As patients’ hearing thresholds rise, what percentage of speech cues do they miss?


Every hearing care professional understands that people’s ability to understand words deteriorates markedly as their hearing threshold increases. When patients can hear sounds near 0 dB, you expect them to have perfect unaided word understanding ability. But when a patient has a flat 80-dB sensorineural hearing loss, you do not expect him or her to understand words very well unaided. In this week’s post I want to discuss this important topic. I also want to do something that is difficult and perhaps a bit presumptuous on my part. I want to give you my opinion about the “exact” values for an “average” patient.

Table 1 shows the speech information available at standard test frequencies to patients with all the levels of hearing loss that you see regularly in a typical hearing aid office. The information in this table is based on three assumptions:

The patient has a sensorineural hearing loss (there are no air-bone-gap components).

The patient has not had any major ear surgery (patients with stapedectomies fall into a different category).

The patient has “average” ear health, i.e., there is no retrocochlear pathology or neuropathy.

To use this chart, you need to understand the Articulation Index (AI), the ANSI standard that quantifies the percentage of overall speech information that is contained in each of five frequency bands.

According to the AI, the portion of all speech information in each of the five standard octave bands is: 8% in the 250-Hz band, 14% in the 500-Hz band, 22% in the 1000-Hz band, 33% in the 2000-Hz band, and 23% in the 4000-Hz band. In other words, about a third of human speech information occurs in the octave band at 2000 Hz. It is important to remember that the ANSI standard on speech intelligibility is derived from people with normal hearing, not people with hearing loss.

Study the numbers in Table 1 and notice how the values are the same near the top of the table. The values for people with a 20-dB hearing loss are identical to those with a 50-dB hearing loss. If a patient’s hearing thresholds fall in the range of 0-70 dB, the amount of speech information they receive (in each frequency band) is about the same as it is for someone with normal hearing. That means there is a good chance that hearing aids will help them understand speech very well.

However, when patients have hearing losses significantly greater than 70 dB, the amount of speech information they receive decreases markedly. The numbers in Table 1 attempt to quantify the amount of this decrease.

Please note that there is an exception to this 0-to-70-dB rule-of-thumb in the octave band centered at 4000 Hz. According to Harvey Dillon, director of research at the National Acoustic Laboratories (NAL) in Australia, hearing in the higher frequencies (as indicated by the threshold at 4000 Hz) is more “sensitive,” i.e., damage in the inner ear produces a greater effect in the higher frequency zone than at lower frequencies.

The percentage of speech cues lost at 70 dB for other frequencies occurs at 60 dB for the 4000-Hz zone. You can see this effect by studying the numbers in the table. According to Dr. Dillon the 4000-Hz zone processes signals at a higher speed, which means that hearing loss in this zone is more “destructive.”

I had a lot of help generating these values. Along with Harvey Dillon, I consulted with Carl Asp, PhD, my professor at the University of Tennessee, Knoxville, regarding these values and this concept.

Now, just because values deteriorate at higher levels does not mean we should stop recommending amplification for patients with severe hearing impairments. In fact, just the opposite is true. We need to redouble our efforts and work even harder to make sure these patients are receiving every single “ounce” of information that may possibly be accessible to them. We must also open our toolbox and make sure we are using all the tricks of our trade to help people with severe impairments.

I will continue my discussion of this topic in future blog posts.

Table 1. The percentage of speech information in each frequency band accessible to “typical” patients with hearing thresholds ranging from 0 to 100 dB.



Leave a Reply

Your email address will not be published.