According to Kochkin, hearing aid adoption in the United States has increased slightly over the last several years to 23% of the hearing-impaired population.1 The three greatest sources of growth in hearing aid usage are: (1) the Veterans Administration (VA), which dispenses hearing aids to qualified veterans free of charge; (2) mail-order sales; and (3) Internet sales.1
According to Schillinger, online purchasing has become a growing trend for web-savvy bargain hunters wishing to get home delivery of hearing aids without the expense and inconvenience of an office visit to a physician and/or a hearing healthcare professional.2
Schillinger believes that Internet hearing aid sales have been growing rapidly. While her data show that “distance purchasing” accounted for only about 1% of the hearing aid market in 1997, a 2002 survey of consumers found that approximately 185,000 of them (roughly 4%-5%) were interested in purchasing a hearing aid online.
Given estimates that consumers spent more than $100 billion on purchases over the Internet in 2007 and that this year about half of Internet users are expected to purchase products online, it is safe to assume that the number of prospective online hearing aid customers has increased greatly since 2002.
Currently, consumers can purchase hearing aids on eBay and other web sites from vendors who may have little or no knowledge of hearing and hearing disorders. Although there are some Internet sites (e.g., hearingplanet.com, AmericaHears.com) that incorporate professional services with the sale of hearing aids, many do not. Yet these sites advertise custom, as well as non-custom, instruments with outputs that range from minimal to potentially dangerously high.
RISKS SEEN IN DIRECT-TO-CONSUMER SALES
Many authors have discussed the risks involved when consumers purchase hearing aids without benefit of in-person professional participation in the selection and fitting of the devices and in the associated counseling.
Chartrand, Glaser, and Beck note that the Internet allows consumers to acquire hearing aids and other amplification devices without the protection of regulatory safety nets and without the benefit of expert professional services.3 The authors also raise concerns regarding the legal and ethical ramifications of online dispensing.
Audiologists have cited reasons to suggest that when a consumer obtains hearing aids from an Internet company in a manner that does not include professional services, the risk of an inappropriate fit is high. For example, Valente contends that verification through an established protocol such as real-ear measurement is the most important aspect of a hearing aid fitting.4 He states that when a hearing aid is programmed to “first fit” and no verification is done, which is the case with many online fittings, there is a high probability of error because the device is not fine-tuned for the patient.
Mueller5 and Killion6 have reported that hearing aid manufacturers' proprietary formulas for calculating amplification requirements (first fit formulas) often fail to meet the needs of hearing aid users, and can leave them unable to hear quiet sounds. Mueller contends that when a patient receives a hearing aid set to first fit, verification through real-ear insertion gain or real-ear aided gain measures should be performed and gain settings should be adjusted to meet specified targets as indicated by a verified prescriptive method, such as NAL-NA1 or DSL.
Even when an online hearing aid purchaser is fortunate enough to get an appropriate fitting without help from a professional, post-fitting problems may still arise. Schum points out that it is normal for a hearing aid purchaser, especially a first-time user, to experience annoyance and fatigue.7 He says that new hearing aid users may be able to manage the adaptation process successfully by adjusting the volume control or by switching from one pre-set program to another to suit the sound environment. However, with many modern hearing aids, neither option exists because they do not have volume controls or a choice of (appropriately) pre-set programs. In such cases, purchasers of inappropriately fitted new hearing aids may have no recourse but to return them to the online company for adjustments.
In addition, if a consumer purchases a hearing aid through the Internet that lacks the self-adjusting adaptation-management capabilities that many high-end instruments offer, the person is unlikely ever to get the full benefit of the device's prescriptive values. Even with hearing aids that switch automatically from one program to another depending on the environment, the pre-set programs are customized for the particular buyer's audiometric data.
THE ONLINE PURCHASE PROCESS
Typically, Internet companies that sell hearing aids directly to consumers ask prospective buyers to do three things: (1) Take an online hearing test, (2) agree to purchase the hearing aid(s) recommended based on the results of the online hearing test, and (3) make their own earmold impressions using materials mailed to them and following instructions from the online company. The consumer sends the impression(s) back to the company in order for the hearing aid(s) to be manufactured. Some online vendors do not even require the consumer to take a hearing test before purchasing hearing aids.
As discussed earlier, authors have expressed deep concern about this method of selling hearing aids that varies so radically from the traditional one in which a credentialed professional draws upon his or her education and experience in testing the patient, selecting appropriate amplification, fitting the device to meet the patient's individual needs, and then providing post-fitting adjustments, while counseling the person at every step along the way.
TESTING THE EFFICACY OF ONLINE PURCHASE
While other articles have discussed the drawbacks of purchasing hearing aids without the involvement of a professional, the study described in this and subsequent articles in this series was designed to evaluate the real-world effectiveness of buying hearing aids over the Internet by having subjects actually follow the purchase process.
Since online hearing testing is in many cases the first step on the way to the purchase of hearing aids online, the first part of this study had subjects take one company's online hearing test and then compared the results with those obtained in a hearing test suite. The company has been selling hearing aids by mail order or via the Internet for decades and is one of the largest such companies.
Testing hearing outside of the clinical setting is not a new concept. Smits et al. designed and implemented a valid automatic speech-in-noise screening test by telephone and the Internet for use in the Netherlands.8,9 While this program was judged successful for identifying potential hearing disorders, those who failed the screening were referred for full diagnostic testing through their primary-care physician, an audiologist, or an ear, nose, and throat specialist.
In contrast, results from online hearing tests, including the one reported on in this study, are used directly to preset (program) the purchased hearing aids. Often, no further diagnostic or medical information is obtained. Consumers are required only to read the posted mandatory Food and Drug Administration guideline recommending that an individual seek a medical evaluation before purchasing a hearing aid and then to sign a medical waiver.
We recruited subjects for this study from undergraduate and graduate classes at Illinois State University, the university speech and hearing clinic, and two local civic organizations in Normal, IL.
A total of 81 subjects in three age groups took part in the research. Group 1 consisted of young adults (N=26, age range 20–29, mean age 21.3 years). Group 2 consisted of subjects 30–64 years of age (N=26, mean age 51.7 years), while those in Group 3 were all over the age of 65 years (N=29, age range 65–80, mean age 70.9 years). Subjects were neither included nor excluded on the basis of their hearing status.
All subjects took both an online hearing test and a hearing test in a hearing test suite. For the online test, subjects were asked to follow the directions as prescribed on the specified web site, which included plugging in headphones (which were provided) and adjusting the computer volume (done through the “run” menu option).
The web site did not specify any recommended type of headphone for use in testing, but subjects wore Aiwa HP-AO91 earphones, a type similar to what is generally used with a media player. The volume setting of the computer was turned to the maximum (full-on) position.
The online hearing test asks consumers to start by logging onto the web site. (We gave each subject a pseudonym for log-in purposes so they would remain anonymous.) Once logged in, those being tested answer questions about their hearing status and then are directed to set up the computer to meet the needs of the hearing test. The online hearing test offers a yes/no option to indicate whether each tone was heard or not heard. A consumer also has the option of asking that the test tone be repeated.
Once a consumer selects the “no” button (“I did not hear”) two times, the test immediately proceeds to the next frequency. When testing at all the frequencies has been completed, an audiogram appears on the screen along with any recommended type and style of hearing aid.
The subjects in this study also received a hearing test in a sound-treated room (IAC) using standard test procedures and calibrated audiometric equipment (Grason-Stadler GSI-61 Audiometer). This test included the same frequencies as the online test. The test was completed by a licensed/certified audiologist or a second-semester audiology graduate (AuD) student under an audiologist's supervision. Approximately half the subjects took the computer test first, while the other half took the test suite evaluation first.
Using paired T-tests at the .05 confidence level, we found statistically significant differences between the results of the online and professional testing conditions across all groups and among individual age groups for the frequency range 250-4000 Hz. Figure 1 shows the mean differences (in dB) of thresholds measured in the two test conditions for each age group and overall at each frequency tested.
Table 1 outlines dB differences for all groups and for each individual group at each frequency tested. For the combined group (Groups 1–3), differences for eight out of ten frequencies were statistically significant and ranged from 1.2 dB to 11.3 dB (mean difference = 6.25 dB). Differences for Group 1 ranged from 1.6 to 12.7 dB (mean = 7.15 dB), for Group 2 from 3.3 to 19.6 dB (mean = 11.45 dB), and for Group 3 from −1.7 dB to 12.7 dB (mean = 6.2 dB).
Group 2 (age 30–64 years) showed the greatest variance between the two testing conditions. For Group 3, negative differences indicated that better threshold levels were obtained via the online test than in the hearing test suite.
Results of this study indicate that the outcomes of one online hearing aid retailer's hearing test differ significantly from the outcomes of traditional audiometric testing. Additionally, subjects in all age groups, but more predominantly those in the elderly group, displayed difficulty performing the online test procedure, especially in the setup portion (e.g., adjusting the overall computer volume). Computer assistance was offered to any participant who requested help in getting started on the online test, but it was the subjects in Group 3, the over-65 age group, who most often required assistance.
Participants in Groups 1 and 2 had a much easier time navigating the site. Because so many of the elderly participants in Group 3 had difficulty with the computer navigation, it seems reasonable to assume that the other groups better reflect the typical ages of those people who are most likely to purchase any products, including hearing aids, online.
The clinical impact of this research appears to be far-reaching. First and foremost, online hearing tests can misrepresent the severity of hearing loss. That, in turn, raises doubts about the prescribed hearing aid settings derived from the online test results.
For example, among individual participants, differences between the online and the test suite tests were as great as 85 dB at some frequencies. In these cases, when the subject inadvertently pressed the “No” button twice, the computer test immediately proceeded to the next test frequency. That led to the “threshold” being marked at 85–95 dB.
Individual differences between the two test conditions ranged on average from 10 to 40 dB for any given frequency. If a consumer was fitted with a hearing aid set to first fit in such cases, the settings would not only be inappropriate, but could place the person at risk for hearing damage.
In other cases, the computer test identified subjects as having mild hearing losses, whereas the booth test indicated normal hearing sensitivity. In these cases, the web site recommended low-gain hearing aids and offered a choice of devices from which the consumer could select. If these people had purchased hearing aids based on the recommendation of the web site, they would probably have wound up with unnecessary or excessive amplification.
Figure 2 shows an example of an audiogram from one subject in each of the testing conditions.
Participants across all groups were startled by the volume level of the online hearing test at its initial level at each frequency. That level appeared to be at least 90 dB based on the dB level indicated on the web site as the testing process begins. Numerous subjects jumped out of surprise or immediately removed the headphones from their ears for the first one or two tones presented at each frequency.
As a result, it seems likely that a testing effect would occur when the subject was anticipating the sound, since subjects initiate the onset of the signals themselves. (When the “yes” [“I heard”] button is pushed, a lower intensity signal is automatically initiated). There is no randomization in the presentation of the test tone via the online hearing test and each frequency begins at the same intensity level and descends in exactly the same manner.
The wide variations between hearing thresholds obtained via the Internet and those obtained in the hearing test suite using traditional techniques suggest that hearing aid fittings based on results obtained online may be inappropriate and at times potentially dangerous to a consumer's hearing health.
Further research is necessary to investigate the procedures at other online hearing aid retailers' web sites for obtaining thresholds to be used in fitting hearing aids.
In addition, other questions regarding the online dispensing process remain. They include how well individuals can make their own earmold impressions, the physical comfort of the aids obtained, and the process for programming and adjusting programming after the initial fit. These will be explored in subsequent articles in this series in The Hearing Journal.