Audiology Science to Practice: Interview with Steven Kramer, PhD
Douglas L. Beck, AuD, spoke with Dr. Kramer about the second edition of his book, Audiology Science to Practice, as well as acoustic reflex testing, auditory neuropathy spectrum disorder, word recognition scores, monitored live voice, recorded tests, phonemic scoring, confidence levels, and more.
Academy: Hi, Steve. Nice to speak with you again.
Kramer: Thanks, Doug. Good to hear your voice, too.
Academy: I really enjoyed your new book Audiology Science to Practice, the second edition. I noticed you have some of my favorite audiologists as contributors, including Gus Mueller and Jim Jerger.
Kramer: Thanks for mentioning that. Their contributions were wonderful and significant and I was honored to have them contribute.
Academy: How long ago did the first edition come out?
Kramer: The first edition came out in 2008. But as you know, there've been so many developments and changes in the profession over the last five years, and there were so many things I didn't get to put into the first edition…so the timing worked out for Plural Publishing and for me, and so we decided to get it done.
Academy: I noticed you've added all new chapters on masking and hearing screenings, and have expanded and updated content in many of the other chapters as well, is that right?
Kramer: Yes, that's correct; we've added new chapters and we've updated chapters while trying to maintain the readability - which was popular in the first edition.
Academy: I agree! Steve, with regard to the importance of the acoustic reflex, you offer a lot of pragmatic insight in the book. I wonder if you can offer the readers a little "greatest hits" version?
Kramer Sure. Well, first of all, the question about how and why we even have an acoustic reflex (AR) remains unknown. Doug, as you pointed out in your Academy article, 2013 Op/Ed on Auditory Neuropathy Spectrum Disorder, it's not very likely the AR developed to protect us from loud noise, as we haven't really had substantial noise in the acoustic environment until perhaps the industrial revolution, and it's not likely the AR would have adapted in the last 250 years!
So, to me, that's a theory that has simply outlived its time. Nonetheless, the AR does help decrease the total loudness and it also helps reduce distortion for high level inputs. Of course the clinical value of the AR has increased lately, too, especially with regard to auditory neuropathy spectrum disorder (ANSD) and the fact that some 95 percent (or more) of the people with ANSD will not have an AR within normal limits…and so, AR has a lot of differential diagnostic value to which we need to pay attention.
Academy: I agree, and I know some clinicians will find it shocking, but one can obtain ARs on infants less than 72 hours of age. Kei (2012) showed that ARs can be found at 82 dB HL (average) at 500 Hz, 71 dB HL at 2000 Hz and 65 dB HL at 4000 Hz, and perhaps most importantly, for broad band noise, 57 dB HL was the average AR. Another hot topic with regard to ARs is correcting the response based on and with respect to the measured ear canal volume—this is an area more than ripe for additional clinical research.
Kramer: I agree, Doug, and of course AR is very simple, very easy to obtain, is already assigned a CPT code, and it clearly deserves more attention in clinical audiology.
Academy: Steve, another area of clinical practice that is quite simple and sadly lacking with regard to day-to-day audiometry is speech audiometry. Let me ask you to review some of the key elements of clinical speech audiometry, and please interject some science and statistics!
Kramer: Sure, Doug. Well, let's start with the basic concept of the word recognition score (WRS), which used to be called "speech discrimination." The core idea is that for the test to have value it has to be equally weighted with regard to speech sound representation and has to be applied scientifically with regard to test length and presentation mode.
Academy: Of course, you're referring to the fact that most people present 25 words, which eliminates the concept of phonetically balanced (PB), which was originally designed into the 50 word lists, and I think you're referring to the presentation mode issue of monitored live voice (MLV) versus recorded presentations?
Kramer: Exactly. When someone presents a 25-word list, it might be okay on some levels, but not okay on others. It all depends on which question you're asking. The presentation level also matters a great deal. Although I prefer to test at multiple levels to define a patient's maximal score (PBmax), it seems to be common practice to do WRS at a single presentation level. If that's the case, then I'm a proponent of testing at high levels (UCL minus 5) to maximize acoustic information to get an estimate of the patient's PBmax.
Another important thing to consider is how "far apart" two scores (e.g., right ear compared to left ear) have to be in order to be statistically significantly different from each other, based on critical differences of the binomial distribution. Using 25-word lists can require huge differences between scores to be significantly different. Let me give you a few examples.
If you only present the WRS stimuli at 40 dB sensation level (SL), or at normal conversational levels (such as 50 to 55 dB HL) to determine a patient's maximal score (PBmax), you may not get their true PBmax or you might exceed their UCL. The research indicates you should present the WRS stimuli at just under the uncomfortable loudness level (UCL) to assure full audibility.
With regard to MLV versus recorded stimuli, there's really no argument. MLV is not standardized and is not repeatable. I can list all the problems with MLV, but the folks doing it know the limits, and they choose to do it, regardless. MLV is not our best (or an advantageous) protocol, and I recommend all WRS testing should be done with recorded stimuli.
Academy: I absolutely agree. Tell me about scoring phonemes versus words?
Kramer: That's a great question…for people who want to use 25-word lists to save time, if they were to score each word for the presence or absence of each phoneme, there are 75 scored items, and the variance is reduced with more scored items. For example the word cat consists of the three phonemes /k/ /a/ and /t/. So if the patient responded "cast," and one were doing 'word scoring' the patient would get no credit. However, if one were to score by phoneme, the patient would get 2 of the 3 phonemes correct, and this would likely better represent what they actually perceived. Arthur Boothroyd has developed a Computer Assisted Speech Perception Assessment Test (CASPA) that uses 10 words per list, yielding 30 phonemes, and so you can get a quick 'WRS' score using ten words.
Academy: That's brilliant. And it takes me back to the California Consonant Test (CCT), same idea, and very efficient and effective. And what about the idea from Thornton and Raffin (1978) about binomial distribution as it relates to using 25- versus 50-word lists?
Kramer: Well, we republished the Boothroyd and Mackersie work on this topic in the second edition. The quick version goes like this….if we're going to use a 25-word list, and the first WRS score is 68 percent, then for the second score (from the other ear) to be statistically different (at a 95 percent confidence level) from the first score, the second score has to be greater than 91 percent. That is 68 percent on the first ear and 91 percent on the second ear (using a 25-word list) are statistically the same.
Now, if the clinician had used the 50-word list and again obtained 68 percent on the first score, the second score would only have to be 86 percent or greater, to be statistically different from the first score; however, 50 words would obviously take longer to administer. But again, my preference is to use the phoneme score to get more score-able items and better representation of the sounds being perceived. As the critical differences chart (in the book) shows, under the 63 item column which represents the critical differences for a 25 word list using phoneme scoring (2.5 phonemes/word), if the lower score is 68 percent, then the upper score only has to be 84 percent or higher, to be significantly different.
Academy: And what about WRS as a function of 3 frequency pure-tone average (3FPTA)?
Kramer: That's another one of those things we often take too casually. There are data by Dubno et al, 1995 (republished in my textbook) that provide better clinical decisions when we're interested in the differential diagnosis of cochlear vs 8th nerve disorders.
For example, if a patient has a 25 dB 3FPTA, the chart shows the lower end of the 95 percent confidence interval for ears with cochlear losses is 80 percent WRS; therefore, if this patient shows a WRS poorer than 80 percent, you would be highly suspicious of an 8th nerve disorder. Likewise, if the 3FPTA is 30 dB, 76 percent WRS represents the lower, and if the 3FPTA is 40 dB, 64 percent represents the lower limit and at 70 dB 3FPTA, the WRS could be as low as 28 percent and still be be consistent with a cochlear hearing loss…and those examples illustrate the latitude of the test and gives audiologists some data to apply evidence-based practice.
Academy: Right, and to me the data (above) are sort of shocking because the WRS test is just not a very good test to begin with, and then when we don't use recorded speech, and we don't use a maximal presentation level, we pay no attention to temporal cues etc, the variation (in results) becomes even greater (than indicated above!). And so there you have it, Audiology from Science to Practice. It's so important to actually apply science to our clinical audiology protocols!
Kramer: I agree. And again that's what the book is all about!
Academy: Thanks, Steve. I think the book is great, it does indeed apply science to audiology and it's a comprehensive and easy-to-read text that makes very good sense!
Kramer: Thanks, Doug. I appreciate your interest in the book.
Steven Kramer, PhD, is a professor at the School of Speech, Language, and Hearing Sciences at San Diego State University, and serves as co-director of the Audiology Joint Doctoral Program at SDSU & UCSD.
Douglas L. Beck, AuD, Board Certified in Audiology, is the Web content editor for the American Academy of Audiology.
Beck, D, Speidel DP. (2013) Acoustic Middle Ear Reflexes: Simply, Under Used, and Critically Important. Hearing Review. 18(39).