Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

The geographic distribution of private health insurance in Australia in 2001

The geographic distribution of private health insurance in Australia in 2001 Background: Private health insurance has been a major focus of Commonwealth Government health policy for the last decade. Over this period, the Howard government introduced a number of policy changes which impacted on the take up of private health insurance. The most expensive of these was the introduction of the private health insurance rebate in 1997, which had an estimated cost of $3 billion per annum. Methods: This article uses information on the geographic distribution of the population with private health insurance cover to identify associations between rates of private health insurance cover and socioeconomic status. The geographic analysis is repeated with survey data on expenditure on private health insurance, to provide an estimate of the rebate flowing to different socioeconomic groups. Results: The analysis highlights the strong association between high rates of private health insurance cover and high socioeconomic status and shows the substantial transfer of funds, under the private health insurance rebate, to those living in areas of highest socioeconomic status, compared with those in areas of lower socioeconomic status, and in particular those in the most disadvantaged areas. The article also provides estimates of private health insurance cover by federal electorate, emphasising the substantial gaps in cover between Liberal Party and Australian Labor Party seats. Conclusion: The article concludes by discussing implications of the uneven distribution of private health insurance cover across Australia for policy formation. In particular, the study shows that the prevalence of private health insurance is unevenly distributed across Australia, with marked differences in prevalence in rural and urban areas, and substantial differences by socioeconomic status. Policy formation needs to take this into account. Evaluating the potential impact of changes in private health insurance requires more nuanced consideration than has been implied in the rhetoric about private health insurance over the last decade. Page 1 of 6 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:19 http://www.anzhealthpolicy.com/content/6/1/19 provide a wealth of information on trends in prevalence Background Private health insurance has been a major focus of Com- and patterns of benefits, and data from the irregular (approximately three-yearly) health surveys conducted by monwealth Government health policy for the last decade. Over the first five years of the Howard government's term the Australian Bureau of Statistics (ABS). In terms of ABS in office (a coalition of the conservative Liberal and surveys, the long gap between surveys militates against National Parties, in office from 1996 to 2007), the Com- their utility in terms of tracking changes in health insur- monwealth Government introduced a new policy to ance. increase the prevalence of private health insurance every 18 months. The government's focus on private health This article's contribution to health service research is the insurance was probably stimulated by the sustained use of a unique data source on the geographic distribution decline in the most commonly used measure of private of private health insurance. health insurance – the proportion of the population cov- ered for private accommodation in public hospitals – that Methods had fallen from around 50% to around 30% over the pre- An unusual source of data on health insurance was infor- vious 10 years [1]. This decline in the headline rate con- mation contained in an answer to a question on notice trasts with the more variable pattern of coverage for asked by Senator Jan McLucas in the Senate supplemen- accommodation in private hospitals that increased in the tary budget estimates hearings for 2002–2003, for the late 1980s to a peak of 39% in 1990 followed by a decline Health and Ageing portfolio (reported on 21 November to around 33% in 1995 when the series was discontinued 2002) [6]. Senator McLucas asked for information on the [2]. take-up of private health insurance by postcode and fed- eral electorate in Australia. This is the lowest geographic The rhetoric of the Howard government, though, was not level of prevalence information that has been released cast in terms of private health insurance per se but rather about private health insurance. This level of release has in terms of increasing the prevalence of private health not been replicated since then. insurance to reduce demand on public hospitals; shifting the load from public hospitals to private hospitals. The The area-based analysis in this article was undertaken most significant of the policy changes in terms of take-up using data from the 30% rebate registration database (an of private health insurance was the introduction of life- administrative collection) in 2001. Due to the once-only time cover taking effect in 1999 [3]. This policy, which registration process – whereby once an individual registers encouraged take-up of insurance at or before age 30, led for the 30% rebate (and details of their postcode of resi- to an increase in health insurance prevalence, with the dence are available) they remain registered – the collec- increase being principally among people who took out tion's suitability for statistical use declines over time. policies with front-end deductibles – policies that Although data for later periods are not available, it is required them to pay the 'front end' of costs (e.g. the first likely that the socioeconomic patterns described in this $500, or $1,000), with insurance covering the tail. report are currently at least as strong, if not stronger, than existed in 2001. For example, insurance cover in 2008 was The most expensive policy was the introduction of the pri- at the same level as in 2001 (from 44.8% in 2001, cover vate health insurance rebate in 1997, which had an esti- declined to 43.0% in 2005 and then returned to the 2001 mated cost of $3 billion per annum [4]. Under this policy, level in 2008) [5]. Further, movement within these overall all Australians eligible for Medicare and covered by a levels is most likely to be in higher cover for people in health insurance policy including inpatient treatment areas of higher socioeconomic status, increasing the dif- offered by a registered health fund are eligible for a rebate ferentials reported below. of 30% of the actual cost of premiums. This policy had rel- atively little impact on private health insurance preva- The two questions asked by Senator McLucas were as fol- lence although the policy itself may have been a necessary lows: political precondition for introduction of the life time cover policy. 1. How many private health insurance contributors, and what proportion of the total, receive the 30% Given the large expenditures and significant proportion of rebate through each of the schemes available for the population affected, it is surprising that private health claiming the rebate? insurance policy is a relative data-free zone, especially in terms of data in the public domain. There are two main 2. With respect to those persons who hold private public data sources on private health insurance: the data health insurance which is eligible for the 30% private published by the Private Health Insurance Administration health insurance rebate and who receive the benefit of Council (PHIAC, http://www.phiac.gov.au) [5], which the rebate through premium reductions: Page 2 of 6 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:19 http://www.anzhealthpolicy.com/content/6/1/19 a) How many persons are covered by private The ABS data included ambulance insurance (where it was health insurance by postcode and by federal elec- separate insurance) and sickness and personal accident torate division? insurance. Together, these items represented 13% of the total expenditure (varying from 7% of expenditure in the b) How many contributor units hold private lowest socioeconomic status areas to 16% in the highest health insurance by postcode and by federal elec- socioeconomic status areas). In addition, the Household torate division? Expenditure Survey excludes households in collection dis- tricts defined as Very Remote, or Indigenous communi- The three schemes referred to in the first question are the ties. The impact of this exclusion is most noticeable in the Premium Reduction Scheme (91.2% of the rebate paid Northern Territory, where the exclusions account for out through this scheme), the Incentive Payment Scheme about 23% of the population. (0.2%) and the Tax Offset (8.6%). Results Data at the postcode level were analysed to illustrate the There were 8,671,106 people who claimed the 30% pri- characteristics of the population covered by private health vate health insurance rebate in Australia at 30 June 2001 insurance. Data were initially allocated to either the capi- (46.1% of the Australian population). Of these, tal city or rest of state/territory for each jurisdiction, based 6,468,996 were residents of capital cities and other major on the postcode of the insured (the contributor unit). The urban centres (74.6%) and 2,202,110 were residents of 'Other Major Centres' – urban centre of 100,000 or more the rest of state/territory areas (25.4%). population at the 2001 Census – were included with the capital city in the same jurisdiction. These were Newcastle Table 1 shows the prevalence of private health insurance and Wollongong in New South Wales; Gold Coast, by socioeconomic status of the postcode of the insured. Townsville-Thuringowa and the Sunshine Coast in Confirming previous studies of prevalence of health Queensland; and Geelong, in Victoria. insurance in Australia, there is a statistically significant [7] socioeconomic gradient for prevalence, with postcodes in Within each of these two major groupings, postcodes were the highest socioeconomic status decile having, on aver- then sorted by socioeconomic status, using the ABS Index age, almost 70% of residents covered by health insurance of Relative Socio-Economic Disadvantage (IRSD), a sum- compared to residents of the most disadvantaged decile, mary measure of socioeconomic status derived from the with a take-up of less than 30%. Not surprisingly, this 2001 Census. Postcodes were ranked by their IRSD score average figure is also confirmed for postcodes falling then grouped into ten groups (deciles), each of approxi- within capital cities and major urban centres, with again mately 10% of the population. The proportion of the pop- about 70% of residents in the wealthiest urban postcodes ulation with private health insurance was then calculated having health insurance compared to fewer than 30% in for each decile. The postcode data were also converted to the most disadvantaged postcodes. Statistical Local Area, to allow a correlation analysis (Pear- son Product Moment Correlation) to be undertaken Table 1 also shows the results for non urban centres, against the IRSD. including regional cities (of less than 100,000 popula- tion) and rural areas. Here we see a quite different pattern At the electorate level, percentage coverage was calculated of coverage. Again, fewer than 30% of residents in the for each federal electorate using electorate populations most disadvantaged decile are covered by private health from the 2001 Census. Each electorate was also allocated insurance but, in contrast to the over 70% prevalence in the political party of the elected member to allow aggre- capital cities and other major urban centres, the coverage gate proportions to be calculated by political party. of private health insurance in the top decile in the rest of the state is much lower, being less than 50%. Although An additional data source was the ABS Household smaller, the difference between these figures is still statis- Expenditure Survey 2003–04. Data were purchased on tically significant. expenditure on private health insurance in five groups (quintiles, each including approximately 20% of the pop- There is also strong correlation at the small area level ulation, based on the IRSD of the Collection District of between the distribution of the population with private the contributor unit's address). The proportional distribu- health insurance cover and socioeconomic disadvantage, tion of household expenditure on private health insur- as measured by the IRSD; a correlation coefficient of 0.60. ance in each quintile was applied to the nominal amount of $3 billion (estimated total cost of the private health There is a similar distinction in terms of coverage analysed insurance rebate) to provide an estimate of rebate funds by party affiliation (Figure 1). Within the overall rate of flowing to households in each quintile. 46.1% of the population covered by private health insur- Page 3 of 6 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:19 http://www.anzhealthpolicy.com/content/6/1/19 Table 1: Private health insurance cover and estimated rebate payments for residents of capital cities and rest of State/Territory , by socioeconomic status, June 2001 Decile Estimated population with private health Quintile Estimated rebate ($m) received by insurance cover in people in Capital cities Rest of State Aust Capital cities Rest State Aust Highest SES 70.8 45.9 68.7 Highest SES 679 73 749 areas areas 262.8 44.6 58.2 3 55.5 45.9 52.0 2 539 163 701 452.1 44.8 47.2 5 48.5 40.6 44.5 3 373 242 615 644.0 41.9 43.7 7 42.3 41.6 41.0 4 231 291 525 840.7 37.0 40.3 9 36.3 38.0 36.4 Lowest SES areas 177 231 410 Lowest SES areas 28.5 27.3 28.1 Total 48.1 40.8 46.0 2000 1000 3000 Rate ratio 2.48 1.68 2.45 3.84 0.32 1.83 Lower 95% 2.47 1.68 2.44 3.35 0.25 1.55 C.I. Upper 95% 2.49 1.69 2.46 4.40 0.40 1.88 C.I. Based on postcode of address of contributor Estimate based on a total rebate amount of $3 billion: allocation to SES areas based on expenditure on private health insurance (incl. accident insurance), by quintiles of socioeconomic status of area, using the ABS Index of Relative Socio-Economic Disadvantage, 2001 Rate ratio is the ratio of value in Highest SES areas to value in Lowest SES areas Calculation of 95% confidence intervals (C.I.) based on the comparison of proportions (Berry & Simpson, 1998) Source: Private health insurance estimates based on data provided by the Senate Community Affairs Legislation Committee (see references). Estimated rebate based on expenditure data purchased from ABS from the Household Expenditure Survey 2003–04 ance with hospital cover, seats held by the Liberal Party [42.93: ± 0.20] across the three electorates held by inde- had an above average coverage of health insurance of 50% pendents. These differences are statistically significant. [CI: 50.23 ± 0.05] and those held by the other parties and independents had below-average rates: 43% [42.90: ± The range of coverage is from an estimated 23% [22.51: ± 0.04] for seats held by the Australian Labor Party; 42% 0.28] in the Labor-held seat of Lingiari, in the Northern [41.71: ± 0.1] for the National Party; 42% [42.12: ± 0.41] Territory to over three and a half times (3.6) higher at 82% for the Country Liberal Party; and an average of 43% [81.89: ± 0.50] in the Liberal-held seat of Bradfield, on Sydney's north shore. There is a notable gap of 20.1 per- centage points between the Liberal- and Labor-held seats Per cent 100 with the highest rates of private health insurance. That is, LP NP/ CLP an estimated 81.9% of the population in the Liberal-held ALP seat of Bradfield were insured in 2001, compared with 70 IND some 24.6% fewer in (the Labor-held seat of) Jagajaga, Average with a rate of 61.8%. 30 Estimates of the allocation of the rebate by socioeconomic status, shown in Table 1), were limited to quintiles (for which the data were available from ABS); these have been aligned with the equivalent deciles. These estimates reflect Electorates, ranked by percentage health insurance cover the marked and statistically significant differences seen in coverage rates for the capital cities and for Australia as a Private health insur June 2001 Figure 1 ance by federal electorate, Australia, 30 whole; differences related to socioeconomic status. Private health insurance by federal electorate, Aus- tralia, 30 June 2001. Source: Compiled from data provided Notably, for capital cities, the estimated per capita rebate by the Senate Community Affairs Legislation Committee [5]. paid to those living in the highest socioeconomic status Page 4 of 6 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:19 http://www.anzhealthpolicy.com/content/6/1/19 areas is nearly four times that paid to those in the lowest deductibles versus products without front-end deducti- socioeconomic status areas (a statistically significant rate bles. An individual taking out a policy with front-end ratio of 3.84). This represents a substantial transfer of deductibles may not intend to use private health insur- funds to the most well-off, and is a substantially wider gap ance as part of a potential hospitalisation, and so again than exists for private health insurance cover, of 2.48. The these data cannot be used to infer information about the difference in these two rate ratios is likely to reflect the demand for private hospital accommodation. larger sums paid for cover, with fewer products purchased with high levels of front-end deductibles, and more prod- Subject to this limitation, there are two clear policy impli- ucts without front-end deductibles, by those in the highest cations of these data. First, although the Labor Prime Min- socioeconomic status areas. Also of interest is the strong, ister Rudd (elected in 2007) has made it clear that support continuous gradient evident across the socioeconomic for the private sector, such as through the rebate, is here to groups, with estimate rebate payments decreasing with stay, policy-makers should not think of private health each increase in socioeconomic disadvantage (Figure 2). insurance policies in a homogenous way. The marked dif- ferences between prevalence in rural and urban areas, and For the non-urban areas, however, the reverse applies, the substantial differences by socioeconomic status, sug- with estimated rebate payments increasing with increas- gest that there is a need for much more nuanced consider- ing disadvantage then declining in the most disadvantage ation of the implications of private health insurance areas. The reasons for this are not clear. One factor con- prevalence. tributing to the low estimate in the highest socioeconomic status areas may be the way in which the quintiles are con- The second policy implication is that private health insur- structed. The areas in the highest socioeconomic status ance has a significant potential to influence the political quintile tend to be the towns and other heavily populated culture of wealthy (urban) electorates. This might explain areas on the fringes of the capital cities, and the insured the very high importance the Howard Liberal government population in these areas may be more likely to purchase accorded private health insurance policy, as it was of products with high levels of front-end deductibles, thus much higher salience in Liberal-voting electorates. reducing the rebate they receive. The implications for policy of the concentration of the insured population in wealthy electorates are difficult to Conclusion Data presented in the answer to the question on notice disentangle. To some extent there may be some circularity did not distinguish the type of private health insurance, here. Private hospitals make a business decision to locate and so, for example, no information is available about the private hospitals in areas where there is a greater market take-up of health insurance with high levels of front-end for private hospital accommodation. In turn, the market for private hospital accommodation would be driven in part by where people have private health insurance, and, Estimated rebate ($m) of course, private hospital insurance take-up is more likely RR= 3.84 if private hospitals are available locally. Policy about pri- vate health insurance thus has greater significance for wealthy (urban) areas across Australia and shapes the uti- lisation of hospital services of the population in these areas. Policies on private health insurance are less impor- tant in poorer (and, to a lesser extent, rural) areas across Australia. This study has shown that the prevalence of private health insurance is unevenly distributed across Australia. Policy formation needs to take this into account. Evaluating the Highest SES Lowest SES Q1 Q2 Q3 Q4 Q5 potential impact of changes in private health insurance on Socioeconomic status public hospitals, or, indeed, given the high correlation between private hospital insurance and ancillary (or gen- eral) insurance, developing or evaluating policies on E insurance cover, Figure 2 stimated rebate payments for capital cities, Austr people wit alia, 30 June 2001 h private health access to allied health and dental services requires more Estimated rebate payments for people with private nuanced consideration than previously implied in the health insurance cover, capital cities, Australia, 30 June 2001. rhetoric about private health insurance over the last dec- ade. Page 5 of 6 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:19 http://www.anzhealthpolicy.com/content/6/1/19 Competing interests The authors declare that they have no competing interests. Authors' contributions JG obtained the private health insurance data and devel- oped the concept of presenting it to show associations with the socioeconomic status of the population, and to which groups the rebate flows. ST undertook the data analysis. SD wrote sections of the paper dealing with the historical and policy context. References 1. Private Health Insurance Administration Council: Coverage of Hospital Treatment Tables, Offered by Health Benefits Funds by State, Insured Per- sons and Percentage of Population [http://www.phiac.gov.au/for-indus try/industry-statistics/membership-statistics/table5ye/]. 2. Duckett SJ: The Australian Health Care System Melbourne: Oxford Uni- versity Press; 2007. 3. Butler JRG: Policy change and private health insurance: Did the cheapest policy do the trick? Aust Health Rev 2002, 25(6):33-41. 4. Australian Institute of Health and Welfare: Health expenditure Australia 2006–07 Health and welfare expenditure series no. 35 [http:// www.aihw.gov.au/publications/index.cfm/title/10659]. 5. Private Health Insurance Council: Annual Coverage Survey [http:// www.phiac.gov.au/for-industry/industry-statistics/annualsurvey/]. 6. Senate Community Affairs Legislation Committee: Consideration of Supplementary Estimates 2002 [http://www.aph.gov.au/hansard/senate/ commttee/s6004.pdf]. 7. Berry G, Simpson J: Comparison of two independent propor- tions. In Handbook of Public Health Methods Edited by: Kerr C, Taylor R, Heard G. Sydney: McGraw-Hill; 1998. Publish with Bio Med Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical researc h in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright BioMedcentral Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp Page 6 of 6 (page number not for citation purposes) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Australia and New Zealand Health Policy Springer Journals

The geographic distribution of private health insurance in Australia in 2001

Loading next page...
 
/lp/springer-journals/the-geographic-distribution-of-private-health-insurance-in-australia-56vvNGm3tx
Publisher
Springer Journals
Copyright
Copyright © 2009 by Glover et al; licensee BioMed Central Ltd.
Subject
Medicine & Public Health; Public Health; Social Policy
ISSN
1743-8462
eISSN
1743-8462
DOI
10.1186/1743-8462-6-19
pmid
19686590
Publisher site
See Article on Publisher Site

Abstract

Background: Private health insurance has been a major focus of Commonwealth Government health policy for the last decade. Over this period, the Howard government introduced a number of policy changes which impacted on the take up of private health insurance. The most expensive of these was the introduction of the private health insurance rebate in 1997, which had an estimated cost of $3 billion per annum. Methods: This article uses information on the geographic distribution of the population with private health insurance cover to identify associations between rates of private health insurance cover and socioeconomic status. The geographic analysis is repeated with survey data on expenditure on private health insurance, to provide an estimate of the rebate flowing to different socioeconomic groups. Results: The analysis highlights the strong association between high rates of private health insurance cover and high socioeconomic status and shows the substantial transfer of funds, under the private health insurance rebate, to those living in areas of highest socioeconomic status, compared with those in areas of lower socioeconomic status, and in particular those in the most disadvantaged areas. The article also provides estimates of private health insurance cover by federal electorate, emphasising the substantial gaps in cover between Liberal Party and Australian Labor Party seats. Conclusion: The article concludes by discussing implications of the uneven distribution of private health insurance cover across Australia for policy formation. In particular, the study shows that the prevalence of private health insurance is unevenly distributed across Australia, with marked differences in prevalence in rural and urban areas, and substantial differences by socioeconomic status. Policy formation needs to take this into account. Evaluating the potential impact of changes in private health insurance requires more nuanced consideration than has been implied in the rhetoric about private health insurance over the last decade. Page 1 of 6 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:19 http://www.anzhealthpolicy.com/content/6/1/19 provide a wealth of information on trends in prevalence Background Private health insurance has been a major focus of Com- and patterns of benefits, and data from the irregular (approximately three-yearly) health surveys conducted by monwealth Government health policy for the last decade. Over the first five years of the Howard government's term the Australian Bureau of Statistics (ABS). In terms of ABS in office (a coalition of the conservative Liberal and surveys, the long gap between surveys militates against National Parties, in office from 1996 to 2007), the Com- their utility in terms of tracking changes in health insur- monwealth Government introduced a new policy to ance. increase the prevalence of private health insurance every 18 months. The government's focus on private health This article's contribution to health service research is the insurance was probably stimulated by the sustained use of a unique data source on the geographic distribution decline in the most commonly used measure of private of private health insurance. health insurance – the proportion of the population cov- ered for private accommodation in public hospitals – that Methods had fallen from around 50% to around 30% over the pre- An unusual source of data on health insurance was infor- vious 10 years [1]. This decline in the headline rate con- mation contained in an answer to a question on notice trasts with the more variable pattern of coverage for asked by Senator Jan McLucas in the Senate supplemen- accommodation in private hospitals that increased in the tary budget estimates hearings for 2002–2003, for the late 1980s to a peak of 39% in 1990 followed by a decline Health and Ageing portfolio (reported on 21 November to around 33% in 1995 when the series was discontinued 2002) [6]. Senator McLucas asked for information on the [2]. take-up of private health insurance by postcode and fed- eral electorate in Australia. This is the lowest geographic The rhetoric of the Howard government, though, was not level of prevalence information that has been released cast in terms of private health insurance per se but rather about private health insurance. This level of release has in terms of increasing the prevalence of private health not been replicated since then. insurance to reduce demand on public hospitals; shifting the load from public hospitals to private hospitals. The The area-based analysis in this article was undertaken most significant of the policy changes in terms of take-up using data from the 30% rebate registration database (an of private health insurance was the introduction of life- administrative collection) in 2001. Due to the once-only time cover taking effect in 1999 [3]. This policy, which registration process – whereby once an individual registers encouraged take-up of insurance at or before age 30, led for the 30% rebate (and details of their postcode of resi- to an increase in health insurance prevalence, with the dence are available) they remain registered – the collec- increase being principally among people who took out tion's suitability for statistical use declines over time. policies with front-end deductibles – policies that Although data for later periods are not available, it is required them to pay the 'front end' of costs (e.g. the first likely that the socioeconomic patterns described in this $500, or $1,000), with insurance covering the tail. report are currently at least as strong, if not stronger, than existed in 2001. For example, insurance cover in 2008 was The most expensive policy was the introduction of the pri- at the same level as in 2001 (from 44.8% in 2001, cover vate health insurance rebate in 1997, which had an esti- declined to 43.0% in 2005 and then returned to the 2001 mated cost of $3 billion per annum [4]. Under this policy, level in 2008) [5]. Further, movement within these overall all Australians eligible for Medicare and covered by a levels is most likely to be in higher cover for people in health insurance policy including inpatient treatment areas of higher socioeconomic status, increasing the dif- offered by a registered health fund are eligible for a rebate ferentials reported below. of 30% of the actual cost of premiums. This policy had rel- atively little impact on private health insurance preva- The two questions asked by Senator McLucas were as fol- lence although the policy itself may have been a necessary lows: political precondition for introduction of the life time cover policy. 1. How many private health insurance contributors, and what proportion of the total, receive the 30% Given the large expenditures and significant proportion of rebate through each of the schemes available for the population affected, it is surprising that private health claiming the rebate? insurance policy is a relative data-free zone, especially in terms of data in the public domain. There are two main 2. With respect to those persons who hold private public data sources on private health insurance: the data health insurance which is eligible for the 30% private published by the Private Health Insurance Administration health insurance rebate and who receive the benefit of Council (PHIAC, http://www.phiac.gov.au) [5], which the rebate through premium reductions: Page 2 of 6 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:19 http://www.anzhealthpolicy.com/content/6/1/19 a) How many persons are covered by private The ABS data included ambulance insurance (where it was health insurance by postcode and by federal elec- separate insurance) and sickness and personal accident torate division? insurance. Together, these items represented 13% of the total expenditure (varying from 7% of expenditure in the b) How many contributor units hold private lowest socioeconomic status areas to 16% in the highest health insurance by postcode and by federal elec- socioeconomic status areas). In addition, the Household torate division? Expenditure Survey excludes households in collection dis- tricts defined as Very Remote, or Indigenous communi- The three schemes referred to in the first question are the ties. The impact of this exclusion is most noticeable in the Premium Reduction Scheme (91.2% of the rebate paid Northern Territory, where the exclusions account for out through this scheme), the Incentive Payment Scheme about 23% of the population. (0.2%) and the Tax Offset (8.6%). Results Data at the postcode level were analysed to illustrate the There were 8,671,106 people who claimed the 30% pri- characteristics of the population covered by private health vate health insurance rebate in Australia at 30 June 2001 insurance. Data were initially allocated to either the capi- (46.1% of the Australian population). Of these, tal city or rest of state/territory for each jurisdiction, based 6,468,996 were residents of capital cities and other major on the postcode of the insured (the contributor unit). The urban centres (74.6%) and 2,202,110 were residents of 'Other Major Centres' – urban centre of 100,000 or more the rest of state/territory areas (25.4%). population at the 2001 Census – were included with the capital city in the same jurisdiction. These were Newcastle Table 1 shows the prevalence of private health insurance and Wollongong in New South Wales; Gold Coast, by socioeconomic status of the postcode of the insured. Townsville-Thuringowa and the Sunshine Coast in Confirming previous studies of prevalence of health Queensland; and Geelong, in Victoria. insurance in Australia, there is a statistically significant [7] socioeconomic gradient for prevalence, with postcodes in Within each of these two major groupings, postcodes were the highest socioeconomic status decile having, on aver- then sorted by socioeconomic status, using the ABS Index age, almost 70% of residents covered by health insurance of Relative Socio-Economic Disadvantage (IRSD), a sum- compared to residents of the most disadvantaged decile, mary measure of socioeconomic status derived from the with a take-up of less than 30%. Not surprisingly, this 2001 Census. Postcodes were ranked by their IRSD score average figure is also confirmed for postcodes falling then grouped into ten groups (deciles), each of approxi- within capital cities and major urban centres, with again mately 10% of the population. The proportion of the pop- about 70% of residents in the wealthiest urban postcodes ulation with private health insurance was then calculated having health insurance compared to fewer than 30% in for each decile. The postcode data were also converted to the most disadvantaged postcodes. Statistical Local Area, to allow a correlation analysis (Pear- son Product Moment Correlation) to be undertaken Table 1 also shows the results for non urban centres, against the IRSD. including regional cities (of less than 100,000 popula- tion) and rural areas. Here we see a quite different pattern At the electorate level, percentage coverage was calculated of coverage. Again, fewer than 30% of residents in the for each federal electorate using electorate populations most disadvantaged decile are covered by private health from the 2001 Census. Each electorate was also allocated insurance but, in contrast to the over 70% prevalence in the political party of the elected member to allow aggre- capital cities and other major urban centres, the coverage gate proportions to be calculated by political party. of private health insurance in the top decile in the rest of the state is much lower, being less than 50%. Although An additional data source was the ABS Household smaller, the difference between these figures is still statis- Expenditure Survey 2003–04. Data were purchased on tically significant. expenditure on private health insurance in five groups (quintiles, each including approximately 20% of the pop- There is also strong correlation at the small area level ulation, based on the IRSD of the Collection District of between the distribution of the population with private the contributor unit's address). The proportional distribu- health insurance cover and socioeconomic disadvantage, tion of household expenditure on private health insur- as measured by the IRSD; a correlation coefficient of 0.60. ance in each quintile was applied to the nominal amount of $3 billion (estimated total cost of the private health There is a similar distinction in terms of coverage analysed insurance rebate) to provide an estimate of rebate funds by party affiliation (Figure 1). Within the overall rate of flowing to households in each quintile. 46.1% of the population covered by private health insur- Page 3 of 6 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:19 http://www.anzhealthpolicy.com/content/6/1/19 Table 1: Private health insurance cover and estimated rebate payments for residents of capital cities and rest of State/Territory , by socioeconomic status, June 2001 Decile Estimated population with private health Quintile Estimated rebate ($m) received by insurance cover in people in Capital cities Rest of State Aust Capital cities Rest State Aust Highest SES 70.8 45.9 68.7 Highest SES 679 73 749 areas areas 262.8 44.6 58.2 3 55.5 45.9 52.0 2 539 163 701 452.1 44.8 47.2 5 48.5 40.6 44.5 3 373 242 615 644.0 41.9 43.7 7 42.3 41.6 41.0 4 231 291 525 840.7 37.0 40.3 9 36.3 38.0 36.4 Lowest SES areas 177 231 410 Lowest SES areas 28.5 27.3 28.1 Total 48.1 40.8 46.0 2000 1000 3000 Rate ratio 2.48 1.68 2.45 3.84 0.32 1.83 Lower 95% 2.47 1.68 2.44 3.35 0.25 1.55 C.I. Upper 95% 2.49 1.69 2.46 4.40 0.40 1.88 C.I. Based on postcode of address of contributor Estimate based on a total rebate amount of $3 billion: allocation to SES areas based on expenditure on private health insurance (incl. accident insurance), by quintiles of socioeconomic status of area, using the ABS Index of Relative Socio-Economic Disadvantage, 2001 Rate ratio is the ratio of value in Highest SES areas to value in Lowest SES areas Calculation of 95% confidence intervals (C.I.) based on the comparison of proportions (Berry & Simpson, 1998) Source: Private health insurance estimates based on data provided by the Senate Community Affairs Legislation Committee (see references). Estimated rebate based on expenditure data purchased from ABS from the Household Expenditure Survey 2003–04 ance with hospital cover, seats held by the Liberal Party [42.93: ± 0.20] across the three electorates held by inde- had an above average coverage of health insurance of 50% pendents. These differences are statistically significant. [CI: 50.23 ± 0.05] and those held by the other parties and independents had below-average rates: 43% [42.90: ± The range of coverage is from an estimated 23% [22.51: ± 0.04] for seats held by the Australian Labor Party; 42% 0.28] in the Labor-held seat of Lingiari, in the Northern [41.71: ± 0.1] for the National Party; 42% [42.12: ± 0.41] Territory to over three and a half times (3.6) higher at 82% for the Country Liberal Party; and an average of 43% [81.89: ± 0.50] in the Liberal-held seat of Bradfield, on Sydney's north shore. There is a notable gap of 20.1 per- centage points between the Liberal- and Labor-held seats Per cent 100 with the highest rates of private health insurance. That is, LP NP/ CLP an estimated 81.9% of the population in the Liberal-held ALP seat of Bradfield were insured in 2001, compared with 70 IND some 24.6% fewer in (the Labor-held seat of) Jagajaga, Average with a rate of 61.8%. 30 Estimates of the allocation of the rebate by socioeconomic status, shown in Table 1), were limited to quintiles (for which the data were available from ABS); these have been aligned with the equivalent deciles. These estimates reflect Electorates, ranked by percentage health insurance cover the marked and statistically significant differences seen in coverage rates for the capital cities and for Australia as a Private health insur June 2001 Figure 1 ance by federal electorate, Australia, 30 whole; differences related to socioeconomic status. Private health insurance by federal electorate, Aus- tralia, 30 June 2001. Source: Compiled from data provided Notably, for capital cities, the estimated per capita rebate by the Senate Community Affairs Legislation Committee [5]. paid to those living in the highest socioeconomic status Page 4 of 6 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:19 http://www.anzhealthpolicy.com/content/6/1/19 areas is nearly four times that paid to those in the lowest deductibles versus products without front-end deducti- socioeconomic status areas (a statistically significant rate bles. An individual taking out a policy with front-end ratio of 3.84). This represents a substantial transfer of deductibles may not intend to use private health insur- funds to the most well-off, and is a substantially wider gap ance as part of a potential hospitalisation, and so again than exists for private health insurance cover, of 2.48. The these data cannot be used to infer information about the difference in these two rate ratios is likely to reflect the demand for private hospital accommodation. larger sums paid for cover, with fewer products purchased with high levels of front-end deductibles, and more prod- Subject to this limitation, there are two clear policy impli- ucts without front-end deductibles, by those in the highest cations of these data. First, although the Labor Prime Min- socioeconomic status areas. Also of interest is the strong, ister Rudd (elected in 2007) has made it clear that support continuous gradient evident across the socioeconomic for the private sector, such as through the rebate, is here to groups, with estimate rebate payments decreasing with stay, policy-makers should not think of private health each increase in socioeconomic disadvantage (Figure 2). insurance policies in a homogenous way. The marked dif- ferences between prevalence in rural and urban areas, and For the non-urban areas, however, the reverse applies, the substantial differences by socioeconomic status, sug- with estimated rebate payments increasing with increas- gest that there is a need for much more nuanced consider- ing disadvantage then declining in the most disadvantage ation of the implications of private health insurance areas. The reasons for this are not clear. One factor con- prevalence. tributing to the low estimate in the highest socioeconomic status areas may be the way in which the quintiles are con- The second policy implication is that private health insur- structed. The areas in the highest socioeconomic status ance has a significant potential to influence the political quintile tend to be the towns and other heavily populated culture of wealthy (urban) electorates. This might explain areas on the fringes of the capital cities, and the insured the very high importance the Howard Liberal government population in these areas may be more likely to purchase accorded private health insurance policy, as it was of products with high levels of front-end deductibles, thus much higher salience in Liberal-voting electorates. reducing the rebate they receive. The implications for policy of the concentration of the insured population in wealthy electorates are difficult to Conclusion Data presented in the answer to the question on notice disentangle. To some extent there may be some circularity did not distinguish the type of private health insurance, here. Private hospitals make a business decision to locate and so, for example, no information is available about the private hospitals in areas where there is a greater market take-up of health insurance with high levels of front-end for private hospital accommodation. In turn, the market for private hospital accommodation would be driven in part by where people have private health insurance, and, Estimated rebate ($m) of course, private hospital insurance take-up is more likely RR= 3.84 if private hospitals are available locally. Policy about pri- vate health insurance thus has greater significance for wealthy (urban) areas across Australia and shapes the uti- lisation of hospital services of the population in these areas. Policies on private health insurance are less impor- tant in poorer (and, to a lesser extent, rural) areas across Australia. This study has shown that the prevalence of private health insurance is unevenly distributed across Australia. Policy formation needs to take this into account. Evaluating the Highest SES Lowest SES Q1 Q2 Q3 Q4 Q5 potential impact of changes in private health insurance on Socioeconomic status public hospitals, or, indeed, given the high correlation between private hospital insurance and ancillary (or gen- eral) insurance, developing or evaluating policies on E insurance cover, Figure 2 stimated rebate payments for capital cities, Austr people wit alia, 30 June 2001 h private health access to allied health and dental services requires more Estimated rebate payments for people with private nuanced consideration than previously implied in the health insurance cover, capital cities, Australia, 30 June 2001. rhetoric about private health insurance over the last dec- ade. Page 5 of 6 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:19 http://www.anzhealthpolicy.com/content/6/1/19 Competing interests The authors declare that they have no competing interests. Authors' contributions JG obtained the private health insurance data and devel- oped the concept of presenting it to show associations with the socioeconomic status of the population, and to which groups the rebate flows. ST undertook the data analysis. SD wrote sections of the paper dealing with the historical and policy context. References 1. Private Health Insurance Administration Council: Coverage of Hospital Treatment Tables, Offered by Health Benefits Funds by State, Insured Per- sons and Percentage of Population [http://www.phiac.gov.au/for-indus try/industry-statistics/membership-statistics/table5ye/]. 2. Duckett SJ: The Australian Health Care System Melbourne: Oxford Uni- versity Press; 2007. 3. Butler JRG: Policy change and private health insurance: Did the cheapest policy do the trick? Aust Health Rev 2002, 25(6):33-41. 4. Australian Institute of Health and Welfare: Health expenditure Australia 2006–07 Health and welfare expenditure series no. 35 [http:// www.aihw.gov.au/publications/index.cfm/title/10659]. 5. Private Health Insurance Council: Annual Coverage Survey [http:// www.phiac.gov.au/for-industry/industry-statistics/annualsurvey/]. 6. Senate Community Affairs Legislation Committee: Consideration of Supplementary Estimates 2002 [http://www.aph.gov.au/hansard/senate/ commttee/s6004.pdf]. 7. Berry G, Simpson J: Comparison of two independent propor- tions. In Handbook of Public Health Methods Edited by: Kerr C, Taylor R, Heard G. Sydney: McGraw-Hill; 1998. Publish with Bio Med Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical researc h in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright BioMedcentral Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp Page 6 of 6 (page number not for citation purposes)

Journal

Australia and New Zealand Health PolicySpringer Journals

Published: Aug 17, 2009

References