The expected benefit of hearing aids as a function of hearing loss

The expected benefit of hearing aids as a function of hearing lossElaine Saunders 1,2, Peter J Blamey 1,3,4

1 Blamey & Saunders Hearing Pty Ltd 2 Faculty of Science & Engineering, Swinburne University of Technology

3 Dept of Medical Bionics and Dept of Audiology and Speech Pathology, The University of Melbourne

4 The Bionics Institute

Research evidence shows that people who start with higher expectations achieve better outcomes with hearing aids

“Higher use time … [and] greater benefit in easy and difficult listening situations … was predicted by higher pre-fitting expectations.”

Jerram & Purdy, JAAA, 2001

“The data also show that positive expectations result in more positive outcome …”

Saunders, Lewis, & Forsline, JAAA, 2009

There is an alternative view amongst clinicians that:

“… excessively high expectations will result in disappointment and thus poor outcome.”

Saunders, Lewis, & Forsline, JAAA, 2009

These two views can be reconciled if we can define “realistic expectations.”

It is desirable that people have positive but not unrealistic expectations of the benefit of hearing aids.

So what is a realistic expectation,……. does it depend on the degree of hearing loss?

In a previous study using the Profile of Hearing Aid Performance* questionnaire, we showed that percentage of problems for communication in quiet decreased by 0.79% per dB of hearing loss.

Blamey, Martin, & Saunders, SST 2010

* Cox & Alexander, 1995

In the current study, we evaluated hearing aid performance and benefit in quiet as a function of hearing loss using a speech perception test (SPT).

• Data were collected for 492 people in the Blamey Saunders East Melbourne clinic

• A 50-word monosyllabic word test1 was performed in aided and unaided conditions with 65 dB SPL presentation level

• The hearing aids all used the ADRO® amplification scheme2 and were fitted with Blamey Saunders IHearYou® software

ADRO® is a Registered Trademark of Cirrus Logic IHearYou® is a Registered Trademark of Blamey Saunders hears

1 Blamey, Blamey, & Saunders. " Journal of telemedicine and telecare 21.8 (2015): 474-4782 Martin, L., et al. Acoustics Australia 29.1 (2001): 21-24.

Example of Speech Perception Test results (the Infogram™)

Infogram™ is a Trademark of BlameySaunders hears

Before hearing aid fitting, the SPT was carried out (unaided condition). A typical sigmoid curve fitted the data well, with 50% point at 35.9 dB HL and width of 11 dB.

80706050403020100

50

40

30

20

10

0

4-frequency PTA in dB HL

Unai

ded

SPT

scor

e

Unaided SPT score vs PTA

50%

After hearing aid fitting, the SPT was carried out in the aided condition. A typical sigmoid curve fitted the data well, with 50% point at 53.4 dB HL and width of 18 dB.

80706050403020100

50

40

30

20

10

0

4-frequency PTA in dB HL

Aide

d SP

T sc

ore

Aided SPT score vs PTA

Maximum average benefit on the SPT was 16.8 words (33.6%) at 52 dB HL. The benefit increased by 0.72% per dB of hearing loss.

80706050403020100

40

30

20

10

0

-10

4-frequency PTA in dB HL

Bene

fit =

Aid

ed -

Unai

ded

SPT

scor

e

Benefit vs PTA

52 dB HL

16.8 words

On average, people made about half as many errors in the aided conditions as they did in the unaided condition.

Conclusions based on word test results

• The results in this study showed 0.72% per dB improvement in SPT score, consistent with 0.79% per dB reduction in problems on the PHAP in an earlier study.

• Hearing aid benefits in quiet increase with hearing loss up to a PTA of 52 dB HL and then reduce gradually.

• On average, hearing aids produce an improvement in SPT scores equivalent to a 17.5 dB improvement in hearing thresholds.

• There is a lot of variability between people, possibly due to factors other than degree of hearing loss.

People are often surprised at how many mistakes they make on the word test. To estimate how people would score on sentences, we transformed the data using the graph below.

Grant, K.W., and Seitz, P.F. (1997)

A typical sigmoid curve fitted the estimated unaided sentence data well, with 50% point at 48.7 dB HL and width of 8.5 dB.

A typical sigmoid curve fitted the estimated aided sentence data well, with 50% point at 76.5 dB HL and width of 15.3 dB.

Maximum average estimated benefit for sentences was 55% at 64 dB HL. Up to 60 dB HL, benefit increased by 0.72% per dB of hearing loss.

64 dB HL

55% benefit

On average, people made about one quarter as many errors for estimated sentences in the aided condition as they did in the unaided condition.

Conclusions based on estimated sentence scores

Sentence scores are much less sensitive to hearing loss than word scores

Estimated hearing aid benefits for sentences in quiet increase with hearing loss up to a PTA of 64 dB HL and then reduce gradually.

On average, hearing aids produce an improvement in estimated sentence scores equivalent to a 27.8 dB improvement in hearing thresholds.

There is a lot of variability between people, possibly due to factors other than degree of hearing loss.

Realistic expectations depend on hearing loss

Realistically, people should expect that hearing aids will: Halve the number of errors they make on difficult listening

tasks in quiet (such as monosyllabic word perception) Reduce the number of errors they make in running speech

(such as sentences) to one quarter of their current rate Allow a person with a mild hearing loss to perform like a

person with normal hearing Allow a person with moderate hearing loss to perform like

a person with mild hearing loss Allow a person with severe hearing loss to perform like a

person with moderate hearing loss.