|Year : 2022 | Volume
| Issue : 4 | Page : 296-300
The credibility of a smartphone-based application for use as a hearing screening tool in underserved areas
Satish Kumar Chokalingam1, Prakash Mathiyalagen2, Kumaran Ramesh Colbert1, Bhanu Vazhumuni1, Sathiyaseelan Murugesan1, Niveda Nagarajan1
1 Department of ENT, Indira Gandhi Medical College and Research Institute, Kathirkamam, Puducherry, India
2 Department of Community Medicine, Indira Gandhi Medical College and Research Institute, Kathirkamam, Puducherry, India
|Date of Submission||01-Apr-2022|
|Date of Acceptance||11-Aug-2022|
|Date of Web Publication||29-Dec-2022|
Dr. Satish Kumar Chokalingam
Department of ENT, Indira Gandhi Medical College and Research Institute, Kathirkamam, Puducherry
Source of Support: None, Conflict of Interest: None
Context: Hearing loss has been rightly regarded as “invisible disability.” Globally, it is the third largest cause for years lived with disability. It is imperative to detect it early to initiate remedial measures. We intended to find a cheaper, quick, and reliable alternative to the traditional audiological services. Aims: The aim of this study was to assess the hearing levels using a self-administered smartphone hearing application and compare these results with pure-tone audiograms performed by an audiologist. Settings and Design: This diagnostic study was conducted at a tertiary care hospital in the State of Puducherry, between 2019 and 2021. Subjects and Methods: One hundred and nineteen participants were recruited; hearing screening was done using the smartphone application “Hearing Test” followed by pure-tone audiometry. Statistical Analysis Used: The results were analyzed for validity and reliability using SPSS software. Results: The mean age was 34.23 ± 9.39 years, and 57.1% had ear complaints at presentation. At the level of statistical significance of P = 0.05, no difference was found between the tests. An absolute difference of <6.712 for each frequency was observed with the average difference being 5.18 dB (95% confidence interval 5.65–4.73) with standard deviation of 3.56. The smartphone application demonstrated a sensitivity of 76.26% and a specificity of 98.99%. Conclusions: We did not find any significant difference between the application and pure-tone thresholds in any frequency. Thus, the Hearing Test application is a valid screening tool to assess hearing loss early.
Keywords: Hearing screening, hearing test, smartphone application
|How to cite this article:|
Chokalingam SK, Mathiyalagen P, Colbert KR, Vazhumuni B, Murugesan S, Nagarajan N. The credibility of a smartphone-based application for use as a hearing screening tool in underserved areas. Indian J Otol 2022;28:296-300
|How to cite this URL:|
Chokalingam SK, Mathiyalagen P, Colbert KR, Vazhumuni B, Murugesan S, Nagarajan N. The credibility of a smartphone-based application for use as a hearing screening tool in underserved areas. Indian J Otol [serial online] 2022 [cited 2023 Mar 29];28:296-300. Available from: https://www.indianjotol.org/text.asp?2022/28/4/296/365961
| Introduction|| |
Hearing loss is a global health-care problem; The World Health Organization (WHO) estimates that around 1.5 billion people suffer from mild hearing loss globally, 430 million people including 34 million children, live with disabling hearing loss, and compromised quality of life secondary to it. The maximum contribution to its prevalence is from Western Pacific Region followed by the South-East Asia region, comprising low- or middle-income countries.
Hearing loss can lead to depression, social isolation, and lack of confidence, as it affects an individual's ability to communicate effectively., Thus, it leads to substantial reduction in academic and economic performance consequent to dismal social and emotional functioning.
Thus, the most important aspect toward improvisation is early detection, conservation, and treatment. This can be easily achieved by audiological evaluation using a pure-tone audiometer, which is considered the “Gold Standard” for hearing assessment. However, access to such services has always been a major hurdle in view of expensive infrastructure, costly equipment, and requirement of trained workforce.
With the research question, “can a mobile-based hearing test serve as a screening tool to detect hearing loss where facilities for pure-tone audiometry do not exist,” our search for alternatives led us to hearing applications. We aimed to compare the accuracy of mobile application-based hearing tests in determining pure-tone thresholds and screening for hearing loss.
| Subjects and Methods|| |
The objective of this study was (a) to assess the hearing levels of a heterogeneous group using a self-administered smartphone hearing application and (b) to compare these results with those obtained from pure-tone audiograms performed by an audiologist using a conventional clinical audiometer as a reference.
This diagnostic study was done on participants attending the ENT outpatient department at a tertiary care hospital in Puducherry, from May 2019 to May 2020. Participants in the age group of 20–50 years with complaints of reduced hearing in at least one ear were included in the study. Those <20 years and >50 years or with profound hearing loss on any side or with history of previous ear surgery or with active ear discharge were excluded from the study. A few volunteers without hearing loss were also included in the study to avoid bias.
Considering the sensitivity as 99%, specificity as 95% for application to detect hearing loss, with the error of margin as 3%, the proportion of chronic suppurative otitis media (CSOM) cases as 25%, the minimum sample size is 212. Due to unfortunate circumstances (COVID pandemic), the study could not be carried out as planned, the sample size of 119 only could be achieved. Hence, to achieve the required sample size, we considered each ear as a test subject, thus doubling the sample size to 238.
This study was performed using a smartphone application named “HEARING TEST” developed by e-audiologia.pl, which is free for use and download on the android platform. It is a self-explanatory diagnostic hearing test. The test can be performed offline once the software is downloaded. An insert type headphone, Samsung® model-EHS61 was used for the test. The frequency range from 250 Hz to 8000 Hz was selected. The ear to be tested was selected and the test was performed from low-to-high frequency by default under the supervision of one audiologist. The results were stored in tabulated as well as the graphic form for analysis and interpretation.
The participants were then sent for a conventional pure-tone audiogram, which was performed using a global real audiometer and the results were stored in both graphic and tabulated form. Both the tests were performed in a sound-treated audiometry room as per the standards by BS/EN/ISO. Another audiologist was used to carry out pure-tone audiometry test to avoid bias.
Data analysis was performed using Statistical Package for the Social Sciences v. 20.0 for Windows (SPSS Inc.; Chicago, IL, USA). The continuous data were expressed as the means and standard deviations (SDs), and categorical data as the frequency and percentage. Normality testing was performed using the Kolmogorov–Smirnov test. For the comparison of repeated measures, the nonparametric counterpart of the Student's paired t-test, i.e., the Wilcoxon signed-rank test was used.
To test the reliability, the interclass correlation coefficient (ICC) with a 95% confidence interval (CI) was calculated. Interpretation of the data was performed according to Cicchetti. An agreement was graded as poor for ICC values <0.40, fair for values between 0.41 and 0.59, good for values between 0.60 and 0.74, and excellent for values between 0.75 and 1.0. The difference between the pure-tone audiogram and smartphone audiogram was calculated for each frequency (ΔdB = | audiogram − smartphone audiogram|). A two-sided P < 0.05 was considered statistically significant.
The agreement between application and pure-tone thresholds was assessed using Cohen's kappa statistics. The Kappa result was interpreted as follows: values ≤0 as indicating no agreement and 0.01–0.20 as none to slight, 0.21–0.40 as fair, 0.41–0.60 as moderate, 0.61–0.80 as substantial, and 0.81–1.00 as almost perfect agreement.
To test the validity, the following parameters were calculated with 95% CI: sensitivity, specificity, positive predictive value, negative predictive value, diagnostic accuracy, likelihood ratio of a positive test, likelihood ratio of a negative test, diagnostic odds, and Cohen's kappa.
Ethical approval for this study (Proposal no. 26/207/IEC-25/PP/2019) was provided by the Institute Ethics Committee of Indira Gandhi Medical College & Research Institute, Puducherry, India on 26 April 2019.
| Results|| |
A total of 119 participants were recruited for the study with a mean age of 34.23 ± 9.39 years (range 18–55 years). There were 67 (56.3%) male and 52 (43.7%) female participants. Since the desired number of participants could not be achieved due to COVID outbreak, each ear was considered a test ear, thus making the sample size to be 238. The vast majority (82.4%) were local residents, and a few were (17.6%) from the neighboring districts of Tamil Nadu [Table 1].
Among the study participants, 68 (57.1%) had ear complaints at presentation. The time of presentation and the onset of complaints varied depending on the nature of the complaints. Those who had acute complaints, such as ear ache or ear discharge presented earlier than those having complaints of ear block or reduced hearing. In the present study group, 43 (36%) had ear complaints for <3 months duration, 5 (4.2%) participants had it for 1–5 years, and only 1 (0.8%) had ear complaints for >5 years [Table 1].
Although many (57%) had presented with ear complaints, only 15 (12.6%) participants had abnormal tympanic membrane findings on clinical examination. Unilateral disease (67%) was more common than bilateral disease (33%), and the left side was more commonly (40%), affected than the right side (27%) [Table 1].
The thresholds obtained by pure-tone audiometry and the hearing application are summarized and categorized (as per Goodman classification of hearing loss grades). The hearing thresholds determined by pure-tone audiometry were compared to those determined by the Hearing Test application on mobile application. The results for the correlation analysis are presented in [Table 2]. The interclass correlation for consistency of a single measurement was measured between 0.642 and 0.854. We observed excellent agreement at 500, 1000, and 2000 Hz, average and good agreement at 4000 Hz and 8000 Hz, respectively.
|Table 2: Correlation between the application and pure-tone thresholds at various frequencies|
Click here to view
The reliability analysis is presented in [Table 3]. At the level of statistical significance of P = 0.05, no significant difference was found between the two groups, and thus further analysis was done. On the assessment of the mean difference between the pure-tone thresholds and mobile application thresholds, we observed an absolute difference of <6.712 for each frequency and the average difference observed was 5.18 dB (95% CI 5.65–4.73) with SD of 3.56 [Table 3]. The same is evident on the scatter plots between the pure-tone and hearing application thresholds [Figure 1].
|Figure 1: Scatter plots between smartphone application and pure-tone thresholds|
Click here to view
|Table 3: Comparison of hearing thresholds determined by a smartphone application and pure-tone audiometry with reliability analysis for each frequency|
Click here to view
The assessment of reliability for hearing loss detection was conducted based on the criterion adopted for this study. Hearing loss was diagnosed when the threshold exceeded 25 dB (on average) at 500, 1000, 2000, and 4000 Hz. The sensitivity obtained was 76.26% (95% CI 6.54–82.57) with a specificity of 98.99% (95% CI 94.5–99.82). The Cohen's Kappa was observed to be 0.719, which deciphers to substantial agreement [Table 4] and [Table 5].
|Table 4: Smartphone application and pure-tone audiogram cross-tabulation|
Click here to view
The difference plots (Bland–Altman plots) between smartphone applications and pure-tone thresholds are depicted in [Figure 2]. The three lines in the plot represent the mean of differences – the center one is called bias and the rest two lines are limits of agreement mean +1.96 SD and mean −1.96 SD. Most of the threshold values determined by smartphone-based application fall within the 5 dB of the standard threshold were determined by pure-tone audiometry.
|Figure 2: Difference plots (Bland–Altman plots) between smartphone application and pure-tone thresholds|
Click here to view
| Discussion|| |
Hearing loss has been rightly regarded as “invisible disability,” not just due to absent visible symptoms, but also due to the associated stigma in communities, and indifference of policymakers toward it. Globally, it is the third-largest cause for years lived with disability.
The world hearing report by the WHO estimates that over 1.5 billion people worldwide experience some form of hearing disability at present, and this might increase to 2.5 billion by 2050. Early detection through audiological screening remains the only solution to this problem. Correction of the preventable causes of hearing loss and early initiation of rehabilitation measures is essential to eliminate this invisible disability.
Unfortunately, audiological screening has a lot of limitations. The biggest barrier to this is limited access to audiological services secondary to a lack of trained professionals and equipment. Although portable and cheaper audiometers are available now, the standard infrastructure requirement for their use has widely limited their use on large scale.
We intended to find an alternative to the traditional audiological services, which could be a cheaper, quick, and reliable method for hearing screening, to eliminate this obstacle. We were fascinated by the use of smartphone devices for hearing screening during the literature review through PubMed and MEDLINE.
The use of smartphones has rapidly increased worldwide off late. They mimic a personal computer with a wide range of functionality. With the recent developments in health-based apps, they have gained more popularity globally. These are empowered with ability to monitor one's heart rate, blood pressure, and various other parameters, thus enabling an individual to monitor his/her health.
With respect to otology, there is a rapid rise in the development of applications to measure noise levels, hearing thresholds, and digital otoscopes. These are basically designed to collect, save, and share data across devices and countries with ease.
There were only a few articles available, which validated smartphone-based applications. The studies done in the initial days mostly used iPod devices, which lacked standard calibration procedures. Swanepoel et al. demonstrated that the conventional audiometry and the smartphone application hearScreen™ were at 97.8% agreement. Kam et al. performed the reliability and validity of application tests and observed a sensitivity of 63% and specificity of 82%. Foulad et al. observed that 94% of threshold values were within 10 dB of formal audiometry thresholds when the test was conducted in a quiet room.
The studies performed by Masalski et al. and Renda et al. used the Hearing Test application, the same application which was used in the present study. The mean absolute difference between the pure-tone and mobile application thresholds was 8.3 dB and 7.8 dB at 95% CI, respectively. We also observed similar results, the average difference in thresholds was 5.8 dB and the difference was <6.7 dB for each frequency.
However, we observed a major difference in the sensitivity and specificity of the Hearing Test application as compared to the earlier study. Masalski et al. obtained 98% of sensitivity and 79% of specificity, whereas we observed it to be 76.3% and 98.9%, respectively.
The tests were performed in a controlled environment (sound-treated room) by trained audiologists, thereby reducing the chances of user error such as accidental skipping of pressing the button and switching the sides of the headphones.
The tests were performed on an android device with bundled and calibrated headphones. Such a facility may not be available for all devices in the market.
The overall sample size and thus the number of affected ears was limited due to the COVID pandemic, which may have affected the sensitivity and specificity of the test. We tested each ear separately to adjust the sample size which may have led to inflated sample size. However, we observed that a similar procedure was followed in previous mobile-based audiometric studies as well.
We also observed that the mobile-based audiometric tests could not differentiate between conductive and sensorineural hearing loss. This may lead to inaccurate results in underserved areas with ambient noise.
| Conclusions|| |
We would like to conclude that the smartphone-based hearing application “hearing test” is a valid and reliable screening tool. It is inexpensive, quick, and hassle-free, without the need for special infrastructure and workforce in places where facilities for audiological evaluation do not exist. However, pure-tone audiometry remains the gold standard test; and may be used as a confirmatory test in view of its expensive proposition.
We would like to thank Dr. Masalski, the developer of the Hearing Test application to have generously allowed us to use his application free of cost for this study.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Lin FR, Yaffe K, Xia J, Xue QL, Harris TB, Purchase-Helzner E, et al.
Hearing loss and cognitive decline in older adults. JAMA Intern Med 2013;173:293-9.
Olusanya BO, Newton VE. Global burden of childhood hearing impairment and disease control priorities for developing countries. Lancet 2007;369:1314-7.
Swami H, Bhargava A, Sabarigirish K, Arvind BM. A comparative study of smartphone based app with free field hearing for possible use as a screening test. Int J Otorhinolaryngol Head Neck Surg 2017;3:710-4.
Cicchetti DV. Guidelines, criteria, and rules of thumb for evaluating normed and standardized assessment instruments in psychology. Psychol Assess 1994;6:284.
Goodman A. Reference zero levels for pure-tone audiometer. ASHA 1965;7:1.
Logan AG, McIsaac WJ, Tisler A, Irvine MJ, Saunders A, Dunai A, et al.
Mobile phone-based remote patient monitoring system for management of hypertension in diabetic patients. Am J Hypertens 2007;20:942-8.
Na Y, Joo HS, Yang H, Kang S, Hong SH, Woo J. Smartphone-based hearing screening in noisy environments. Sensors (Basel) 2014;14:10346-60.
Swanepoel de W, Myburgh HC, Howe DM, Mahomed F, Eikelboom RH. Smartphone hearing screening with integrated quality control and data management. Int J Audiol 2014;53:841-9.
Kam AC, Gao H, Li LK, Zhao H, Qiu S, Tong MC. Automated hearing screening for children: A pilot study in China. Int J Audiol 2013;52:855-60.
Foulad A, Bui P, Djalilian H. Automated audiometry using apple iOS-based application technology. Otolaryngol Head Neck Surg 2013;149:700-6.
Masalski M, Kipiński L, Grysiński T, Kręcicki T. Hearing tests on mobile devices: Evaluation of the reference sound level by means of biological calibration. J Med Internet Res 2016;18:e130.
Renda L, Selçuk ÖT, Eyigör H, Osma Ü, Yılmaz MD. Smartphone based audiometric test for confirming the level of hearing; is it useable in underserved areas? J Int Adv Otol 2016;12:61-6.
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]