Objective: To model the incremental costs and benefits of a universal antenatal HIV screening programme in New Zealand (NZ).
Design: Cost effectiveness analysis, including only health service costs, using secondary data sources and expert opinion. Uncertainty assessed in multi-way sensitivity analyses.
Setting: The NZ Health Care System.
Subjects: Antenatal population of NZ.
Intervention: Universal antenatal HIV screening programme.
Main outcome measures: incremental cost per true-positive HIV case detected in mothers; incremental cost per HIV case avoided in babies; and incremental cost per discounted life-year gained, for mothers and babies, due to screening.
Results: Using base case values the application of universal screening would cost an additional $NZ 723 607 ($US 307 917) and would lead to the identification of an additional 6.25 true-positive women. After terminations have been excluded, the screening programme would detect five HIV exposed babies. There would be 1.15 avoided cases of HIV infection in babies and a net gain of 41.97 discounted life-years, for mothers and babies combined. The cost per incremental HIV-positive woman detected was $NZ 115 859 ($US 49 301), HIV infected baby avoided $NZ 629 669 ($US 267 944) and discounted life-year gained $NZ 17 241 ($US 7336).
Conclusion: The discounted cost per life gained in NZ compares favourably to estimates reported in studies of similar interventions in other developed countries and other health care interventions in NZ. The decision of whether to implement universal screening in NZ would be clarified if the prevalence of antenatal HIV infection was known and policy makers identified their willingness to pay for an additional life-year gained.
From the University of Auckland, Auckland, New Zealand, the aSchool of Public Health, Queensland University of Technology, Queensland, Australia, and the bDepartment of Public Health and Policy, London School of Hygiene and Tropical Medicine, London, UK.
See also p755
Correspondence to Dr Dale Bramley, Dept. of Community Health, University of Auckland, Private Bag 92019, Auckland, New Zealand. Tel: +64 9 373 7599; Fax: +649 373 7494; e-mail: firstname.lastname@example.org
Received: 15 March 2002; revised: 13 September 2002; accepted: 11 November 2002.
If HIV-positive women are identified early in their pregnancy and follow the appropriate treatment regimes, deliver by Cesarean section and avoid breast-feeding [1–3] the risk of vertical transmission of HIV to their unborn children can be reduced from 14–33% [1,4] to 1–4% [4–7]. Universal antenatal screening for HIV in pregnancy is now advocated in a number of countries [8–10] and whether or not this should be introduced to New Zealand (NZ) is currently being debated [11–15]. Existing screening activities are based on a circular letter sent in 1997 from the NZ Ministry of Health to health care professionals . This recommends a number of screening questions and suggests that those deemed to be at high risk are offered an HIV antibody test. However, research suggests that health care professionals have not followed these guidelines . As a result 10 infants, four of whom were born in NZ, have been diagnosed with HIV since 1998  (N. Dickson, 2002, personal communication). The aim of this research is to develop a model of the incremental costs and benefits associated with the introduction of universal antenatal screening, compared to the current screening activities. The proposed intervention of universal routine screening will require that all pregnant mothers are offered an HIV test as part of their antenatal care and that treatments be instigated for any true positives to minimize the risk of HIV transmission to their babies. The incremental cost per true-positive HIV case detected in mothers, incremental cost per HIV infection avoided in babies, and the incremental cost per discounted life-year gained, for the sum of the discounted life-years gained for mothers and babies, are estimated. A systematic review of the literature on the cost effectiveness of public health interventions for primary HIV prevention programs  found wide variation in the quality of published studies; its findings were used to inform the design of this study.
Estimates are derived of the incremental cost of testing all pregnant mothers for HIV, re-testing true and false positives and the cost of health care services supplied to true positives and their babies. The savings in the lifetime care costs due to an avoided case of HIV in a baby are deducted from these to derive a net cost of the programme. Against the net incremental cost is balanced the incremental benefit that arises from the identification of true positives. These are the net discounted life-years gained from the avoidance of HIV infection in babies, the life-years gained from the earlier diagnosis of the HIV-positive mothers and the net discounted life-years gained from the early onset of treatment for HIV-positive babies. All assumptions made relate to the base year for the analysis, 1999. Only costs incurred by the health service were included in the analysis and future costs were discounted at 5% and life-years at 2% in line with existing guidance [19,20]. The values used for the variables included in the model are included in Table 1. The model was developed in the MS Excel spreadsheet software.
Identification of incremental benefits
It was assumed that only a proportion of pregnant women would agree to enter the screening programme. HIV prevalence and false-positive rates were applied to the screened population to predict the number of true and false positives. Rates of false positives were derived from the NZ blood donor population false-positive HIV antibody rate (K. Croxson, 2002, personal communication). Exact HIV prevalence is not known in NZ and so expert opinion was used to identify a likely range of values. The net true positives likely to be identified from universal screening activities are the total true positives less the number identified from existing screening activities. Estimates of the transmission rates with and without treatment were applied to predict the likely number of cases of HIV infection avoided. Net life-years gained were assumed to be the gain in life expectancy for the babies who avoided HIV infection due to the universal screening intervention plus the gain in life expectancy for HIV mothers and babies due to earlier onset of treatment. It was assumed that treatment with antiretroviral drugs causes no adverse effects to the health of uninfected infants [21,22]. Treatment is based on the HIV paediatric protocol for Auckland, which recommends that babies be given zidovudine for the first postnatal 6 weeks . There is currently minimal data available regarding the potential long-term side effects of in utero exposure to combivir and nevirapine, the standard antenatal treatment for women infected with HIV in NZ .
Identification of incremental costs
It was assumed that all those who accepted entry to the screening programme would receive pre-test counselling, an HIV antibody test, and would attend for a consultation to receive the result. All true and false positives would require two further HIV antibody tests and a Western Blot test . True positives would attend for post-test counselling and a proportion would choose to terminate the pregnancy . True positives would be started on appropriate medical therapy, combivir and nevirapine from 28 weeks’ gestation, and would receive the appropriate medical investigations . They would attend for outpatient appointments and would deliver by Cesarean section during which time intravenous zidovudine is administered. After delivery they would continue to receive appropriate therapeutic and diagnostic interventions. Clinical protocols for the care of these women were developed by HIV infectious disease experts at Auckland Hospital and reflects the standard care given to HIV infected woman post-delivery. The babies born to the true positives would be provided with the appropriate medical treatment, Zidovudine with or without Lamivudine, and diagnostic interventions and would attend for outpatient consultations. These treatments are based on NZ specific paediatric clinical protocols . Costs for treating babies and mothers were counted for a defined period after birth to a point when it was assumed that both mothers and babies would have been identified regardless of a universal screening programme . The discounted lifetime care costs of treating a case of HIV infection were deducted from the gross costs of the screening and treatment programmes to derive the net cost of the intervention.
Base case, favourable and unfavourable values and the sources used for all variables are summarized in Table 1. If the application of the ‘favourable’ or ‘unfavourable’ value caused a change of > 5% in the cost per outcome the variable was included in subsequent sensitivity analyses.
Using base case values the application of universal screening would cost an additional $NZ 723 607 ($US 307 917) and would lead to the identification of 14.25 true-positive women per year. It was estimated that eight of these would have been detected anyway due to the current screening procedures; this is based on the average number of women identified over the last 2 years (L. Wilkinson, 2002, personal communication). Thus, 6.25 cases would be the net true positives of whom 1.25 would choose to terminate the pregnancy. The screening programme will identify 5.00 babies exposed to HIV. Appropriate treatment would result in 1.15 cases of avoided HIV infection in babies and a net gain of 41.97 discounted life-years. The incremental cost per case detected would be $NZ 115 859 ($US 49 301), per case of HIV infection avoided in babies $NZ 629 669 ($US 267 944), and per discounted life-year gained $NZ 17 241 ($US 7336).
Favourable and unfavourable values for the variables marked by the footnote ‘a’ in Table 1 were applied to the model and the results are presented in Table 2.
The results were most sensitive to changes in the assumptions over the prevalence of HIV. Selected two-way sensitivity analyses were undertaken. In particular, values for the cost per discounted life-year gained were plotted for all prevalence rates between the favourable and unfavourable values for: the incidence of transmission without treatment and the incidence of transmission with treatment (Fig. 1); the percentage of the antenatal population who accept screening and the life expectancy of an HIV baby (Fig. 2); the duration of pre-test counselling and the lifetime care costs for an HIV baby (Fig. 3).
Due to the paucity of data regarding some key variables, assumptions were made that were tested in the sensitivity analyses. Varying the prevalence rate had a dramatic effect on the cost per discounted life-year gained, with a best case of $NZ 10 433 ($US 4439) and a worst case of $NZ 53 554 ($US 22 789) per discounted life-year gained. More precise results could be elicited from this model if the prevalence of HIV infection among pregnant women was known. While this would be relatively easy through the anonymous unlinked testing of antenatal maternal blood or neonatal blood spots it has not been undertaken in NZ due to a variety of legal, ethical and financial concerns. Further scenarios were tested in the two-way sensitivity analyses included in Figs 1, 2 and 3.
The results of this analysis should be considered in the light of findings from similar studies. A study of universal screening in the UK  found that 6.39 discounted life-years were gained at a net cost of £ 14 833 ($NZ 49 643, $US 21 125) yielding a cost per discounted life-year gained of £ 2321 ($NZ 7658, $US 3305). The authors assumed that a discounted life-year gained was worth £ 10 000 ($NZ 33 511, $US 14 242) and concluded that universal screening should be adopted throughout the UK, as long as the cost and uptake of the test could be maintained within certain ranges. Another UK study  found the cost per discounted life-year gained to be £ 4000 ($NZ 13 404, $US 5696) but for areas with comparatively low prevalence rates this increased to £ 20 000 ($NZ 66 000, $US 28 484). A study in the USA  of the cost effectiveness of prenatal screening estimated the cost per discounted life-year gained to be $US 8900 ($NZ 21 831). A further USA study  that valued each discounted life-year gained at $US 50 000 ($NZ 115 814), not surprisingly concluded that universal screening of pregnant women in Chicago was cost effective. The results reported here compare favourably with these findings.
To set these results in a NZ context, studies of other public health interventions, that reported a cost per discounted quality-adjusted life-year (or QALY) gained, are reviewed. A study of the cost effectiveness of therapies for end stage renal disease found the cost per life-year gained for health centre dialysis, home haemodialysis, continuous ambulatory peritoneal dialysis (CAPD) at hospital, and transplantation lay within the range $NZ 18 463–35 270 . A study of the cost effectiveness of mammography screening  found a cost per QALY of between $NZ 12 668 ($US 5390) and $NZ 14 500 ($US 6170). The cost effectiveness of a policy of compulsory bicycle  helmets found that the cost per life-year gained for the age group 5–12 years was $NZ 100 000 ($US 42 552), for 13–18 years $NZ 75 000 ($US 31 914) and for adults, $NZ 950 000 ($US 404 244). If we knew the policy makers’ maximum willingness to pay for an additional year of life, we could use this as a basis for accepting or rejecting interventions. For example, if a discounted life-year gained was judged to be worth $NZ 40 000, then all scenarios that yielded a cost per discounted life-year gained of less than $NZ 40 000 would be accepted as cost effective. A consensus over the value of a discounted life-year is emerging with £ 30 000 ($NZ 100 533) [35,36] suggested for the UK, $CAN 20 000–100 000 ($NZ 29 551–147 558)  suggested for Canada, $US 50 000 ($NZ 117 498)  suggested for the USA, $AU 42 000–76 000  suggested for Australia, and the lowest estimate from a comprehensive review of the value of life literature was $US 24 777 ($NZ 58 225) . To date, no estimates of the value of a life-year gained have been proposed for NZ, however alternative benchmarks exist. For example, an intervention that results in a cost per life-year gained less than the per capita gross national product of the study's country, is sometimes considered to be cost effective. In NZ, for the year 2000, gross national product per capita was estimated to be $NZ 30 951 . By returning to Figs 1, 2, and 3, and reading across from a value of $NZ 30 951 on the ordinate, the cost effective scenarios can be identified. These are as follows: if transmission without treatment was 33% or 14% then prevalence would have to be 0.0222% and 0.0308 respectively; if transmission with treatment was 1% or 4% then prevalence would have to be 0.0238% and 0.0266% respectively; if 70% or 100% accepted screening then prevalence would need to be 0.0278% or 0.0216% respectively; if the life expectancy of an HIV baby was 4.71 years or 19.52 years then the prevalence would need to be 0.0232% and 0.0274% respectively; if the lifetime care costs for an HIV child were $NZ 80 923 or $NZ 404 617 then the prevalence would need to be 0.0246% and 0.0228% respectively; and if pre-test counselling took 1.2 min or 20 min then the prevalence would need to be 0.0236% or 0.0306% respectively. Similar analyses can be performed for other values of the cost per discounted life-year gained in order to identify the parameter values required to produce such a result.
This cost analysis adopts a limited perspective including only the direct economic costs incurred by the health care sector. Many other costs may be relevant, e.g., the health care sector may incur costs in setting up, managing, evaluating and publicizing a universal screening programme. There may well be additional non-market costs borne by those participating in the programme, such as stress and anxiety associated with taking the HIV test and waiting for the results. It would also be important to have a clear and appropriate policy for the management of false positives. Regarding the uptake of screening, the model relies on an assumption that the prevalence of HIV is consistent across those who decline and those who agree to participate in the screening programme. Finally, while issues around informed consent for HIV testing also need to be addressed, a recent survey of 51 NZ women found that routine antenatal HIV screening was acceptable .
The analysis included here is unique, offers the best available estimates for NZ and renders the issues transparent. The estimates of the cost per discounted life-year gained in NZ compare favourably to the estimates reported in other studies. Ultimately policy makers will have to decide whether universal antenatal HIV screening represents an attractive investment. Attempts have been made to indicate the scenarios under which the intervention might be considered cost effective given different decision rules.
The authors thank N. Dickson (AIDS Epidemiology Group), C. Paul (AIDS Epidemiology Group), L. Wilkinson (Auckland District Health Board), R. Ellis-Pegler (Auckland District Health Board), L. Voss (Auckland District Health Board), R. Jackson (University of Auckland), B. Boyd (Ministry of Health), K. Croxson (Auckland District Health Board), and H. Brady (Environmental Science and Research, Wellington).
Sponsorship: For the conduct of this research N. G. was on sabbatical from the London School of Hygiene and Tropical Medicine, at the Department of Community Health, University of Auckland, New Zealand. The NHS Executive, London, R & D Training and Development Programme supported N. G. during this academic sabbatical. D. W. is a member of the Health Economics and Financing Programme, which is supported by funds from the UK Department for International Development (DFID).
1.Connor E, Sperling R. Reduction of maternal–infant transmission of HIV type I with zidovudine treatment. New Engl J Med
2.Mofenson L, McIntyre J. Advances and research direction in the prevention of mother to child HIV-1 transmission. Lancet
3.Rogers MF, Fowler M, Lindgren M. Revised recommendations for HIV screening in pregnancy. MMWR
4.Mofenson L. Epidemiology and determinants of vertical HIV transmission. Semin Paediatr Infect Dis
5.The European Mode of Delivery Collaboration. Elective caesarean-section versus vaginal delivery in prevention of vertical HIV-1 transmission: a randomised clinical trial. Lancet
6.The International Perinatal HIV Group. The mode of delivery and the risk of vertical transmission of human immunodeficiency virus type I. New Engl J Med
7.Lallemant M, Jourdain G, Le-Coeur S. A trial of shortened zidovudine regiments to prevent mother-to-child transmission of human immunodeficiency virus type 1. New Engl J Med
8.Department of Health. Reducing Mother to Baby Transmission of HIV
. London: Department of Health; 1999.
9.Patrick D, Money D, Forbes J, Dobson S, Rekart M. Routine prenatal screening for HIV in a low prevalence setting. Can Med Assoc J
10.Lindgren S, Bohlin A, Forsgen M. Screening for HIV -1 antibodies in pregnancy: results from the Swedish national programme. BMJ
11.Chambers S, Heckert K, Bagshaw S, Ussher J, Birch M, Wilson M. Maternity care providers attitudes and practices concerning HIV testing during pregnancy; results of a survey of the Canterbury and upper South Island. N Z Med J
12.Dickson N. HIV infection in pregnancy. N Z Med J
13.Heckert K, Bagshaw A, Fursman L, Kipa M, Wilson M, Braiden V, et al
. Women's acceptability of screening for HIV in pregnancy. N Z Med J
14.Eberhart-Phillips J, Dickson N, Williams S, Clarke R, Fonua L, Kini GP, et al
. Asking pregnant women about HIV risk [letter]. N Z Med J
15.Bramley D, Graves N. Antenatal screening for HIV. N Z Med J
16.Ministry of Health. HIV in Pregnancy: Risk Screening Guidelines for Health Professionals
. Wellington: Ministry of Health; 1997.
17.Taylor B, Dickson N, Carter M. Third Annual Report of the New Zealand Paediatric Surveillance Unit
. Dunedin: University of Otago, 2000.
18.Schrappe M, Lauterbach K. Systematic review on cost-effectiveness of public health interventions for primary HIV-prevention programmes. AIDS
19.Commonwealth Department of Human Services and Health. Guidelines for the Pharmaceutical Industry on the Preparation of Submissions to the Pharmaceutical Benefits Advisory Committee
. Canberra: Commonwealth Department of Human Services and Health; 1995.
20.Department of Health. Policy Appraisal and Health
. London: Department of Health; 1995.
21.Culnane M, Fowler M, Lee S. Lack of long term effects of in utero exposure to zidovudine among uninfected children born to HIV infected women. JAMA
22.Sperling R, Shapiro D, McSherry G. Safety of the maternal-infant zidovudine regimen utilised in the paediatric AIDS clinical trial group 076 study. AIDS
23.Auckland Starship Hospital Protocol. Guidelines for the paediatric management of infants born to HIV positive pregnant women.
Auckland: Auckland District Health; 2001:1–2.
24.Center for Disease Control and Prevention. Public Health Service Task Force Recommendations for the use of anti-retroviral drugs in pregnant women infected with HIV-1 for maternal health and for reducing peri-natal HIV-1 transmission in the United States. MMWR
25.Brady H. Standard Protocol for Confirmation of Positive HIV Antibody Tests in New Zealand
. Wellington: ESR; 2001.
26.Gibb D, MacDonagh S, Tookey P, Duong T. Uptake of interventions to reduce mother to child transmission of HIV in United Kingdom and Ireland. AIDS
27.Richardson M, Sharland M. Late diagnosis of paediatric HIV infection in south west London. BMJ
28.Ades A, Sculpher M, Gibb D. Cost effectiveness analysis of antenatal HIV screening in the United Kingdom. BMJ
29.Postma M, Beck E, Mandalia S. Universal HIV screening of pregnant women in England: Cost-effectiveness analysis. BMJ
30.Zaric G, Bayoumi A, Brandeau M. The cost effectiveness of voluntary prenatal and routine newborn HIV screening in the United States. J Acquir Immune Def Syndr
32.Croxson BE, Ashton T. A cost-effectiveness analysis of the treatment of end stage renal failure. N Z Med J
33.Szeto KL, Devlin NJ. The cost effectiveness of mammography screening: evidence from a microsimulation model for New Zealand. Health Policy
34.Hansen P, Scuffham PA. The cost effectiveness of compulsory bicycle helmets in New Zealand. Austral J Publ Health
34.Surrey Advertiser. Patients’ lives ‘are given a price'
. Guilford: Surrey Advertiser; 01 February 2002.
37.Laupacis A, Feeney D, Detsky AS, Tugwell P. How attractive does a new technology have to be to warrant adoption and utilization? Can Med Assoc J
38.Owens D. Interpretation of cost-effectiveness analyses. J Gen Intern Med
39.George B, Harris A, Mitchell A. Cost-effectiveness analysis and the consistency of decision making. Evidence from pharmaceutical reimbursements in Australia. Pharmacoeconomics
40.Hirth RA, Chernew ME, Miller E, Fendrick M, Weussert WG. Willingness to pay for a quality adjusted life year: in search of a standard. Medical Decision Making
43.Lewis R, O'Brien J, Ray D, Sibai B. The impact of initiating a human immunodeficiency virus screening programme in an urban obstetric population. Am J Obstet Gynecol
44.Lindsey M, Feng T, Peterson H, Slate B. Routine human immunodeficiency virus screening in unregistered and registered inner city patients. Obstet Gynecol 1991, 77:599–603.
45.Cozen W, Mascola L, Enguidanos R, Bauch S, Giles M. Screening for HIV and Hepatitis B in Los Angeles County prenatal clinics. J Acquired Immune Defic Syndr
46.Faden R, Gielen A, Kass N, O'Campo P, Anderson J. Prenatal HIV antibody testing and the meaning of consent. AIDS Public Policy J 1994, 9:151–159.
47.Barbacci M, Repke J. Routine prenatal screening for HIV infection. Lancet
48.Croxson K. False Positive Rate for HIV Antibody Testing in the Blood Donor Population in New Zealand. Auckland: Auckland District Health Board; 2001.
49.Mofenson LM, Lambert JS, Stiehm ER. Risk factors of perinatal transmission of human immunodeficiency virus type 1 in women treated with Zidovudine. N Engl J Med
51.Mauskopf J, Paul J, Wichman D. Economic impact of treatment of HIV positive pregnant women and their newborns with zidovudine. J Am Med Assoc
52.Simpson W, Johnstone F, Goldberg D, Gormley S, Hart G. Antenatal HIV testing: assessment of a routine voluntary approach. BMJ
53.Chrystie I, Wolfe C, Kennedy J, Zander L, Tilzey A, Banatvala L. Voluntary named testing for HIV in a community based antenatal clinic: a pilot study. BMJ
54.Richardson M, Sharland M. Late diagnosis of paediatric HIV infection in south west London. BMJ
54.Auckland District Health Board Finance Department. Cost of HIV Investigations. 2001.
54.Pharmac. Pharmac Interactive Drug Schedule. http://www.pharmac
54.Counties Manukau District Health Board - Finance Department. Cost of Investigations. 2001.
58.Diagnostic Medlab. Cost Pricing for HIV Related Lab Tests. Auckland: Diagnostic Medlab; 2001.