DR. SUMIT DHAR: Hello Dr. Humes, thank you for agreeing to answer a few questions about your recent publication related to different modes of hearing aid delivery (Humes, 2017). I have many questions that probably are on the minds of readers as well.
It is not a surprise to me that the study is meticulously designed to isolate one variable of interest. You could not have timed this publication better, with all of the current interest on this topic. I am very familiar with your prolific publication history on a variety of topics from basic psychophysics, to speech perception, to amplification, to multisensory perception in aging. This, however, is your first foray into hearing health-care delivery. So, how did this work and the resultant publication come about?
DR. LARRY HUMES: Yes, although I have been involved in clinical research, including hearing aid outcomes, for many years, this was my first foray into clinical trials research involving hearing health-care delivery. The impetus for my involvement in this area, which led to the recently completed clinical trial, was my participation as a member of a working group put together by Amy Donahue at the National Institute on Deafness and Other Communication Disorders (NIDCD), addressing the accessibility and affordability of hearing health care.
It was clear from the discussion that many working group members and presenters were deeply concerned about the decades-long impasse in getting hearing help to millions of adults with mild-to-moderate hearing loss. The presentations and subsequent discussion led to the identification of a wide range of areas that, with additional research, might one day lead to improved hearing health care for these millions of Americans.
Great, sounds like the National Institutes of Health (NIH) workshop in 2009 was instrumental in starting a movement toward many fundamental changes in hearing health care. The resultant editorial by Donahue, Dubno, and Beck (2010) lists multiple research needs for the field. What attracted you to this particular question?
We actually came away from this working group meeting with a number of possible research problems that could be addressed by our group at Indiana University (IU). After doing some more research, we decided to focus on two topics: (1) pros and cons of service-delivery models for the provision of hearing aids and (2) improved auditory training for those who purchase hearing aids. Each was designed to be addressed through a randomized clinical trial and both trials have now completed data collection. Thus far, only the clinical trial on service-delivery models has been published, with the analyses and publication of the other trial still in progress.
Before asking the next set of questions, let me take a quick break and ask all of our readers to go find your article (Humes, 2017) and read it before proceeding any further. Knowing about your design and findings will make the following discussion more meaningful.
Let me try to summarize the study. Participants were randomly assigned to one of three groups, one fit by audiology best practice (AB), the second using essentially an over-the-counter model that you call “consumer decides” (CD), and finally a placebo group (P) who received no gain in their devices. You could have named the third group “essentially fake” or EF, which would have led to the “ABCDEF study.” Every group received the same device and the question was whether the mode of delivery made a difference.
Both AB and CD groups showed benefit on many of the outcome measures, which was statistically different from the placebo group. Half the people paid $3,600 for the hearing aids, the other half $600. Price, it turns out, had no bearing on the results. In the end about 80 percent of those in the AB group, as compared to approximately 50 percent in the CD group, said they would like to keep the hearing aid. If that is a fair summary, let’s charge ahead.
The design of the study was critical for isolating one particular question. What were some of the difficult decisions that you had to navigate to arrive at the final design?
As noted, I’d already been conducting clinical research for many years, but I had never previously done true clinical-trials research. It proved to be trickier than realized initially to both propose the research and to conduct it. First, we had to whittle down the “over-the-counter” alternative to audiology best practices to a manageable, but key component of that model. We chose to focus solely on the service-delivery model, keeping all else constant across the branches. We also chose to use factorial combinations of three service-delivery groups and purchase price. I believe that most people who have advocated for “over-the-counter” approaches envision that the devices provided would be simpler, less expensive devices, whereas hearing aids provided through best practices would be more feature-rich and expensive by comparison.
However, given no prior experimental research, which had manipulated purchase price and available consumer research for other products, it was unclear what the impact of purchase price alone might be on measured outcomes, especially self-reported measures of benefit and satisfaction.
One could imagine a scenario in which higher prices yielded higher expectations and, unable to meet those expectations, more negative outcomes. Alternatively, one could imagine the reverse: higher prices imply higher quality and a greater personal (financial) investment on the part of the consumer, which might lead to greater effort to make things work and to higher reported benefit and satisfaction.
In the end, we opted to use both purchase prices, $600/pair and $3,600/pair, with each of the three service-delivery groups in a full-factorial design so that we could tease out both the main effect of purchase price as well as any interactions of purchase price with service-delivery group. To do this effectively, while maintaining full blinding of participants and collaborators, deception was involved. That is, the participants were not informed of the difference in purchase price until the entire study had been completed. Those who had paid the higher price received a refund of the difference in purchase price when they were debriefed at the conclusion of the entire study. Also, the four research audiologists involved thought that this was randomly manipulated from participant-to-participant throughout the study. In fact, they were deceived too and debriefed at study conclusion. The first half of the study participants paid the higher price and the second half all paid the lower price. In the end, after working out all of these procedural details, the results turned out to be very simple: there were no effects of purchase price on outcomes.
Along similar lines, several other questions could easily be addressed by slight modifications of the design. Did you consider any other questions? Why did you target the delivery model and not one of the other questions?
Yes! That’s where the “whittling down” came into play. As in any research, there is always a trade-off between internal validity, how well a study is done and could be replicated, and external validity, generalization to other populations, devices, or practices. Moreover, the default design for clinical trials assumes one primary outcome measure and one secondary outcome measure. To maintain sufficient statistical power, one must keep the number of independent variables involved to a minimum.
As noted, we toyed with just co-varying the purchase price with the service-delivery model, as this is what is typically envisioned. That is, those fitted through audiology best practices would pay $3,600 and those fitted through the CD over-the-counter approach would pay $600. If, however, outcome differences were observed, one would wonder whether this was due to purchase price, service delivery, or the interaction of the two factors.
By moving to a full-factorial combination, we could address this directly in a single study, but this now meant running a total of six different groups (three service-delivery groups times two purchase prices). Adding additional variables, perhaps a third purchase price or different devices, would have increased the number of subjects needed substantially as well as the complexity of the design. We chose the service-delivery process as the focus in this study, with only two values for purchase price, but future research will be needed to enable one to generalize these findings more broadly to other devices, other purchase prices, and other populations.
I could read the article from various points of view and come away with very different conclusions. If I were a manufacturer of an over-the-counter device, I would take away the fact that there were no statistical differences between the audiologist-delivered and consumer-decides outcomes for several of the measures used. Could one interpret this finding to be a full validation of the either model of hearing aid delivery?
No. As noted, this is just one clinical trial and all studies have limits on external validity or generalization. Somewhat surprisingly, very few such clinical trials have been conducted on the efficacy of hearing aids in older adults with mild-to-moderate hearing loss. The “gold standard” for clinical evidence is a meta-analysis of the results of several such clinical trials. I think this study provides a “proof of concept” for this particular over-the-counter approach, but even this needs further study with other devices and populations. It is a good start, I believe, but just that: a start.
Looking at the results from the audiologist’s point of view, doesn’t the fact that only half of those in the CD arm indicated that they would keep the device suggest that an audiologist needs to be involved if we want outcomes beyond chance? Was the half that did not want to keep their devices negatively primed to not utilize hearing health care for a protracted period of time?
Yes, a higher percentage of those in the audiology best practices branch retained their hearing aids, compared to the other two service-delivery branches. In addition, most of those who completed the six-week trial in the consumer-decides over-the-counter branch opted for a follow-up four-week trial using audiology best practices and significant improvements in outcomes were observed. I think both of these general findings speak to the potential value of audiology best practices.
One way to think about this might be as follows. Assume that there are currently 15 million older adults with mild-to-moderate hearing loss in the United States. Over the past several decades, industry data shows that about 20 percent, or three million, seek help and purchase hearing aids. This leaves 12 million who are not currently seeking out or purchasing hearing aids, but who could potentially benefit from them. These 12 million are not being served by audiologists or other professionals. If over-the-counter devices and delivery systems are developed and about half of those 12 million unserved, or 6 million, are now trying and purchasing such devices, but only half keep them based on the percentage who retained their hearing aids in our study, that is still three million more people served than at present; a 100 percent increase in number served! Of course, these numbers are just hypothetical and for illustration purposes, but I think they illustrate the potential gains from alternative service-delivery models for hearing aids.
Moreover, keep in mind, as noted earlier, that subsequent audiology best-practices improved outcomes significantly. As those who purchase over-the-counter devices enter the hearing aid market and experience difficulties or decide that their needs are now greater or just need additional help, they will most likely turn to professionals for that help and follow-up. There is evidence in this study that such help will improve outcomes.
Very clearly said, Dr. Humes. This points me to a couple of things that probably should be thought through as we enter this era of multimodal service delivery. First, I think we should invest in trying to understand candidacy for over-the-counter hearing aids rather deeply. Your results tell me that there is more to it than degree of hearing loss. But how we measure the pertinent variables and deliver the information to the consumer are interesting challenges.
Second, the whole over-the-counter model should probably have strong messaging about seeing an audiologist at the appearance of some signs. I do not mean to suggest the small print that outlines all of the possible side effects in a pharmaceutical commercial, but some meaningful guidance to the consumer will go a long way in making the whole program successful.
Finally, what about the third of the participants in the placebo group who wanted to keep their devices?
As you may recall from the article, we had two different implementations of “placebo” treatment in our study. Both had flat insertion gain to make the devices acoustically transparent in that regard, but one used omni-directional microphones and the other used directional microphones. We were uncertain about the need to restore the directionality of the pinnae, which led to these two versions of the placebo. Basically, no significant differences were observed in outcomes and we merged these two subgroups into one placebo group for all subsequent analyses.
However, I thought that the directionality differences might underlie this observation that about one-third of the placebos indicated that they were likely to retain their hearing aids at the end of the six-week trial. This proved not to be case, however, as about one-third of each of the placebo subgroups opted to keep their hearing aids; the directionality of the devices had no impact.
Of course, it is well known that placebo effects are quite strong in medical research, ranging from placebo drugs to placebo surgeries. Not only were one-third likely to keep their hearing aids, on average, the placebos wore their hearing aids as much as both the other two groups. I suppose this also speaks well of the double blinding in that the wearers truly didn’t know what they were wearing, resulting in positive outcomes for placebos regarding usage and retention of the devices, and to some extent, satisfaction as well.
Let us switch gears and talk about hearing health care in general. The current interest in alternate models of hearing health-care delivery seems to be driven by the ever-increasing need and the sheer lack of qualified personnel to provide the necessary care. What else needs to happen before we can feel assured that an over-the-counter model may be efficacious?
As noted earlier, basically, for every older adult with mild-to-moderate hearing loss seeking help and purchasing hearing aids, there are four others with similar needs that are unmet by the current approach to service provision. Things need to change and options should be explored on several fronts—self-assessment of hearing loss, over-the-counter devices, telehealth for assessment, and delivery of counseling services, etc. Too many adults with mild-to-moderate hearing loss who could benefit from hearing aids are not receiving the services they need in the United States.
I would like to comment, though, that I have an issue with the oft-cited analogy of over-the-counter hearing aids and reading glasses. It is just not that simple with hearing aids. Reading glasses address relatively simple refractory problems in vision, comparable to simple conductive losses in hearing. If we fit conductive losses routinely, instantly achieving “20/20” hearing would be a reachable goal for most hearing aid purchasers. This, of course, is not the case.
Audiologists and other hearing health-care professionals deal with sensorineural hearing loss when providing hearing aids. This is akin to macular degeneration or retinopathy in vision, and I doubt few would argue in favor of meeting the vision needs of all or most of those individuals through over-the-counter delivery. Still, that is why it is important to note that the focus of the over-the-counter alternative to hearing-aid provision is for adults with mild-to-moderate sensorineural hearing loss; not all sensorineural hearing loss.
Let’s change gears again—a randomized controlled trial (RCT) seems like a very involved undertaking. The expectation for scientists nowadays seems to involve a regular and frequent schedule of publications. How does such a publication schedule work with an RCT that might result in publications at the very end of the study?
Over the years, I’ve had the opportunity to mentor a number of PhD students, postdocs, and junior faculty. In each case, I would not advise undertaking a clinical trial. It is very time consuming to do this research and the number of “products” disseminated is not commensurate with the amount of time required. So, I would actively discourage students and junior (pre-tenure) colleagues from pursuing clinical-trials research as a result. I don’t think that this is a serious limitation, though, in that considerable research experience is needed to even prepare to conduct clinical-trials research.
Despite active involvement in clinical research for many years prior to these first randomized clinical trials, I would have been unable to complete the grant applications and the subsequent research without the excellent advice of the late Dr. Gordon B. Hughes, who was overseeing the clinical-trials program of the NIDCD at the time and who I had the opportunity to meet at the NIDCD working group on affordability and accessibility of hearing health care.
Finally, are there any other observations, experiences, advice that you want to share with readers? Do you have any recommendations about how readers can contribute to the open questions about most efficient and efficacious modes of delivering hearing health care?
First, I’d like to note that our particular version of the consumer-decides, over-the-counter service-delivery approach is just one of many versions that could be developed. As new options, some undoubtedly superior, emerge, evidence of comparable quality should be obtained to support those options. Likewise, we had used one particular version of audiology best-practices service delivery. Others can also be evaluated. In fact, it is important to note that this best-practices model included two key components advocated by both the American Academy of Audiology and the American Speech-Language-Hearing Association: real-ear measurement for verification of the hearing aid fitting and an hour-long hearing-aid counseling or orientation session.
Unfortunately, many surveys of clinical practices by audiologists over the past 20 years or so, including some as recently as 2015, indicate that these aspects of best practices are not “common practices.” Thus, it is unknown how the “common practices” hearing aid-delivery model would fare against the consumer-decides over-the-counter model or the “best practices” model examined in our study. It is imperative that our field obtain such evidence to justify the services provided and the payment for those services by others.
Finally, 80 percent of adults with hearing loss in the United States who could benefit from hearing aids are not seeking help and are not purchasing hearing aids. This is a problem or impasse that we should all own and resolve to fix. Collectively, our approach to addressing the needs of this 80 percent have failed consistently over 30+ years. We all must redouble our efforts, evaluate our assumptions, and rethink the solutions.
If we should be so fortunate to be 100 percent successful in addressing the needs of this 80 percent, this would be a 500 percent increase in the number of hearing-impaired adults served. There is more than enough work to go around to keep all involved working diligently toward this goal for many years. Critical at this juncture is good quality research to guide our choices and inform our decisions. Hopefully, this randomized controlled trial is the first of many that will emerge and the paths to achieving our goal of delivering affordable devices to 100 percent of those in need will be met in the near future.
Thank you, Dr. Humes, for your generosity in sharing your views with us. I certainly feel more informed about the whole issue than I would have been after just reading your American Journal of Audiology article. Also, being an audiologist and being involved in clinical science for many years, I want to take this opportunity to thank you for taking on such a difficult but important question. As you said at the very beginning of this conversation—let us hope this is the start.
Donahue A, Dubno JR, Beck L. (2010) Guest editorial: accessible and affordable hearing health care for adults with mild to moderate hearing loss. Ear Hear 31(1):2–6.
Humes LE, Rogers SE, Quigley TM, Main AK, Kinney DL, Herring C. (2017) The effects of service-delivery model and purchase price on hearing aid outcomes in older adults: A randomized double-blind placebo-controlled clinical trial. Am J Audiol 26(1):53–79.