Previous PageTable Of ContentsNext Page

4 Public Health Programs to Reduce HIV/AIDS

4.1. Introduction

Following the outbreak of human immunodeficiency virus (HIV) infection in human populations in the early 1980s, many countries including Australia experienced a rapid rise in incidence of the disease for several years. In the mid-1980s, however, a turning point occurred and HIV incidence in Australia began to fall. Accordingly, the number of deaths attributable to the syndrome caused by the virus – acquired immunodeficiency syndrome (AIDS) – has not been particularly large in Australia. The cumulative total number of deaths caused by AIDS up to the end of 1999 was 5850 (National Centre in HIV Epidemiology and Clinical Research 2000, Table 1.1.5).

Figure 4.1 shows the crude death rate from AIDS in Australia over the period 1982 to 1999. Two points about this Figure warrant emphasis. First, because of the time lag from HIV infection to death from AIDS, the initial increase in the death rate and the peak in the death rate occur some years after the corresponding rise in the number of HIV infections. Second, even when the crude death rate from AIDS was at its peak of 4.1 deaths per 100 000 population in 1994, AIDS did not represent a major cause of death in Australia. By way of comparison, the crude death rates per 100 000 population from some other diseases in that year were as follows: ischaemic heart disease 171.2; lung cancer 37.5; female breast cancer 29.6; and road trauma 11.0.1

These data on AIDS mortality should not be interpreted as trivialising HIV/AIDS as a public health problem in Australia. Rather, they indicate the success of the efforts of a variety of individuals and organisations, including the public health response by governments and communities, to contain the epidemic. This chapter outlines the public health response to the outbreak of HIV infection in Australia, estimates the costs of that response and the numbers of deaths averted and life-years saved attributable to that response, and provides an economic evaluation of the public health strategies adopted to prevent HIV infection. As with the other case studies presented in this report, health outcomes in the absence of the public health intervention cannot be observed and must be estimated using modelling techniques. Estimates of the number of HIV infections that would have occurred in the absence of a public health response are also provided in this chapter.

4.2. Public Health Programs to Reduce HIV/AIDS

The public health response to HIV/AIDS in Australia can be characterised as non-coercive, non-partisan and cooperative. It has relied primarily on the provision of information about the virus and the consequences of infection to achieve behavioural change. While stronger measures such as incarceration of infected individuals and mandatory testing for HIV infection were proposed by some in the early stages of the epidemic, the public health response by governments has been based upon education and voluntary rather than compulsory behaviour change.

Figure 4.1 Annual crude death rate from AIDS, Australia, 1982 to 1999

Sources: Numbers of deaths from AIDS: 1991 to 1999 - National Centre in HIV Epidemiology and Clinical Research (2000, Table 1.1.5); 1990 - National Centre in HIV Epidemiology and Clinical Research (1999, Table 1.1.5); 1989 - National Centre in HIV Epidemiology and Clinical Research (1998, Table 1.1.5); 1982 to 1988 - National Centre in HIV Epidemiology and Clinical Research (various, Table 2.9 or 4.9). (Numbers of deaths for 1989 onwards have been adjusted for reporting delays). Australian population: Estimated resident population as at 30 June each year provided by the Australian Bureau of Statistics (2000a).

This approach to the public health issues raised by HIV/AIDS was stated clearly in a policy paper released by the Commonwealth Government in 1989 (Commonwealth of Australia 1989). The paper stated that ‘behaviour change through education is the principal means of preventing the spread of the virus’ (p.10) and that ‘education has been the primary focus of Commonwealth Government activity on HIV’ (p.27). On HIV testing, its position was that ‘... an assessment of the personal benefit versus the personal cost [of testing for HIV antibody status] is a matter for informed individual choice (or, if the individual is not legally competent, by his or her guardian)’ (p.38). The paper argued that mandatory testing should take place only in specific circumstances (pp.40-1) and that research aimed at monitoring the spread of the virus did not require that individuals be identified (pp.74-6).

The policy paper both described the philosophy that underpinned the public health response to HIV/AIDS up to 1989 and set the stage for the continuation of that philosophy thereafter. The underlying principles of non-coercion and cooperation were embedded in the first National HIV/AIDS Strategy that was introduced in 1989-90 with funding for four years to 1992-93. This was followed by the second National HIV/AIDS Strategy for three years from 1993-94 to 1995-96, the third for three years from 1996-97 to 1998-99, and the fourth for five years from 1999-00 to 2003-04.

Table 4.1 shows some selected milestones in the public health response to HIV/AIDS in Australia since the diagnosis of the first case in 1982. The main public health responses can be summarised as securing the blood supply, introducing needle and syringe programs for injecting drug users, and educating the population about the virus and the consequences of infection. Education campaigns targeted both high-risk sub-groups in the population (e.g. men who have sex with men (MSM) and injecting drug users) as well as the general population. Perhaps the most high profile education activity aimed at the general population was the Grim Reaper campaign in 1987.2

While an emphasis on education and prevention characterised all four national strategies as well as the period before the first national strategy, the tendency in recent times has been to move away from education for the general population and towards more concentration of effort on education for high risk groups. The evaluation of the second national strategy, for example, recommended ‘that the Commonwealth and the States and Territories continue to provide education to the general community but that the funding level be decreased’ (Feachem, 1995, p.12, Recommendation 9). It also recommended that homosexually active men, injecting drug users, and Aboriginal and Torres Strait Islander communities remain priority groups for education and prevention activities.

4.3. Expenditure on HIV/AIDS Public Health Programs

In general, expenditures on HIV/AIDS programs in Australia can be classified according to whether they are aimed at education and prevention, treatment and care, or research. Like many expenditure classification systems, the boundaries between these categories can be fuzzy. For the most part, expenditures on education and prevention are targeted at people who are not infected with the virus, attempting to educate them about the virus, the consequences of infection, and measures that can be taken to reduce the risk of infection.

However, some activities targeted at those already infected are also concerned with prevention, e.g. a treatment program might contain an education component targeting infected individuals and containing information about measures to improve their own health status as well as reducing the risk of infecting others. In this study, HIV/AIDS public health programs are considered to be those programs with a primary focus on education and prevention.

Funding of HIV/AIDS education and prevention programs is predominantly the responsibility of the Commonwealth and State/Territory governments. While it is not possible to estimate the extent of private sector funding of these programs, the extent of such funding is generally thought to be minimal. In this report, the expenditure estimates include government expenditures only.

Commonwealth government expenditures on HIV/AIDS programs fall into two main categories – programs undertaken directly by the Commonwealth such as nationwide advertising campaigns, and programs undertaken by the States and Territories which receive Commonwealth funding. The latter programs may involve a matched funding requirement, as with the Matched Funding Program for HIV/AIDS in operation from 1985-86 to 1996-97. The Commonwealth has also funded HIV/AIDS activities through grants to the States and Territories without a matched funding requirement, such as the funding for Aboriginal and Torres Strait Islander programs that commenced in 1993-94.3

Table 4.1 Selected milestones in the Australian public health response to HIV/AIDS

Year

Milestone

1982

First person with AIDS diagnosed in Australia.

1983

National Health and Medical Research Council (NHMRC) established a working party on AIDS.

1984

National AIDS Task Force established, replacing the NHMRC working party.
A ministerial advisory committee, the National Advisory Committee on AIDS (NACAIDS) formed to advise on educational, social and legal issues.
Donor Declaration Forms introduced at all blood banks.
Commonwealth commenced funding of State-based non-government organisations providing education, counselling and support services to higher risk groups.

1985

Australian Health Ministers’ Conference endorsed the National Health Strategy for AIDS Control.
HIV test became available and Australia became one of the first countries to have universal blood donor screening.
Australian Federation of AIDS Organisations (AFAO) established to represent AIDS Councils at the national level.
Commonwealth and States agree to 50:50 cost sharing of funding for State-based non-government organisations under the new Matched Funding Program.

1986

NACAIDS adopted a national education strategy on AIDS that provided the framework for its national education campaign.
National AIDS Task Force recommended the provision of needles and syringes to intravenous drug users.

1986

Needle and Syringe Exchange Programs commenced.
Grim Reaper campaign commenced on Passion Sunday in April.
Commonwealth required at least 50per cent of amounts spent under the Matched Funding Program to be directed to prevention, education and counselling.

1988

The Australian National Council on AIDS (ANCA) formed to replace NACAIDS and the National AIDS Task Force.
The Commonwealth AIDS Research Grants (CARG) program commenced.

1989

First National HIV/AIDS Strategy implemented with funding for four years.

1993

Second National HIV/AIDS Strategy implemented with funding for three years, supporting a change in emphasis from Commonwealth to State/Territory responsibility for direct program planning and delivery.

1995

Review of second National HIV/AIDS Strategy reinforced the importance of education and prevention campaigns for high risk groups.

1996

Third National HIV/AIDS Strategy implemented for three years.
Australian National Council on AIDS and Related Diseases (ANCARD) is formed to replace ANCA.
Matched Funding Program terminated at the end of 1996-97, with Commonwealth specific purpose payments to the States for eight public health areas including HIV/AIDS being broadbanded in the new Public Health Outcome Funding Agreements (PHOFA) in 1997-98.

1999

Review of Third National HIV/AIDS Strategy recommends that educational and advertising material should be directed at particular groups or populations.

2000

Fourth National HIV/AIDS Strategy implemented for five years.

The States/Territories have contributed funding to HIV/AIDS public health programs according to the conditions of the Matched Funding Program, but have commonly committed funds in excess of those required to match Commonwealth funding. Butler (1996, Table 4.1) found that, from the early 1990s, State funding began to exceed that required by the matched funding provisions. By 1995-96, States were contributing 66 per cent of the total funding under the Matched Funding Program, implying that States were expending $1.94 for each $1.00 of Commonwealth funding received under this Program.

The Matched Funding Program was terminated in 1996-97, and in 1997-98 the Public Health Outcome Funding Agreements (PHOFAs) were introduced. Three important features of the funding arrangements under the PHOFAs are as follows. First, the PHOFAs broadbanded a number of public health programs including HIV/AIDS, so that identification of expenditures on HIV/AIDS programs within the PHOFAs is not possible. Second, the PHOFAs do not have a matched funding requirement. Third, the allocation of PHOFA funds between the public health programs covered by them is at the discretion of the States/Territories. Taken together, these features make it very difficult to estimate that part of PHOFA funding committed to HIV/AIDS programs, and also to estimate the extent of additional funding of HIV/AIDS programs by States/Territories over and above the funding they receive under the PHOFAs.

Expenditure on HIV/AIDS education and prevention programs

Within overall expenditure on HIV/AIDS programs, it is necessary for the purposes of this study to estimate expenditures on education and prevention programs, as opposed to expenditures on treatment and care programs or research. For the years 1984-85 to 1995-96, this study employs the expenditure estimates developed by Butler (1996). In that study, all expenditures on Commonwealth Programs apart from expenditure on research and the National Serology Reference Laboratory were assumed to be for education and prevention. For projects funded under the Matched Funding Program, that study exploited a register of projects that was maintained for the duration of the first and second National HIV/AIDS Strategies to estimate expenditure on education and prevention. The resulting expenditure series is shown in Table 4.2. This Table also provides estimates of expenditure on education and prevention for the three years 1996-97 to 1998-89 that have elapsed since the publication of the earlier study.4 In constant price terms, expenditures on HIV/AIDS education and prevention programs lie within the range $47 to $55 million each year since 1990-91.

Education and prevention expenditure by target group

These expenditure estimates can be disaggregated according to the target groups for the education and prevention programs. The earlier study by Butler (1996) provides estimates for expenditures on four groups (Aboriginal and Torres Strait Islander peoples, MSM, injecting drug users, and others). Only relatively small numbers of AIDS cases have arisen in sub-groups other than MSM. Therefore, in the present study, in addition to undertaking analyses for all sub-groups combined, sub-group analyses will be undertaken only for the MSM sub-group. However, to undertake such sub-group analyses, it is necessary to obtain estimates of expenditure on education and prevention for the MSM sub-group.

Butler (1996, Table 4.16) provides a series for education and prevention expenditures on the MSM sub-group for the period 1984-85 to 1995-96. Because of the uncertainties involved in allocating expenditures to particular target groups, three series are constructed – a baseline (most likely) series, a series of lower estimates, and a series of upper estimates. These three series in current price terms are shown in Table 4.3. The baseline estimates indicate that from 21 per cent to 25 per cent of all education and prevention expenditures have been targeted at the MSM sub-group. The lower estimates put this range at 3 per cent to 6 per cent, and the upper estimates 24 per cent to 34 per cent.

To obtain expenditure estimates for 1996-97 to 1998-99 in Table 4.3, the mean proportions of education and prevention expenditures on the MSM sub-group for the three years 1994-95 to 1995-96 (6.0 per cent, 24.4 per cent and 34.0 per cent for the lower, baseline and upper estimates respectively) have been applied to the total expenditure on HIV/AIDS for the corresponding years in Table 4.2.

Table 4.4 presents the constant price expenditure series (year 2000 prices) corresponding to the three current price series shown in Table 4.3.

Table 4.2 Estimated expenditures on HIV/AIDS education and prevention activities under Commonwealth programs and the Matched Funding Program(a), 1984-85 to 1998-99 ($’000)

 

Commonwealth

Matched

Total(b)

 
 

Programs
(1)

Funding Program
(2)

Current prices
(3)

Constant 2000 prices
(4)

1984-85

1,363

3,794

5,157

9,485

1985-86

1,914

6,965

8,879

15,064

1986-87

3,889

8,279

12,168

18,873

1987-88

7,404

11,795

19,199

27,742

1988-89

7,710

15,790

23,500

31,646

1989-90

6,256

24,349

30,605

38,164

1990-91

13,001

29,865

42,866

50,763

1991-92

10,509

31,448

41,957

48,761

1992-93

8,543

38,286

46,829

53,871

1993-94

6,268

36,050

42,318

47,799

1994-95

9,082

36,081

45,163

49,445

1995-96

9,349

37,497

46,846

49,214

1996-97

11,653

37,207

48,875

50,663

1997-98

12,101

37,260

52,416

54,333

1998-99

11,561

38,206

49,767

50,952

(a) ‘Commonwealth Programs’ refers to HIV/AIDS programs provided directly by the Commonwealth government. Activities funded under the Matched Funding Program are undertaken by States/Territories but are partially Commonwealth funded. The Matched Funding Program was terminated at the end of 1996-97. The figures under this heading for 1997-98 and 1998-99 are estimates of the combined Commonwealth and State funding for HIV/AIDS activities under the new Public Health Outcome Funding Agreements.

(b) Column (3) = column (1) + column (2).
Column (4) = column (3) converted to constant prices using the Consumer Price Index (weighted average of eight capital cities, all groups, financial year).
Sources: Expenditure data in current prices: 1984-85 to 1995-96 – Butler (1996, Tables 4.9, 4.14); 1996-97 to 1998-99 – see Attachment 1.1. Consumer Price Index: Australian Bureau of Statistics (2000b).

Table 4.3 Estimated expenditures on HIV/AIDS education and prevention activities targeting men who have sex with men, 1984-85 to 1998-99 (current prices)

Table 4.4 Estimated expenditures on HIV/AIDS education and prevention activities targeting men who have sex with men, 1984-85 to 1998-99 (AUD ’000, constant 2000 prices)

4.4. The Impact of HIV/AIDS Public Health Programs on HIV Incidence

Estimating HIV incidence

The sequence of events that lead to the development of AIDS begins with infection by the HIV virus. While those infected may experience flu-like symptoms a short time after infection, these may not be thought to be associated with HIV infection. When these symptoms subside, the individual may remain asymptomatic for some months or years. For this reason, the exact time when an individual became infected with the virus may not be known, as this may have occurred some time before he/she was diagnosed as being infected. HIV incidence must, therefore, be estimated from the available data on AIDS incidence.

A method commonly employed to obtain estimates of HIV incidence is back-projection. Suppose the time of an AIDS diagnosis is known but that the time of HIV infection is unknown. The rate of progression from HIV infection to AIDS varies between cases. However, if a probability density function is specified for the time from HIV infection to AIDS (referred to as the incubation distribution), the probability of an AIDS case diagnosed at time t having been infected with HIV at time t-1, t-2, etc. can be estimated. In this way, HIV incidence can estimated by projecting backwards in time from the date of AIDS diagnosis. A discussion of this method can be found in Becker and Watson (1991).

Back-projection estimates of HIV incidence become less reliable the closer is the projection year to the current year (or the most recent year for which AIDS incidence data are available). The reason for this is that shorter periods of time have elapsed for infections in more recent years, and fewer cases will have progressed to AIDS.

In back-projection modelling, the underlying incubation distribution should reflect the treatment effect, i.e. the effect of treatment on the progression rate from HIV infection to AIDS. Beginning with an incubation distribution that reflects the natural history of the disease, it is necessary to incorporate the effect of treatment on the incubation distribution to improve the accuracy of the incidence estimates. Treatment can be expected to reduce the progression rate and lengthen the time from HIV infection to AIDS.

Estimates of HIV incidence in Australia have been constructed by the National Centre in HIV Epidemiology and Clinical Research (NCHECR) as part of the review of the second National HIV/AIDS Strategy. These estimates, which show estimated annual HIV incidence for all exposure categories (homosexual intercourse, injecting drug use, etc.) are shown in Figure 4.2. A number of different series were estimated, reflecting differing assumptions about the incubation distribution and the treatment effect. Figure 4.2 shows the results of the two preferred methods – ‘method a’ modelled the incubation distribution as a Weibull distribution, while ‘method d’ used a log-logistic distribution. Both methods assumed a moderate treatment effect of 0.8 (i.e. the effect of treatment was a 20 per cent reduction in the progression rate).

Two features of these series warrant mention. First, although the incidence estimates vary between the two methods with the peak number of infections being around 2500 with method a and 3000 with method d, both methods show the peak occurring in 1984. Second, both methods show a levelling off of HIV incidence from 1991 onwards at a rate of 480 new infections per year.

Figure 4.2 Estimated annual HIV incidence, all exposure categories, 1981 to 1993

Source: National Centre in HIV Epidemiology and Clinical Research (1996, Table 1.5).

The Centre also provided estimates of HIV incidence by exposure category – homosexual/IDU (injecting drug user), IDU, and female heterosexual. The latter two exposure categories had relatively small numbers, the reason being that HIV infection has been spread in Australia predominantly through male homosexual contact. In its recent Annual Surveillance Report, the Centre states:

Transmission of HIV in Australia continues to be overwhelmingly through sexual contact between men ... Approximately 85 per cent of all HIV transmissions in Australia were estimated to have been by this route. Similarly, most reported diagnoses of newly acquired HIV infection were in men who were exposed through homosexual contact. (NCHECR 2000, p.9).

For the present study, the NCHECR has provided back-projection estimates for each exposure category using more recent data on AIDS incidence (see Table 4.5).5 Total annual HIV incidence is now estimated to have levelled off at 430 infections per year since 1993, with 380 of these infections arising in the homosexual/IDU exposure category. Further details of the basis on which these estimates have been constructed are provided in Annex E.

Table 4.5 Estimated annual HIV incidence, by exposure category, 1980 to 1994

(a) Annual estimates are sums of quarterly estimates. +/- = with or without.
(b)
Column (4) = column (1) + column (2) + column (3). Any discrepancies between totals and sums of components are due to rounding.
Source: Data provided by NCHECR.

Estimating the reduction in incidence due to HIV/AIDS education and prevention programs

How many HIV infections would there have been each year in the absence of education and prevention programs? Such programs aim to change behaviour so as to reduce transmission between infected and susceptible individuals. Transmission rates corresponding to the estimated HIV incidence shown in Table 4.5 can be estimated from the following simple epidemic model:

where Ht is HIV incidence in time period t, It is the number of infected individuals at the beginning of each time period t, st is the proportion of the relevant population that is susceptible at the beginning of time period t, and βt is the transmission rate in time period t or the number of newly infected persons per infectious person adjusted for the proportion of the population that is susceptible in time period t.

The values of βt underlying the estimated HIV incidence in Table 4.5 include the effect of education and prevention programs. In the absence of such programs, these values of βt would have been greater. By using inflated values of βt to capture the effect of education and prevention programs, estimates of HIV incidence in the absence of such programs, and hence the reduction in the number of HIV infections attributable to these programs, can be obtained.

The first step is to obtain estimates of βt underlying estimated HIV incidence including the effect of education and prevention programs. NCHECR has provided quarterly estimates of HIV incidence for the MSM sub-group, and annual estimates for the IDU and female heterosexual sub-groups. Since there is no cure for HIV infection, It is measured as cumulative HIV incidence up to time t less the number of individuals who have developed AIDS up to that time.6 The term st can be obtained by estimating the size of the relevant population (e.g. MSM, IDU, etc.) and subtracting from this the number of individuals in that population already infected. Further details are provided in Annex E.

The quarterly transmission rates estimated using the simple epidemic model and based on the estimated HIV incidence for the MSM sub-group are shown in Figure 4.3. The transmission rate fell rapidly in the early years of the epidemic. Prior to the commencement of government funding of HIV/AIDS education and prevention programs in 1984-85, the transmission rate fell 0.59 at the commencement of the epidemic to 0.22 in the second quarter 1984. By the second quarter of 1989 it had fallen to 0.01 and has remained at that level since then.

Figure 4.3 Quarterly transmission rates implied by estimated HIV incidence, MSM, 3rd Quarter 1982 to 4th Quarter 1996

The second step is to re-estimate HIV incidence excluding the effect of education and prevention programs by using higher values for the transmission rates. A difficulty arises here because the decline in the transmission rate in Figure 4.3 may be due to either or both of two factors - changes in HIV risk behaviour and the extent of infection in the population at risk. The second factor reflects an underlying feature of infectious disease dynamics – as more people are infected by a disease, there are fewer left to become infected, leading to a fall in the transmission rate.7 This factor is unaffected by education and prevention programs which have their impact on the first factor, HIV risk behaviour.

Hurley et al (1996a) observe that, between 1985 and 1993, HIV risk behaviour in MSM changed in that the prevalence of unprotected anal intercourse with casual partners declined by 25 per cent. They then assumed that 25 per cent of the decline in the transmission rate over this time period was attributable to changes in HIV risk behaviour. However, not all of this behaviour-induced change in the transmission rate can be attributed to government-funded education and prevention programs. Hurley et al (1996a) assumed that 50 per cent of this 25 per cent decline was attributable to such programs. On this basis, they estimated that there would have been 1397 more cases of HIV infection over the period 1985 to 1993 in the absence of education and prevention programs.

In the present study, the increase in the transmission rate in the absence of education and prevention programs has been selected so as to result in approximately the same increase in the number of HIV infections over the time period 1985 to 1993 as estimated by Hurley et al (1996a). The resulting increase in the transmission rate is 25 per cent (see Annex E). The resulting HIV incidence series for the MSM sub-group is shown in column (3) in Table 4.6. The difference between this series and that including the effect of education and prevention programs is the number of HIV infections prevented in the MSM sub-group (see column (5) in Table 4.6). Projections from 1993 to 2010 are based on a fixed annual HIV incidence in the MSM sub-group of 380 with, and 567 without, the effect of education and prevention programs. Over the 31 years from 1980 to 2010, an estimated 5488 HIV infections (undiscounted) will have been avoided in the MSM sub-group as a result of education and prevention programs targeted at that sub-group.

The simple epidemic model and underlying assumptions just discussed have also been used to obtain estimates of the transmission rates and counterfactual estimates of HIV incidence for the IDU and female heterosexual sub-groups (see Attachment E). Adding these to the MSM data gives the total for all exposure categories (see columns (2) and (4) in Table 4.6) and also provides an estimate of the number of HIV infections across all exposure categories (column (6)). Overall, it is estimated that 6973 HIV infections (undiscounted) will have been avoided by all government-funded HIV/AIDS education and prevention programs over the period 1980 to 2010.

These results are based on an assumed 25 per cent increase in the transmission rate in the absence of education and prevention programs. To test the sensitivity of the results to this assumption, HIV incidence series have also been estimated using 15 per cent and 35 per cent increases in the transmission rate in the absence of education and prevention programs. The resulting series, together with the baseline series and the series that incorporates the effect of education and prevention programs for all exposure categories are shown in Figure 4.4. The total numbers of HIV infections avoided under the alternative assumptions over the period 1980 to 2010 are shown in Table 4.7.

Table 4.6 Estimated HIV incidence with and without education and prevention programs, and number of HIV infections avoided, MSM and all groups, 1980 to 2010

Notes: E&P = education and prevention
Sources: Table 4.5 and results from epidemic model.

Figure 4.4 Estimated HIV incidence with and without education prevention programs, all exposure categories

Table 4.7 Total numbers of HIV infections avoided for the period 1980 to 2010 under alternative assumptions on the increase in the transmission rate without education and prevention programs, MSM sub-group and all exposure categories

4.5. Benefits of Reduced HIV Incidence

Avoiding HIV infection produces three benefits:

  • avoidance of a period lived in less than perfect health;
  • avoidance of years of life lost due to premature mortality; and
  • avoidance of treatment costs during the time lived in less than perfect health.

Morbidity and mortality from HIV/AIDS

The time lived in less than perfect health due to HIV/AIDS, i.e. the time from HIV infection to death from AIDS, is generally characterised by four stages: asymptomatic disease; AIDS-related complex (ARC); AIDS; and AIDS terminal phase. The Victorian Burden of Disease study of morbidity provided mean durations for each of these four stages, and corresponding disability weights for use in calculating years lived with disability (YLDs) (Government of Victoria 2000). These data are shown in Table 1.8, which also shows the YLDs attributable to one HIV infection.

The years of life lost due to premature mortality (YLL) for an infected person are calculated as the difference between the age at death of the person and the age at which death would have been expected had HIV infection been avoided. Using data on the number of deaths from HIV/AIDS in Australia in 1996 and YLLs from those deaths, the mean discounted YLLs per death is 22.8 (Mathers et al 1999, Annex Tables E, F - the authors report only discounted and not undiscounted results for YLLs). This figure requires modification for use in the present study for two reasons. First, it is obtained using a discount rate of 3 per cent whereas the baseline discount rate in this report is 5 per cent. Second, it discounts YLLs back to the year in which the death occurs. The year in which the death occurs is some 15 years after HIV infection (see Table 4.8). In the present study which is a cost-benefit analysis of HIV prevention, the YLLs are discounted back to the year in which the HIV infection is prevented.

Table 4.8 Calculation of years lived with disability (YLDs) for one HIV infection using Victorian Burden of Disease disability weights and disease durations

(a) undisc = undiscounted.
disc. = discounted.
Discount rate is 5per cent per annum and continuous time discounting is employed.
(b)
The Victorian Burden of Disease study assumes that, for half those infected, diagnosis of HIV infection does not occur until two years after infection. For this group the disability weight for the first two years is zero.
Source: Disease durations and disability weights taken from Government of Victoria (2000).

Figure 4.5 Time line for HIV infection (T0), death from AIDS (T1) and death from other causes (T2)

This second point is illustrated in Figure 4.5. HIV infection occurs at time T0, while death from AIDS occurs at T1. If HIV infection is avoided, death occurs at time T2. When a prevention program is implemented, avoidance of HIV infection then occurs at T0 but years of life saved do not begin to accrue until T1. For the purposes of economic evaluation, these life-years saved should be discounted back to T0 which is when the prevention program occurs. The estimates of YLLs from the study by Mathers et al (1999) are discounted back to T1 which is when death occurs.

The mean discounted YLLs of 22.8 translate into an undiscounted equivalent of 38.4 years.8 Allowing for the 15.5 year time lapse between HIV infection and death from AIDS, and discounting back to the year of HIV infection, the mean discounted YLLs from one HIV infection using a 5 per cent discount rate is 7.9 years.

Combining the results for YLDs and YLLs, on average one HIV infection results in 44.0 DALYs in undiscounted terms (= 5.6 YLDs + 38.4 YLLs). In discounted terms, one HIV infection results in 11.4 DALYs (= 3.5 YLDs + 7.9 YLLs) using a discount rate of 5 per cent. Using the value of $60 000 per DALY adopted in the present study (see Chapter 1), the monetary value of the morbidity and premature mortality avoided by preventing one HIV infection is $2.64 million undiscounted or $684 000 discounted.

Increase in treatment efficacy post-1996

There is considerable uncertainty around the DALY calculations because of developments in the treatment of HIV/AIDS over the last five years. The use of protease inhibitors in conjunction with two other antiretroviral compounds (triple therapy) may reduce the viral load in many HIV/AIDS patients. While the long-term efficacy of this therapy is uncertain, it is possible that the time from HIV infection to development of AIDS and consequent death from AIDS will be considerably lengthened. The NCHECR (2000, p.86) estimates that, in the context of back-projection modelling, the treatment effect will change from an 11 per cent reduction in the progression rate to AIDS in the 1st quarter 1995 to an 82 per cent reduction in the 4th quarter 1999.

Improved treatment efficacy will reduce the DALY burden of HIV/AIDS and hence the benefits of HIV/AIDS prevention. While the long-term impact of triple therapy remains uncertain, it is useful to test the sensitivity of the economic evaluation to possible improvements in treatment efficacy. Accordingly, the NCHECR estimates of the reduction in the progression rate from HIV infection to AIDS have been used to model the change in the mean time from HIV infection to AIDS. The model projects AIDS progression out to 50 years after HIV infection. The newer therapies are likely to have the most benefit in persons infected after 1996, but even those infected earlier than this time can be expected to obtain some benefit. The model predicts that, in post-1996 infections, 62 per cent of cases will progress to AIDS in 50 years. Because the average age at infection is around 30 years old, this effectively means that 38 per cent of those who are infected post-1996 will not develop AIDS by age 80. The model estimates that, in this post-1996 group, the mean time to AIDS for the 62 per cent of cases who develop AIDS is 14.4 years. The weighted mean time to AIDS for the whole group is 31.2 years, about three times the mean time from HIV infection to AIDS shown in Table 4.8.9 The mean time to AIDS declines for cases infected earlier than 1997, falling to 21.8 years for HIV infections that occurred in 1984. These projections are incorporated into the DALY calculations using the disability weights in Table 4.8, on the assumptions that the time from AIDS to death remains unchanged (5.5 years) and that the increase in incubation time is split evenly between the time spent in the asymptomatic state and the ARC state.

Treatment costs

Duckett (1987) estimated that the mean cost of treatment of an AIDS case in 1987 was $38 000. Hurley et al (1996b) estimated that the present value of lifetime treatment costs in 1992-93 was in the range $70 000 to $93 000 depending upon the time of diagnosis (these results use a discount rate of 5 per cent).

The introduction of new antiviral therapies in 1997 increased the drug cost of HIV/AIDS treatment. Anis et al (1998) found that the average incremental cost of adding a protease inhibitor to a 2-drug antiretroviral regimen was Cnd$2318. However, the authors also found that savings in hospitalisation costs offset most, if not all, of the additional drug cost. Gebo et al (1999) found that total Medicaid payments tended to be slightly lower in patients receiving a protease inhibitor regimen, again reflecting hospitalisation cost savings that offset the higher cost of drug therapy.

Some studies dispute the conclusion that the new protease inhibitor combination therapies will save costs. Pinkerton and Holtgrave (1999), using an economic model capable of estimating long-term medical care costs under different disease progression scenarios, argue that projections of long-term savings in medical care costs are ‘probably illusory’. Moore (2000) has undertaken a cost-effectiveness analysis of a three-drug regimen containing a protease inhibitor compared with a one- or two-drug regimen without a protease inhibitor, finding incremental direct costs per life-year saved in the range US$10 000 to just over US$13 000. The positive net cost per life-year saved indicates that protease inhibitor therapy is not cost saving, that is, the hospitalisation cost savings do not more than offset the increased drug cost.

In the present study, the following assumptions are made regarding treatment costs:

  • For HIV incident cases in the six-year period 1984 to 1989, the treatment cost estimate of $38 000 per case from Duckett (1987) is used. This becomes $57 000 in year 2000 prices.
  • For HIV incident cases in the seven-year period 1990 to 1996, the treatment cost estimate of $93 000 per case from Hurley et al (1996b) is used. This becomes $107 000 in year 2000 prices.
  • For HIV incident cases from 1997 onwards, the treatment cost of $107 000 is inflated by around 10 per cent to $118 000 to reflect a possible increase in net cost of treatment due to triple therapy since 1997.10

4.6. Evaluation of Public Health Programs to Reduce HIV Incidence

This evaluation of public health programs to reduce HIV incidence covers the period 1984 to 2010. While the estimates of HIV incidence indicate that the epidemic started around 1979 or 1980, the first government expenditures on education and prevention programs arose in the 1984-85 financial year. The costs of HIV/AIDS public health programs for the MSM sub-group and for all exposure categories are shown in Table 4.9. From 1999 onwards, it is assumed that expenditures each year are the average annual expenditures over the five-year period 1994 to 1998.

The total benefits of the HIV/AIDS public health programs are also shown in Table 4.9. The benefits comprise the value of DALYs avoided at $60 000 per DALY and treatment cost savings. Disaggregating the total benefits into these two components (not shown in Table 4.9), of the total benefits of $2464 million for the MSM sub-group, the value of DALYs avoided is $2173 million and the treatment cost savings $291 million with a discount rate of 5 per cent. Of the total benefits for all exposure groups of $3105 million, the corresponding amounts are $2732 million and 373 million. Hence the treatment cost savings amount to about 12 per cent of the total benefits with a 5 per cent discount rate.

The estimated net present value of education and prevention programs for HIV/AIDS amounts to $2319 million for the MSM sub-group and $2498 million for all exposure groups, using a discount rate of 5 per cent. Net benefits remain substantially positive for both groups if the discount rate is increased to 7 per cent. For all three discount rates for both groups, the costs of the public health programs are considerably less than the benefits.

Sensitivity analysis

The sensitivity of net benefits to changes in key underlying assumptions is shown in Table 4.10. For all exposure groups, even with a smaller increase of 15 per cent in the transmission rate and allowing for a substantial post-1996 treatment effect, net benefits remain positive at $651 million. For the MSM sub-group, even allowing for the combined effects of the smaller increase in transmission rate, the higher estimated cost of public health programs for MSM, and an increase in the treatment effect post-1996 (all of which tend to reduce net benefits), the net benefits remain positive at $821million.

Table 4.9 Evaluation of public health programs to prevent HIV infection ($ million, constant 2000 prices)(a)

 

Costs(b)

Benefits(c)

Net benefits

 

MSM

All Groups

MSM

All groups

MSM

All groups

1984

1.1

4.7

0

0

-1.1

-4.7

1985

2.9

12.3

545.0

569.3

542.0

557.0

1986

3.9

17.0

1,187.1

1,270.7

1,183.1

1,253.7

1987

5.0

23.3

1,046.8

1,184.4

1,041.8

1,161.1

1988

6.4

29.7

728.4

911.9

722.0

882.2

1989

8.3

34.9

580.0

820.2

571.8

785.3

1990

10.5

44.5

557.8

849.1

547.3

804.6

1991

11.4

49.8

557.8

860.1

546.4

810.3

1992

12.4

51.3

552.3

791.4

539.9

740.1

1993

12.7

50.8

513.9

651.3

501.2

600.4

1994

12.0

48.6

513.9

656.7

501.9

608.1

1995

11.9

49.3

513.9

656.7

501.9

607.4

1996

12.1

49.9

513.9

656.7

501.8

606.8

1997

12.4

52.5

515.9

659.4

503.5

606.9

1998

12.5

52.6

515.9

659.4

503.5

606.7

1999

12.2

50.6

515.9

659.4

503.7

608.8

2000

12.2

50.6

515.9

659.4

503.7

608.8

2001

12.2

50.6

515.9

659.4

503.7

608.8

2002

12.2

50.6

515.9

659.4

503.7

608.8

2003

12.2

50.6

515.9

659.4

503.7

608.8

2004

12.2

50.6

515.9

659.4

503.7

608.8

2005

12.2

50.6

515.9

659.4

503.7

608.8

2006

12.2

50.6

515.9

659.4

503.7

608.8

2007

12.2

50.6

515.9

659.4

503.7

608.8

2008

12.2

50.6

515.9

659.4

503.7

608.8

2009

12.2

50.6

515.9

659.4

503.7

608.8

2010

12.2

50.6

515.9

659.4

503.7

608.8

NPV @ 5 per cent

144.5

607.2

2,463.6

3,105.1

2,319.1

2,497.9

NPV @ 0 per cent

281.6

1,178.6

15,033.5

19,109.8

14,751.9

17,931.2

NPV @ 3 per cent

185.5

778.1

4,695.1

5,939.2

4,509.6

5,161.2

NPV @ 7 per cent

115.0

484.1

1,423.5

1,787.3

1,308.5

1,302.2

(a) Any discrepancies between net benefits and the difference between benefits and costs are due to rounding.

(b) Costs for calendar years have been obtained by pro-rating the financial year estimates in Table 1.2 and 1.4. For the MSM sub-group, the baseline estimates of expenditure are employed.

(c ) Benefits are based on a 25per cent increase in the HIV/AIDS transmission rate and no increase in the treatment effect post-1996.
Sources: Tables 4.2, 4.4 and model results.

Table 4.10 Sensitivity of net benefits to changes in key parameters ($ million)(a)

(a) All results based on discount rate of 5 per cent.
Source: Model results.

Financial return to government

If the measure of benefit is confined to treatment cost savings, i.e. excluding the value of DALYs avoided, net benefits remain positive for the MSM sub-group. At a discount rate of 5 per cent, the treatment cost savings of $291 million exceed the costs of education and prevention programs of $145 million, giving net benefits of $146 million. Net benefits remain positive for discount rates of 3 per cent and 7 per cent also. However, this conclusion does not hold for all exposure groups. The costs of education and prevention programs ($607 million) exceed the treatment cost savings ($373 million). Net benefits are also negative for 3 per cent and 7 per cent discount rates.

4.7. Conclusions

HIV/AIDS is a viral disease against which there is no vaccine and for which there is no cure. The epidemic in Australia, along with many countries, commenced in the early 1980s. In 1984, government funding for HIV/AIDS education and prevention programs commenced. Australia’s public health response to preventing the spread of this serious disease has always been based on a non-coercive, non-partisan and cooperative approach. Education programs targeted both high-risk groups and the general population, and by the second half of the 1980s HIV incidence was falling. Back-projection estimates of HIV incidence suggest the annual number of new infection has stabilised at 380 per year in the MSM group and 430 in all exposure groups.

This chapter has provided an evaluation of Australia’s public health response to HIV/AIDS over the period back to 1984 and forward to 2010. The forward estimates are predicated on the assumption that annual expenditure on education and prevention programs remains at the same level in real terms as over the five-year period 1994 to 1998. Analyses have been conducted for the MSM sub-group and all exposure groups combined. Key assumptions in the Central Case analysis were:

  • the HIV/AIDS transmission rate would have been 25 per cent higher in the absence of education and prevention programs;
  • there will be no increase in the effect of treatment on the incubation period post-1996 when triple therapies including a protease inhibitor were introduced; and
  • the baseline estimates of expenditure on education and prevention programs for the MSM sub-group are the most plausible estimates.

As with other public health programs analysed in this report, Australia’s investment in HIV/AIDS public health programs has returned substantial positive net benefits. For all exposure groups, the present value of expenditures on education and prevention programs in 2000 prices discounted back to 1984 is $607 million. The present value of the benefits derived from these programs is $3105 million, with net benefits then being equal to $2498 million.

The conclusion of substantial positive net benefits from HIV/AIDS education and prevention programs is robust to changes in the key underlying assumptions. Even with simultaneous and substantial changes in all three key underlying assumptions, the net benefits of the programs are $651 million for all exposure groups and $821 million for the MSM sub-group.

In terms of public finances, the present value of treatment cost savings exceeds the cost of education and prevention programs for the MSM sub-group for 3 per cent, 5 per cent and 7 per cent discount rates. However, the opposite is true for all exposure groups taken together.

1 The total female population is used in the denominator for calculating the crude death rate for female breast cancer.

2 Eight weeks after the launch of the hard-hitting TV commercial that was the centrepoint of this campaign, 97 per cent of 610 people surveyed in Sydney and Melbourne had seen the Reaper (Taylor 1988, p.547). However, other research has questioned the impact of this campaign on knowledge about AIDS (Rigby et al 1989).

3 Two other examples of unmatched specific purpose payments are Medicare grants that include funding for AIDS hospital treatment, and the Highly Specialised Drugs grants that include funding for high cost drugs used in hospital-based AIDS treatment. As such grants are related to treatment and care of AIDS patients, they are not considered further here.

4 Details of the construction of these estimates are provided in Annex D.

5 These estimates use a Weibull-with-levelling distribution, a treatment effect of 0.9 (i.e. a 10 per cent reduction in the progression rate) and allow for a change in the AIDS case definition in 1988. See Annex E for further details.

6 The number of individuals who have developed AIDS is subtracted in this calculation on the assumption that, once a person develops AIDS, he/she is no longer infectious.

7 In the limit, when the whole population is infected, the transmission rate falls to zero.

8 This undiscounted result has been retrieved using the continuous time discounting formula adopted in Mathers et al (1999, p.16).

9 The weighted mean uses 50 years as the time to AIDS for the 32per cent who do not develop AIDS after 50 years. That is, 31.2 = [(0.68 x 14.4) + (0.32 x 50)].

10 Taking the lifetime treatment cost of $107,000 and dividing by the time from HIV infection to death in Table 1.8 (15.5 years) gives an average annual expenditure per case of $6,903. In 1996 there were an estimated 11,080 people living with AIDS so the estimated annual expenditure on HIV treatment and care in that year is $76.5 million (year 2000 prices). The AIHW has reported a current price estimate of $54.1 million as the total direct cost of HIV/AIDS in 1989-90 (AIHW 1994, Table 2.3). This is $67.5 million in year 2000 prices, which is comparable with the figure of $76.5 million given developments in treatment that occurred in the years between the two studies (1989-90 and 1992-93).

Previous PageTop Of PageNext Page