When Research Can’t Keep Pace with Technological Innovation



Trying to Better Understand How Clinicians Make Decisions
An Interview with Melinda Anderson, Ph.D.

Most audiologists have some knowledge of evidence-based decision making and its application to the construction of best-practice clinical protocols, but their awareness of these principles does not equate to implementation. To compound the problem, the most often cited published clinical guidelines, such as the American Academy of Audiology’s Clinical Guidelines for Adult Amplification were written more than a decade ago, before the advent of many modern hearing aid features such as frequency lowering and wireless streaming.

To better understand how audiologists make decisions in their clinic today – in a era when technology is changing more rapidly than ever – and how potential blind spots in their decision-making process could have a detrimental effect on patient outcomes, Dr. Brian Taylor sat down with Dr. Melinda Anderson of the University of Colorado School of Medicine in Aurora, Colorado. Dr. Anderson is a clinical audiologist and an associate professor there. She received her master’s degree in Audiology from Vanderbilt University in 2002 and her Ph.D. in Hearing Science from the University of Colorado-Boulder in 2010.

BT: Dr. Anderson, thank you so much for agreeing to sit down with me and talk about audiologist’s attitudes and behaviors toward the use of evidence-based decision making iwn a world of disruptive and rapidly changing hearing aid technology.

MA: Great to spend some time with you between patients.

BT: In a recent paper, published about a year ago in the Journal of the American Academy of Audiology, you surveyed approximately 250 audiologists about their fitting practices. Could you explain the motivation behind your survey?

MA: Our goal with the survey, which was co-authored by Pam Souza of Northwestern University and Kathyn Arehart of the University of Colorado, was to identify the techniques used by clinical audiologists to fit adult hearing aids. As clinical and research audiologists, we place a high value on evidence-based practice. But as any clinician can tell you, the evidence base is far behind what is commercially available. Given this gap, we wanted to identify how clinicians were making decisions when the evidence base did not provide guidance.

BT: The results of that survey are eye-opening. For example, 81% of respondents said specific signal-processing features are very important, yet beyond the audiogram and speech testing scores, few audiologists utilize other clinical strategies and tools in the initial fitting process to gain a better understanding of who might benefit from these specific features. What does this disconnect say about how audiologists make clinical decisions with respect to selecting hearing aid features?

MA: From our perspective, the results show that the current literature does not provide enough guidance in how to use additional resources and tools to make good clinical decisions. The survey findings also show that what evidence is available hasn’t persuaded many audiologists to adopt additional testing tools (such as speech in noise assessments). One contributing factor may be the fact that much of the research and development on signal processing algorithms focuses on algorithms in isolation rather than in combination. But commercial hearing aids combine many types of processing to the end-user, so the effects of algorithms really need to take into account how one algorithm interacts with others. An additional factor may also be the disconnect between laboratory measures and clinically-feasible measures. As the director of a very busy multi-site clinical practice, I value tools that can provide good usable information in short periods of time. In my laboratory, I’ve got the freedom to spend longer amounts of time with individuals to very explicitly examine their responses to signal processing changes. Each of these factors may pose a barrier to the adoption of new clinical testing tools.

BT: When it comes to the fine-tuning process, it seems like 100% of the audiologists rely on patient reports (i.e., base their decision to adjust the hearing aid on what the patient tells them), yet other tools and strategies are “often used” 50% or less of the time, according to your survey? What is driving the relatively low use of these other more scientifically-derived tools like speech in noise testing?

MA: Based on the responses to the survey, and from ongoing discussions with clinical audiologists, I believe there are perceived time constraints and limited understanding of how to apply the evidence base to daily clinical practice. Audiologists ask questions like: What do I do with the information collected? How can I apply the results of this test to this patient? If there is not an obvious way to use the information, providers are unlikely to be willing to spend the clinical time capturing the data. One long-term goal of our work is to provide a roadmap for clinicians on how to apply test results to clinical fittings.

BT: Last November, you co-authored another article with those same colleagues in the International Journal of Audiology (IJA). In that IJA article, in which you evaluated two different types of signal processing found in hearing aids, could you explain the study design and the difference between mild and strong signal processing that you employed in the study?

MA: Using a randomized, double-blind crossover trial design, adults with mild-to-moderate symmetrical sensorineural loss were fit with behind-the-ear hearing aids. To make sure there was effective blinding, the clinical audiologists completing the fitting were not involved in the experimental data collection. The experimenters collecting data were not informed of the processing condition. Participants were kept blind about the specifics of the processing conditions until the end of the study, although participants could hear there had been a change in signal processing at the transition time point. The devices were programmed for strong signal modification (fast-acting wide dynamic range compression (WDRC) plus frequency compression) or mild signal modification (slow-acting WDRC, with no frequency compression). Each processing condition was worn for an average of 5 weeks, during which adherence data were collected to ensure consistent use of the devices. The order of processing conditions was randomized across participants.

BT: Let’s talk about the results of this study. It seems that participants perceived a difference in those two signal processing strategies on both the EAR and SSQ (two different measures of self-reported benefit). However, it seems that, for some patients seen in a busy clinic, participants reported results that were poorer in the aided condition than the baseline unaided condition, especially for the strong signal processing. Do I have that right? What are the clinical implications of this finding?

MA: By design, the frequency compression characteristics in the strong condition had very distinct acoustic differences compared to the mild condition. The consequence of our parameter-selection is that the amount of frequency compression was stronger than might typically be used in a clinical setting. For clinical fittings, frequency compression should be set based on the needs of the individual listener and should be based on assessment of audibility of high-frequency sounds, using probe microphone measures as well as assessment of speech perception abilities. In this study, all participants received the same amount of frequency compression regardless of audibility of high frequency sounds without frequency compression. However, even with this constraint, there are implications for clinical practice. For example, this study shows that is possible to use the EAR and SSQ scales for documentation of perceived performance with hearing aids, and that perceived performance does change as a function of signal processing modifications.

BT: What are your recommendations for clinicians when making decisions about selecting a signal processing strategy. How might you go about making those decisions based on your two papers?

MA: In our clinic, we routinely complete standard audiometric evaluations, probe microphone measures, speech-in-noise assessments, and measures of patient benefit and satisfaction. Each of these tools plays a role in the determination of signal processing strategies. In clinical fittings, you aren’t selecting a single strategy. We must be aware of the interactions of the many types of signal processing implemented in commercial devices and consider how the output of a hearing aid with all signal processing in combination may impact a listener.

Also, in our clinic, we use speech-in-noise assessments to guide decision making regarding directional microphones and coupling systems. We use probe microphone measures and speech assessments to determine the gain-frequency response, compression ratios, and use of frequency lowering. Pre and post measures of perceived performance can help guide the decisions regarding noise management. We use each of these tools to consider how an individual is responding to the device as a whole in the context of their primary listening environments. In clinical practice, we seek to document and justify the decisions we make for patients.

Without the use of evidence-based tools, like the speech in noise tests or validated self-reports of patient outcome such as the SSQ, it is difficult to provide justification for the use of amplification. Ongoing collaborative laboratory and clinical work at the University of Colorado Health Hearing and Balance Clinics, the University of Colorado Boulder, and Northwestern University continues to explore how clinical practice may benefit from understanding the impacts of combined signal processing on hearing aid outcomes. Be looking for that work over the next few years!

BT: Thanks for your insights, Dr. Anderson.

MA: My pleasure.    
Melinda Anderson, Ph.D. is Assistant Professor in the Department of Otolaryngology at the University of Colorado School of Medicine in Aurora, CO. She can be contacted at melinda.anderson@ucdenver.edu.
Citations
Anderson, M. C., Arehart, K. H., & Souza, P. E. (2018). Survey of current practice in the fitting and fine-tuning of common signal-processing features in hearing AIDS for adults. Journal of the American Academy of Audiology, 29(2), 118-124.

Anderson M, Rallapalli V, Schoof T, Souza P, Arehart K. (2018) The use of self-report measures to examine changes in perception in response to fittings using different signal processing parameters. Int J Audiol. 57(11), 809-815. Epub 2018 Jul 27.