Real Ear Measurements in Fitting Hearing Aids for Children

Choi, Chul-Hee; Rhee, Chung-Ku

Review

Hearing aids

Korean Journal of Audiology 1998;2(2):109-117.

Real Ear Measurements in Fitting Hearing Aids for Children

¹Department of Audiology, University of Arizona, Tucson, Az
²Department of Otolaryngology-Head Neck Surgery, College of Medicine, Dankook University, Cheoan, Korea

소아에서 보청기의 맞춤을 위한 실이 측정

최철희^¹^, 이정구^²

^¹애리조나대학교 청각학교실
^²단국대학교 의과대학 이비인후-두경부외과학교실

Introduction Audiologists must consider many factors when selecting and fitting hearing aids for children with hearing impairments. For example, the important factors that must be considered include hearing aid style, microphone type, hearing aid arrangement, gain and frequency response, and maximum output. Among these factors, the two most important decisions made when choosing the electroacoustic characteristics of a hearing aid concern selection of gain/frequency response and maximum power output (SSPL90)(Hawkins, 1992). There are several clinical procedures such as behavioral threshold, 2 cc coupler, loudness discomfort level and real ear measurements for setting the gain/ frequency response and maximum output of hearing aids for children. However, traditional approaches have several disadvantages when they are used with children. Behavioral threshold tests are typically not appropriate with young children because of required cooperation from children, extensive testing time, and poor test-retest reliability (Hawkins & Northern, 1992). The 2 cc coupler test does not take into account the external auditory canal volume difference between adults and children, pinna or concha effect and head or body baffle effect (Matkin, 1983；Hawking,1992). Furthermore, loudness comfort or discomfort judgments may be beyond the child's conceptual abilities (Hawkins et al., 1989). Real ear measurements offer many advantages over other traditional approaches with children. Hawkins et al.(1989) recommended real ear measurements in selecting and evaluating hearing aids based on the following justifications：(1) the opportunity to assess the real ear response of hearing aids with relative ease and reasonable reliability, (2) realization of some of the practical limitations of behavioral aided threshold measurements, (3) limitations of speech-based evaluation procedures, and (4) the objective nature of the measurements, that is, a minimum amount of cooperation is required of the client. Furthermore, Tecca (1994) suggested five reasons to verify real ear measurements (REMs)：(1) the performance of a hearing aid on any given ear is not well predicted from average group data, (2) REMs provide information essential for evaluating and/or improving a hearing aid fitting, (3) they provide baseline documentation for future revisions to a fittings if necessary, (4) they can provide an excellent education or counseling tool, and (5) they provide a means of demonstration for professionals, from other disciplines that audiology is an accountable profession. Based on the above justifications, audiologists should utilize real ear measurements for fitting hearing aids on children. This paper reviews an appropriate hearing aid selection method for children describing how real ear measurements can be useful in fitting hearing aids on children and some of the studies obtained about hearing aid fitting from real ear measurements. Selection of Gain and Frequency Response The hearing aid should be adjusted to approximate the predetermined target gain values for each ear. In other words, the selection of the gain and frequency response of a hearing aid is obtained through the comparison of a desired target real ear insertion response to real ear probe-microphone measurements using a speech input. Real ear insertion response (REIR) is determined by subtracting the real ear unaided response (REUR) from the real ear aided response (REAR). While the REUR refers to the sound pressure level (SPL), as a function of frequency, at a specified point in the unoccluded ear canal for a specified soundfield, the REAR represents the SPL, as a function of frequency, at a specified point in the ear canal for a specified soundfield with the hearing aid in place and turned on. Therefore, the REIR refers to the amount of gain that is obtained from the hearing aid when it is inserted in the ear. Since the REIR is a measure of the resonance characteristics of the ear canal and the concha, how the REUR differs in children and how it changes as a function of age is very important when considering the REIR (Hawkins, 1992). 1. Ear canal SPL (or the REUR) The REUR represents the acoustic properties (resonance) of the ear canal and the concha. The REUR often reflects abnormalities of the ear canal or the middle ear. Other parts of the external ear and the child's head and body also can influence the REUR measure (Mueller, 1992). Kruger and Ruben (1987) measured the REURs of two adults and 13 children under the age of 40 months. According to their study, while the REUR peak for the newborn is at 7200 Hz, the REUR peak for theadults is at 2700 Hz. Kruger (1987) indicated that as an child's age increases, his/her REUR peak frequency decreases rapidly and reaches the typical adult peak of 2700 Hz by 12-24 months ofage. According to Hawkins and Northern (1992), Bentler (1989) reported that wide variability can be seen in the location of the peak in the REUR across children ages 3-13 years and the REURfor children in this age range is similar to that expected for adults. Hawkins and Northern (1992) suggested that the age-related changes in the REUR have important implications for hearing aid selection using a REIR approach. When selecting hearing aids that achieve the desired REIR, audiologists must take into account the REUR. In an infant, the REUR in 2500-3000 Hz region will gradually decrease as the child ages. That is, as the resonant frequency slowly changes downward in frequency, the hearing aid gain decreases. For example, a hearing aid fit on an infant that has 40 dB of REIR at 2500 Hz may have only 25 dB of REIR by the time the child is 2 years old (Hawkins and Northern, 1992). In conclusion, children below 1~2 years of age may show REURs that have resonant peaks in higher frequencies than older children or adults. Therefore, audiologists must take into account this factor when assessing REIR. It has been suggested that the SPLs produced in the ear canals in children may be greater than the SPLs in adult ears. This is due to the fact that the ear canal volume in children (0.66 cm3) is smaller than that in adults (1.26 cm3) (Bratt, 1980). However, Nelson Barlow, Auslander, Rines, and Stelmachowicz (1988) measured real ear SPLs in both children and adults. No significant differences in SPLs in both children and adults were observed. No systematic real ear coupler differences (RECDs) were observed between the values obtained for adults and those obtained for children between 3 and 15 years of age. These results suggested that other factors such as ear canal length and middle ear impedance may interact with volume to affect the SPL measured at the eardrum. This complex interaction would reduce the value of volume measures as a predictor of RECDs. Bernstein and Kruger (1986) reported the estimates of ear canal length in children from birth to 3 years of age. Their estimates suggested that the greatest changes in ear canal length occur before 2 years of age. Thus, RECDs in very young children may be greater than for adults even though differences were not observed for older children. Feigin, Kopun Stelmachowicz and Gorga (1989) indicated that RECDs for children under 5 years of age exceed those for adults, but systematically decrease with increasing age. The results of their study indicated that RECDs in young children typically are greater than in adults. They concluded that in cases where real ear measures are not feasible, the magnitude of this difference should be considered when limiting the output of hearing aids for infant and young children. Furthermore, the RECD can be used as a mathematical correction for the prescriptive 2 cc coupler gain. The RECD information allows audiologists to determine the actual SPL that is produced in the ear canal by the hearing aid for each child. When disregarding this information, audiologists can not monitor the possible overamplification and additional hearing loss for each child. Therefore, this information is very useful for setting the SSPL90 of the hearing aid, as well as determining the gain and frequency response (Hawkins and Northern 1992). 2. Real ear insertion response (REIR) The REIR is the mathematical difference between the REUR and the REAR. The primary purpose of conducting REIR measures is to compare these REIR findings to the desired target gain that has been derived from some prescriptive fitting procedure. The clinical use of the REIR is to verify that a predetermined prescriptive target insertion gain has been achieved (Mueller, 1992). Therefore, without a theoretical target, REIR measures may become rather meaningless. The REIR is the electroacoustic (or close equ-ivalent) of functional gain (Mueller, 1992). Therefore, REIR measurements can be used as a substitute for functional gain to verify a prescriptivetarget. Mueller (1992) suggested that probe microphone measurements have several practical advantages over the behavioral functional gain procedure：(1) information is obtained across frequency range, not just at discrete frequencies, (2) it is not necessary to mask the non-test ear, (3) the masking noise of the hearing aid does not prevent determination of real ear gain, (4) REIR can be determined for patients unable to provide behavioral response, (5) the effects of the input level on real ear gain can be assessed, (6) determination of real ear gain for individuals with profound hearing loss is not prevented by the upper limits of the soundfield equipment, and (7) there is significantly improved test-retest reliability. Currently, the gain and frequency response approaches which are popularly being used for computing a desired target gain can be categorized in several ways. The procedures are divided into two general categories：threshold-based procedures and suprathreshold-based procedures. The threshold based procedures include the methods of Byrne and Tonisson, Berger, NAL, POGO and Libby. The suprathreshold-based procedures include MCL-based procedures (CID and Shapiro), which bisect the dynamic range (MSUv3 and Bragg) and the Levitt procedure. There is considerable difference of opinion as to whether the above prescriptive techniques result in significant differences in speech recognition, sound quality or clarity. Byrne (1988) supports that if various prescriptive target formulas are strictly adhered to in the selection of hearing aids, no two procedures would be equally effective for all individuals. However, Humes (1988) suggests that no single prescriptive method emerges as being clearly superior and relative performance of the procedures varies as a function of output level. Regardless of the specific procedures used to select gain and frequency response, audiologists must pay attention to the following areas (Hawkins, 1986)：(1) the audibility of the long term speech spectrum (2) the range of the frequency response of the hearing aid, (3) the amount of gain in the low frequencies, and (4) the peakiness of the frequency response curve. It should be noted that most of the above procedures were developed for adults with hearing losses of less than 80 dB HL and are not appropriate for children with severe and severe-to-profound hearing impairments (Hawkins & Northern, 1992). Although audiologists have applied the procedures that were produced for adults to children whether the procedures are appropriate for children has not been determined. Selection of Maximum Output The primary purpose limiting maximum output of the hearing aid is to protect the child from loudness discomfort or potential damage to the ear from amplified sound. If the output of the hearing aid exceeds the child's loudness discomfort level (LDL), the child may be able to complain that the output of the hearing aid is uncomfortably loud. The possibility of the hearing aid causing temporary or permanent hearing threshold shift can not be dismissed. Therefore, selecting an appropriate maximum output level may be the characteristic most related to user acceptance of the hearing aid (Hawkins, 1992). When setting an appropriate maximum output for children, audiologists must know what output levels exist in the ear canal of the child and obtain the child's LDL. 1. Real ear saturation response (RESR) The RESR is the SPL, as a function of frequency, at a specified measurement point in the ear canal with the hearing aid in place and turned on. The measurement is obtained by using a 90 dB SPL input level and adjusting the hearing aid's volume control wheel (VCW) to a point without feedback (Mueller, 1992). This measurement is the real ear counterpart to the 2 cc ccouplermeasure of saturation sound pressure level 90 (SSPL90). How to determine an appropriate RESR and how to verify the RESR with probe microphone measurements must be addressed when fitting hearing aids for children. Hawkins and Northern (1992) suggested that an appropriate RESR should meet the following four criteria：(1) low enough to prevent loudness discomfort, (2) low enough to prevent additional hearing 1 loss from over amplification, (3) high enough to leave an adequately wide dynamic range, and (4) high enough that normal conversational speech does not constantly saturate the hearing aid. When the RESR is too high and exceeds the child's LDL, Hawkins (1992) suggests that one of several unfortunate consequences may occur：(1) the child may constantly change the VCW to adjust for different input levels, (2) to avoid this constant adjustment, the child may simply leave the VCW at a lower than optimal position to minimize the number of times that the output produce loudness discomfort, (3) the hearing aid is worn only in relatively quiet environments where input levels are low, or (4) the child may reject the hearing aid, as discomfort experienced outweighs the communication benefits received. According to Matkin (1983), Rintleman and Bess (1977) recommended that hearing aids with SSPL90 approaching 130 dB SPL should be used with extreme caution, and then only with children having severe or profound losses, and an SSPL 90 of 120 dB SPL can be used with young children with mild and moderate hearing losses. The University of Arizona Hearing Clinics have used the following guidelines (Matkin, 1983)：an output of 125 dB SPL be the maximum output used, and only with children having profound impairments. In contrast, 120 dB SPL is used with young children having moderate to severe impairment, while 115 dB SPL is the maximum output recommended for those with mild losses. However, RESR may exceed these values because the RECD is larger in younger children than in adults. Cox (1985) proposed that the SSPL90 could be predicted by the simple equation：SSPL90-100＋1/4hearing loss (dB HL). For example, if the hearing threshold of a child is 60 dB HL, the SSPL90 would be set to 115 dB SPL (100＋60/4＝115). It also should be kept in mind that these values are measured in a 2 cc coupler, not the ear canal. Audiologists need to keep in mind that each child's ear canal variability may make it difficult to predict what the RESR will be for a particular child given a recommended 2 cc coupler SSPL90 (Hawkins & Northern, 1992). 2. Loudness discomfort level (LDL) LDLs have been advocated as a means for selecting the SSPL90 setting of an individual's hearing aid (Stuart et al, 1991). The procedure for obtaining LDLs was developed in detail by Hawkins et al (1987). It is important to note that the procedures was developed for adults, not children. Kawell et al. (1988) developed an LDL procedure for children and evaluated its reliability on a group of 20 children ages 7-14 with hearing impairments. The procedure was based on the adult LDL method produced by Hawkins et al. (1987), which utilized loudness categories such as comfortable, comfortable but slightly loud, loud but ok, uncomfortably loud, and hurt. Each subject was given oral instructions provided with pictorial representations of the loudness categories. Subjects were instructed to point to or describe the loudness of the signal by referring to one of the five categories. Kawell et at. (1988) reported that it is possible to obtain reasonably reliable LDL measures on children as young as 7 year of age and there are no systematic differences in LDLs between adults and children in this age range. Stuart et al. (1991) measured LDLs with children ages 7~14 years based on the Kawell et al procedure. They placed a probe tube connected to a probe microphone system into the ear canal to measure the SPL at the point of loudness discomfort. Their study indicated that probe tube microphone measures of LDL are a feasible and reliable procedure for use with children with 7 to 14 years of age and the advantage lies in procedural conveniences and the ability to compare real ear audiometric measures and hearing aid performance. Therefore, the probe-measured SPL in the ear canal at the point of loudness discomfort becomes the target RESR for the hearing aid (Hawkins & Northern 1992). The Kawell et al procedure may not be appropriate for use with all children with hearing impairment because many children with hearing impairment have significant language delays. Based on the Piagetian theory which indicates that around the age of 5 or 6 years children can recognize equivalencies between two distinct orderings of magnitude, Macpherson Elfenbein, Schum and Bentler (1991) measured thresholds of discomfort with children ages 3~5 years. This study asked children to match the loudness of a signal to the brightness of a light numerical value, or length of a line. These study results indicated that children with mental ages of 5 years and older can perform the task but that children with mental ages between 4 and 5 years may or may not be able to complete the tasks. Finally, regardless of how the RESR is determined and set, it is imperative that the audiologists monitor the child closely for possible evidence of loudness discomfort and/or overamplification (Hawkins & Northern, 1992). It is recommended that a brief hearing evaluation and hearing aid check be performed every 3 months during the first year of hearing aid use with all children. The Desired Sensation Level (DSL)Selection Method Seewald and colleagues (1985, 1987, 1988, 1992) proposed the desired sensation level (DSL) selection method as a comprehensive hearing aid fitting model for children. A primary goal of the DSL method is to provide an amplified speech signal which is audible, comfortable, and undistorted across the broadest relevant frequency range possible (Seewald & Ross, 1987). The theoretical basis for this goal is derived from the concept of the Articulation Index (Al) which has been applied to people with hearing impairments. Seewald, Ross & Spiro (1985) mentioned two specific issues which are especially relevant to the selection of amplification characteristics for preverbal children with hearing impairments. First, the relationship between distance and the resulting speech spectrum which reaches the microphone of the hearing aid must be considered. In other words, the relationship between distance and speech spectrum characteristics as it pertains to the young child with hearing impairments must be accounted for in the selection of an idealized spectrum of speech. Second, the selection of the idealized speech spectrum must account for the complete role audition plays for the preverbal child. Therefore, this DSL approach specifies DSLs across frequencies for the long-term speech spectrum. Seewald et al. (1985) explained the DSL method in four steps：(1) specification of electroacoustic characteristics predicted to be appropriate for auditory threshold data, (2)selection of a hearing aid and earmold coupling system which should provide the best electroacoustic approximation to the predicted specifications, (3) measurement of real ear performance with the selected hearing aid and earmold, and (4) modification of the preliminary selection as based upon observed differences between predicted and real ear performance. Hawkins, Morrison, Halligan, and Cooper (1989) described a procedure in which probe-microphone measurements are utilized in conjunction with the DSL approach. Auditory thresholds, amplified levels of the long-term speech spectrum, and RESR are all expressed as SPL in the ear and measured with a probe-microphone system. As a result, there are no conversions from sound field to earphones, couplers to real ears, and so forth. Since measurements of SPL in the ear canal are being made, the probe tube is placed as close to the tympanic membrane as is practical. The procedure is carried out in the following three steps： 1) Determine the SPL in the ear canal at auditory threshold. To do this, the earphone is placed over the ear and a steady state signal is introduced at threshold. The SPL in the ear canal at threshold is measured with the probe tube system. This measurement defines the lower end of the dynamic range. An alternative method of delivering the signal to the ear is an Etymotic ER3-A insert earphone attached to the child's own earmold. 2) Select the gain and frequency response. Gain and frequency response are chosen such that the in situ gain (not insertion gain) of the hearing aid will amplify the long term speech spectrum to the desired sensation levels using the long term speech spectrum described by Cox and Moore (1988) (See Fig. 4). The target levels for this amplified speech spectrum are obtained by adding the desired sensation levels to the thresholds in the ear canal. The difference between the Cox and Moore speech spectrum and the target levels is the required in situ gain of the hearing aid. The actual procedure is as follows：(1) based upon the auditory threshold in dB HL, determine the desired SLs (Fig. 1), (2) add the desired SLs to the auditory thresholds in dB SPL in the ear canal to determine the target levels for the amplified speech spectrum (3) subtract the Cox and Moore speech spectrum levels (Fig. 4) from the target levels to determine the desired in situ gain, and (4) with the probe tube still in the ear canal from the threshold and RESR measurements, place the hearing aid on the child. Select a 60 dB SPL input and determine the in situ gain of tile hearing aid. Adjust the VCW and tone control until the best match to the desired in situ gain across frequencies is obtained. Upon obtaining the best match, the in situ gain can be added to the Cox and Moore speech spectrum to determine the final sensation levels. 3) Set the RESR. In defining the upper end of the dynamic range, or the RESR, three criteria should be satisfied：(1) loudness discomfort is not experienced, (2) risk of damage to the auditory system is minimized by avoiding excessive output levels, and (3) the width of the dynamic range is sufficient such that peaks of the amplified speech signal at normal conversational levels do not saturate the hearing aid. Since loudness discomfort measures often cannot be obtained with children, audiologists can try to satisfy the second and third criteria and hope that the first is also achieved. Determining the desired target RESR is the first step in setting the maximum output of the hearing aid. These target levels can be determined through examination of Fig. 2. Initial adjustment of the RESR control on the hearing aid to yield these target values can serve as a starting point to reduce the possibility of the child experiencing any loudness discomfort and thus becoming lightened and agitated. The RESR is determined by placing the hearing aid on the child and turning the VCW up to its maximum position before feedback. 4 swept pure tone signal of 90 dB SPL is introduced via the loudspeaker and the output in the ear canal is measured. It is assumed that the 90 dB input plus the gain of the hearing aid at the VCW position just below feedback will be sufficient to saturate the hearing aid, thus providing information on its maximum output capability. The output control can be adjusted until the closest match to target RESR is obtained. With this DSL procedure, audiologists can confirm the following：(1) the maximum output the hearing aid is capable of delivering to the child's ear (RESR), (2) the width of the dynamic range across frequency, (3) the sensation level of speech in each frequency region, and (4) the recommended VCW position and tone and output control settings to achieve the desired results (Hawkins & Northern, 1992). Seewald et al. (1991) proposed that REIR values are useful if audiologists prefer the REIR procedure to the DSL method and provided REIR values that produce similar results to the DSL procedure. Hawkins et al., (1989) presented four case studies with children and reviewed the usefulness of real ear measurements and how they can be applied within the DSL method. They demonstrated how well the hearing aid amplifies and packages the long term speech spectrum into the child's residual dynamic range. They concluded that the DSL method provides audiologists with a clear visualization of what the hearing aid is capable of acoustically providing the child. Furthermore, expressing all the measurements in SPL in the ear canal was appropriate and parsimonious. Conclusion Real ear measurements provide audiologists with a powerful, reliable, useful, and objective tool in selecting and fitting hearing aids for children. This paper reviewed how real ear measurements can be used when setting gain / frequency response and maximum output of the hearing aids for children. Furthermore, when using real ear measurements with children for fitting hearing aids, audiologists must consider many factors such as REUR, REIR, RESR, LDL and the long term speech spectrum. A desired REIR can be determined and the hearing aid adjusted to match closely the desired REIR. An appropriate RESR can be determined and the hearing aid adjusted to mmatchthe value without exceeding the LDL and producing feedback. The target long term speech spectrum can be determined and the hearing aid adjusted to amplify to the resulting desired sensation levels. There remains a need of systematic and objective investigation and thoughtfulness as we have yet to reach a consensus regarding the above factors (Seewald & Ross, 1987). In conclusion, real ear measurements offer four basic advantages: (1) better frequency resolution, (2) enhanced reliability, (3) efficiency, and (4) lower level of cooperation in selecting and verifying hearing aid performance characteristics for children with hearing impairments (Seewald, 1991). These advantages provide suggestions for further use of these measurements with children.
REFERENCES
1) Bess F, Gravel 1, & Tharpe A. 1995 Committee; Amplification for infants and children with hearing losss. Preprint., 1995 2) Byme D. Recent hearing instrument selection research suggests what? The Hearing Instruments, 1988;39(12):22-3. 3) Feigin JA, Kopun JG, Stelmachowicz PC & Gorga MP. Probe-tube microphone measures of ear canal sound pressure levels in infants and children. Ear and Hearing 989;10(4):254-8. 4) Hawkins DB, Morrison TM, Halligan PLW & Cooper WA. Use of probe tube microphone measurements in hearing aid selection for children: Some initial clinical experiences. Ear and Hearing 1989;10(5):281-7. 5) Hawkins DB & Northern JL. Probe microphone measurements with children. In H. G. Mueller DB. Hawkins, & JL Northern. (Ed), Probe microphone Measurements: Hearing aid selection and assessment(pp.159-182). San Diego: Singular Publishing Group, Inc, 1992 6) Humes LE. Reply to Denis Byrne. The Hearing Instruments, 1988;39(12):24-5. 7) Kawell ME, Kopun JG & Stelmachowicz PC. Loudness discomfort levels in children. Ear and Hearing 1988;9(3):133-6. 8) Macpherson BJ, Elfenbein JL, Schum RL, Bentler RA. Thresholds of discomfort in young children. Ear and Hearing 1991;12(3):184-9. 9) Matkin ND. Wearable amplification: a litany of persisting problems. In J. Jerger (ed.), Pediatric Audiology (pp.125-146), 1983 10) Mueller HG. Terminology and procedures. In HG Mueller, DB Hawkins & JL Northern. (Ed.), Probe microphone Measurements: Hearing aid selection and assessment (pp.41-66). San Diego: Singular Publishing Group, Inc, 1992. 11) Nelson Barlow NL, Auslander MC, Rines D & Stelmachowicz PG. Probe-tube microphone measures in hearing-impaired children and adults. Ear and Hearing 1988;9(5):243-7. 12) Seewald RC. The desired sensation level method for fitting children: Version 3.0. The Hearing Journal 1992;45(4):36-41. 13) Seewald RC & Ross M. Amplification for young hearing-impaired children. In M. Pollack (Ed.), Amplification for the Hearing-impaired(pp.213-271). Orlando, FL: Grune & Stratton, 1988 14) Seewald RC & Ross M. Selecting and verifying hearing aid performance characteristics for young children. Journal of the Academy of Rehabilitative Audiology 1987;20:25-37. 15) Seewald RC, Ross M & Spiro MK. Selecting amplification characteristics for young hearing-impaired children. Ear and Hearing, 1985;6(1):48-53. 13) Seewald RC & Ross M. Amplification for young hearing-impaired children. In M. Pollack (Ed.), Amplification for the hearing impaired(pp.213-271). Orlando, FL: Grune & Stratton, 1988. 14) Seewald RC & Ross M. Selecting and verifying hearing aid performance characteristics for young children. Journal the Academy of Rehabilitative Audiology 1987;20:25-37. 15) Seewald RC, Ross M & Spiro MK. Selecting amplification characteristics for young hearing-impaired children. Ear and Hearing 1985;6(1):48-53. 16) Stelmachowicz pG, Mace AL, Kopun JG & Carney E. Long term and short term characteristics of speech: Implications for hearing aid selection for young children. Journal of Speech and Hearing Research 1993;36:609-20. 17) Stuart A, Durieux-Smith A & Stenstorm R. Probe tube microphone measures of loudness discomfort levels in children. Ear and Hearing, 1991;12(2):140-3. 18) Tecca JE. Use of real ear measurements to verify hearing aid fitting. In M. Valence (Ed.), Strategies for selecting and verifying hearing aid fittings. NY: Thieme Medical Publishers, Inc, 1994.