Private Practice Audiology on the Cutting Edge

An Interview with Dr. Ron Leavitt
By: Brian Taylor, Au.D. and Ron Leavitt, Au.D

If you’re an avid reader of the trade publications, you’re likely familiar with some of the work of Ron Leavitt, Au.D. Dr. Leavitt, who owns a successful private practice in Corvalis, Oregon has published several pertinent articles on the use of evidence-based practice to improve patient outcomes. In addition, he is a savvy business owner who has been using audiology assistants for several years to lower the overall cost of care for a broader number of patients. I had the opportunity to exchange some emails with Dr. Leavitt a few months ago. As you will read in this exchange, Ron has many useful insights about patient care and the use of scientific principles to improve patient outcomes and business productivity.

Brian Taylor: Over the last few years you and some of your colleagues, such as Professor Carol Flexer have conducted a few timely studies with a great deal of clinical significance. First, can you tell us about your practice in Corvallis, Oregon, and second, how do you find time to conduct research with Dr. Flexer and others while running a private practice?

Ron Leavitt: At Corvallis Hearing Center, I make it a point to only do the tasks that require my specific expertise. I rely heavily on audiology assistants for all other activities. Audiology assistants in my practice do most everything except conducting and interpreting diagnostic tests, and the programming of hearing aids. Because audiology assistants complete prepare all test materials, make copies of needed paperwork, check off billing, outline the reports on each patient, provide the appropriate cable interfaces, and make sure those interfaces are working before I even start programming hearing aids, it frees up more time for me to do those things which I am trained to do – conducting and interpreting diagnostic tests, programming hearing aids, counseling patients, and conducting research.

BT: Let’s discuss some of the clinical research you have contributed to the field. Much of your clinical research seems to involve audiologists’ overreliance on first-fit settings. Could you tell us what motivated you to study a manufacturer’s so-called first-fit algorithms?

RL: I have been practicing audiology for 39 years. During that time I have repeatedly seen that aided audibility results in improved recognition of speech in quiet and noise. Back in 1976, at University of Arizona, we were using functional gain measurements and striving for Dr. David Pascoe’s Uniform Hearing Level fittings. As our knowledge evolved, thanks in part to the work of Dr. Margo Skinner and colleagues at Washington University, it became apparent that we had to consider more than aided threshold audibility.

For example, I recall a conversation with Dr. Wayne Staab in 1976. He had suggested we needed to measure supra threshold audibility in individual hearing aid users. Neither one of us at the time knew how that would be accomplished but his argument seemed compelling. Around 1983, I saw the Rastronix real-ear system at Dr. Robert Sandlin's International Hearing Aid Seminar in San Diego. I remember Dr. Mead Killion and I stood together looking at this innovative piece of hardware saying that it would change everything about hearing aid fitting science. Unfortunately, it appears we were wrong – it remains puzzling that the majority of audiologists still do not routinely use real ear measures to verify real-ear aided audibility has been achieved.

Impressed as I was with that device, we obtained a real-ear system soon thereafter and began experimenting with it. We quickly could see the advantage over functional gain and were able to, in a less formalized way compared to today’s standards, make supra threshold measurements.

Unfortunately, at the time, the Speech Intelligibility Index (SII) didn’t exist in a form we could use, as Dr. Popelka had not yet completed his pioneering work on the subject. Nor did we have access to the NAL or DSL formulae. Despite these limitations, real ear-aided measurements became standard practice on every patient we saw in the 1980s, and my interest in real ear-verified aided audibility was born.

BT: One study you completed with Dr. Flexer that caught my attention a few years back was when you evaluated performance for a speech intelligibility-in-noise task using several high-end devices, in relation to the patient’s older hearing aids. Please explain the design of that study, and the question you were trying to address?

RL: As for the design of this study, we obtained premium-level hearing aids from the so-called Big Six manufacturers and used their first-fit software to program their premium-level hearing aids on our research participants. We then measured their real ear-aided responses and sentence recognition in noise using the Quick Speech in Noise (SIN) test.

It was our thesis that real ear-aided audibility would make more difference for these subjects than the advanced features these premium hearing aids offered. For comparative purposes, we employed an older single-microphone, analog-programmable hearing aid with no noise reduction features, programmed to a real-ear verified NAL target for all subjects. We then measured sentence recognition in noise using the Quick SIN.

As hypothesized, the analog hearing aids outperformed all of the Big Six premium hearing aids when the premium level hearing aids were programmed to the manufacturer’s first-fit starting point. Looking at the real ear-aided measures, it was apparent that each of the manufacturer’s first-fit settings were giving much less aided audibility than suggested by the independently validated, original NAL target. Not surprisingly, Quick SIN scores were best with the analog hearing aid that provided the most aided speech cue audibility - that is, the highest aided SII.

One more point: It was also our thesis that advanced features on the premium hearing aids were of value, provided the hearing aids first obtained sufficient audibility as verified by the real ear-measured NAL target. Not surprisingly, with one exception, all premium hearing aids outperformed the analog hearing aids when the premium hearing aids were programmed to the NAL target.

Logically, you would think that the Big Six manufacturers would really get behind this idea of real ear-verified aided audibility based on our findings, particularly when they have put so much thought and money into the possibility of good speech recognition in noise. Unfortunately, this has not happened.

BT: Based on your findings, what is the take-home message to clinicians for this study?

RL: The initial title of this article was “Audibility Trumps All Other Hearing Aid Features.” At that time, my friend and mentor Dr. James Jerger thought a better title was needed and I deferred to his suggested title, of which I cannot completely recall, but it was more academic sounding. Although Dr. Jerger would likely disapprove of me saying it again, “audibility really does trump all other hearing aid features and benefits.” You cannot rely on all the features that today’s premium level hearing aids provide, until you have built a foundation of audibility that has been verified with real ear measures.

BT: Another aspect of your work that may have flown under the radar, but related to the work we’ve been discussing, was an Audioscan “Best Fit Challenge” contest you won in 2013. As I recall, the objective of the challenge was to highlight an evidence-based approach to objectively and subjectively document hearing aid benefit. What are the components of your approach? For those who think they don’t have ample time to complete these measures, please share with us how long each component in your fitting protocol takes, on average.

RL: I agree with you that this work remains one of the best kept secrets in audiology. This disappoints me and my co-author, as it was our intention to provide a one-page form that would employ our best science to show that people are doing better with their hearing aids than without them.

As for the components of this approach, I agree with Dr. Wayne Staab’s blinded review of the article when he said there’s really nothing new here. It is just that all this basic science has never been distilled into a single page form. While Dr. Staab’s review of our article makes us sound a lot less brilliant, his comment hits the nail on the head. These ideas are not new, but they do seem to be under-utilized. Our one-page form is composed of the Count-the-Dots audiogram, created by Killion and Mueller in the 1990’s and revised about five years ago. In addition, our form includes, a portion of the APHAB from Dr. Robyn Cox and colleagues at the University of Memphis, the SII computations provided by Bill Cole at Audioscan; and the Quick SIN scores obtained in the unaided, first-fit and NAL prescriptive target conditions. Take a look at the actual form.

Figure 1. Evidence-Based Practice Hearing Aid Performance Report

Click to Enlarge

The APHAB is completed in our overflow room to avoid using clinic time. The audiology assistant supervises APHAB completion in the NOAH program before the testing described below begins. If there is a large discrepancy between the patient self-report of hearing loss and our observations, we will often ask the spouse to complete the APHAB from the spouse’s perspective. Oftentimes, this confirms our results when we show the patient the actual severity of hearing loss.

The form (Figure 1) that we created is easy for clinicians to use. Basically, you plot the thresholds on the Count-the-Dots audiogram, take the aided SII calculations from the real ear measurement printout, obtain the SNR loss score using the Quick SIN, and, finally compare the APHAB unaided to aided score in noise to better understand the patient’s self-report of benefit.

As I said earlier, we employ audiology assistants so that we can do a comprehensive audiological evaluation including pure tone air and bone conduction, immittance with ipsilateral and contralateral reflexes at 500 to 4000 Hz, word recognition in quiet, tone decay. We also conduct pure tone UCLs if hearing aids are warranted as well as a sound field Quick SIN at 50 dB HL in the unaided condition (provided the PTA is better than 50 dB HL). This entire battery of tests can usually be completed in 30 minutes.

I agree with those who have argued that single word recognition in quiet is not real world so we use the NU-6 ordered-by-difficulty word list. Thus, I can stop at the first 10 words if the patient gets them all correct.

As soon as I complete this testing, the audiology assistant changes the EAR inserts for the next patient, cleans the patient’s hearing aids, draws a colorful audiogram showing the missing speech sounds (and computed unaided SII to show the patient), enters the patient’s audiogram into NOAH (and sends it to the real-ear equipment in preparation of obtaining real ear data), types the report based on the case history information (and adds information about the pure tone, speech & immittance results) and, finally, provides the front desk with billing charges.

For patients who have hearing aids, when we are done with unaided testing, the assistant has clean probe tubes ready on the real ear system and the patient is directed to the chair in front of the real-ear system. The cleaned hearing aid is inserted in the patient’s ears and we obtain a real ear aided response (REAR) at 50, 60 and 75 dB SPL inputs for the recorded Carrot Passage. If programming is needed, the assistant has the appropriate interface ready and has opened the appropriate software. This eliminates the need for the audiologist to do a lot of computer watching, so I can immediately start programming. This entire procedure I just described takes 10 minutes. Then, we repeat binaural aided sound field Quick SIN at 50 dB HL. In this way, we have proven to ourselves and, maybe more importantly, to the patients that they do better with the hearing aids in noise than without them --- an age-old complaint that needs to be laid to rest on the very first day we provide hearing aids.

BT: Let’s talk a little more about your use of audiology assistants. It seems they are an integral part of virtually every clinical process in your practice. Do I have that correct?

RL: Yes. For audiologists, who believe they cannot afford assistants, I maintain we cannot afford not to have them. As I mentioned, they are involved in just about all aspects of patient care. Since most of my assistants are lifelong hearing aid wearers, they do a very good job of hearing aid counseling. Also, because all my assistants use wireless hearing aid technology themselves, they are better than me, or the tech team at the hearing aid manufacturer facility, in making elaborate wireless connections.

For example, yesterday one of my assistants programmed a ReSound Linx2 hearing aid with two external microphones. Both microphones are used in binaural listening mode. Then each microphone can be used as mono listening devices, so that one can be set on one end of the table, and another remote microphone at the further end of the table. Since my assistants use a similar system, they can very effectively counsel the patient on how to use this set-up, especially when the iPhone microphone settings are appropriately names in the Resound app on the iPhone.

The final article that I wanted to discuss with you was published in June of 2017. In that study, as I recall, you used probe microphone to evaluate how close a group of almost 100 patients, fitted at a variety of locations in the state Oregon, were to the NAL NL-2 prescriptive fitting target. Please tell us what you found.

BT: By way of background information, both Dr. Gus Mueller and I were suspicious of the self-report of real-ear use provided by audiologists and hearing instrument dispensers that he and his co-author Erin Picou published in 2010.

RL: In our study, we cut out the middle man, so to speak. We did not ask these audiologists and hearing instrument dispensers if they were using real ear measures. We just looked at their real ear-aided fittings at 50, 60 and 75 dB SPL inputs using the Carrot Passage after we ascertained the hearing aids were functioning properly. From these data, it was quite apparent that few of the 176 hearing aids examined had been programmed to an NAL target, as only 2.3% came within 5 dB of an NAL target and a whopping 72% were within 10 dB.

In my post-publication conversations with Dr. Mueller, he questioned, if perhaps, some of these subjects might have found the output of an NAL target fitting too loud, and so had requested that their hearing aids be turned down, or that perhaps the patients simply turned the hearing aids down themselves with a manual volume control. Several facts argue against the latter interpretation that the patients manually adjusted the volume lower.

First, we have been re-programming these hearing aids to an NAL target and no one has yet requested we put them back to their “first-fit” setting.

Secondly, our multi-variate statistical analysis showed that the age of hearing aid had no effect on the results. In other words, a person using hearing aids for 10 years was just as likely to be programmed below an NAL target as a person using hearing aids for less than two months. While possible, it is hard for me to believe that a number of eight t-o 10-year experienced hearing aid users would continue to insist that their hearing aids be consistently set 10 to 15 dB below an NAL target, as such a fitting makes most voiceless consonants inaudible. Further, new hearing aids were not significantly further off target than were 10-year old hearing aids, suggesting that this was not simply older hearing aid degradation. The bottom line is that hearing aid users tend to stick to the settings their audiologist gave them at the start.

BT: I am not too surprised by your findings. I think you did an excellent job of shining a light on a significant problem plaguing the industry: Too many patients are under fit relative to a validated prescriptive target, and they end up staying that way. In light of the recent OTC legislation, what do the findings of this study say about how audiologists and hearing instrument specialists fit and adjust hearing aids?

RL: I believe this question raises several issues. The first issue is whether achieving a real ear- verified aided NAL target results in better patient speech recognition in noise. In the past this question was a lot more debatable.

We now have data showing better Quick SIN scores are obtained with a real ear-verified NAL fitting. Let me provide a cursory review of some of the relevant research findings. I would encourage all clinicians to spend some time getting familiar with how these studies relate to clinical practice. You can do that with a simple Pub Med or Google Scholar search. For example, we have physiological data from Dr. Anu Sharma, and colleagues at the University of Colorado, showing better Quick SIN scores are associated with maintenance of proper brain resource allocation and poorer Quick SIN scores are not protective of brain resource reallocation. This is an argument in favor of providing amplification to individuals with mild hearing losses.

We have data from Dr. Larry Humes, and colleagues at Indiana University, suggesting their 99 elderly subjects with hearing loss performed as well on many cognitive, visual and speech in noise tasks as subjects in their twenties with normal hearing, provided that high-frequency consonant audibility was restored on the elderly subjects.

I am not saying that all patients who receive an NAL-verified real ear-aided response will score well on the Quick SIN, but I do believe such real ear-aided responses increase the likelihood of this outcome.

The second question is whether a good Quick SIN or any other speech in noise test score translates into real-world sentence-recognition ability. Brian, you addressed this question in your Audiology Online article in 2011 quoting research by Dr. Sergei Kochkin and commentary from Carole Rogin, of the Hearing Industries Association (HIA), supporting the face validity of speech in noise testing. As you may know, Drs. Doug Beck and Michael Nilsson, in 2013, draw similar conclusions.

The third question, assuming Quick SIN real-world validity, is whether patients that perform well in noise are happy (satisfied) hearing aid users. Again quoting the work of Dr. Sergei Kochkin, we can say the answer is “yes.” However, as you know, patient happiness (satisfaction) is a multivariate phenomenon. Aspects beyond the performance of the hearing aids, like time spent waiting to see you and your ability to be a good listener, are part of the satisfaction equation. That being said, a lot of these patient variables like level of life satisfaction, satisfactory social interaction, chronic pain issues and chronic health problems are beyond our control. What we can control, to some extent, through real ear-verified speech cue audibility is the increased likelihood of improved speech recognition in noise and an increased likelihood of deflecting unfavorable brain resource allocation.

Now, to your question. We can conclude that an OTC model will not likely address this brain resource reallocation or speech in noise dilemma reported by Kochkin as there will be no formal speech in noise litmus test performed in the unaided and aided conditions. There will exist no evidence that there are not huge 10 to 15 dB gaps in the real ear-aided response at frequencies crucial for speech understanding in quiet and in noise. There will be no evidence that the patient’s UCLs have not been exceeded as they were never measured. There will be no evidence that the patient has a normal middle ear system. There will be no evidence that there is not a major cerumen impaction.

If clinicians choose to ignore our best scientific practices, as highlighted by numerous researchers, and summarized on the American Academy of Audiology website’s best practice guidelines, it does not change the fact that there is an ever-increasing body of knowledge supporting these procedures we have been discussing. It is my hope that our article will not only shine a light on this problem, as you have stated, but may suggest that spending more time with the licensed professional is worth the money. I would, however, be naïve to suggest that the money savings associated with bypassing the medical/audiological model will not be attractive to lawmakers. We see numerous historic examples of best-practice abandoned in the name of cost savings. This will likely be just one more example.

I do have a suggestion that would save money and still allow utilization of today’s hearing aids and best science, without resorting to a buyer beware self-treatment and self-diagnosis of hearing and balance problems methodology.

BT: Curious to know what that suggestion is, but before we get there, let me ask you one more thing. In that June 2017 article we’ve been discussing, which was incidentally published in the Hearing Review, in the discussion section, you allude to the series of University of Memphis studies, most of them published in Ear and Hearing that indicate there are no clinically (or statistically ) significant differences in patient outcomes between premium and basic hearing aid technology. How do you use these findings in your clinic when discussing amplification options with your patients?

RL: It might be helpful to your readers for me to first describe the studies. The researchers took a group of older hearing impaired adults, with medically uncomplicated hearing loss, and using best-practice clinical guidelines fit them both basic and premium technology. Each participant used each technology level for about a month. Both the participant and clinicians were blinded to the level of technology being used. At the end of each 30-day period, a series of laboratory and real world measures of benefit were obtained. Because so much data was collected, the researchers split the study into three parts, published recently in three separate issues of Ear and Hearing. I believe this series of studies from the University of Memphis shows essentially the same thing that Dr. Flexer and I found about five years ago: “Aided audibility trumps all other hearing aid features.”

Since Dr. Cox and colleagues programmed both premium and basic level hearing aids to the same real ear-aided NAL prescriptive target, basic-level technology performed comparably to premium-level technology. That being said, I am a big fan of wireless connectivity in hearing aids. As most audiologists know, wireless technology cannot always been found in basic hearing aids.

This wireless feature is often not provided in basic hearing aid technology. For this reason, when the patient’s listening demands warrant it, and the measured SNR loss suggests a frequent need for improved speech recognition in noise, we move beyond a recommendation for basic technology. Currently, I rely on the SNR measurements reported by Olsen in an article published in the American Journal of Audiology (AJA), in 1998, to determine when a wireless recommendation is appropriate. (Here is the article: Average Speech Levels and Spectra in Various Speaking/Listening Conditions: A Summary of the Pearson, Bennett, & Fidell 1977 Report by Wayne Olsen in AJA).

Also, I do believe in the future we will rely on the real time ecological assessment of SNR being developed by Dr. Wu and colleagues at University of Iowa. (For a review of ecological assessment, see Wu Y-H. (2017, June). 20Q: EMA methodology - research findings and clinical potential. AudiologyOnline, Article 20193. Retrieved from

BT: Dr. Leavitt, we really appreciate your contributions to audiology and private practice in general. Looking at the big picture and all of the changes happening within the industry, what will it take for private practice audiology to thrive in the next 10 to 20 years?

RL: As I said above, I believe hearing aid costs will be addressed with or without our professional input. It behooves us to show the value of our services, as we have discussed today. Beyond this, we must embrace a large-scale group purchase option. For hearing aids, the VA already has a large group purchase option, as does Costco and other large retailers. Medicare and Medicaid already have one for medical procedures and prescription drugs. It is standard operating procedure all around us.

I think it unrealistic to educate the masses about the value of audiological services. I say this as a person who has conducted consumer education workshops throughout Alaska, Arizona, Idaho, Montana, Oregon, and Washington as part of a two-year grant through Western Oregon University.

Our own Oregon Association for Better Hearing Group has conducted consumer education workshops every second Thursday of the month at our local hospital here in Corvallis for the past 32 years. I have probably presented the benefits of real ear-aided measures at this meeting on over 100 occasions, in this 32-year time span, to thousands of people with hearing loss. Yet, right here in this town, I have seen 97 successive patients from facilities throughout Oregon with dramatically under-fit hearing aids. For publication purposes, we stopped counting at 97 but we currently could boost this number to over 150 patients.

If it is possible for all these patients to have never heard about the benefits of real ear-aided speech audibility, right in our back yard, what chance do we have of achieving such education on a national level? I believe without a competitive price advantage, private practitioners will be in trouble with or without real ear-aided measures.

In closing, I want to thank you for your excellent questions, your interest in our research and the good work you have done for our profession over these many years. I would also like our national associations to become more interested in cost-containment through group purchasing power. I hope this will become a frequent item of discussion in future professional conferences.    
Dr. Ron Leavitt is a private practice owner and accomplished clinical researcher. He can be reached at