Concept Report
of the “Societal Impact of Hearing Impairment” Study Group
Costs and Benefits Analysis
of
Fitting Hearing Aids in the Netherlands
A Cost-Effectiveness Study
Follow-up 18 months
Commissioned by the
European Hearing Instrument Manufacturers Association
(EHIMA)
SIHI, Maastricht, May 2000
Copyright © 2000 by SIHI, Maastricht, The Netherlands
All rights reserved
Citation of this report: SIHI (Societal Impact of Hearing Impairment) study group 2000. Costs and Benefits Analysis of Fitting Hearing Aids in The Netherlands. A Cost-Effectiveness Study. Follow-up 18 months. Commissioned by EHIMA (European Hearing Instrument Manufacturers Association). SIHI, Maastricht, The Netherlands. RM/2000/05.01.
Contractor:
Stichting ter bevordering van wetenschappelijk onderwijs en onderzoek in de Keel-, Neus- en Oorheelkunde, Werkgroep Audiologie, Maastricht (Foundation for the promotion of scientific research and education in Oto-Rhino-Laryngology, Task group Audiology, Maastricht)
P. Debyelaan 25, 6202 AZ Maastricht.
SIHI study group:
Department of Otorhinolaryngology, Head & Neck Surgery, University Hospital Maastricht; Department of Clinical Epidemiology and medical Technology Assessment (CEMTA), University Hospital Maastricht; Department of Quantitative Economics, Faculty of Economics and Business Administration, University Maastricht; MHERCA, Maastricht; Hoensbroeck Audiological Centre (Venlo, Venray); Health Centre Neerbeek.
Table of contents
Introduction 4
Study population 5
Improvement in Objective Measurement 7
Experienced Hearing Improvement 7
Hearing Aid use10
Hearing Aid Evaluation11
After Care14
Social Contacts16
Quality of Life, the EuroQol18
Hearing Quality of Life Questionnaire20
Costs per QALY22
Conclusion25
Literature27
Introduction
In July 1999, a report of the clinical study of the benefits of hearing aid fitting by the Maastricht SIHI group was completed. A follow-up study was launched to determine to what extent the benefits found by the new hearing aid subjects partaking in the first study were still present one and a half-year after hearing aid fitting.
The time interval between intervention and evaluation necessary to measure benefit of intervention is still subject to research. ‘Quality adjusted life years’ indicate the gain in utility due to the intervention and the expected time the effect will last. This is depicted in Figure 1. The grey area represents the effect of intervention, while the white area below it represents utility in case of no intervention.
Figure 1. ‘Quality adjusted life years’
In the case of hearing aid fitting, a reduction in self-perceived disability, and therefore an increase in utility can be observed after intervention. Long-term use of hearing aids is likely to result in stabilisation and reduction of perceived gain in utility. The utility score is also influenced by a number of factors; variability in subjects’ health condition and self-perception, variability in hearing aid performance, decreasing hearing levels and adaptation to the benefit all result in a varying utility gain observed after hearing aid fitting. Important to notice is the fact that hearing users need to be refitted with a hearing aid after time. A more realistic way of presenting this is depicted in Figure 2.
Figure 2 ‘Quality adjusted life years’
Also shown in Figure 2 are moments of measurement in the SIHI study: Baseline: the fitting of hearing aids in first time users, T2: the second evaluation after 16 weeks, and T3, the evaluation after 18 months and subject of this follow-up study. Not depicted here is utility at T1, the first evaluation after 6 weeks. Of interest in the present study are the items: perceived hearing ability, quality of life, utility, quality and quantity of social contacts. The original questionnaire was enhanced with some additional questions regarding the hearing aid dispenser, the quality of after-care, experienced hearing aid comfort and discomfort.
Study Population
The number of participants was 126 at baseline. This reduced to 88 after 6 weeks (T1), and 66 after 16 weeks (T2). Results obtained in these 66 were included in the first report on the clinical study of the benefits of hearing aid fitting. Those 66 persons were contacted first for the follow up measurements at 1.5 years. Results today show 41 participants and at least 16 dropout cases. Inquiry into the reasons for this dropout yielded the following: 5 deceased, 4 having discarded use of the hearing aid altogether, and the rest feeling too old or that it was too much effort to make the visit for the measurements. Measurements were obtained in a total of 41 persons and are included in this report. The average age of the follow up population was 67.75 at baseline compared to 68.59 in the original study. The Fletcher index best ear was 47.66 dB, which was also similar to 47.55 dB in the larger group. (See Table 1 for details).
There remains a group of potential participants, which will be included in the follow up study. Another 10-20 participants are expected to be measured in the coming weeks. Eventually, papers and articles will deal with results obtained in the total group.
Table 1. Background descriptives of the participants at 16 weeks and 1.5 years.
Gender16 weeks1.5 years
Male56.1%58.5%
Female43.9%41.5%
Marital Status
Married, living together66.7%78%
Widow/widower23.3%12.2%
Unmarried7.6%7.3%
Unmarried, living together3.0%2.4%
Employment
Yes15.2%4.9%
No84.8%95.1%
Social Economic Status
Unskilled labour6.6%2.4%
Skilled labour6.6%4.9%
Lower/intermediate29.5%29.3%
Self-employed small business6.6%4.9%
Intermediate/higher position6.6%9.8%
Self-employed, higher staff/higher work44.3%43.9%
Residence form
Alone and self-sufficient93.9%95.1%
Retirement home1.5%
Nursing home1.5%
Otherwise1.5%4.9%
Domestic situation
Living alone21.2%9.8%
Living with others78.8%90.2%
Children at Home
Yes15.2%12.2%
No84.8%87.8%
Mean age in years21.825.2
Improvement in Objective Measurement
Participants were submitted to an audiometric examination at 18 months, aided and unaided thresholds were compared to the baseline audiometric measurements. Loss in dB was measured at 500, 1000, 2000, 4000 Hz and averaged. The average in the best ear was taken as an objective measure of hearing. The average improvement in hearing between baseline (unaided) and 18 months (aided) was 13.93 dB.
Figure 3 shows the average hearing thresholds in dB and a range of two standard errors at baseline and 18 months.
Figure 3. Average hearing thresholds at baseline and 18 months.
Experienced Hearing Improvement
The Amsterdam inventory, a hearing specific Quality of Life measurement, which derives scores for five aspects of hearing, was administered on the basis of using the hearing aid. This had also been done at 6 and 16 weeks and compared to baseline where there is not a hearing aid in use. The figures following illustrate the trends in the specific scores at the various moments of measurement. Baseline measurements are taken as the reference level and the improvement in the Amsterdam Inventory scores at the following measurement moments are shown. At 18 months the improvement when compared to baseline is still substantial. There is a diminished improvement compared to earlier follow up measurements. This could be attributed to either an increased awareness of disability because of hearing aid use, i.e., one is now more aware of what one was missing, or possibly to decreased functioning of the ears themselves [Taylor 1993]. He also found a decrease after 3 months with a stabilisation at 6 months and one year. We cannot know whether there was the same stabilisation at 6 months in this study as the participants were measured at 4 months. Therefore the negative trend seen between 16 weeks and 18 months might possibly be attributable to the fact that a levelling off occurred somewhere after 16 weeks and stabilisation occurred. As could be expected, the experienced improvement is particularly maintained in sound distinction.
Figure 4.1 Experienced hearing improvement, auditory localisation
Figure 4.2 Experienced hearing improvement, sound detection
Figure 4.3 Experienced hearing improvement, sound distinction
Figure 4.4 Experienced hearing improvement, intelligibility in noise
Figure 4.4 Experienced hearing improvement, intelligibility in quiet
Hearing Aid Use
A number of questions were posed as to how much, where and when the participants used their hearing aids. Thirty of the 41 reported using their hearing aid every day, five on most days and 6 sometimes or on some days. On these days, 22 reported using the aid the whole day, 9 the greatest part of the day, 8 half the day, and two only for short periods at a time. Thirty-six wore their aids within the home and when going out. Four reported using the aid mainly when going out and one primarily at home.
When asked questions about how “use now” compared to “use when it was just new”, 30 reported using it just as often, five less often and 3 more often. As to how many hours per day they wore their aid compared to the beginning, 29 reported wearing it just as long, 4 less, and 8 longer.
Figure 5.1. Number of participants turning their hearing aid off in particular situations
Figure 5.1 shows the number of participants who always (or usually, often, occasionally, never) turn their hearing aid off in particular situations. It is important in interpreting the following graphs to consider the fact that the population considered here is elderly. Thus situations such as being in a group in a noisy environment are going to be rare and possibly uncomfortable for these persons anyway. It would seem the activities most prevalent in this population would be talking to one person or watching television.
Figure 5.2 gives an indication of how many participants rate the hearing aid as very good, good, moderate, poor or worthless in various situations.
Figure 5.2 Hearing aid ratings in various situations.
.
Hearing Aid Evaluation
Also assessed were the advantages and disadvantages of hearing aid use. The bar chart (Figure 6.1) shows how many participants expressed various complaints about the hearing aid. Figure 6.2 shows the number of participants expressing certain benefits they experienced from their hearing aids.
Figure 6.1 Participants expressing complaints.
Figure 6.2. Participants expressing benefits.
The participants were asked to rate a number of statements on a VAS scale from 0 to 100 whereby 100 would mean the participant totally agrees with the statement. Figure 6.3 gives the mean scores and two standard errors for a number of negative statements regarding hearing aids. Figure 6.4 does the same but for a number of positive statements.
Figure 6.3 Negative statements regarding hearing aids
Figure 6.4. Positive statements regarding hearing aids
After Care
The participants were asked how many times (Figure 7.1) and forwhat reason (Figure 7.2) they visited the hearing aid dispenser and the prescription giver (audiologist or ENT staff) since the last meeting.
Figure 7.1. After care, number of visits
Figure 7.2. After care, reason for visits
When asked how they would rate (on a scale of 0 to 100) the hearing aid dispenser and the prescription giver in after-care and relaying related information the results in Figure 7.3 were obtained. It must be kept in mind here that the time the dispenser has to explain matters relative to the prescription giver is many times larger, i.e., the dispenser usually spends three to four hours with a client, whereas the prescription giver spends on average only 15 minutes. Only 7 persons judged the after-care of the prescription giver due to the fact that only a small portion comes back to the prescription giver for consultation.
Figure 7.3 Rating of information by dispenser and prescription giver
Social Contacts
Participants were asked how many persons visited them (Figure 8.1) and how many visits they paid others (Figure 8.2) per month.
Figure 8.1 Visits received
Figure 8.2 Visits paid
Also asked was to what extent (Figure 8.3) and how often (Figure 8.4) their physical or emotional health had hindered normal social interaction with family, friends and neighbours.
Figure 8.3 Extent of social interaction hindered by health state
Figure 8.4 Amount of social interaction hindered by health state
Results show that hearing aid use contributes to a substantial improvement in quality and quantity of social contacts.
Quality of Life, the EuroQol
Participants were asked to rate their present state of health on a scale from 0 ‘worst imaginable health’ and 100 ‘best imaginable health, just as they were asked at baseline, 6 weeks, and 16 weeks. The trend in this score is given in the following graph. The initial significant gain in perceived general health is maintained at 18 months.
Figure 9.1 EuroQol VAS scores
The participants were also asked to answer five questions relating to mobility, pain and other complaints, self-care ability, ability to perform daily activities, and feeling of anxiety and depression. The Dolan algorithm is then applied to obtain a Dolan EuroQol score that can take on values between 0 and 1. The higher the score, the better the quality of life is. The graph below shows the trend in median scores at the various measurement times. The median scores are given as the Dolan score values do not appear to be normally distributed. The population is a relatively healthy group and therefore the values are relatively high to begin with. Nevertheless a significant gain is obtained at 6 weeks and this is maintained for at least 18 months.
Figure 9.2 Dolan median scores
The Hearing Quality of Life Questionnaire
The Hearing Quality of Life Questionnaire (H-QOL) is a newly developed instrument to measure the influence of hearing on quality of life. The H-QOL offers a description of the influence of hearing on quality of life, a valuation of one’s sense of hearing, as well as a valuation of the influence of hearing on the overall health state.
The H-QOL consists of two parts, a description of the health state regarding hearing and a valuation of the health state regarding hearing. The health state description consists of five questions, each question represents a factor of hearing handicap derived from the Amsterdam Inventory, a questionnaire that measures hearing handicap [Kapteyn et al, 1995]. To valuate the hearing related health state a Visual Analogue Scale (VAS) ranging from 0 (‘deaf’) to 100 (‘perfect sense of hearing’) is used. This VAS is referred to as a personal hearing VAS. The value of hearing related health state is transformed to a general health visual analogue scale ranging from 0 (‘dead’) to 100 (‘perfect health’) using linear regression.
The results of the health state description in the H-QOL are shown in the section of the Amsterdam Inventory in this report.The results of the valuation of hearing related health state, the scores on the personal hearing VAS, are shown in Figure 10.1.
Figure 10.1 Valuation of hearing related health state. Specific utility.
These results show that the valuation of the sense of hearing initially strongly increases after the fitting of the hearing aid (T1 and T2), while at 18 months after baseline (T3) a slight decrease in improvement is observed. However there is still a significant increase when compared to baseline.
There is a linear relation between the two Visual Analogue Scales, and therefore the gains on the personal hearing VAS were transformed by a linear regression model to the generic VAS: Gain on generic VAS = 0.659 * gain on personal hearing VAS
The gain in utility after hearing aid fitting is approximately 18% at the measurements T1 and T2. At the measurement T3, 18 months after hearing aid fitting, the utility gain has decreased to 12%. This is to be expected, as hearing impaired persons are likely to become adapted to the benefits of their hearing aid. The results of the transformation are shown in Figure 10.2
Figure 10.2 Utility gain from linear regression
Costs per QALY
The costs/QALY were calculated using the utility values from the patient study. Since the average baseline age in the present follow-up study is 67.75 years, the costs of the cohort aged 65 to 69 years were taken from the model. Costs/QALY calculations are always discounted because of time preference. The standard discount rate is set at 5%. The costs per QALY in the group aged 65-69 years using the utility gain measured at T1, T2 and T3 are compared in Table 2. The costs per QALY outcomes using the utility gain from the EQ-5D are relatively stable across the measurements. The costs per QALY-outcome based on the Hearing-QOL shows an increase at T3.
Table 2: Costs/QALY-outcomes using different utility values
Questionnaire / Utility gainBaseline–T1 / Costs/
QALY / Utility gain
Baseline–T2 / Costs
/QALY / Utility gain
Baseline–T3 / Costs
/QALY
EuroQol
VAS / 3 / 15,000 / 2 / 23,000 / 3 / 15,000Dolan Algorithm / 2 / 23,000 / 4 / 11,500 / 2 / 23.000
Hearing QOL
linear regression / 18 / 2,600 / 18 / 2,600 / 12 / 3,800Figure 11 shows the age specific costs per QALY outcomes using the utility gain form the EQ-5D (2,3 and 4%). Hearing aid fitting is more cost-effective in younger age categories. This is due to higher mortality and a larger proportion of people becoming dissatisfied with or unable to adjust to a hearing aid in older age categories.
Figure 11. Age specific costs per QALY
To put these figures in perspective, Laupacis et al [1992] calculated cost/QALY cut-off points to interpret the outcomes of cost-effectiveness studies (see Figure 12). The cut-off points are based on information about the costs of interventions that are reimbursed and not reimbursed. Laupacis found that governments are rarely willing to reimburse interventions that cost more than 100,000 Euro per QALY.
Interventions costing between 20,000 Euro and 100,000 Euro per QALY are sometimes reimbursed, and sometimes not. Interventions costing below 20,000 Euro per QALY are almost always reimbursed. The highest calculated costs\QALY for fitting hearing aids in this study was 23,000 Euro per QALY, just over 20.000 Euro per QALY. Taking into account that this outcome was influenced by a number of factors such as adaptation, the conclusion of this study is that hearing aid fitting should be reimbursed.
Figure 12. Outcomes of cost-effectiveness studies
Another way of placing the costs/QALY outcome in perspective is a QALY league table. In this table different interventions are ranked from the lowest costs/QALY to the highest.
Table XXX: QALY league table
Intervention / Costs/QALYPKU screening1 / < 0
GP advising to stop smoking7 / 500
Anti-hypertensive therapy to prevent stroke (45- 64 years old) 7 / 1,700
CABG (left main vessel disease, severe angina) 7 / 4,000
Neonatal intensive care (1 to 1,5 kg weight) 1 / 7,500
Breast cancer screening7 / 10,300
T4 (thyroid) screening1 / 10,500
Hearing aid (65 to 69 years old, utility gain = 4) / 11,500
Hearing aid (65 to 69 years old, utility gain = 3) / 15,000
Cochlear implant3 / 15,000
Repair of asymptomatic intracranial aneurysms4 / 17,400
20,000 Euro/QALY
Hearing aid (65 to 69 years old, utility gain = 2) / 23,000
Implantable defibillator5 / 27,500
Cardiac transplant2 / 36,400
Oestrogen replacement therapy (for postmenopausal symptoms in women without hysterectomy) 1 / 45,000
Knee replacement6 / 46,500
Coronary bypass for single vessel disease with moderate angina7 / 60,500
School tuberculin testing program1 / 72,800
Peritoneal dialysis2 / 78,000
Hospital hemodialysis1 / 90.000
100,000 Euro/QALY
Erythropoietin treatment for anaemia in dialysis patients (ass. 10% mortality reduction) 7 / 100,700
Neurosurgery malignant intracranial tumors7 / 191,900
Erythropoietin treatment for anaemia in dialysis patients (ass. no increase in survival) 7 / 233,900
Primary prevention of hyperlipidemia (using Lovastatin in a population of 35-44 years old) 4 / 1,900,000
1Torrance & Zipursky, 1984, 2Kupperman et al, 1990, 3Wyatt et al, 1996, 4King et al, 1995, 5Larsen et al, 1992, 6Drewett et al, 1992, 7Drummond, 1993.