«THE ECONOMIC IMPACT AND COST OF HEARING LOSS IN AUSTRALIA A report by Access Economics Pty Ltd February 2006 Listen Hear! The economic impact and ...»
Hearing loss has been described as an under-estimated health problem (Wilson et al, 1992). Adult hearing loss is associated with an increased risk for a variety of health
diabetes (Wilson et al, 1992; Mitchell, 2002);
stroke (Mitchell, 2002);
elevated blood pressure (Wilson et al, 1992);
heart attack, particularly those rating their hearing as poor (Hogan et al., 2001);
psychiatric disorder, particularly those rating their hearing as poor (Hogan et al., 2001);
affective mood disorders (Ihara, 1993; Mulrow et al, 1990);
poorer social relations (Mulrow et al, 1990);
higher sickness impact profiles (physical and psycho-social (Bess et al, 1989);
reduced health related quality of life, particularly those with more severe hearing loss (Wilson, 1997).
Wilson (1997:104-108) investigated two health conditions (diabetes and tinnitus) and two risk factors (overweight and risky drinking) in relation to people with hearing loss.
However, only the relationship with diabetes was found to be statistically significant at all levels of hearing loss. The Blue Mountains Hearing Study (Mitchell, 2002) also reported that older people with hearing loss (25 dB) were more likely to report diabetes (OR 1.5 CI 95% 1.1-2.1), as well as a history of stroke (OR 1.7 CI 95% 1.1and currently smoking (OR 1.4 CI 95% 0.9 – 2.1).
Mortality: While hearing loss has been associated with a number of conditions that are life threatening (eg diabetes) and with social isolation that may also lead to premature mortality, no direct causality has been found between hearing loss and increased mortality or injury rates. One study by Appollonio et al (1996) did report an elevated mortality rate in men aged 70-75 years with unmanaged hearing loss. However, as Harvey Dillon observes (pers.com, 2005) there may be a number of possible
explanations for this outcome:
a cascade of benefits results from effective hearing interventions, so people not receiving assistance may fare worse;
differences in attitudes and behaviours associated with health service utilisation including willingness to seek treatment for medical conditions resulting in differing health outcomes; and health professionals or patients trading off the need to treat hearing loss in the context of managing seemingly more serious conditions, focus on the other conditions.
Each of these questions warrants further investigation.
Children: The research shows that it is quite common for children to have a disability in addition to hearing loss. This factor, referred to as dual diagnosis, is a particular issue for children with hearing loss as it has profound implications for their educational placement and subsequent life chances. Fortnum et al (2002:176) report that 27.4% of children with hearing loss have at least one other disability. From a sample of 17,169 children with hearing loss, there were reports of 4,709 children having an additional disability, with 7,581 disabilities (or an average of 1.6 disabilities other than hearing loss per disabled child) reported. The most common additional disabilities were learning difficulties (11.1%) and visual impairment (5.7%) followed by a series of conditions with a prevalence of 2%-3% (developmental delays, cerebral palsy, speech and language, musculoskeletal, psychosocial and neuromotor). Additional educational and community services costs are incurred when children with hearing loss have more than one disability.
HEALTH UTILISATION AND DISABILITY2.5.2
Wilson (1997:104-108) investigated whether people with hearing loss were higher users of health and care services (Table 2-3). However, apart from the greater use of medications and, for people with severe hearing loss only, elevated utilisation of GP services, health system service utilisation was not significantly greater for people with hearing loss. However the need for help was significantly greater for all levels of hearing severity.
On average, people with hearing loss delay seeking help for their disability for six years from when they realise they are experiencing difficulties. There are two key factors that motivate a person to seek help for their hearing loss. First, their hearing problems become so unmanageable that they can no longer deny they have a hearing problem, and second, family members, tired of communication difficulties, bring pressure to bear on them to do something about their hearing problem. (Kochkin, 1999). The adverse impacts of hearing loss on inter-personal relationships have been established as have health effects associated with hearing loss. Employment impacts are discussed in Section 5.1.1. Early intervention in hearing loss may serve to avert these difficulties or minimise their impact.
3. PREVALENCE OF HEARING LOSS AND HEARINGDISABILITY 3.1 DATA SOURCES In Australia, solid data exist on the measured loss of hearing in children and adults.
For decades, the government-funded service, Australian Hearing, has been recognised as a world leading service for children with hearing loss. For this study, their data on measured hearing loss by severity for children aged up to 15 years has been used.
These data also have a high level of consistency with prevalence rates reported in international studies. Prevalence rates for congenital (pre-lingual) and child acquired hearing loss were therefore readily derived and applied to population data, with historical series drawn from the Australian Bureau of Statistics estimates and future population projections drawn from Access Economics projections, which in turn are based on ABS Series B data.
David Wilson and his then team at the Behavioural Epidemiology Unit within the South Australian Health Commission conducted a measured study of hearing loss in adults in the mid to late 1990s (Wilson, 1997; Wilson et al,1998). This study was based on the methodology of the renowned British Hearing Study (Davis, 1989). The South Australian study yielded prevalence data that were quite consistent with international
studies. Wilson (1997:93) observes:
“(T)he overall estimates for the South Australia population are largely in agreement with those of the MRC National Study of Hearing and at each level of severity the confidence intervals overlap”.
Estimates of educational, employment and socio-economic outcomes were subsequently derived from the South Australian data (Hogan et al, 1999).
The Wilson study was a representative population sample which consisted of a “multistaged, clustered, self-weighting, systematic area sample of persons aged 15 years or older who live(d) in metropolitan Adelaide and major country centres... (H)otels, motels, hospitals, nursing homes and other institutions are excluded” (Wilson,1997:63).
Random samples of people who did and did not report hearing loss subsequently underwent audiological assessment to provide measured prevalence data. The base sample size was N=9,027 with a participation rate of 75.2%. The sample had double the number of respondents (n=926) required to meet power requirements at the 95% level for detecting differences in hearing loss and almost three times the size required for data on quality of life.
Data from the Beaver Dam study (Cruickshanks et al, 1998) are not directly comparable with the Australian data since the study focused on older people, but also because of slightly different age groupings. Nonetheless, some similarities were evident, particularly among males. For measures of the worse ear, the South Australian study reported for men aged 60-69 years a prevalence of 63.8% compared with the Beaver Dam study of 61.8%. Similarly for males aged 70-79 years, the South Australian study reported a prevalence of 87.7% compared with the Beaver Dam study of 83.0%. The similarities were not so apparent in females, with the South Australian study reporting for females aged 60-69 years a prevalence of 53.8% compared with the Beaver Dam study of 28.1%. Similarly for females aged 70-79 years, the South
Australian study reported a prevalence of 63.8% compared with the Beaver Dam study of 54.6%.
Alternate data choices were available but were limited in their usefulness for this project. First, the Blue Mountains Hearing Study (Mitchell, 2002) is a respected local data source and its outcomes are comparable to the South Australia Study. However, its focus was limited to older adults. In addition, very detailed data on the study could not be accessed in the required timeframe. The second alternate data source was the Survey of Disability, Ageing and Carers (SDAC; ABS, 2003a,b). However, in comparison with the data sets selected, SDAC is particularly limited.
First, SDAC defines hearing loss as “loss of hearing where communication is restricted or an aid to assist with, or substitute for, hearing is used” (ABS, 2003b:12). This definition of hearing loss used may only capture people with more severe levels of hearing loss.
Second, the levels of reported disability in SDAC (ABS, 2003a) relate to perceived communication difficulties and/or the need for assistance rather than measured hearing loss or disability. The profound classification was defined in terms of people being unable to perform the core activity i.e. to communicate.
The severe classification was defined as having difficulty understanding family or friends or that the person communicated more easily in sign language. The moderate classification suggested that the person needed no assistance with communication. These measures do not concord well with categories used to measure hearing loss or hearing handicap and in fact distort the common clinical understanding of hearing loss. A person can readily have difficulties understanding a conversation and only have a mild or moderate hearing loss.
Third, the methodology is based on reports from the opinion of the first responsible adult contacted within the household, that a person in that household has a hearing loss. Wilson et al (1999) showed that the false positive rate in self reports of hearing loss was 46% and the false negative rate was 17%. Self reported hearing loss then is a poor indicator of prevalence. Further, with regard to reporting the hearing loss of others in the household, it is feasible that the person reporting other peoples’ experience of hearing loss may also mis-report their hearing status both by nature and severity. The SDAC data report the prevalence of childhood hearing loss for children aged less than 15 years as approximately 19,000, almost double that reported below based on internationally consistent measured studies. For adults, the reverse occurs – SDAC reports an estimated 901,000 adults aged 15 years or more with hearing loss (ABS, 2003c), compared with estimates based on measured studies of 3.5 million people (Wilson, 1997). This suggests that SDACS under estimates the population by a factor of 1:3.
In sum, SDAC may under-report the prevalence and over-report the severity of hearing loss, mistaking communication difficulties and hearing handicap with hearing impairment. In studies concerned with the cost and burden of disease, it is important that the data can be segmented by severity of impairment, as the cost of interventions differs considerably. A hearing aid intervention for an adult with mild loss, for example, may cost as little as $8087 (Commonwealth Department of Health, 2004) whereas a cochlear implant program can cost up to $45,000 in the first year (Carter and Hailey, 7 Email from Office of Hearing Services 23/11/05 cites a cost of $793.03 per client, adjust for 6 months health inflation.
1999). Consequently, the data sets selected to estimate prevalence for this study enable the segmentation of the population based on proven measured rates of severity.
General population forecasts used in this study are from the Access Economics Demographic Model (AE-DEM) of the Australian population. AE-DEM uses a combination of fertility, mortality and migration rates forecasts to project the future Australian population. Base fertility and mortality profiles for each age and gender (for mortality) are sourced from Productivity Commission (2005), and adjusted over time to match the projection for the total value. Migration rates are forecast in line with the assumptions in the ABS (forthcoming), with adjustments for changes to Australia’s migration program (that is, an additional 20,000 migrants each year). There are also some initial adjustments to reflect the latest actual migration (international and interstate) results.
The following sections estimate prevalence in children and adults, as well as providing overall estimates of the prevalence of hearing loss in Australia in 2005, projected to mid-century.
3.2 PREVALENCE IN CHILDREN Table 3-1 reports the prevalence rates of measured hearing loss in children from a range of studies in Australia and overseas.
TABLE 3-1: STUDIES OF HEARING LOSS PREVALENCE RATES IN CHILDREN
The two Australian studies (Australian Hearing, 2005; Upfold and Ipsey, 1982) report clients seen by the country’s national service for children with hearing loss. The service is thought to cover 99% of the sector and the data cover children who use any form of hearing device. The data would not include children with such a mild loss that they did not need a device. The original Australian Hearing data report rates for children aged to 17 years. The grouped rates reported here were re-calculated from the Australian Hearing data for children aged less than 15 years. These data suggest a prevalence of pre-lingual (0-4 years) hearing loss of 1.2/1,000 live births and of child acquired loss (4-14 years) as 3.2/1,000 live births.
The American studies (Yoshinaga-Itano et al, 2000 and Mehl and Thomson, 2002) are limited in that they only report data on neonates and as such do not include children with later onset hearing loss. The United Kingdom study (Fortnum et al, 2001) reports a prevalence rate of 2.05/1,000 once adjusted for under-ascertainment. However, the study does not include children with mild hearing loss. Australian Hearing reports 36% of children with hearing loss in the mild hearing loss range of 0-30dB.