‘objective’ outcome measure is speech recognition in quiet and/or in noise (Gatehouse, 1998; Gurgel et al., 2012). In case of evaluating bilaterally fitted devices, other objective tests can be added, e.g., directional hearing and/or speech recognition in noise with spatially separated speech and noise sources (Snik et al., 2015).
To quantify the outcome of an intervention in conductive or mixed hearing loss, surgeons often use (sound-field) hearing thresholds to evaluate the effectiveness of reconstructive surgery (Toner & Smyth, 1993; Gurgel et al., 2012). When an auditory implant is applied, then the difference between pre- and post-intervention hearing thresholds is of importance as a measure of the gain (amplification) provided by the implanted device. However, although this approach is widely used, it has certain limitations.
A2.2 Validating the gain of an auditory implant in patients with sensorineural hearing loss
The most widely used measure to assess the gain of a hearing device is the ‘functional gain’, which is, by definition, the difference between unaided and aided hearing thresholds. Essentially, functional gain assumes linear amplification.
Nowadays, most auditory implants don’t use linear amplification but instead non-linear amplification or compression. The most frequently used type of compression amplification (wide dynamic range compression; WDRC) implies that for soft sounds the gain provided by the device is relatively high, while for louder sounds, the gain is gradually reduced. This kind of compression deals with pathological ‘accelerated’ loudness growth in sensorineural hearing loss (Dillon, 2012). Thus, compared to linear amplification, compression ‘boosts’ low-level sounds; consequently, for those frequencies, ‘functional gain’ is boosted as well. When using compression, the ‘functional gain’ assesses just the gain of the device for soft sounds and overestimates the gain at higher input levels like normal conversational levels.
Snik and Cremers (2001) discussed two options to estimate the gain of an auditory implant at more realistic listening levels (thus at supra-tone threshold levels). Their first option was the determination of ‘gain at most-comfortable-listening level (MCL)’. In short, the patient’s MCL is determined at octave frequencies twice, once unaided and once with the hearing implant active, and the results are subtracted. In this way, the ‘gain at MCL’ is determined, frequency specific. Second option was the ‘gain in speech reception threshold (SRT)’; the SRT is determined twice, once unaided and once aided. The two SRT values are subtracted; the difference is the (non-frequency specific) ‘gain in SRT’. Essentially, the SRT typically lies 20-25 dB above the (mean) pure-tone threshold, therefore, these two measures (gain in MCL or in SRT) are less affected by compression. Snik and Cremers compared gain measures in 14 VSB users with 304 processor using WDRC. Averaged at 0.5, 1, 2 and 4 kHz, the ‘gain at MCL’ was indeed approx. 10 dB higher than the ‘gain-at-threshold-level’. The ‘gain in SRT’ and ‘gain at MCL’ were comparable. Rameh et al. (2010) published ‘gain-at-threshold’ data and aided and unaided SRT values in 112 auditory implant users with sensorineural hearing loss. Using their data, a similar difference (10 dB) was calculated between ‘gain-at-threshold-level’ and ‘gain in SRT’. So, ‘the gain-at-threshold level’ or the ‘functional gain’ is not a representative gain measure whenever compression is being used.
Further, it should be noted that aided hearing threshold measurements might be affected by advanced signal processing: viz. by expansion used to prevent the patient from hearing the microphone noise of the device or by noise reduction algorithms (to prevent that the test signal is treated as ‘noise’).
To evaluate hearing aid fittings in sensorineural hearing loss, the ‘functional gain’ is often used. However, when evaluating devices with non-linear amplification ‘functional gain’ doesn’t have the usual significance. Better option is to report the ‘gain in SRT’ or ‘gain at MCL level’.
A2.3 Validating gain in patients with conductive or mixed hearing loss
To assess benefit of hearing devices in patients with mixed or conductive hearing loss, the ‘functional gain’ is not an appropriate measure. It should be realized that middle ear implants (with their actuator coupled to one of the cochlear windows) as well as bone-conduction implants directly stimulate the cochlea; consequently, the mal-functioning middle ear doesn’t play a role. However, the status of the middle ear (width of the air-bone gap) does affect the outcome of a ‘functional gain’ measurement.
To illustrate this, let us assume a patient with pure conductive hearing loss of 65 dB HL (bony aural atresia) using a device that stimulates the cochlea rather perfectly (aided thresholds at 15 dB HL). Then, following the definition, the ‘functional gain’ is 65-15=50 dB. However, if the air-bone gap is 40 dB instead of 65 dB (partial atresia) while using the same device, then the ‘functional gain’ is (only) 40-15=25 dB. This example shows that owing to its definition, the ‘functional gain’ depends on the width of the air-bone gap, and, therefore, it is not an adequate measure of the gain (amplification) provided by the implant.
If a patient with pure conductive hearing loss is fitted with a powerful bone conductor and if some every-day sound of 40 dB is indeed perceived as 40 dB by the patient, evidently, the gain is 0 (as the gain is the difference between the input and output sound levels). In other words, the bone conductor has effectively compensated the conductive hearing loss. In the audiogram, this leads to aided thresholds coinciding with the cochlear (thus bone-conduction) thresholds. This gain, the difference between aided thresholds and the bone conduction thresholds, has been referred to as the ‘bone-conduction gain’ (Carlsson and Hakansson, 1997) or the ‘effective gain’ (Bush et al., 2017). The ‘effective gain’ might be negative, indicating that the aided thresholds are worse than the bone-conduction thresholds. That means that every-day sounds are (still) perceived attenuated and, in the audiogram, a persistent although reduced air-bone gap is seen. When a patient has mixed hearing loss with an obvious sensorineural hearing loss component (SNHLc) of e.g., 50 dB HL, obviously an ‘effective gain’ of 0 is not enough; part of the SNHLc should be ‘compensated’ as well, to make speech sufficiently audible. The ‘effective gain’ should be positive, e.g., 22 dB (calculated target gain using the adapted NAL-rule; see Chapter 4); then, a 28 dB HL sound is just audible for his patient, using that amplification.
Note that by definition, the ‘functional gain’ equals the ‘effective gain’ plus the air-bone gap. Therefore, in contrast to the ‘effective gain’, the ‘functional gain’ is not a measure of the effectiveness/quality of the device fitting.
Two limitations of measuring aided thresholds have to be discussed. In case of a predominant conductive hearing loss, ambient noise level in the test booth should be low; if not, the noise might affect the threshold measurements falsely. Furthermore, expansion and/or compression might have been applied. Compression, to deal with a relatively low maximum output (Chapter 3), results in low (thus favorable) aided thresholds and expansion (to deal with device noise) results in higher (worse) aided thresholds (Dillon, 2012; chapter 10), respectively over- and under-estimating the gain at normal speech level, provided by the device.
In the previous section (A2.2) we discussed the use of alternative gain measures (viz. ‘gain in SRT’ and ‘gain at MCL level’). However, these measures cannot be applied easily in conductive or mixed hearing loss. This is caused by limitations of the bone-conduction transducer (part of the audiometer) used to determine the bone-conduction thresholds. The standard B71 bone-conduction transducer has a relatively low maximum output (might be close to the patient’s MCL levels instead of LDL levels) affecting the MCL measurements. Furthermore, its frequency response is poor in high and low frequencies (Alomari, 2014) affecting proper amplification of broadband sounds like speech. Additionally, the distortion of low-frequency sounds by the B71 is significant. The relatively new B81 bone-conduction transducer is the better option although the maximum output is still limited (Alomari, 2014).
If auditory implants are used in patients with conductive or mixed hearing loss, the ‘functional gain’ doesn’t provide any useful information on the amplification provided by the implant. Instead, the ‘effective gain’ should be used, as it is a fitting-quality measure. The ‘effective gain’ is by definition the difference between bone-conduction thresholds and aided thresholds and, therefore, might be negative.
A2.4 References, Appendix 2
Alomari HM. Binaural hearing with bone conduction stimulation. Eprints.soton.ac.uk/370832/1/Hala%20Alomari%20thesis.pdf. Thesis, 2014, University of Southampton
Busch S, Lenarz T, Maier H. Comparison of Alternative Coupling Methods of the Vibrant Soundbridge Floating Mass Transducer. Audiol Neurootol. 2016; 21(6):347-355
Carlsson PU, Håkansson BE. The bone-anchored hearing aid: reference quantities and functional gain. Ear Hear. 1997;18(1):34-41
Dillon H. Hearing aids. 2012. Second edition, Thieme Verlag, New York
Gatehouse S. Speech tests as measure of outcome. Scand Audiol Suppl. 1998;49:54-60
Gurgel RK, Jackler RK, Dobie RA, Popelka GR. A new standardized format for reporting hearing outcome in clinical trials. Otolaryngol Head Neck Surg. 2012;147(5):803-807
Humes LE. Dimensions of hearing aid outcomes. J Am Acad Audiol. 1999; 10:26-39
Rameh C, Meller R, Lavieille JP, Deveze A, Magnan J. Long-term patient satisfaction with different middle ear hearing implants in sensorineural hearing loss. Otol Neurotol2010;31:883-892
Snik AFM, Agterberg M, Bosman A. How to quantify binaural hearing in patients with unilateral hearing using hearing implants. Audiol Neurootol. 2015;20 Suppl 1:44-47
Snik AFM, Cremers CWRJ. Vibrant semi implantable hearing device with digital sound processing. Effective gain and speech perception. Arch Otolaryngol Head Neck Surg 2001;127:1433-1437
Toner JG, Smyth GD. Comparison of methods of evaluating hearing benefit of middle ear surgery. J Laryngol Otol. 1993;107(1):4-5
Appendix 3: History of the blog and acknowledgements
In September 2014, Franco Trabalzini organised the EAONO meeting in Siena and asked me to organize a panel on Consensus on Auditory Implants. Preparing that panel, I wrote a document what can be considered as the bleu print of the present blog. Later that year, an updated version was sent to my colleagues from Auronet (Auronet is a private initiative to develop a core set of patient-centered outcome measures which enable the search for the best treatment options for a patient with conductive or mixed hearing loss; Tysome et al., 2015). Comments by my Auronet colleagues, especially Bill Hodgetts and Penny Hill, encouraged me to go on. Bill added valuable comments.
Chapters 1 to 3 were published on the internet on April 22nd, 2015. Chapters 4 and 5.1-5.2 were added on May 6th and June 16th.
During the Osseo meeting in Lake Louise, Alberta (May 21-23, 2015), the website was introduced to the public for the first time.
On July 28th, chapter 5.3 and 6 were uploaded. Martijn Agterberg critically read this part of the blog before it was published. Chapter 7 was added on August, 18th.
Owing to comments by Arjan Bosman, Martijn Toll, Tove Rosenbom and Marc Flynn, the text of chapters 1 to 4 was revised and re-published on September 7th and October 8th, 2015. During the latter update it was decided to skip the option for the readers to respond online. From the start until early October, more then 90 responses were collected and all of them were spam. On March 29th 2016, chapters 5 and 6 were revised and Appendix 2 was added.
On August 12th 2016, chapters 5 to 7 and Appendix 2 were updated and revised. The remarks made by Hannes Maier are acknowledged.
Chapter 8, on the use of auditory implants in patients with pure sensorineural hearing loss, was added on August 20th 2016.
January 20th 2017, data regarding relatively new devices was added (Chapter 3). This concerned the Sophono Alpha 2 processor, the Cochlear Baha Attract and the Med-El BoneBridge.
On May 2nd 2017, chapter 1, 2, 3 and 5 were updated and revised. Amongst others, new data was added concerning new super-power percutaneous bone conductors. Appendix 2 was rewritten and Appendix 3 updated.
In the month of November 2017, this website was updated and fully re-designed.
August 2018, Chapters 3, 4 and Appendix 2 have been re-written and updated.
December 2018, Chapters 6 and 8 are being re-written and updated. Blog 2018-4 added.
April 2019, Blog 2019-1 added.
October 2019, Tables 2.1, 4.1 are updated and Figure 4.2. Blog 2019-1 updated.
December 2019, Blog 2019-2 added.
October 2020: Chapter 7 was updated. Blog 2019-2 was removed (concerned a PPT presentation at the Osseo conference).
November 2020: Chapter 8 was updated and revised as two of the discussed auditory implants were taken off the market end 2019. In addition, appendices 1 to 3 were updated.
4 February 2021: Chapter 6 was updated and Chapter 8 was extended.
9 February 2021: Chapter 2, 4 and Paragraph 5.3 adapted.
10 March 2021: Added summary of the website, others: Ad Snik, Martijn Agterberg
A3.1 Reference Appendix 3:
Tysome J. et al., The Auditory Rehabilitation Outcomes Network: an international initiative to develop core sets of patient-centered outcome measures to assess interventions for hearing loss. Clin Otolaryngol. 2015;40:512-515