International Journal of Pediatric Otorhinolaryngology 78 (2014) 290–295
Contents lists available at ScienceDirect
International Journal of Pediatric Otorhinolaryngology journal homepage: www.elsevier.com/locate/ijporl
A new hearing screening system for preschool children Wu Wenjina,b,1, Lu¨ Jingronga,b,1, Li Yuna,b, Anna Chi Shan Kam c, Michael Chi Fai Tong c, Huang Zhiwua,b,*, Wu Haoa,b,* a
Department of Otolaryngology-Head and Neck Surgery, Xinhua Hospital, Shanghai Jiaotong University School of Medicine, Shanghai 200092, China Ear Institute, Shanghai Jiaotong University School of Medicine, Shanghai 200092, China c Department of Otorhinolaryngology, Head and Neck Surgery and Institute of Human Communicative Research, The Chinese University of Hong Kong, Hong Kong, China b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 31 July 2013 Received in revised form 23 October 2013 Accepted 24 November 2013 Available online 1 December 2013
Objectives: This study aimed to investigate the practical application of Smart Hearing, a new hearing screening system for preschool children. Methods: The screening system was applied to 6288 preschool children. The system auto-tested hearing thresholds at three frequencies: 1 kHz, 2 kHz, and 4 kHz; a 30 dB hearing level (HL) was the critical intensity for passing. Children with positive results were referred for audiological evaluation (pure tone audiometry, tympanometry and distortion product otoacoustic emissions assessment, etc.). To evaluate the test accuracy, 312 children (5%) were randomly selected to receive audiology assessment. Results: In this study, 582 children (9.3%) tested positive in the screening, and the referral rate of the four age groups from 3 to 6 years old was 18.8%, 11.9%, 6.5% and 4.0%, respectively. A total of 463 children underwent audiological assessment, of which 12 cases (1.91%; 95% CI: 0.83%, 2.99%) were diagnosed with permanent hearing loss, and 75 cases (1.19%; 95% CI: 0.92%, 1.46%) were diagnosed with temporary conductive hearing loss. No mixed hearing loss was found in this study. The specificity of the system was 92.6% and the sensitivity was only 37.5%. Conclusions: This screening system is suitable for the universal hearing screening of preschool children above 4 years old, and further improvements of the system are needed to increase its sensitivity. ß 2013 Published by Elsevier Ireland Ltd.
Keywords: Pure tone screening Permanent hearing impairment Conductive hearing impairment Multimedia
1. Introduction Congenital hearing loss is detected approximately in 1.1–1.4 per 1000 births through universal newborn hearing screening programs in the United States [1,2]; however, the prevalence of hearing loss in children can reach 14.9% [3]. In addition to congenital hearing loss, hearing impairment can occur at any stage of growth and development in childhood, including late-onset, acquired, and progressive loss that cannot be identified through universal newborn hearing screening [4,5]. The preschool years, particularly between the ages of 3 and 6 years, are important for
Abbreviations: CI, confidence interval; dBA, decibels A-weighted; HL, hearing level; OAE, otoacoustic emissions; PTA, pure tone audiometry. * Corresponding authors at: Department of Otolaryngology-Head and Neck Surgery, Xinhua Hospital, Shanghai Jiaotong University School of Medicine, 1665 Kongjiang Road, Shanghai 200092, China. Tel.: +86 21 55570010; fax: +86 21 65152394. E-mail addresses: [email protected] (Z. Huang), [email protected], [email protected] (H. Wu). 1 These authors contributed equally to this work and share the first authorship. 0165-5876/$ – see front matter ß 2013 Published by Elsevier Ireland Ltd. http://dx.doi.org/10.1016/j.ijporl.2013.11.026
speech, language and cognitive development. Studies have shown that even slight or mild hearing loss can exert an adverse impact on children’s behavior, development, education and overall wellbeing [6,7]. Early identification and management of hearing loss caused by various reasons is necessary for the sake of children’s language, social interaction, learning ability, quality of life and long-term prospects [8]. Up to now, the hearing screening methods for preschool children have mainly included pure tone screening, tympanometry and otoacoustic emissions (OAE). However, with relatively few studies comparing these methods for different populations, there has been no clear consensus about a unified standard for screening. A systematic review comparing various screening protocols indicates that questionnaires and otoscopes should be used only as adjuncts to other methods of hearing screening [9]. Furthermore, tympanometry and OAEs require specialized devices and personnel, and are technically not tests of hearing thresholds. Tympanometry is used to evaluate middle-ear function, while OAE measures the function of the outer hair cells. Hence, some children having mild hearing loss or having an auditory synchrony problem may be missed by screening using OAE alone [10]. Pure tone
W. Wu et al. / International Journal of Pediatric Otorhinolaryngology 78 (2014) 290–295
screening is highly recommended and currently thought to be the most suitable test for preschool hearing screening [9], but this technique also has its own problems. Its result relies on the testers’ subjective judgment of children’s reactions, and sometimes the children, especially those under 4 years of age, are unable to cooperate with the screening instructions, thereby interfering with the procedures and results. All these problems suggest the need to seek a new kind of hearing screening system that is easier and more practical for use in preschool children. The Smart Hearing Screening system has been developed as a new type of behavioral hearing screening test based on the principle of traditional pure tone audiometry (PTA). The software conducts pure tone screening automatically and uploads real-time results to the network. This study applied the new screening system to preschool children, aiming to evaluate the feasibility of its use with regard to its screening performance, sensitivity and specificity.
291
6,372 children involved
76 children refused to take part in the screening; 8 children had been diagnosed with learning disabilities 6,288 eligible children
5,706 children passed
582 children positive on the screening
2. Materials and methods 119 children lost to follow-up
2.1. The screening system The screening system is a behavioral audiological test combining pure tone screening with multimedia and networked computers. The software, containing pure tone audio files and a set of cartoon animations for response to voice, can deliver pure tones (1 kHz, 2 kHz and 4 kHz) ranging in intensity from 20 dB HL to 60 dB HL, with adjustments at intervals of 5 dB HL. The software is installed on a Samsung GP-P6800 Smart Tablet, the platform of which is the Android 3.2 (Honeycomb) operating system with supporting Internet connection through a subscriber identity module (SIM) card. During the test, pure tones are presented via Bose QuietComfort 15 active noise-canceling headphones that effectively reduce low-frequency (100–1000 Hz) noise. The Internet is also indispensable in this system for uploading results. All the facilities have to conduct acoustic calibration before they are used. The complete screening procedure includes three parts: guidance, formal screening and uploading of results. Children passing the guidance would undergo the formal hearing test independently, and the results should be uploaded to the network immediately following a standard presentation by the device manufacturer before the next round of screening. Children with thresholds above 30 dB HL at any frequency of 1 kHz, 2 kHz and 4 kHz in either ear were deemed to have tested positive in the screening. Those who did not pass the guidance should also be classified as referral cases (see Appendix). 2.2. Subjects This was a cross-sectional study with a sample of 6372 preschool children enrolled from 41 kindergartens spread throughout Yangpu District in Shanghai (a total of 106 kindergartens in this area). Among the total sample, 76 children refused to take part in the screening, and another eight children were diagnosed with learning disabilities such as growth retardation, autism and cerebral hypoplasia before by designated medical institutions. These 84 children were excluded from this study. Therefore, 6288 eligible (98.7%) preschool children, consisting of 3282 (52.2%) boys and 3006 (47.8%) girls with a mean age of 5.05 years at testing (range = 3.01–6.92, standard deviation = 0.71), participated in this study (Fig. 1). All of these participants were not diagnosed with mental abnormality before their inclusion in this project, and all had normal vision or normal corrected visual acuity. Written informed consent from children’s parents was obtained prior to the screening. This study was approved by the Medical Ethics Committee of the Xinhua Hospital affiliated with
463 children referral
376 children with normal hearing
12 cases of permanent hearing loss (bilateral in 7 and unilateral in 5) 75 cases of conductive hearing loss (bilateral in 30 and unilateral in 45)
Fig. 1. Flow diagram of participants through the study.
Shanghai Jiaotong University School of Medicine (No. XHEC-C2012-006). 2.3. Screening procedures All the screening personnel, with normal hearing, had undertaken specific training in the screening procedures. Children were tested in relatively quiet rooms within each school, in which the ambient noise did not exceed 55 dBA (between 39.8 and 50.5 dBA) as measured by a TES 1357 sound level meter. The testers began the hearing screening following routine calibration of the device each day. A program of daily calibration was set in the Smart Hearing Screening system, which aims to detect system malfunctions and prevent audio outputs that produce significant bias. Through the self-test settings, testers had to ensure that pure tones of 30 dB HL could be clearly heard symmetrically on both sides by headphones in the test environment to prevent issues with device connection. Children who did not pass the test in the initial screening were immediately rescreened. Those who failed in the rescreening and in the guidance were referred for diagnostic assessment (Fig. 1). In order to estimate the sensitivity and specificity of the new screening system, 312 children (5%) were randomly selected from the 6288 eligible children before the screening. Regardless of the screening results, all of these 312 children (mean age = 5.06, standard deviation = 0.72) received audiological assessment by pure tone audiometry (PTA) or play PTA supplemented by tympanometry and distortion product OAE.
W. Wu et al. / International Journal of Pediatric Otorhinolaryngology 78 (2014) 290–295
292
2.4. Hearing loss assessment procedures The children referred to the referral center (Shanghai Children’s Hearing and Speech Center) had received otological examination and audiological evaluation, including otoscope, pure tone audiometry (MADSEN, CONERA), tympanometry (GSI Tympstar) and distortion product OAEs (MADSEN, CAPELLA, f2:f1 = 1.22, L2/ L1 = 55/65 dB SPL). Play PTA or visual reinforcement audiometry was applied to the small proportion of children that could not complete the PTA. Hearing loss was categorized as mild (26–40 dB HL), moderate (41–60 dB HL), severe (61–80 dB HL), or profound (>80 dB HL) based on the average value of threshold at 0.5, 1, 2 and 4 kHz [11]. 2.5. Data analysis SPSS software (version 17.0 for Windows, SPSS Inc., Chicago, IL, USA) was used for all statistical analyses. The x2 test was used to compare the categorical data to test for trends in referral rates across ages. The corresponding 95% confidence interval (CI) was calculated with normal distribution. For all of the statistical analyses, a probability value of P < 0.05 was considered statistically significant. 3. Results 3.1. Screening test results A total of 6288 children aged from 3 to 6 years were enrolled for the screening, of which 582 children (9.3%) met the failure criteria (Fig. 1). There were 375 cases (6.0%) of unilateral referral and 183 cases (2.9%) of bilateral referral, and 24 cases (0.4%) failed in the guidance. The results proportion of the screening test did not differ significantly between boys and girls (x2 = 3.275, P = 0.070). However, a trend x2 test showed a significant influence of age on screening performance (x2 = 112.238, P < 0.01). The referral rate of the screening showed a noticeable trend of reduction with increasing age (Table 1).
unilateral mild hearing loss, and one each with unilateral moderate, severe, or profound hearing loss. All of the children with permanent hearing loss diagnosed in our study had sensorineural hearing loss. There were 75 children (1.19%; 95% CI: 0.92%, 1.46%) that had temporary conductive hearing loss, and all of them were diagnosed with otitis media effusion, comprising bilateral hearing loss in 30 children and unilateral hearing loss in 45 children, ranging from mild to moderate. No cases of mixed hearing loss were found. In this study, 86.2% (75/87) of the hearing impairment was caused by otitis media effusion. Of the children who were referred, 376 had been confirmed to have normal hearing. It is worth mentioning that 146 children (2.32%; 95% CI: 1.95%, 2.69%) among them were diagnosed with impacted cerumen. Their wax had been removed before they received the diagnostic audiologic test. Another 52 children were diagnosed with otitis media effusion with average pure tone air conduction thresholds between 15 dB HL and 25 dB HL, so the prevalence of otitis media effusion in this study was 2.02%. 3.3. Sensitivity and specificity For the 312 children selected in advance for the accuracy study, the sensitivity was approximately 37.5% (95% CI: 16.3%, 64.1%) with a specificity of about 92.6% (95% CI: 88.8%, 95.2%). And the predictive value of a positive test was 21.4% (95% CI: 9.0%, 41.5%), while the negative predictive value was 96.5% (95% CI: 93.4%, 98.2%) (Table 2). Among the 312 children, normal hearing was confirmed in 296 children; 16 children were diagnosed with conductive hearing impairment through the detailed audiological examination. Among these, six children tested referred in the screening (four had bilateral temporary conductive hearing loss and two had unilateral temporary conductive hearing loss); 10 children passed the screening test (one was diagnosed with bilateral temporary conductive hearing loss and nine with unilateral temporary conductive hearing loss) (Table 3). These 16 children underwent hearing retests by pure tone audiometry after 3 months, which showed that they had normal hearing. 4. Discussion
3.2. Confirmation of hearing impairment
Refer
Pure tone screening is most recommended for preschool children’s hearing screening. Traditional pure tone screening requires that children respond to sound stimuli by raising their hands. However, there is such variability in children’s abilities to express themselves, understand instructions and cooperate within the population aged 3–6 years that judgment of the test can be difficult [12]. The new type of automatic screening method was developed based on the principle of pure tone screening and was therefore tested for feasibility and effectiveness. This study found that the average referral rate was 9.3%, with the referral number in an acceptable range that will not put excessive pressure on the follow-up institutions, especially for those children above the age of 4 years. We also found that the referral rate increased significantly with decreasing age, and only 81.2% of 3-year-old children passed the screening (Table 1). Table 2 Test performance measured for the screening system in comparison with pure tone audiometry results.
Referral diagnostic assessment was recommended to 582 cases, while 463 children (79.6%) actually received audiological evaluation in the referral center and the other 119 cases (20.4%) were lost to follow-up (Fig. 1). A total of 87 children (prevalence of 1.38%; 95% CI: 1.09%, 1.67%) were confirmed to have hearing impairment. Of those, 12 children (1.91%; 95% CI: 0.83%, 2.99%) were diagnosed with permanent hearing loss, among which six children presented with bilateral mild hearing loss and one child with bilateral moderate hearing loss; two children presented with Table 1 Distribution of the screening results according to gender and age. Characteristics
Total
Pass No. (%)
Gender Boys Girls Age 3y 4y 5y 6y a
3282 3006 x2 = 3.275, P = 0.070
2999 (91.4) 2707 (90.1)
283 (8.6) 299 (9.9)
495 2347 2868 578 x2 = 112.238, P < 0.01a
402 2068 2681 555
93 279 187 23
No. (81.2) (88.1) (93.5) (96.0)
(18.8) (11.9) (6.5) (4.0)
Trend x2 test was applied by comparison across the four age groups.
Hearing impairment Normal Validity indices (%) Diagnosed by PTA
Screening 6 Referred Passed 10
Sensitivity Specificity PPV 22 274
37.5
92.6
PPV, positive predictive value; NPV, negative predictive value.
NPV
21.4 96.5
W. Wu et al. / International Journal of Pediatric Otorhinolaryngology 78 (2014) 290–295
293
Table 3 Characteristics of the 16 children confirmed with conductive hearing loss. Patient
Gender
Screening system
Age
Degree of hearing loss
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
F M M M M F F F M F F M M M M F
R R R R R R P P P P P P P P P P
3.64 4.24 4.56 4.59 4.97 5.20 4.59 4.32 4.19 5.45 4.78 5.30 4.88 4.99 4.76 4.83
Bilateral mild Bilateral mild Unilateral mild Bilateral mild Bilateral mild Unilateral mild Unilateral mild Bilateral mild Unilateral mild Unilateral mild Unilateral moderate Unilateral mild Unilateral mild Unilateral mild Unilateral moderate Unilateral mild
Thresholds (dB HL) Left
Right
30 32.5 30 27.5 36.25 32.5
32.5 42.5
32.50 33.75
27.5 31.25 28.75 28.75 35
46.25 31.25 40 36.25 41.25 26.25
M, male; F, female; P, pass; R, referral.
Although the main procedures of the new system can be completed by children independently, similar to traditional pure tone screening, the accuracy of the screening results is also influenced by how well the subjects cooperate with the test. Children less than 4 years old were more likely to be uncooperative in the screening, leading to a higher referral rate of 18.8%. However, the screening performance was much better in children more than 4 years old, especially for 5- and 6-year-old children, who had relatively high pass rates of 93.5% and 96.0%, respectively. Therefore, the screening system was more suitable for children aged more than 4 years. The new system could also be used for younger children, but combined screening might be a better alternative to increase the acceptability and accuracy of the screening. Through this screening project, 1.38% of preschool children were diagnosed with hearing loss, which is lower than the 1.8% reported by a study in the United States [13]. Children with hearing loss caused by impacted cerumen accepted the otologist’s check first when referred to the referral center and their cerumen had been removed before the diagnostic test; therefore, audiological assessment results for this group were in the normal range and were not included in the statistical incidence of hearing loss. It is reported that earwax embolism accounts for approximately 8.9% of the incidence of hearing loss [14]. The prevalence of otitis media with effusion was 2.02%, which is close to the lowest rates reported previously (2.2–9.86%) [15–18]. Otitis media with effusion ranked first as a cause of hearing loss for preschool children, accounting for 86.2% of the children with hearing impairment, which is consistent with the previous study (Fig. 1) [19]. In our study, permanent hearing impairment was found in 12 cases with an incidence rate of 1.91%; a finding consistent with that in other epidemiological reports [9,20,21]. The percentage of unilateral permanent hearing impairment was 0.80% (5/6288), higher than the proportion reported in our previous studies of preschool children’s hearing screening using pediatric audiometry [22]. The pure tone was presented through headphones in our system so that the hearing conditions could be detected for the left and right ear, respectively, avoiding their mutual interference similar to that in a pediatric audiometer screening. As for the 119 children lost to follow-up, there may be a part of hearing loss cases that were not counted into the statistics regarding the final prevalence. Therefore, it is believed that the actual prevalence of hearing loss may be slightly higher than the statistical data. This screening project was conducted using a new device without a gold standard. Owing to the heavy workload of conducting audiological evaluation for all the children involved, approximately 5% of the children were selected randomly for the
accuracy study. According to the data calculated, the sensitivity (37.5%) of the new system was lower than that of the pure tone sweep test, tympanometry, and transient evoked OAE in most of the reported studies, but the specificity (92.6%) was better than that of transient evoked OAE and close to that of pure tone screening [9]. In this study, a few false positive cases were found, leading to a lower positive predictive value, but more importantly, some children with hearing impairment were undetected in the screening (Table 2) [5]. There may be several reasons for the false negative and positive cases. First, the screening system tests thresholds at only three frequencies (1 kHz, 2 kHz and 4 kHz); however, otitis media with effusion had a greater impact on the low frequency, so those who had hearing loss only at low frequencies may be overlooked in this screening. Second, as mentioned previously, children in our project did not accept external ear canal cleaning before our screening, and a considerable proportion of children may have been detected in the screening with hearing loss caused by earwax embolism. The accuracy of these children’s screening results is difficult to verify in this study. Third, the average interval between screening and referral date was more than 1 week, so the diagnostic hearing assessment cannot fully represent the hearing level in the screening. We also found that only one child whose hearing loss went undetected during screening was diagnosed with mild bilateral conductive hearing loss and most children were diagnosed with unilateral conductive hearing loss (Table 3). Our project is the first study for the new screening system, and the sensitivity data obtained based on a random group of children is much too low. This is likely to have been a result of the unskilled nature of the personnel conducting the screening and imperfect system settings. Further research and improvement is needed to increase the sensitivity of the system in the future. For example, the screening system should include more automatic checks for user noncooperation instead of human supervision, and pure tones should be played irregularly at unequal intervals in case children tap the screen regularly without hearing a sound, which lead to the occurrence of false negative cases. However, in practice, the screening system also showed some advantages. All screening records can be uploaded and queried repeatedly on the network, and the results are judged automatically by the system itself, reducing data loss and human error in the process. The ease of operation made it possible for teachers in kindergarten to complete the screening following simple training. The costs of the new system were much less than that of pediatric audiometry, tympanometry or otoacoustic emissions. Purchasing a Smart Hearing Screening device set, which includes a tablet and
294
W. Wu et al. / International Journal of Pediatric Otorhinolaryngology 78 (2014) 290–295
noise-canceling headphones, only costs USD $800 or less, and no extra consumables are required during the screening process. In terms of efficiency, as calculated by the test time for the initial screening of each child, the entire screening can be accomplished by a single individual in 4.4 min on average, while the formal test takes approximately 2.5 min. In addition, as children can complete the formal test by themselves, one tester can conduct the test for several subjects simultaneously, which dramatically reduces the average test time. If four children are screened at the same time, for example, it takes 2 min on average to explain the instructions for each person (8 min in total), and the formal test takes an extra 2.5 min. Therefore, the screening process takes only 2.6 min for one child. The time taken for older children is shorter, especially in terms of the length of guidance required before the formal test starts. In this study, we spent 3 months to complete screening for 6288 preschool children. Two testers were required with eight sets of devices and each worked for 3 h each day. The monthly wage was USD $1000 per person, and therefore the total cost of labor and the devices was approximately USD $12,400. The per capita cost of our initial screening program was less than USD $2, and these costs will reduce further over time. Therefore, the Smart Hearing Screening system is expected to become one of the distinctive screening methods for the early detection of hearing loss in preschool children. Further developments of the software and device will be required to improve the sensitivity of the screening system. We propose that this screening system is feasible on the basis of it being automated, easy to use, cheap and highly efficient. In addition, its interesting properties can attract preschool children to actively participate with the screening. At the same time, the portability of the device and the simple principles of the system make it readily accepted by kindergarten teachers and preschool education institutions alike. This study was our first attempt to apply the Smart Hearing Screening system to preschoolers, but we believe that this system will become increasingly feasible with continual software and device improvements. Further research is needed to work out the problems such as the accuracy of the new system in different age groups, the frequency specificity of the screening results related to the final diagnosis, the impact of earwax on preschool children’s hearing, the accuracy of combined screening methods, and it is also desirable to test the application of this new system in children diagnosed with a hearing impairment. 5. Conclusion Our study is the first application of the screening system in preschoolers. This study found a considerable proportion of hearing impairment in preschool children. The new screening system has advantages that include ease of use, a multimedia platform that holds the attention of children, and automatic judgment and recording mechanisms. At present, the Smart Hearing Screening system is more suitable for children older than 4 years. Further studies are warranted to improve the system to make it more practical as a method for hearing screening in preschool children. Acknowledgements We are grateful to the doctors from Yangpu District Women and Child Health Institution in Shanghai for their assistance throughout this work. We also thank all the participating subjects, their parents and teachers. This work was supported by grants from the Health Ministry Special Fund, China (grant number 201202005) and the 12th Five-Year National Key Technologies R&D Program (2012BAI12B01).
Appendix Instruction manual for the Smart Hearing Screening system: A.1. Guidance First, the screening personnel should explain the directions of the screening test to the child and complete the training process. The child will be told to touch the cartoon hat on the screen immediately when hearing a pure tone, the sound like ‘‘DiDi’’. If the response is consistent with the tone presentation, a cartoon rabbit with different random facial expressions and in different colors will appear on the screen. If the child touches the screen without a sound stimulus, the cartoon rabbit will not appear. To ensure that the child understands the instructions and can be familiarized with the task, the screening personnel will put on the headphone for the child and begin to practice in the guidance mode. (1) The personnel first generate a 1000 Hz pure tone of 60 dB HL to the child’s right ear, at the same time observing whether the child responds correctly to the tone. (2) If the reaction is correct, the personnel will continue to generate a 1000 Hz pure tone of 40 dB HL and continue the training of the left ear. (3) If the child does not respond to the 60 dB HL pure tone on two occasions, it is considered that the child has not understood the guidance and is classified as an incomplete case. A.2. Formal screening Children demonstrating understanding of the guidance will undergo the hearing test in the formal screening mode. The entire screening test will be completed by children independently. (1) Starting with the right ear at a 1000 Hz pure tone of 40 dB HL, the Smart Hearing Screening system changes the frequency and intensity of the tones presented automatically based on the child’s reactions, thus, simulating the procedure of a traditional pure tone audiometry. (2) Following a satisfactory positive response, the system will reduce the level of the tone in 10 dB steps until no further response is obtained followed by increases in the level of the tone in 5 dB steps until a response is obtained. After the first response using an ascending approach, the system will continue the 10-dB-descending, 5-dB-ascending sequence until the subject responds twice at the same level; this represents the hearing threshold level. (3) If there is no response at the start, the system will increase the level to 60 dB HL. If the tone is still inaudible, the symbol ‘‘-’’ will be recorded. (4) Next, the system will proceed to 2000 Hz, 4000 Hz in that order and the process repeated for the left ear. The duration of the presented tone is approximately 1.5 s and the interval between the tones is approximately 2 s. (5) During the formal screening, personnel must observe on one side of the subject in order to avoid them indiscriminately touching the screen. A.3. Results evaluation and upload At the end of the screening test, the tablet will display the subject’s pure tone air conduction thresholds at three frequencies for both ears. Subjects with all auditory thresholds less than or equal to 30 dB HL
W. Wu et al. / International Journal of Pediatric Otorhinolaryngology 78 (2014) 290–295
are deemed to have tested negative for hearing loss and pass the screening. (1) Personnel have to click the ‘‘Done’’ button to upload the results to the network terminal before proceeding to the next round of screening. (2) As long as the threshold at one frequency is above 30 dB HL, a positive result will be given and the system will automatically be prompted to continue. The last attempt results should be taken as valid. (3) All the test results uploaded are available on the specified website (http://shshearing.com:8080/shs_webportal/login) at any time by any computer with Internet access. The Smart Hearing Screening system provides data analysis and generates a detailed hearing screening report for every child.
References [1] S. Mehra, R.D. Eavey, D.G. Keamy Jr., The epidemiology of hearing impairment in the United States: newborns, children, and adolescents, Otolaryngol. Head Neck Surg. 140 (2009) 461–472. [2] US Centers for Disease Control and Prevention, Summary of 2009 National CDC EHDI Data, Version A, 2009 Available from: http://www.cdc.gov/ncbddd/hearingloss/2009-Data/2009_EHDI_HSFS_Summary_508_OK.pdf (revised January 2012). [3] A.S. Niskar, S.M. Kiezak, A. Holmes, E. Esteban, C. Rubin, D.J. Brody, Prevalence of hearing loss among children 6–19 years of age, J. Am. Med. Assoc. 279 (1998) 1071–1075. [4] Joint Committee on Infant Hearing, Year 2007 position statement: principles and guidelines for early hearing detection and intervention programs, Pediatrics 120 (2007) 898–921. [5] H. Skarzynski, A. Piotrowska, Screening for pre-school and school-age hearing problems: European Consensus Statement, Int. J. Pediatr. Otorhinolaryngol. 76 (2012) 120–121. [6] M. Khairi Md Daud, R.M. Noor, N.A. Rahman, D.S. Sidek, A. Mohamad, The effect of mild hearing loss on academic performance in primary school children, Int. J. Pediatr. Otorhinolaryngol. 74 (2010) 67–70.
295
[7] M. Wake, Z. Poulakis, E.K. Hughes, C. Carey-Sargeant, F.W. Rickards, Hearing impairment: a population study of age at diagnosis, severity, and language outcomes at 7–8 years, Arch. Dis. Child. 90 (2005) 238–244. [8] C. Yoshinaga-Itano, A.L. Sedey, D.K. Coulter, A.L. Mehl, Language of early- and later-identified children with hearing loss, Pediatrics 102 (1998) 1161–1171. [9] J. Bamford, H. Fortnum, K. Bristow, J. Smith, G. Vamvakas, L. Davies, et al., Current practice, accuracy, effectiveness and cost-effectiveness of the school entry hearing screen, Health Technol. Assess. 11 (2007) 1–168. [10] C.I. Berlin, T. Morlet, L.J. Hood, Auditory neuropathy/dyssynchrony: its diagnosis and management, Pediatr. Clin. North Am. 50 (2003) 331–340. [11] Report of the first informal consultation on future programme developments for prevention of deafness and hearing impairment, WHO, Geneva, 1997, Available from: http://whqlibdoc.who.int/hq/1997/WHO_PDH_97.3.pdf. [12] D.R. Halloran, T.C. Wall, H.H. Evans, J.M. Hardin, A.L. Woolley, Hearing screening at well-child visits, Arch. Pediatr. Adolesc. Med. 159 (2005) 949–955. [13] Y.C. Serpanos, F. Jarmel, Quantitative and qualitative follow-up outcomes from a preschool audiologic screening program: perspective over a decade, Am. J. Audiol. 16 (2007) 4–12. [14] P.R. Adobamen, F.O. Oqisi, Hearing loss due to wax impaction, Niger. Q. J. Hosp. Med. 22 (2012) 117–120. [15] O.C. Erdivanli, Z.O. Coskun, K.C. Kazikdas, M. Demirci, Prevalence of otitis media with effusion among primary school children in Eastern Black Sea, in Turkey and the effect of smoking in the development of otitis media with effusion, Indian J. Otolaryngol. Head Neck Surg. 64 (2012) 17–21. [16] E. Okur, I. Yildirim, M. Akif Kilic, S. Guzelsoy, Prevalence of otitis media with effusion among primary school children in Kahramanmaras, in Turkey, Int. J. Pediatr. Otorhinolaryngol. 68 (2004) 557–562. [17] H.C. Rushton, M.C. Tong, V. Yue, P.J. Wormald, C.A. Hasselt, Prevalence of otitis media with effusion in multicultural schools in Hong Kong, J. Laryngol. Otol. 111 (1997) 804–806. [18] M.C. Tong, V. Yue, P.K. Ku, P.S. Lo, C.A. van Hasselt, Screening for otitis media with effusion to measure its prevalence in Chinese children in Hong Kong, Ear Nose Throat J. 79 (2000) 626–630. [19] M.A. Al-Rowaily, .A.I. AlFayez, M.S. AlJomiey, A.M. AlBadr, M.A. Abolfotouh, Hearing impairments among Saudi preschool children, Int. J. Pediatr. Otorhinolaryngol. 76 (2012) 1674–1677. [20] H.M. Fortnum, A.Q. Summerfield, D.H. Marshall, A.C. Davis, J.M. Bamford, Prevalence of permanent childhood hearing impairment in the United Kingdom and implications for universal neonatal hearing screening: questionnaire based ascertainment study, Br. Med. J. 323 (2001) 536–540. [21] C. MacAndie, H. Kubba, M. McFarlane, Epidemiology of permanent childhood hearing loss in Glasgow, 1985–1994, Scott. Med. J. 48 (2003) 117–119. [22] J. Lu¨, Z. Huang, T. Yang, Y. Li, L. Mei, M. Xiang, et al., Screening for delayed-onset hearing loss in preschool children who previously passed the newborn hearing screening, Int. J. Pediatr. Otorhinolaryngol. 75 (2011) 1045–1049.
Contents lists available at ScienceDirect
International Journal of Pediatric Otorhinolaryngology journal homepage: www.elsevier.com/locate/ijporl
A new hearing screening system for preschool children Wu Wenjina,b,1, Lu¨ Jingronga,b,1, Li Yuna,b, Anna Chi Shan Kam c, Michael Chi Fai Tong c, Huang Zhiwua,b,*, Wu Haoa,b,* a
Department of Otolaryngology-Head and Neck Surgery, Xinhua Hospital, Shanghai Jiaotong University School of Medicine, Shanghai 200092, China Ear Institute, Shanghai Jiaotong University School of Medicine, Shanghai 200092, China c Department of Otorhinolaryngology, Head and Neck Surgery and Institute of Human Communicative Research, The Chinese University of Hong Kong, Hong Kong, China b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 31 July 2013 Received in revised form 23 October 2013 Accepted 24 November 2013 Available online 1 December 2013
Objectives: This study aimed to investigate the practical application of Smart Hearing, a new hearing screening system for preschool children. Methods: The screening system was applied to 6288 preschool children. The system auto-tested hearing thresholds at three frequencies: 1 kHz, 2 kHz, and 4 kHz; a 30 dB hearing level (HL) was the critical intensity for passing. Children with positive results were referred for audiological evaluation (pure tone audiometry, tympanometry and distortion product otoacoustic emissions assessment, etc.). To evaluate the test accuracy, 312 children (5%) were randomly selected to receive audiology assessment. Results: In this study, 582 children (9.3%) tested positive in the screening, and the referral rate of the four age groups from 3 to 6 years old was 18.8%, 11.9%, 6.5% and 4.0%, respectively. A total of 463 children underwent audiological assessment, of which 12 cases (1.91%; 95% CI: 0.83%, 2.99%) were diagnosed with permanent hearing loss, and 75 cases (1.19%; 95% CI: 0.92%, 1.46%) were diagnosed with temporary conductive hearing loss. No mixed hearing loss was found in this study. The specificity of the system was 92.6% and the sensitivity was only 37.5%. Conclusions: This screening system is suitable for the universal hearing screening of preschool children above 4 years old, and further improvements of the system are needed to increase its sensitivity. ß 2013 Published by Elsevier Ireland Ltd.
Keywords: Pure tone screening Permanent hearing impairment Conductive hearing impairment Multimedia
1. Introduction Congenital hearing loss is detected approximately in 1.1–1.4 per 1000 births through universal newborn hearing screening programs in the United States [1,2]; however, the prevalence of hearing loss in children can reach 14.9% [3]. In addition to congenital hearing loss, hearing impairment can occur at any stage of growth and development in childhood, including late-onset, acquired, and progressive loss that cannot be identified through universal newborn hearing screening [4,5]. The preschool years, particularly between the ages of 3 and 6 years, are important for
Abbreviations: CI, confidence interval; dBA, decibels A-weighted; HL, hearing level; OAE, otoacoustic emissions; PTA, pure tone audiometry. * Corresponding authors at: Department of Otolaryngology-Head and Neck Surgery, Xinhua Hospital, Shanghai Jiaotong University School of Medicine, 1665 Kongjiang Road, Shanghai 200092, China. Tel.: +86 21 55570010; fax: +86 21 65152394. E-mail addresses: [email protected] (Z. Huang), [email protected], [email protected] (H. Wu). 1 These authors contributed equally to this work and share the first authorship. 0165-5876/$ – see front matter ß 2013 Published by Elsevier Ireland Ltd. http://dx.doi.org/10.1016/j.ijporl.2013.11.026
speech, language and cognitive development. Studies have shown that even slight or mild hearing loss can exert an adverse impact on children’s behavior, development, education and overall wellbeing [6,7]. Early identification and management of hearing loss caused by various reasons is necessary for the sake of children’s language, social interaction, learning ability, quality of life and long-term prospects [8]. Up to now, the hearing screening methods for preschool children have mainly included pure tone screening, tympanometry and otoacoustic emissions (OAE). However, with relatively few studies comparing these methods for different populations, there has been no clear consensus about a unified standard for screening. A systematic review comparing various screening protocols indicates that questionnaires and otoscopes should be used only as adjuncts to other methods of hearing screening [9]. Furthermore, tympanometry and OAEs require specialized devices and personnel, and are technically not tests of hearing thresholds. Tympanometry is used to evaluate middle-ear function, while OAE measures the function of the outer hair cells. Hence, some children having mild hearing loss or having an auditory synchrony problem may be missed by screening using OAE alone [10]. Pure tone
W. Wu et al. / International Journal of Pediatric Otorhinolaryngology 78 (2014) 290–295
screening is highly recommended and currently thought to be the most suitable test for preschool hearing screening [9], but this technique also has its own problems. Its result relies on the testers’ subjective judgment of children’s reactions, and sometimes the children, especially those under 4 years of age, are unable to cooperate with the screening instructions, thereby interfering with the procedures and results. All these problems suggest the need to seek a new kind of hearing screening system that is easier and more practical for use in preschool children. The Smart Hearing Screening system has been developed as a new type of behavioral hearing screening test based on the principle of traditional pure tone audiometry (PTA). The software conducts pure tone screening automatically and uploads real-time results to the network. This study applied the new screening system to preschool children, aiming to evaluate the feasibility of its use with regard to its screening performance, sensitivity and specificity.
291
6,372 children involved
76 children refused to take part in the screening; 8 children had been diagnosed with learning disabilities 6,288 eligible children
5,706 children passed
582 children positive on the screening
2. Materials and methods 119 children lost to follow-up
2.1. The screening system The screening system is a behavioral audiological test combining pure tone screening with multimedia and networked computers. The software, containing pure tone audio files and a set of cartoon animations for response to voice, can deliver pure tones (1 kHz, 2 kHz and 4 kHz) ranging in intensity from 20 dB HL to 60 dB HL, with adjustments at intervals of 5 dB HL. The software is installed on a Samsung GP-P6800 Smart Tablet, the platform of which is the Android 3.2 (Honeycomb) operating system with supporting Internet connection through a subscriber identity module (SIM) card. During the test, pure tones are presented via Bose QuietComfort 15 active noise-canceling headphones that effectively reduce low-frequency (100–1000 Hz) noise. The Internet is also indispensable in this system for uploading results. All the facilities have to conduct acoustic calibration before they are used. The complete screening procedure includes three parts: guidance, formal screening and uploading of results. Children passing the guidance would undergo the formal hearing test independently, and the results should be uploaded to the network immediately following a standard presentation by the device manufacturer before the next round of screening. Children with thresholds above 30 dB HL at any frequency of 1 kHz, 2 kHz and 4 kHz in either ear were deemed to have tested positive in the screening. Those who did not pass the guidance should also be classified as referral cases (see Appendix). 2.2. Subjects This was a cross-sectional study with a sample of 6372 preschool children enrolled from 41 kindergartens spread throughout Yangpu District in Shanghai (a total of 106 kindergartens in this area). Among the total sample, 76 children refused to take part in the screening, and another eight children were diagnosed with learning disabilities such as growth retardation, autism and cerebral hypoplasia before by designated medical institutions. These 84 children were excluded from this study. Therefore, 6288 eligible (98.7%) preschool children, consisting of 3282 (52.2%) boys and 3006 (47.8%) girls with a mean age of 5.05 years at testing (range = 3.01–6.92, standard deviation = 0.71), participated in this study (Fig. 1). All of these participants were not diagnosed with mental abnormality before their inclusion in this project, and all had normal vision or normal corrected visual acuity. Written informed consent from children’s parents was obtained prior to the screening. This study was approved by the Medical Ethics Committee of the Xinhua Hospital affiliated with
463 children referral
376 children with normal hearing
12 cases of permanent hearing loss (bilateral in 7 and unilateral in 5) 75 cases of conductive hearing loss (bilateral in 30 and unilateral in 45)
Fig. 1. Flow diagram of participants through the study.
Shanghai Jiaotong University School of Medicine (No. XHEC-C2012-006). 2.3. Screening procedures All the screening personnel, with normal hearing, had undertaken specific training in the screening procedures. Children were tested in relatively quiet rooms within each school, in which the ambient noise did not exceed 55 dBA (between 39.8 and 50.5 dBA) as measured by a TES 1357 sound level meter. The testers began the hearing screening following routine calibration of the device each day. A program of daily calibration was set in the Smart Hearing Screening system, which aims to detect system malfunctions and prevent audio outputs that produce significant bias. Through the self-test settings, testers had to ensure that pure tones of 30 dB HL could be clearly heard symmetrically on both sides by headphones in the test environment to prevent issues with device connection. Children who did not pass the test in the initial screening were immediately rescreened. Those who failed in the rescreening and in the guidance were referred for diagnostic assessment (Fig. 1). In order to estimate the sensitivity and specificity of the new screening system, 312 children (5%) were randomly selected from the 6288 eligible children before the screening. Regardless of the screening results, all of these 312 children (mean age = 5.06, standard deviation = 0.72) received audiological assessment by pure tone audiometry (PTA) or play PTA supplemented by tympanometry and distortion product OAE.
W. Wu et al. / International Journal of Pediatric Otorhinolaryngology 78 (2014) 290–295
292
2.4. Hearing loss assessment procedures The children referred to the referral center (Shanghai Children’s Hearing and Speech Center) had received otological examination and audiological evaluation, including otoscope, pure tone audiometry (MADSEN, CONERA), tympanometry (GSI Tympstar) and distortion product OAEs (MADSEN, CAPELLA, f2:f1 = 1.22, L2/ L1 = 55/65 dB SPL). Play PTA or visual reinforcement audiometry was applied to the small proportion of children that could not complete the PTA. Hearing loss was categorized as mild (26–40 dB HL), moderate (41–60 dB HL), severe (61–80 dB HL), or profound (>80 dB HL) based on the average value of threshold at 0.5, 1, 2 and 4 kHz [11]. 2.5. Data analysis SPSS software (version 17.0 for Windows, SPSS Inc., Chicago, IL, USA) was used for all statistical analyses. The x2 test was used to compare the categorical data to test for trends in referral rates across ages. The corresponding 95% confidence interval (CI) was calculated with normal distribution. For all of the statistical analyses, a probability value of P < 0.05 was considered statistically significant. 3. Results 3.1. Screening test results A total of 6288 children aged from 3 to 6 years were enrolled for the screening, of which 582 children (9.3%) met the failure criteria (Fig. 1). There were 375 cases (6.0%) of unilateral referral and 183 cases (2.9%) of bilateral referral, and 24 cases (0.4%) failed in the guidance. The results proportion of the screening test did not differ significantly between boys and girls (x2 = 3.275, P = 0.070). However, a trend x2 test showed a significant influence of age on screening performance (x2 = 112.238, P < 0.01). The referral rate of the screening showed a noticeable trend of reduction with increasing age (Table 1).
unilateral mild hearing loss, and one each with unilateral moderate, severe, or profound hearing loss. All of the children with permanent hearing loss diagnosed in our study had sensorineural hearing loss. There were 75 children (1.19%; 95% CI: 0.92%, 1.46%) that had temporary conductive hearing loss, and all of them were diagnosed with otitis media effusion, comprising bilateral hearing loss in 30 children and unilateral hearing loss in 45 children, ranging from mild to moderate. No cases of mixed hearing loss were found. In this study, 86.2% (75/87) of the hearing impairment was caused by otitis media effusion. Of the children who were referred, 376 had been confirmed to have normal hearing. It is worth mentioning that 146 children (2.32%; 95% CI: 1.95%, 2.69%) among them were diagnosed with impacted cerumen. Their wax had been removed before they received the diagnostic audiologic test. Another 52 children were diagnosed with otitis media effusion with average pure tone air conduction thresholds between 15 dB HL and 25 dB HL, so the prevalence of otitis media effusion in this study was 2.02%. 3.3. Sensitivity and specificity For the 312 children selected in advance for the accuracy study, the sensitivity was approximately 37.5% (95% CI: 16.3%, 64.1%) with a specificity of about 92.6% (95% CI: 88.8%, 95.2%). And the predictive value of a positive test was 21.4% (95% CI: 9.0%, 41.5%), while the negative predictive value was 96.5% (95% CI: 93.4%, 98.2%) (Table 2). Among the 312 children, normal hearing was confirmed in 296 children; 16 children were diagnosed with conductive hearing impairment through the detailed audiological examination. Among these, six children tested referred in the screening (four had bilateral temporary conductive hearing loss and two had unilateral temporary conductive hearing loss); 10 children passed the screening test (one was diagnosed with bilateral temporary conductive hearing loss and nine with unilateral temporary conductive hearing loss) (Table 3). These 16 children underwent hearing retests by pure tone audiometry after 3 months, which showed that they had normal hearing. 4. Discussion
3.2. Confirmation of hearing impairment
Refer
Pure tone screening is most recommended for preschool children’s hearing screening. Traditional pure tone screening requires that children respond to sound stimuli by raising their hands. However, there is such variability in children’s abilities to express themselves, understand instructions and cooperate within the population aged 3–6 years that judgment of the test can be difficult [12]. The new type of automatic screening method was developed based on the principle of pure tone screening and was therefore tested for feasibility and effectiveness. This study found that the average referral rate was 9.3%, with the referral number in an acceptable range that will not put excessive pressure on the follow-up institutions, especially for those children above the age of 4 years. We also found that the referral rate increased significantly with decreasing age, and only 81.2% of 3-year-old children passed the screening (Table 1). Table 2 Test performance measured for the screening system in comparison with pure tone audiometry results.
Referral diagnostic assessment was recommended to 582 cases, while 463 children (79.6%) actually received audiological evaluation in the referral center and the other 119 cases (20.4%) were lost to follow-up (Fig. 1). A total of 87 children (prevalence of 1.38%; 95% CI: 1.09%, 1.67%) were confirmed to have hearing impairment. Of those, 12 children (1.91%; 95% CI: 0.83%, 2.99%) were diagnosed with permanent hearing loss, among which six children presented with bilateral mild hearing loss and one child with bilateral moderate hearing loss; two children presented with Table 1 Distribution of the screening results according to gender and age. Characteristics
Total
Pass No. (%)
Gender Boys Girls Age 3y 4y 5y 6y a
3282 3006 x2 = 3.275, P = 0.070
2999 (91.4) 2707 (90.1)
283 (8.6) 299 (9.9)
495 2347 2868 578 x2 = 112.238, P < 0.01a
402 2068 2681 555
93 279 187 23
No. (81.2) (88.1) (93.5) (96.0)
(18.8) (11.9) (6.5) (4.0)
Trend x2 test was applied by comparison across the four age groups.
Hearing impairment Normal Validity indices (%) Diagnosed by PTA
Screening 6 Referred Passed 10
Sensitivity Specificity PPV 22 274
37.5
92.6
PPV, positive predictive value; NPV, negative predictive value.
NPV
21.4 96.5
W. Wu et al. / International Journal of Pediatric Otorhinolaryngology 78 (2014) 290–295
293
Table 3 Characteristics of the 16 children confirmed with conductive hearing loss. Patient
Gender
Screening system
Age
Degree of hearing loss
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
F M M M M F F F M F F M M M M F
R R R R R R P P P P P P P P P P
3.64 4.24 4.56 4.59 4.97 5.20 4.59 4.32 4.19 5.45 4.78 5.30 4.88 4.99 4.76 4.83
Bilateral mild Bilateral mild Unilateral mild Bilateral mild Bilateral mild Unilateral mild Unilateral mild Bilateral mild Unilateral mild Unilateral mild Unilateral moderate Unilateral mild Unilateral mild Unilateral mild Unilateral moderate Unilateral mild
Thresholds (dB HL) Left
Right
30 32.5 30 27.5 36.25 32.5
32.5 42.5
32.50 33.75
27.5 31.25 28.75 28.75 35
46.25 31.25 40 36.25 41.25 26.25
M, male; F, female; P, pass; R, referral.
Although the main procedures of the new system can be completed by children independently, similar to traditional pure tone screening, the accuracy of the screening results is also influenced by how well the subjects cooperate with the test. Children less than 4 years old were more likely to be uncooperative in the screening, leading to a higher referral rate of 18.8%. However, the screening performance was much better in children more than 4 years old, especially for 5- and 6-year-old children, who had relatively high pass rates of 93.5% and 96.0%, respectively. Therefore, the screening system was more suitable for children aged more than 4 years. The new system could also be used for younger children, but combined screening might be a better alternative to increase the acceptability and accuracy of the screening. Through this screening project, 1.38% of preschool children were diagnosed with hearing loss, which is lower than the 1.8% reported by a study in the United States [13]. Children with hearing loss caused by impacted cerumen accepted the otologist’s check first when referred to the referral center and their cerumen had been removed before the diagnostic test; therefore, audiological assessment results for this group were in the normal range and were not included in the statistical incidence of hearing loss. It is reported that earwax embolism accounts for approximately 8.9% of the incidence of hearing loss [14]. The prevalence of otitis media with effusion was 2.02%, which is close to the lowest rates reported previously (2.2–9.86%) [15–18]. Otitis media with effusion ranked first as a cause of hearing loss for preschool children, accounting for 86.2% of the children with hearing impairment, which is consistent with the previous study (Fig. 1) [19]. In our study, permanent hearing impairment was found in 12 cases with an incidence rate of 1.91%; a finding consistent with that in other epidemiological reports [9,20,21]. The percentage of unilateral permanent hearing impairment was 0.80% (5/6288), higher than the proportion reported in our previous studies of preschool children’s hearing screening using pediatric audiometry [22]. The pure tone was presented through headphones in our system so that the hearing conditions could be detected for the left and right ear, respectively, avoiding their mutual interference similar to that in a pediatric audiometer screening. As for the 119 children lost to follow-up, there may be a part of hearing loss cases that were not counted into the statistics regarding the final prevalence. Therefore, it is believed that the actual prevalence of hearing loss may be slightly higher than the statistical data. This screening project was conducted using a new device without a gold standard. Owing to the heavy workload of conducting audiological evaluation for all the children involved, approximately 5% of the children were selected randomly for the
accuracy study. According to the data calculated, the sensitivity (37.5%) of the new system was lower than that of the pure tone sweep test, tympanometry, and transient evoked OAE in most of the reported studies, but the specificity (92.6%) was better than that of transient evoked OAE and close to that of pure tone screening [9]. In this study, a few false positive cases were found, leading to a lower positive predictive value, but more importantly, some children with hearing impairment were undetected in the screening (Table 2) [5]. There may be several reasons for the false negative and positive cases. First, the screening system tests thresholds at only three frequencies (1 kHz, 2 kHz and 4 kHz); however, otitis media with effusion had a greater impact on the low frequency, so those who had hearing loss only at low frequencies may be overlooked in this screening. Second, as mentioned previously, children in our project did not accept external ear canal cleaning before our screening, and a considerable proportion of children may have been detected in the screening with hearing loss caused by earwax embolism. The accuracy of these children’s screening results is difficult to verify in this study. Third, the average interval between screening and referral date was more than 1 week, so the diagnostic hearing assessment cannot fully represent the hearing level in the screening. We also found that only one child whose hearing loss went undetected during screening was diagnosed with mild bilateral conductive hearing loss and most children were diagnosed with unilateral conductive hearing loss (Table 3). Our project is the first study for the new screening system, and the sensitivity data obtained based on a random group of children is much too low. This is likely to have been a result of the unskilled nature of the personnel conducting the screening and imperfect system settings. Further research and improvement is needed to increase the sensitivity of the system in the future. For example, the screening system should include more automatic checks for user noncooperation instead of human supervision, and pure tones should be played irregularly at unequal intervals in case children tap the screen regularly without hearing a sound, which lead to the occurrence of false negative cases. However, in practice, the screening system also showed some advantages. All screening records can be uploaded and queried repeatedly on the network, and the results are judged automatically by the system itself, reducing data loss and human error in the process. The ease of operation made it possible for teachers in kindergarten to complete the screening following simple training. The costs of the new system were much less than that of pediatric audiometry, tympanometry or otoacoustic emissions. Purchasing a Smart Hearing Screening device set, which includes a tablet and
294
W. Wu et al. / International Journal of Pediatric Otorhinolaryngology 78 (2014) 290–295
noise-canceling headphones, only costs USD $800 or less, and no extra consumables are required during the screening process. In terms of efficiency, as calculated by the test time for the initial screening of each child, the entire screening can be accomplished by a single individual in 4.4 min on average, while the formal test takes approximately 2.5 min. In addition, as children can complete the formal test by themselves, one tester can conduct the test for several subjects simultaneously, which dramatically reduces the average test time. If four children are screened at the same time, for example, it takes 2 min on average to explain the instructions for each person (8 min in total), and the formal test takes an extra 2.5 min. Therefore, the screening process takes only 2.6 min for one child. The time taken for older children is shorter, especially in terms of the length of guidance required before the formal test starts. In this study, we spent 3 months to complete screening for 6288 preschool children. Two testers were required with eight sets of devices and each worked for 3 h each day. The monthly wage was USD $1000 per person, and therefore the total cost of labor and the devices was approximately USD $12,400. The per capita cost of our initial screening program was less than USD $2, and these costs will reduce further over time. Therefore, the Smart Hearing Screening system is expected to become one of the distinctive screening methods for the early detection of hearing loss in preschool children. Further developments of the software and device will be required to improve the sensitivity of the screening system. We propose that this screening system is feasible on the basis of it being automated, easy to use, cheap and highly efficient. In addition, its interesting properties can attract preschool children to actively participate with the screening. At the same time, the portability of the device and the simple principles of the system make it readily accepted by kindergarten teachers and preschool education institutions alike. This study was our first attempt to apply the Smart Hearing Screening system to preschoolers, but we believe that this system will become increasingly feasible with continual software and device improvements. Further research is needed to work out the problems such as the accuracy of the new system in different age groups, the frequency specificity of the screening results related to the final diagnosis, the impact of earwax on preschool children’s hearing, the accuracy of combined screening methods, and it is also desirable to test the application of this new system in children diagnosed with a hearing impairment. 5. Conclusion Our study is the first application of the screening system in preschoolers. This study found a considerable proportion of hearing impairment in preschool children. The new screening system has advantages that include ease of use, a multimedia platform that holds the attention of children, and automatic judgment and recording mechanisms. At present, the Smart Hearing Screening system is more suitable for children older than 4 years. Further studies are warranted to improve the system to make it more practical as a method for hearing screening in preschool children. Acknowledgements We are grateful to the doctors from Yangpu District Women and Child Health Institution in Shanghai for their assistance throughout this work. We also thank all the participating subjects, their parents and teachers. This work was supported by grants from the Health Ministry Special Fund, China (grant number 201202005) and the 12th Five-Year National Key Technologies R&D Program (2012BAI12B01).
Appendix Instruction manual for the Smart Hearing Screening system: A.1. Guidance First, the screening personnel should explain the directions of the screening test to the child and complete the training process. The child will be told to touch the cartoon hat on the screen immediately when hearing a pure tone, the sound like ‘‘DiDi’’. If the response is consistent with the tone presentation, a cartoon rabbit with different random facial expressions and in different colors will appear on the screen. If the child touches the screen without a sound stimulus, the cartoon rabbit will not appear. To ensure that the child understands the instructions and can be familiarized with the task, the screening personnel will put on the headphone for the child and begin to practice in the guidance mode. (1) The personnel first generate a 1000 Hz pure tone of 60 dB HL to the child’s right ear, at the same time observing whether the child responds correctly to the tone. (2) If the reaction is correct, the personnel will continue to generate a 1000 Hz pure tone of 40 dB HL and continue the training of the left ear. (3) If the child does not respond to the 60 dB HL pure tone on two occasions, it is considered that the child has not understood the guidance and is classified as an incomplete case. A.2. Formal screening Children demonstrating understanding of the guidance will undergo the hearing test in the formal screening mode. The entire screening test will be completed by children independently. (1) Starting with the right ear at a 1000 Hz pure tone of 40 dB HL, the Smart Hearing Screening system changes the frequency and intensity of the tones presented automatically based on the child’s reactions, thus, simulating the procedure of a traditional pure tone audiometry. (2) Following a satisfactory positive response, the system will reduce the level of the tone in 10 dB steps until no further response is obtained followed by increases in the level of the tone in 5 dB steps until a response is obtained. After the first response using an ascending approach, the system will continue the 10-dB-descending, 5-dB-ascending sequence until the subject responds twice at the same level; this represents the hearing threshold level. (3) If there is no response at the start, the system will increase the level to 60 dB HL. If the tone is still inaudible, the symbol ‘‘-’’ will be recorded. (4) Next, the system will proceed to 2000 Hz, 4000 Hz in that order and the process repeated for the left ear. The duration of the presented tone is approximately 1.5 s and the interval between the tones is approximately 2 s. (5) During the formal screening, personnel must observe on one side of the subject in order to avoid them indiscriminately touching the screen. A.3. Results evaluation and upload At the end of the screening test, the tablet will display the subject’s pure tone air conduction thresholds at three frequencies for both ears. Subjects with all auditory thresholds less than or equal to 30 dB HL
W. Wu et al. / International Journal of Pediatric Otorhinolaryngology 78 (2014) 290–295
are deemed to have tested negative for hearing loss and pass the screening. (1) Personnel have to click the ‘‘Done’’ button to upload the results to the network terminal before proceeding to the next round of screening. (2) As long as the threshold at one frequency is above 30 dB HL, a positive result will be given and the system will automatically be prompted to continue. The last attempt results should be taken as valid. (3) All the test results uploaded are available on the specified website (http://shshearing.com:8080/shs_webportal/login) at any time by any computer with Internet access. The Smart Hearing Screening system provides data analysis and generates a detailed hearing screening report for every child.
References [1] S. Mehra, R.D. Eavey, D.G. Keamy Jr., The epidemiology of hearing impairment in the United States: newborns, children, and adolescents, Otolaryngol. Head Neck Surg. 140 (2009) 461–472. [2] US Centers for Disease Control and Prevention, Summary of 2009 National CDC EHDI Data, Version A, 2009 Available from: http://www.cdc.gov/ncbddd/hearingloss/2009-Data/2009_EHDI_HSFS_Summary_508_OK.pdf (revised January 2012). [3] A.S. Niskar, S.M. Kiezak, A. Holmes, E. Esteban, C. Rubin, D.J. Brody, Prevalence of hearing loss among children 6–19 years of age, J. Am. Med. Assoc. 279 (1998) 1071–1075. [4] Joint Committee on Infant Hearing, Year 2007 position statement: principles and guidelines for early hearing detection and intervention programs, Pediatrics 120 (2007) 898–921. [5] H. Skarzynski, A. Piotrowska, Screening for pre-school and school-age hearing problems: European Consensus Statement, Int. J. Pediatr. Otorhinolaryngol. 76 (2012) 120–121. [6] M. Khairi Md Daud, R.M. Noor, N.A. Rahman, D.S. Sidek, A. Mohamad, The effect of mild hearing loss on academic performance in primary school children, Int. J. Pediatr. Otorhinolaryngol. 74 (2010) 67–70.
295
[7] M. Wake, Z. Poulakis, E.K. Hughes, C. Carey-Sargeant, F.W. Rickards, Hearing impairment: a population study of age at diagnosis, severity, and language outcomes at 7–8 years, Arch. Dis. Child. 90 (2005) 238–244. [8] C. Yoshinaga-Itano, A.L. Sedey, D.K. Coulter, A.L. Mehl, Language of early- and later-identified children with hearing loss, Pediatrics 102 (1998) 1161–1171. [9] J. Bamford, H. Fortnum, K. Bristow, J. Smith, G. Vamvakas, L. Davies, et al., Current practice, accuracy, effectiveness and cost-effectiveness of the school entry hearing screen, Health Technol. Assess. 11 (2007) 1–168. [10] C.I. Berlin, T. Morlet, L.J. Hood, Auditory neuropathy/dyssynchrony: its diagnosis and management, Pediatr. Clin. North Am. 50 (2003) 331–340. [11] Report of the first informal consultation on future programme developments for prevention of deafness and hearing impairment, WHO, Geneva, 1997, Available from: http://whqlibdoc.who.int/hq/1997/WHO_PDH_97.3.pdf. [12] D.R. Halloran, T.C. Wall, H.H. Evans, J.M. Hardin, A.L. Woolley, Hearing screening at well-child visits, Arch. Pediatr. Adolesc. Med. 159 (2005) 949–955. [13] Y.C. Serpanos, F. Jarmel, Quantitative and qualitative follow-up outcomes from a preschool audiologic screening program: perspective over a decade, Am. J. Audiol. 16 (2007) 4–12. [14] P.R. Adobamen, F.O. Oqisi, Hearing loss due to wax impaction, Niger. Q. J. Hosp. Med. 22 (2012) 117–120. [15] O.C. Erdivanli, Z.O. Coskun, K.C. Kazikdas, M. Demirci, Prevalence of otitis media with effusion among primary school children in Eastern Black Sea, in Turkey and the effect of smoking in the development of otitis media with effusion, Indian J. Otolaryngol. Head Neck Surg. 64 (2012) 17–21. [16] E. Okur, I. Yildirim, M. Akif Kilic, S. Guzelsoy, Prevalence of otitis media with effusion among primary school children in Kahramanmaras, in Turkey, Int. J. Pediatr. Otorhinolaryngol. 68 (2004) 557–562. [17] H.C. Rushton, M.C. Tong, V. Yue, P.J. Wormald, C.A. Hasselt, Prevalence of otitis media with effusion in multicultural schools in Hong Kong, J. Laryngol. Otol. 111 (1997) 804–806. [18] M.C. Tong, V. Yue, P.K. Ku, P.S. Lo, C.A. van Hasselt, Screening for otitis media with effusion to measure its prevalence in Chinese children in Hong Kong, Ear Nose Throat J. 79 (2000) 626–630. [19] M.A. Al-Rowaily, .A.I. AlFayez, M.S. AlJomiey, A.M. AlBadr, M.A. Abolfotouh, Hearing impairments among Saudi preschool children, Int. J. Pediatr. Otorhinolaryngol. 76 (2012) 1674–1677. [20] H.M. Fortnum, A.Q. Summerfield, D.H. Marshall, A.C. Davis, J.M. Bamford, Prevalence of permanent childhood hearing impairment in the United Kingdom and implications for universal neonatal hearing screening: questionnaire based ascertainment study, Br. Med. J. 323 (2001) 536–540. [21] C. MacAndie, H. Kubba, M. McFarlane, Epidemiology of permanent childhood hearing loss in Glasgow, 1985–1994, Scott. Med. J. 48 (2003) 117–119. [22] J. Lu¨, Z. Huang, T. Yang, Y. Li, L. Mei, M. Xiang, et al., Screening for delayed-onset hearing loss in preschool children who previously passed the newborn hearing screening, Int. J. Pediatr. Otorhinolaryngol. 75 (2011) 1045–1049.
