• Login
• Print Friendly

• Recent Publications
• Other Publications
• Public Economics - Principals and Practice

5.1 Introduction

Immunisation is an important dimension of public health policy. This chapter evaluates the impact of childhood immunisation coverage in reducing the prevalence of two vaccine preventable diseases: measles and haemophilus influenzae type b disease (Hib disease).¹ The evaluations cover the period 1970-2003 for measles and 1993-2003 for Hib.

Measles was selected for evaluation because of the importance of the disease and its infectiousness and because there is almost no Australian work on the longer-term impact of immunisation upon measles. Immunisation for Hib disease was selected because of its comparatively recent introduction and the availability of comparable national time series data on the incidence of disease, before and after the introduction of Hib vaccine in 1993.

Unlike other public health programs considered in this study, there are good clinical data to support a direct causal relationship between health gain and a public health program (Bower et al 1998; Mast et al 1990). Although a successful immunisation program requires people to participate in it, the impact of immunisation on health depends less on behavioural change than on the effectiveness of the vaccine. The object of most immunisation programs is to promote coverage of at least 90-95 per cent. At such levels, most vaccine-preventable infections stop spreading and herd immunity is likely to protect those who are not successfully vaccinated. Moreover, once herd levels of immunisation coverage have been achieved and sustained, even if coverage later falls temporarily below required herd levels of coverage, susceptibility to disease outbreaks is reduced.

In this chapter, we start with a description of the nature of the immunisation programs, their funding, and data availability. Section 5.3 examines the incidence of measles and the impacts of immunisation programs for measles. Section 5.4 discusses the benefits and costs of these programs. Section 5.5 and 5.6 discuss the impacts of immunisation programs on Hib diseases and the benefits and costs of these programs respectively.

5.2 Immunisation Programs and Data for Measles and Hib

There are striking differences between the modes of transmission and patterns of measles and Hib disease. Measles is highly contagious and liable to sudden, sharp and unpredictable outbreaks, especially where immunisation coverage is low (Herceg et al., 1994; Srirajalingam and Sheridan, 1998). It is a highly infectious disease that spreads rapidly through susceptible populations (Russell, 1988). On the other hand, Hib disease is uncommon but endemic.

Sir Richard May has said of measles that:

'… ultimately one may conclude that the dynamics of measles transmission are chaotic, best explained as a nonlinear system, which cannot be accurately anticipated by the best mathematical logic and epidemiologic investigation' (quoted in Thacker et al. 1991).

The average duration of the acute phase of childhood measles is five days. About 10 per cent of measles cases are admitted to hospital for re-hydration and 2 per cent are admitted to intensive care (Shiell et al, 1998). About one per 1000 cases develops into measles encephalitis and about one in 25 000 develops into sub acute schelerosing panencephalitis². Because it is associated with the risk of severe, albeit rare, complications, measles must be regarded as a serious disease.

In relation to Hib disease, McIntyre et al (2000, p.9) report that:

'… prior to the introduction of Hib vaccination the most common manifestation of invasive Hib disease was meningitis, with children aged less than 18 months most at risk. Survivors of Hib meningitis commonly had neurological sequelae such as deafness and intellectual impairment. Epiglottitis was the other major category of infection, most often occurring in children over the age of 18 months. Less common manifestations of Hib disease included cellulitis, septic arthritis, pneumonia, pericarditis, osteomyelitis and septicaemia'.

To achieve full immunisation coverage for measles in accordance with the Australian Standard Vaccination Schedule (NHMRC, 2000), two childhood doses of vaccine are required—one at 12 months and another at 4–5 years (4 years since 2000). The second measles infant dose was introduced in 1998. This brought forward an adolescent dose, following the 1993/94 measles outbreak in Australia. With this change to the schedule, the Measles Control Campaign was implemented in order to ensure that children aged 5-12 years received their second dose. The vaccine for measles is a multivalent vaccine, administered with vaccines for mumps and rubella, and usually referred to as 'MMR'.

For Hib disease, vaccine may be administered as a HibTITER or Pedvax dose. The former is given at 2, 4, 6 and 18 months; the latter at 2, 4 and 12 months. Almost all Hib vaccinations in Australia were of the HibTITER type and our modelling below allows only for the HibTITER dose. From May 2000 the Schedule changed to Pedvax only. This required three shots instead of four. However the total cost of completing the Hib schedule remained the same. The major chronologies associated with the adoption of measles and Hib immunisations in Australia are shown in Tables 5.1 and 5.2.

Immunisation funding

Before 1988, the Commonwealth directly funded States/Territories for childhood vaccines for distribution to providers in the public sector. Private practitioners who gave vaccines were required to issue prescriptions for the private supply of vaccines by a pharmacist.

In July 1988 the Commonwealth withdraw from direct funding of vaccines and instead increased funding to States/Territories by way of Finance Assistance Grants (FAGs) and Hospital Funding Grants (HFGs). The FAG funding was designed to cover the bulk of the vaccines on the National Health and Medical Research Council (NHMRC) Immunisation Schedule, whether publicly or privately provided (DHAC, 2000a).

Table 5.1 Vaccination programs for measles

1968	Measles vaccine first gained marketing approval in Australia through the Australian Drug Evaluation Committee.
1970	Measles vaccine became available without charge.
	Incorporation of measles vaccine into the first NHMRC Recommended Immunisation Schedule.
1982	Conjugate measles/mumps (MM) vaccine was introduced to the NHMRC Schedule.
1989	Conjugate measles/mumps/rubella (MMR) vaccine replaced MM vaccine on the NHMRC Schedule.
1994	MMR adolescent dose incorporated into the Schedule, replacing the Rubella dose for schoolgirls.
1998	MMR adolescent dose moved forward to 4 -5 years.
	Measles Control Campaign.
1999	MMR vaccine became available free of charge in the improved PRIORIX version (less injection pain).
2000	MMR infant dose moved from 4 -5 years to 4 years.

Table 5.2 Vaccination programs for Hib disease

1992	Hib vaccine first gained marketing approval in Australia through the Australian Drug Evaluation Committee.
1993	Hib vaccine first incorporated into NHMRC Schedule and available free of charge.
	Direct funding from the Commonwealth to the States/Territories for co-ordinating the central purchasing of Hib vaccine. In the year following this became a model for Specific Purpose Payments to States/Territories for the purchase of other vaccines.

However, these arrangements resulted in a fragmented delivery of immunisation services and wide variations in the prices that States/Territories paid for vaccines, with the smaller jurisdictions paying higher prices (McIntrye, et al, 2000). The NHRMC considered that these arrangements contributed to a low level of immunisation. In January 1993, a new Hib vaccine became available. As there were no existing funding arrangements for purchase of this vaccine by States and Territories, the Commonwealth introduced a Specific Purpose Payment. Commonwealth funding was conditional on vaccines being provided to all public and private practitioners and was formalised in Bilateral Agreements with each State and Territory.

From 1997–1998 the Commonwealth started to fund certain vaccines through the Public Health Outcome Funding Agreements (PHOFAs) and by 1999–2000, PHOFA funding included coverage for all vaccines on the NHMRC Schedule. This funding is a 'special appropriation' under s9B of the National Health Act for vaccine to cover 105 per cent of each eligible population cohort, allowing for 100 per cent coverage and 5 per cent wastage (DHAC, 2000a).

Funding for the delivery and reporting of immunisation services is met at both State and Commonwealth levels. The Commonwealth funds the overhead expenses of the Australian Childhood Immunisation Register (ACIR) and shares the ACIR reporting payments with the States. Immunisation providers receive $6.00 for correctly reporting each immunisation encounter to the ACIR (except in Queensland where it is $3.00, with $3.00 going directly to VIVAS, which Queensland continues to operate as a sub register).³

States provide funding for provision of immunisation services by local councils (as predominantly occurs in Victoria), public health units and community health centres. The Commonwealth funds services by general practitioners through Medicare and by special incentive payments under the General Practice Immunisation Incentive (GPII) scheme. This scheme was introduced in July 1988. It includes a Service Incentive Payment of $18.50 for completing an age appropriate series immunisation according to the Schedule and a Practice Outcomes Payment that is designed to increase the overall coverage of children seen by a practice (even if they are not immunised in that practice). Separate funding ($8 million) was provided to divisions of general practice to educate GPs about the GPII scheme and to encourage their participation.

The introduction of the GPII scheme led to an increased share of immunisation services in the hands of general practitioners, largely at the expense of local councils. Some local councils are now quitting involvement in immunisation. Between the mid-1990s and 1999, the proportion of childhood immunisation services delivered by general practitioners grew from 56 per cent to 69 per cent (Human Capital Alliance, 1997; HIC, 2000).

In addition to vaccine and delivery costs, immunisation also attracts various one-off program costs, such as the Commonwealth’s Measles Control Campaign, mounted between August and November 1998 at a cost of $30 million (DHAC, 2000b).

Data on immunisation programs

The paucity of effective time series data on disease notification and vaccine coverage poses a challenge for researchers seeking to estimate the historical impact of immunisation coverage on morbidity associated with vaccine–preventable diseases.

Before 1991, data on the occurrence of vaccine preventable disease was not collated nationally. In that year the States and Territories amended their respective Public Health Acts to ensure systematic collection and proper definition and notification of data. As described below, in order to estimate the incidence of measles disease, we draw on some comparable disease data from the United Kingdom.

The NHRMC’s Australian Standard Vaccination Schedule contains a list of the vaccines for children 0-6 years and prescribes ages at which they should be administered for a child to be considered 'fully immunised'. However, before the ACIR was established in 1996, data on coverage was maintained on local registers such as VIVAS in Queensland and in a fragmentary way, in smaller local government and public health systems and by some Divisions of General Practice. The Australian Bureau of Statistics undertook surveys of national immunisation coverage in 1983, 1989/90 and 1995. There are, nevertheless, inconsistencies between ABS estimates of coverage and the early data captured on the ACIR. Some commentators consider that because some providers do not always report to the ACIR, its records underestimate actual levels of coverage (Human Capital Alliance, 2000).

Since 1996, the ACIR has maintained comprehensive national records of immunisation coverage. This is a joint Commonwealth / State project and is centrally administered by the Health Insurance Commission (HIC). The centralised administration and management of statistical information by the ACIR enables national tracking of immunisation coverage.

Coupled with the Immunise Australia: Seven Point Plan (1997), which included several initiatives to increase immunisation rates as well as the eradication of measles, the aim of the ACIR is to promote age appropriate immunisation throughout Australia and to improve the level of immunisation service delivery through provision of payments and information. This enables recognised immunisation service providers and parents to determine the immunisation status of individual children. The ACIR and the HIC also administer financial incentives (see above) to providers to monitor, promote and provide appropriate immunisation services to children under seven in their practices and to report immunisations to the ACIR.

Concurrent with the implementation of the ACIR, the introduction of a national educational campaign on cold chain management improved the handling of vaccines by providers and considerably enhanced the effectiveness of vaccines. Cold chain improvement grew out of the work of the National Immunisation Committee and the publication, Keep it cool (1994).

5.3 The Impact of Immunisation Programs on Measles

Evaluation of the economic impact of measles would ordinarily involve identifying the trend in notifications before and after the introduction of measles vaccine in 1970. By extrapolating the trend in notifications that would likely have occurred without the vaccine, it would be possible to develop a counterfactual scenario that could be compared with what actually occurred with the vaccine.

This approach is not feasible because data on notifications commenced in 1991, some 20 years after the introduction of a vaccine for measles. The best available indicator of the impact of measles vaccine is therefore the series on measles deaths, which started in 1905 (AIHW).

The trend in measles deaths since 1940 (Figure 5.1) reveals a secular decline. This reflects a reduction in case fatality associated with a general improvement in health status as well as the introduction of antibiotics in the late 1940s (Russell, 1988). By fitting a trend to measles deaths for the period 1940–69 and extrapolating it from 1970 onwards, we can estimate the deaths that would have occurred without immunisation (Figure 5.1). A trend can also be fitted to actual deaths that occurred with immunisation (Figure 5.2). The difference between these two trend curves is our estimate of the lives saved because of immunisation (Table 5.4)⁴.

We also estimate a data series for Australian measles notifications by relating Australian deaths to known deaths and notifications in an environment comparable to that in Australia. There are excellent data for this purpose collected in England and Wales, which introduced measles immunisation at the same time as Australia.

Using the case fatality rate for England and Wales we construct a series for Australian measles notifications. This is done in Table 5.3 by multiplying Australian measles deaths by the reciprocal of the England and Wales case fatality rate.

Source: AIHW

Figure 5.1 Australia, measles deaths, 1940-2003

Source: AIHW

Figure 5.2 Australia, measles deaths after immigration, 1970-2003

Source: Table 5.4.

Figure 5.3 Australia, imputed measles notifications immunisation, 1940-2003

Source: Table 5.4.

Figure 5.4 Australia, imputed measles notifications after immunisation, 1970-2003

There is some presumption that secular reductions in case fatality in Australia and England and Wales would have occurred pari passu with comparable overall health improvements occurring simultaneously in each of these countries. The residual case fatality could therefore be attributed to the impact of immunisation. It is accordingly used as tracer for Australian notifications. The difference between the trends fitted to notifications, before and after the subsidised availability of measles immunisation in 1970, is used to represent the impact of immunisation (Figures 5.3 and 5.4 and Table 5.4).

Because of the deficiency in hard data on which the overall projection relies, the forward projection is restricted to 2003. This is five years ahead of the last year for which actual data are available.

Based on our method, between 1970 and 2003, the introduction of measles immunisation could be expected to have saved some 95 lives and to have averted 4.0 million cases (Table 5.4). These values are derived from trend data that smooth considerable year-to-year volatility associated with sporadic outbreaks. Two further points should be borne in mind:

• Imputing the Australian trend data relies on a relatively small number of deaths to generate a much larger number for notifications. In conjunction with random year-to-year variations in the case fatality rate for England and Wales, this may be responsible for considerable instability in estimated Australian notifications. The potential standard errors are significant.⁵

The gradient of the trend in measles notifications is sensitive to the base year selected. This study picked 1940 as base year because it yielded a time series of significant duration before the advent of vaccine. The longer the time series, the greater it's influence in mitigating adverse or propitious base year selection.

To evaluate the economic impact of measles immunisation, lives saved and cases averted, the benefits accruing need to be related to the increasing level of measles coverage, the number of doses involved and their associated cost.

Table 5.3 Estimated Australian measles notifications drawing on England and Wales data

	Australian deaths						Estimated Australian Notifications	Actual
Year	Males	Females	Total	Notifications	Deaths	Rate	Notificationsa	Notifications
1940	69	57	126	409,521	857	0.00209	60,210
1941	11	8	19	409,715	1,145	0.00279	6,799
1942	123	122	245	286,341	458	0.00160	153,174
1943	52	55	107	376,104	773	0.00206	52,061
1944	6	5	11	158,479	243	0.00153	7,174
1945	14	11	25	446,796	729	0.00163	15,322
1946	50	59	109	160,402	204	0.00127	85,705
1947	23	13	36	393,787	644	0.00164	22,013
1948	33	47	80	399,606	327	0.00082	97,763
1949	31	25	56	385,935	307	0.00080	70,399
1950	28	32	60	367,725	221	0.00060	99,835
1951	28	20	48	616,182	317	0.00051	93,302
1952	16	16	32	389,502	141	0.00036	88,398
1953	12	15	27	545,050	242	0.00044	60,811
1954	24	19	43	146,995	45	0.00031	140,462
1955	15	16	31	693,803	174	0.00025	123,609
1956	23	21	44	160,556	28	0.00017	252,302
1957	10	5	15	633,678	94	0.00015	101,119
1958	10	4	14	259,308	49	0.00019	74,088
1959	9	11	20	539,524	98	0.00018	110,107
1960	8	8	16	159,364	31	0.00019	82,252
1961	20	10	30	763,531	152	0.00020	150,697
1962	7	4	11	184,895	39	0.00021	52,150
1963	10	16	26	601,255	127	0.00021	123,092
1964	8	6	14	306,801	73	0.00024	58,839
1965	6	14	20	502,209	115	0.00023	87,341
1966	4	13	17	343,642	80	0.00023	73,024
1967	13	11	24	460,407	99	0.00022	111,614
1968	6	9	15	236,154	51	0.00022	69,457
1969	15	19	34	142,111	36	0.00025	134,216
1970b	4	5	9	307,408	42	0.00014	65,873
1971	11	7	18	135,241	28	0.00021	86,941
1972	4	2	6	145,916	29	0.00020	30,190
1973	2	4	6	152,578	33	0.00022	27,741
1974	6	6	12	109,636	20	0.00018	65,782
1975	4	1	5	143,072	16	0.00011	44,710
1976	7	4	11	55,502	14	0.00025	43,609
1977	4	1	5	173,361	23	0.00013	37,687
1978	3	5	8	124,067	20	0.00016	49,627
1979	3	8	11	77,363	17	0.00022	50,058
1980	3	2	5	139,487	26	0.00019	26,824
1981	2	4	6	52,979	15	0.00028	21,192
1982	2	3	5	94,195	13	0.00014	36,229
1983	1	1	2	103,700	16	0.00015	12,963
1984	2	2	4	62,079	10	0.00016	24,832
1985	4	1	5	97,408	11	0.00011	44,276
1986	3	2	5	82,054	10	0.00012	41,027
1987	1	5	6	42,158	6	0.00014	42,158
1988			6	86,001	13	0.00015	39,693
1989			2	26,222	3	0.00011	17,481
1990			4	13,302	1	0.00008	53,208c
1991			3	9,680	0	0.00000	1,815	1,380
1992			3	10,268	1	0.00010	1,925	1,425
1993			1	9,612	0	0.00000	601	4,536
1994			3	16,375	1	0.00006	1,535	4,895
1995			3	7,447	0	0.00000	698	1,324
1996			0	5,614	0	0.00000	526	498
1997			0	3,962	0	0.00000	371	852
1998			0	3,728	0	0.00000	350	306

a. Australian deaths multiplied by the reciprocal of the England and Wales death rate
b. Immunisation introduced
c. Outlier value ignored in trend analysis.
Source: AIHW; Office of Population Censuses and Surveys (UK); Department of Health (UK); NCIRS

Table 5.4 Trend data for measles deaths and notifications

	Deaths			Notifications
Year	Without	with	Estimated lives	Without	with	Estimated cases
	Immunisationa	immunisationb	savedc	immunisationd	Immunisatione	avertedc
1940	74			21,331
1941	70			29,741
1942	67			36,124
1943	63			41,467
1944	60			46,150
1945	57			50,366
1946	54			54,230
1947	52			57,815
1948	49			61,175
1949	46			64,344
1950	44			67,353
1951	42			70,223
1952	40			72,970
1953	38			75,610
1954	36			78,153
1955	34			80,610
1956	32			82,987
1957	31			85,293
1958	29			87,533
1959	28			89,713
1960	26			91,837
1961	25			93,908
1962	24			95,931
1963	22			97,909
1964	21			99,844
1965	20			101,740
1966	19			103,598
1967	18			105,420
1968	17			107,209
1969	16			108,966
1970	16	10	6	110,693	77,000	33,693
1971	15	10	5	112,391	67,459	44,932
1972	14	10	4	114,061	59,100	54,961
1973	13	10	3	115,706	51,777	63,929
1974	13	9	4	117,325	45,362	71,964
1975	12	9	3	118,921	39,741	79,180
1976	11	9	2	120,494	34,817	85,677
1977	11	8	3	122,044	30,503	91,542
1978	10	8	2	123,574	26,723	96,851
1979	10	7	3	125,083	23,412	101,671
1980	9	7	2	126,573	20,511	106,062
1981	9	7	2	128,044	17,969	110,075
1982	8	6	2	129,497	15,743	113,754
1983	8	6	2	130,932	13,792	117,140
1984	8	6	2	132,351	12,083	120,268
1985	7	5	2	133,753	10,586	123,167
1986	7	5	2	135,139	9,274	125,865
1987	6	4	2	136,511	8,125	128,385
1988	6	4	2	137,867	7,118	130,749
1989	6	4	2	139,209	6,236	132,973
1990	6	4	2	140,537	5,464	135,074
1991	5	3	2	141,852	4,787	137,065
1992	5	3	2	143,153	4,194	138,960
1993	5	2	3	144,442	3,674	140,768
1994	5	2	3	145,719	3,219	142,500
1995	4	2	2	146,983	2,820	144,163
1996	4	1	3	148,236	2,470	145,765
1997	4	1	3	149,477	2,164	147,313
1998	4	0	4	150,707	1,896	148,811
1999	3	0	3	151,927	1,661	150,266
2000	3	0	3	153,136	1,455	151,680
2001	3	0	3	154,335	1,275	153,059
2002	3	0	3	155,523	1,117	154,406
2003	3	0	3	156,702	979	155,723

a From trend fitted to actual deaths prior to immunisation b From trend fitted to actual deaths after the introduction of immunisation c Period and cohort effects not taken into consideration
d From trend fitted to notifications imputed from Australian deaths prior to immunisation
e From trend fitted to notifications imputed from Australian deaths after the introduction of immunisation

Measles immunisation coverage

Apart from the ABS surveys on measles coverage, no systematic national data on measles coverage were collected between 1970 and 1996. Moreover, such data as are available from the ABS conflict with those collected by the ACIR after 1996. For instance, the level of coverage estimated by the ABS in 1995 (91.5 per cent at 24 months, ABS, 1996) exceeds that revealed by ACIR records for 1997 and 1998. Although, as remarked above, the ACIR may have underestimated coverage, it provides the first systematic, ongoing national collection on measles immunisation coverage. This source is therefore used as the anchor for this study.

We accordingly project numbers of measles doses with reference to ACIR recorded coverage for the period 1996 to 1999—the first four years of operations by the ACIR for which full-year data are available. Between 1970 and 1995 we estimate measles coverage by assuming the uptake of the vaccine adopted a monotonic trend over the period 1970 – 1999 (Figure 5.5).⁶ We then estimate actual doses delivered by applying coverage to the age appropriate population cohorts. For the period 1999–2003 we make a simple linear forward projection of doses from doses recorded between 1997 and 1999. The estimated overall trend in measles vaccine doses delivered over the period 1970–2003 is set out in Table 5.5 and Figure 5.6. The striking rise in doses provided between 1994 and 1998—when immunisations grew from 206,000 to 253,000—may be related to the effect of bringing the second measles dose forward in 1999 to 4–5 years, which was part of the impact of the 1998 measles control campaign.

Source: 1996 – 1999, ACIR

Table 5.5 Estimated trend in measles vaccines doses

Year	No	Year	No	Year	No	Year	No
1970	19,480	1979	105,411	1988	168,771	1997	257,876
1971	32,384	1980	113,042	1989	175,240	1998	253,109
1972	43,597	1981	120,490	1990	181,623	1999	246,002
1973	53,837	1982	127,774	1991	187,926	2000	240,455
1974	63,408	1983	134,910	1992	194,152	2001	234,518
1975	72,478	1984	141,911	1993	200,307	2002	228,581
1976	81,152	1985	148,788	1994	206,394	2003	222,644
1977	89,500	1986	155,552	1995	212,416
1978	97,574	1987	162,211	1996	229,922

5.4 Benefits and Costs of Immunisation for Measles

Benefits of immunisation programs

To estimate the benefits from the reduction in the number of measles cases attributable to immunisation we attach a valuation to lives saved, disability averted and resulting treatment savings at constant mid-1990s prices. As we assume that all the deaths are child deaths, a death saved represents the gain of a whole life, which in this study we value at $1 million (Chapter 1).

The estimated cost of treating a measles case without complications is $2982 (see Table 5.6) with a disability weight of 0.152. There are long ongoing costs associated with each of measles encephalitis and sub acute schelerosing panencephalitis, for which the respective (lifetime) disability weights are an estimated 0.338 and 0.938 respectively.

Table 5.6 Characteristics of measles notifications, their sequelae and their cost (mid 1990 $s)

	Prevalence per notification	Estimated case cost	DALY
Uncomplicated measles	0.99896a	$2,982b	0.152c
Measles encephalitis	0.00101a	$10,000 for 20 years @ 5% = $124,622a	0.338c
Sub acute schelerosing panencephalitis	0.00004a	$55,480 for 5 years, 10 years after onset @ 5% = $140,440a	0.938c

Source: (a) Personal communication, Gerard Fitzsimmons, AIHW
(b) Sheill et al (1998); (c) Mathers et al 1999, p 187.

These prevalences, health care costs and disability weights are applied to lives saved and notifications averted (as calculated in Table 5.4). The estimated economics benefits accruing to the community as a whole between 1970 and 2003, discounted back to 1970 at 5 per cent at mid-1990s prices is $9204 million (Table 5.7).

A high proportion of these benefits accrues to government in the form of reduced health care treatment costs. Using a five per cent discount rate, the estimated present value of the savings to government is $8531 million.

Table 5.7 also shows the estimated total benefits and the savings to government with a zero, 3 and 7 per cent discount rate.

Costs of immunisation programs

The cost of providing measles immunisation consists of the following elements:

• Cost of delivering age appropriate dose I (12 months), 1970–2003 by general practitioners, for 56 per cent of doses 1970–98 and for 69 per cent of doses 1999–2003 @ $23.28 per MBS item 23 (standard consultation) (assumes that 80 per cent of services were bulk billed @ Schedule Fee of $26.45⁷) × level of coverage of relevant population cohort⁸

Cost of delivering age appropriate dose II (4- 5 years), 1994–2003 by general practitioners under same conditions as for dose I

Cost of delivering age appropriate dose I (12 months), 1970–2003 by non-GP immunisation providers for 44% of doses, 1970–98 and for 31% of doses, 1999–2003 @ $11.64 (an estimated half the cost of a GP service) × level of coverage of relevant population cohort

Cost of delivering age appropriate dose II (4-5 years), 1994–2003 by non-GP immunisation providers under same conditions as for dose I

Cost of vaccine for doses 1 and II @ $2.02, assuming the measles component of the MMR dose @ ¹/₃ of unit cost used by DHAC for purposes of PHOFA vaccine appropriations (DHAC, 2000a) × level of coverage of relevant population cohort.

Measles component of the cost of ACIR administration, ACIR reporting and GPII SIP and outcome payments, and payments to divisions of general practice, weighted in accordance with the share that measles vaccine bears in relation to the cost of all childhood vaccine doses listed on the Australian Standard Vaccination Schedule

The program costs associated with the Measles Control Campaign, August–November, 1998

A vaccine reaction rate of 4.1 per million immunisations, costed at $100 per reaction⁹

Some campaigns initiated at State/Territory levels may not be captured by the costs identified above. However, no data for these costs are centrally collected, let alone capable of attribution to measles. The above calculations allow for most State/Territory delivery costs and the cost of VIVAS in Queensland is captured in 50 per cent of Queensland’s ACIR reporting fee.

The costs of the immunisation programs are summarised in Table 5.8. Their estimated NPV, discounting back to 1970 at 5 per cent, is approximately $55 million. Table 5.8 also shows the estimated total costs of the programs with a zero, 3 and 7 per cent discount rate.

Overall evaluation of benefits and costs: central case

Adopting a 5 per cent discount rate, and drawing on our most likely assumptions, the estimated net benefit of the measles immunisation programs to the community is about $9.1 billion (Table 5.9).

Drawing on the same inputs, the estimated net benefit of the measles immunisation programs to the government sector is in the order of $8.5 billion.

Sensitivity analysis

There are uncertainties in the cost elements associated with variability in doses administered (due to year to year variability in coverage) and with variability in proportions of immunisations delivered by general practitioners and local councils. To account for this, we allow that costs may have been 25 per cent higher or lower than in our central case. Thus total discounted costs of the programs, with a 5 per cent discount rate, could be between $39 million and $66 million (see Table 5.8).

On the benefit side, there are potentially significant errors arising because of the way in which we have imputed the decline in the Australian morbidity from measles from the case fatality rate in England and Wales. Despite the likely similarities between the two sets of case fatality rates, there would inevitably be random discrepancies between the rates. There may also have been variations in measles case severity and their associated treatment costs. These variations are allowed for by varying the estimated benefits by 25 per cent in Table 5.7. Thus the total discounted benefits could be in the range $6.9 billion to $11.5 billion.

Clearly, the net benefit of the measles immunisation program has been high whatever the possible variations in assumptions that might be made (see Table 5.9).

Table 5.7 Estimated benefits from measles immunisation, $m (mid-1990 $s)

				Disability savings
	Treatment Savings	Death savings	Sequelai Savings	No complications	Complications	Total	Savings to government
1970	100.5	5.7	4.2	70.2	12.8	193.3	174.9
1971	134.0	4.9	5.7	93.6	17.1	255.2	233.2
1972	163.9	4.1	6.9	114.4	20.9	310.3	285.2
1973	190.6	3.4	8.1	133.1	24.3	359.5	331.8
1974	214.6	3.7	9.1	149.8	27.4	404.6	373.5
1975	236.1	3.1	10.0	164.9	30.1	444.1	411.0
1976	255.5	2.5	10.8	178.4	32.6	479.7	444.7
1977	273.0	2.9	11.5	190.6	34.8	512.8	475.1
1978	288.8	2.3	12.2	201.7	36.8	541.8	502.7
1979	303.2	2.8	12.8	211.7	38.7	569.2	527.7
1980	316.3	2.3	13.4	220.8	40.3	593.1	550.5
1981	328.2	1.9	13.9	229.2	41.8	615.0	571.3
1982	339.2	2.4	14.3	236.9	43.2	636.1	590.4
1983	349.3	2.0	14.8	243.9	44.5	654.5	608.0
1984	358.6	1.6	15.2	250.4	45.7	671.5	624.2
1985	367.3	2.2	15.5	256.5	46.8	688.3	639.3
1986	375.3	1.8	15.9	262.1	47.9	703.0	653.3
1987	382.8	2.5	16.2	267.3	48.8	717.6	666.3
1988	389.9	2.2	16.5	272.2	49.7	730.5	678.6
1989	396.5	1.9	16.8	276.9	50.6	742.6	690.1
1990	402.8	1.6	17.0	281.2	51.4	754.0	701.1
1991	408.7	2.3	17.3	285.4	52.1	765.8	711.4
1992	414.4	2.0	17.5	289.3	52.8	776.1	721.2
1993	419.8	2.8	17.7	293.1	53.5	786.9	730.6
1994	424.9	2.5	18.0	296.7	54.2	796.3	739.6
1995	429.9	2.3	18.2	300.2	54.8	805.3	748.2
1996	434.7	3.1	18.4	303.5	55.4	815.0	756.5
1997	439.3	2.9	18.6	306.7	56.0	823.5	764.6
1998	443.8	3.7	18.7	309.9	56.6	832.6	772.4
1999	448.1	3.5	18.9	312.9	57.1	840.5	779.9
2000	452.3	3.3	19.1	315.8	57.7	848.2	787.2
2001	456.4	3.1	19.3	318.7	58.2	855.7	794.4
2002	460.4	3.0	19.5	321.5	58.7	863.1	801.4
2003	464.4	2.8	19.6	324.2	59.2	870.3	808.2
PV@5%	4,901.6	48.2	207.1	3,422.5	624.9	9,204.3	8,531.1
+25%	6,127.0	60.2	258.9	4,278.0	781.1	11,505.2	10,663.9
-25%	3,676.2	36.1	155.3	2,566.8	468.7	6,903.1	6,398.3
PV@ 0%	11,863.6	94.9	501.2	8,283.6	1,512.5	22,255.7	20,648.4
PV@3%	6,780.7	61.0	286.5	4,734.5	864.5	12,727.2	11,801.7
PV@7%	3,671.8	39.5	155.1	2,563.8	468.1	6,898.3	6,390.7

Table 5.8 Cost of providing measles immunisation, $s’000s (mid-1990 $s)

	Measles dose delivery costs				Vaccine
	I GPs	I Others	II GPs	II Others	I	II	ACIR, etc	Campaign	Scale effects	TOTAL
1970	254	99	0	0	39	0	0	0	0.008	393
1971	422	165	0	0	65	0	0	0	0.008	653
1972	568	223	0	0	88	0	0	0	0.008	879
1973	701	275	0	0	108	0	0	0	0.008	1,086
1974	826	324	0	0	128	0	0	0	0.008	1,279
1975	944	371	0	0	146	0	0	0	0.008	1,462
1976	1,058	415	0	0	163	0	0	0	0.008	1,637
1977	1,166	458	0	0	180	0	0	0	0.008	1,806
1978	1,272	499	0	0	197	0	0	0	0.008	1,968
1979	1,374	539	0	0	212	0	0	0	0.008	2,127
1980	1,473	579	0	0	228	0	0	0	0.008	2,281
1981	1,570	617	0	0	243	0	0	0	0.008	2,431
1982	1,665	654	0	0	258	0	0	0	0.008	2,578
1983	1,758	691	0	0	272	0	0	0	0.008	2,722
1984	1,850	726	0	0	286	0	0	0	0.008	2,863
1985	1,939	762	0	0	300	0	0	0	0.008	3,002
1986	2,027	796	0	0	314	0	0	0	0.008	3,138
1987	2,114	830	0	0	327	0	0	0	0.008	3,273
1988	2,200	864	0	0	340	0	0	0	0.008	3,405
1989	2,284	897	0	0	354	0	0	0	0.008	3,536
1990	2,367	930	0	0	366	0	0	0	0.008	3,664
1991	2,450	962	0	0	379	0	0	0	0.008	3,792
1992	2,531	994	0	0	392	0	0	0	0.008	3,917
1993	2,611	1,025	0	0	404	0	0	0	0.008	4,041
1994	2,690	1,057	1,183	465	416	366	0	0	0.008	6,180
1995	2,769	1,087	1,225	481	429	379	0	0	0.008	6,372
1996	2,997	1,177	1,265	497	464	392	162	0	0.008	6,957
1997	3,361	1,320	1,305	512	520	404	274	0	0.008	7,701
1998	3,299	1,296	1,345	528	511	416	391	30,000	0.008	37,789
1999	3,951	887	1,706	383	496	429	1,130	0	0.008	8,984
2000	3,862	867	1,846	414	485	464	1,416	0	0.008	9,358
2001	3,767	846	2,071	465	473	520	1,409	0	0.008	9,553
2002	3,671	824	2,032	456	461	511	1,549	0	0.008	9,508
2003	3,576	803	1,978	443	449	496	1,624	0	0.008	9,370
PV @5%	26,360	9,687	3,711	1,118	3,968	1,032	1,696	7,288	0.130	54,865
+ 25%	32,950	12,109	2,932	727	4,960	761	2,121	9,110	0.162	65,673
- 25%	19,770	7,265	1,759	436	2,976	456	1,272	5,466	0.097	39,404
PV@0%	71,383	24,876	10,977	2,692	10,510	2,839	7,958	30,000	0.273	161,239
PV@3%	38,155	13,756	4,306	1,063	5,698	1,116	3,117	12,730	0.170	79,945
PV@7%	18,893	7,056	1,293	322	2,863	336	935	4,216	0.103	35,918

Table 5.9: Summary of results for measles immunisations $m (mid-1990 $s)

	Costs	Benefits	Net benefits
NPV @ 5%	54.9	9,204.3	9,149.4
+ 25%	65.7	11,505.2	11,439.5
- 25%	39.4	6,903.1	6,863.7
NPV @ 3%	79.9	12,727.2	12,647.3
NPV @ 7%	35.9	6,898.7	6,862.6

5.5 The Impact of Immunisation Programs on Hib Disease

Because the vaccine for Hib disease was adopted after systematic national data collection on notifiable diseases commenced, the task of making forward projections is less speculative than for measles.

One consideration in identifying Hib cases is that the number and rate of hospitalisations have been consistently higher than for notifications. McIntyre et al.(2000, p.12) report that:

… notifications probably underestimate Hib cases (especially cases of epiglottitis, which are often diagnosed clinically) while hospitalisations probably overestimate them. This is because the hospitalisation codes (a) lack specificity (as non-Hib cases may be included) and (b) may count cases twice if the patient is transferred to another hospital (more frequent for epiglottitis). Overall it is likely that notifications, because they are usually linked to laboratory identification of Hib, more closely represent the true incidence of Hib disease than hospitalisations.

This study accordingly uses notification data as the basis for evaluating the impact of Hib immunisation. Moreover, following McIntyre et al (2000), only children 0–14 years of age are included. This is because vaccines are targeted at reducing disease in children and the diagnosis of children for Hib infection is likely to be more specific.

Monthly plots of Hib disease notifications are set out in Figures 5.7 and 5.8, commencing in 1991 and projected forward until 2003. There is a noticeable fall in notifications after the introduction of Hib vaccine in February 1993—from some 39 cases per month between 1991–92 to about 9 cases between 1993–98. To develop a counterfactual proposition, separate trends are fitted to the monthly series both before and after the introduction of Hib vaccine. In the latter case, a forward projection is made until the year 2003. These trend data are collapsed into annual observations in Table 5.10, from which it will be seen that, between 1993–98, 3614 Hib cases are likely to be averted and about 87 lives saved.

Source: NCIRS

Source: NCIRS

Table 5.10 Numeric trend data for Hib disease deaths and notifications before and after immunisation

	Notifications		Apparent cases	Deaths		Apparent lives
	Before	After	Averted1	Before	After	Saved1
1991	432			10
1992	434			10
1993	436	230	206	10	6	5
1994	439	136	303	11	3	7
1995	441	129	312	11	3	7
1996	443	122	321	11	3	8
1997	445	116	330	11	3	8
1998	448	110	338	11	3	8
1999	450	104	346	11	2	8
2000	452	99	354	11	2	8
2001	455	93	361	11	2	9
2002	457	88	368	11	2	9
2003	459	84	375	11	2	9

¹ Period and cohort effects not taken into consideration
Hib disease immunisation coverage

Increasing levels of Hib immunisation coverage underpin these favourable trends in mortality and morbidity. The trend in Hib doses associated with increasing levels of coverage is linearly projected backwards to 1993 and forwards to 2003, from observations for the four full-years of ACIR reporting between 1996 and 1999 (Figure 5.9). The number of Hib doses peaked in 1997 and thereafter the number has slightly diminished, as ‘catch up’ campaigns were completed.

Source: 1996 – 1999, ACIR

5.6 Benefits and Costs of Immunisation for Hib Disease

Benefits of reduced Hib incidence

We estimate the benefits from the reduction in Hib morbidity by allocating projected cases averted to three classes of Hib disease, to death, and to four classes of disability (severe disability, less severe disability, hearing aid and less severe hearing loss) in accordance with point prevalences in Table 5.11. As for measles, because all deaths are child deaths, deaths averted are valued in this study at $1 million. We also ascribe quality of life indices to the four classes of disability to account for time avoided at less than full health.

The estimated benefits are summarised in Table 5.12. Discounted back to 1993 at 5 per cent, the present value of the estimated benefits to the community is approximately $165 million. A high proportion of these benefits is attributed to averted deaths and improved quality of life. The present value (in 1993) of the estimated savings to government is $45 million. The estimated savings are also shown with discount rates of 0, 3 and 7 per cent.

Table 5.11 Point prevalence per notification of Hib type disease, their sequelae and their cost

	Prevalence	Estimated case cost	Quality of life index
Type
Meningitis	0.4608	$7,380
Epiglottitis	0.36	$5,789
Other	0.1792	$2,130
Sequelae
Death	0.024	$1.0m	0
Severe disability	0.008	$55,480 for 50 years @ 5% = $1.013m	-0.12
Less severe disability	0.0368	$45,400 for 20 years @ 5% = $65,000	0.80
Hearing aid	0.0176	(Cochlear) $16,000	0.86
Less severe hearing loss	0.0176	(Assessment) $1,470	0.91

Source: McIntyre et al (1994).

Table 5.12 Benefits from Hib disease immunisation, $’000s

	Cases	Treatment	Death	Sequelae	Improved quality	TOTAL	Total savings
	averted	savings	savings	savings	of life	benefits	to govta
1993	206	1,208	4,944	2,225	4,328	12,707	3,434
1994	303	1,775	7,264	3,270	6,358	18,668	5,045
1995	312	1,830	7,488	3,371	6,555	19,245	5,201
1996	321	1,883	7,704	3,468	6,743	19,799	5,351
1997	330	1,933	7,911	3,561	6,925	20,332	5,495
1998	338	1,982	8,111	3,651	7,100	20,845	5,633
1999	346	2,029	8,303	3,737	7,268	21,339	5,767
2000	354	2,075	8,489	3,821	7,430	21,816	5,896
2001	361	2,118	8,668	3,901	7,587	22,275	6,020
2002	368	2,161	8,840	3,979	7,738	22,719	6,140
2003	375	2,201	9,007	4,054	7,884	23,148	6,256
PV @5%	-	15,713	64,285	28,937	56,271	165,207	44,651
+25%		19,642	80,356	36,172	70,338	206,509	55,814
-25%		11,785	48,214	21,703	42,203	123,905	33,488
PV @0%		21,200	86,733	39,042	75,920	222,897	60,243
PV@3%		17,634	72,145	32,475	63,151	185,407	50,110
PV@7%		14,080	57,602	25,929	50,421	148,033	40,009

Costs of immunisation programs for Hib disease

The cost of providing Hib immunisations consists of the following elements:

• Cost of delivering age appropriate doses I–IV of HibTITER vaccine (2, 4, 6 and 18 months) by general practitioners for 56 per cent of doses 1993–98 and for 69 per cent of doses 1999–2003 @ $23.28 per MBS item 23 (standard consultation). This assumes that 80 per cent of services were bulk billed @ Schedule Fee of $26.45) × ¹/₃ (because HibTITER doses are delivered concurrently with other Schedule vaccines) × number of doses from Figure 5.9.
• Cost of delivering age appropriate doses I–IV of HibTITER vaccine (2, 4, 6 and 18 months), 1993–2003 by non-GP immunisation providers for 44% of doses, 1993–98 and for 31 per cent of doses, 1999–2003 @ $11.64 (an estimated half the cost of a GP service) under the same conditions as general practitioners
• Cost of vaccine for doses I-IV @ $9.20 per dose, the price used by DHAC for purposes of PHOFA vaccine appropriations (DHAC, 2000a), × number of doses from Figure 5.9.
• Hib component of the cost of ACIR administration, ACIR reporting and GPII SIP and outcome payments, weighted in accordance with the share that Hib vaccine bears in relation to the cost of all vaccine doses listed on the Australian Standard Vaccination Schedule
• A vaccine reaction rate of 0.0126 per immunisation, costed at $25 per reaction (McIntyre et al., 1994) × number of doses.

These costs are set out in Table 5.13. Discounted back to 1993 at 5 per cent, their present value is approximately $153 million. Estimated present values with 0, 3 and 7 per cent discount rates are also shown.

Table 5.13 Cost of providing Hib disease immunisation, $‘000s

	GPs	Others	Vaccine	ACIR, etc	Side effects	TOTAL
1993	2,978	1,169	6,304	0	215.9	10,668
1994	3,275	1,286	6,934	0	237.4	11,734
1995	3,573	1,403	7,564	0	259.0	12,800
1996	3,870	1,520	8,194	1,248	280.6	15,115
1997	4,168	1,637	8,824	2,103	302.1	17,035
1998	4,136	1,624	8,756	3,001	299.8	17,818
1999	5,116	1,149	8,790	8,663	301.0	24,020
2000	5,096	1,144	8,756	10,858	299.8	26,156
2001	5,086	1,142	8,739	10,804	299.2	26,071
2002	5,076	1,140	8,722	11,875	298.6	27,113
2003	5,066	1,138	8,705	12,456	298.1	27,664
PV@5%	34,873	10,934	67,270	39,959	2,303	155,341
+25%	43,591	13,667	84,088	49,949	2,879	194,176
-25%	26,154	8,200	50,453	29,969	1,727	116,506
PV @ 0%	47,443	14,358	90,293	61,012	3,091	216,198
PV@3%	39,265	12,142	75,343	47,169	2,579	176,500
PV@7%	31,148	9,897	60,396	33,998	2,067	137,508

Overall evaluation of benefits and costs: central case

Adopting a 5 per cent discount rate, and drawing on our most likely assumptions, the estimated net benefit of the Hib immunisation programs to the community as a whole is approximately $10 million (see Table 5.14).

However, because a high proportion of the benefits accrued in averted deaths and improved quality of life, the to the community, the present value of the estimated savings to government ($45 million) were below the present vale of the costs of the programs ($155 million).

Table 5.14: Summary of results for Hib immunisations $m (mid-1990 $s)

	Costs	Benefits	Net benefits
NPV @ 5%	155.3	165.2	9.9
+ 25%	194.2	206.5	11.3
- 25%	116.5	123.9	7.4
NPV @ 3%	176.5	185.4	8.9
NPV @ 7%	137.5	148.0	10.5

Sensitivity analysis

Unlike the analysis for measles, the results are sensitive to the assumptions made. For the costs of the programs we allow conservatively for a 25 per cent range due to possible variations in doses administered and in proportions of immunisations delivered by general practitioners and local councils. With a 5 per cent discount rate, the present value of the costs ranges from $117 million to $194 million.

There is also uncertainty about the Hib related diseases that would have occurred in the absence of immunisation (Figure 5.7). The projection of the ‘no-immunisation’ case relies on a small period of pre-immunisation experience. The explanatory power of the trend fitted in Figure 5.7 is therefore low. McIntye et al. (1994) also remark on the uncertainties associated with Hib diseases and their severity. Again, we allow for a 25 per cent variation in the estimated benefits. As shown in Table 5.14, the present value of the total economic benefits varies from $124 million to $207 million.

5.7 Conclusions

The study reports on the evaluation of two immunisation programs: immunisations for measles from 1970 to 2003 and immunisations for Hib disease from 1991 to 2003.

Between 1970 and 2000, measles notifications fell from around 100 000 cases a year to under 2000 cases a year. The introduction of subsidised immunisations for measles in 1970 saved an estimated 95 lives and averted approximately 4.0 million cases between 1970 and 2003. The estimated net present value of these immunisation programs for measles is about $9.2 billion.

On the other hand, the estimated present value of the costs of the measles immunisation programs was only $52 million. Therefore the net benefit of the program exceeded $9.1 billion. The overall result is not sensitive to the assumptions made.

In the case of measles immunisation programs, savings in health treatment expenditures accounted for a high proportion of the estimated benefits. Accordingly, the government made very substantial savings as a result of these immunisation programs.

Hib notifications also fell substantially after the introduction of the Hib immunisation programs. Our study estimates that approximately 350 cases were averted annually in the 1990s. We estimate that the introduction of immunisation for Hib-related diseases saved 78 lives and averted 3600 cases between 1993 and 2003.

In the central case evaluation, the estimated present value of the Hib immunisation programs is $165 million. On the other hand, the present value of the expenditure on the programs is $155 million. The estimated net present value of the immunisation programs for Hib disease is a marginal $10 million. However, this result is sensitive to the assumptions made, especially to the estimates to the benefits of the Hib immunisation programs.

In the case of Hib immunisation programs, savings in health treatment expenditures accounted for about a third of the estimated benefits. Accordingly, the government recovered less than it outlaid in expenditures on these immunisation programs.

¹ These two immunisation programs were chosen after consultation with the National Centre for Immunisation Research and Surveillance and agreed by the Department of Health and Aged Care following the consultant's Study Method and Design Report (Applied Economics, 1999).

² Personal communication, Gerard Fitzsimmons, AIHW.

³ Although the ACT and NT also maintain their own registers, ACT and NT providers receive the full $6.00 reporting fee.

⁴ The implicit assumption of the counterfactual projection in the left hand column in Table 5.3 is that measles mortality increases with population growth, hence that the age distribution of the population from 1970 onwards is roughly comparable with that for 1940– 69. The effect of the post- war baby boom may exaggerate lives saved; low birth rates during the war would have had the opposite effect.

⁵ Note, for example, that a significant outlier value for imputed measles notifications in 1990 (Table 5.3) is excluded for purposes of fitting the trend to estimated notifications after immunisation (Figure 5.4).

⁶ The extrapolation in Figure 5.5 is not intended to imply that immunisation coverage is expected to attain 100 per cent. The highest attainable coverage is probably about 95 per cent.

⁷ Medicare Benefit Schedule Fee as at October 2000.

⁸ This costing is an overestimate in so far as it ignores immunisations given ‘opportunistically’ during consultations occasioned for reasons unrelated to immunisation. The DHAC advises that 23 per cent of immunisations in children below 15 years of age are given at a consultation where another problem was addressed.

⁹ During the Australian measles control campaign 1998, 1.7 million doses of MMR vaccine were used. There was a very low incidence of adverse effects during the campaign with only 7 per cent of children with anaphylaxis / anaphylactoid reactions, each of whom recovered fully following administration of adrenalin (NHMRC, 2000). Based on this result, a reaction rate of 4.1 per million was adopted in this report. The costing of each reaction rate at $100 to cover the cost of administering adrenalin and associated care is a notional figure adopted by the consultant in the absence of other cost information. This reaction cost does not necessarily reflect the views of the DHAC. The consultant has been advised that hypersensitive reactions such as wheal and flare reaction and uriticaria at the injection site can also occur because of MMR vaccinations, However such reactions are mild and rare and are not included in this analysis.