Correspondence: Professor C. Normand, Department of Public Health and Policy, London School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, UK.
Objective To consider the organisation cost and effectiveness, of universal, community-based antenatal screening for the haemoglobinopathies, and to estimate the cost-effectiveness of programmes at different levels of prevalence and mix of haemoglobinopathy traits.
Design Retrospective review of laboratory and Sickle Cell and Thalassaemia Centre worksheets with costing of capital equipment, consumables, salaries and overheads, and estimation of costs in a range of circumstances.
Setting A haematology department, including a Sickle Cell and Thalassaemia Centre, providing antenatal and neonatal screening programmes in Inner London.
Participants Two thousand one hundred and one women booking at the antenatal clinic whose samples were referred for screening during 1994.
Main outcome measures and Results In addition to assessing the cost-effectiveness of antenatal haemoglobinopathy screening in a number of settings, the following specific financial information was assembled for the service in Brent: 1. cost of identifying abnormal haemoglobin in mother (£209); 2. cost of identifying at-risk fetus before confirmation by prenatal diagnosis (£2455); 3. cost of providing genetic information and counselling to mother with abnormal haemoglobin (£109); 4. programme savings from cases averted (£61,000).
Conclusions Antenatal screening with follow up counselling can be self-financing at most prevalences of haemoglobinopathy traits, with greater savings where a high proportion of the traits are β thalas saemia. There is a net financial cost (£1350) only at prevalences below 2.5% of traits if these are mainly for sickle cell disease. Since there are other benefits is it likely that antenatal screening will be considered cost-effective even at quite low levels of trait prevalence.
We report the costing of a universal community-based antenatal screening programme for haemoglobinopathies. This allows an assessment of the cost-effectiveness in terms of the cost of providing full genetic choice to mothers and couples, and the cost per clinically significant haemoglobinopathy averted.
These data have not previously been available and should help health service planners in their reassessment of the UK Standing Medical Advisory Committee's recommendation that, where more than 15% of the local population is from the ethnic minorities, universal antenatal screening should be established1. NHS purchasers and trusts considering establishing universal programmes will be able to base their decisions on the best estimates of the cost-effectiveness of such programmes.
The work described has been carried out at the Central Middlesex Hospital, a district general hospital in an inner London health district (Brent), with around 2000 births per year and 45% of the population from ethnic minorities. The Central Middlesex Hospital antenatal screening programme is universal and community based and was established 11 years ago.
The full process of antenatal screening and follow up is described in Fig. 1. Blood samples are collected at the antenatal booking clinic and forwarded to the haematology laboratory at Central Middlesex Hospital. On receipt, the laboratory performs an automated full blood count, including haemoglobin (Hb), mean cell volume, and mean corpuscular haemoglobin, using a Coulter STK-S (STKS Coulter Electronics, Hialeah, Florida, USA). The laboratory records are checked to ascertain if the haemoglobin phenotype has been previously performed on each patient, and where a previous result is available (∼ 9% of cases), this is provided to the clinic with no further haemoglobinopathy testing being undertaken. The other samples are subjected to isoelectric focusing using Isolab neonatal gels to screen for structural haemoglobin abnormalities, and high performance liquid chromatography, using a Shimadzu Haemoglobin Analyser (Shimadzu, Kyoto, Japan) for confirmation of variant haemoglobins as well as quantification of the haemoglobins, including HbA2 to diagnose the β thalas-saemia trait2–4. The presence of HbS is confirmed by a solubility test5 and other variant bands are identified by repeat isoelectric focusing with known variants as adjacent controls. This is followed, if required, by acid gel electrophoresis (pH 6.2) or cellulose acetate electrophoresis (pH 8.6) for diagnostic reasons6. However, only nine and two, respectively, of these extra tests were performed during the sample year. The fj thalassaemia trait is diagnosed by the finding of HbA2 of > 3.5%. All diagnosed and possible thalassaemia trait samples (β and α) are subjected to zinc protoporphyrin assay, using a protofluor-2 (Helena Laboratories).
All patients with definite haemoglobinopathy traits or disease, as well as those with putative α thalassaemia, are referred to the Brent Sickle Cell and Thalassaemia Centre. The nurse specialist there makes up to three attempts by post, as well as telephoning, to contact the woman in order to inform her of her haemoglobinopathy phenotype and to invite both her and her partner for counselling. The counselling session includes education about inheritance, the implications of a positive result in the partner and risk to the fetus, as well as the option of prenatal diagnosis. Arrangements are made for specimen collection from the partner, and if his result indicates that the fetus is at risk of a clinically significant haemoglobinopathy, then prenatal diagnosis (or DNA analysis in the case of risk of possible α thalassaemia major or haemoglobin H disease, followed by prenatal diagnosis as appropriate)4 is offered to the couple. Since 1990, any woman with the sickle cell trait who has a partner from an ethnic group at risk of the sickle cell trait, who is unavailable for, or refuses testing, has also been offered prenatal diagnosis.
Following prenatal diagnosis at a perinatal medicine centre, the Brent Sickle Cell and Thalassaemia Centre nurse specialist supports the mother, or parents, in taking the decision whether to terminate an affected pregnancy and, in the case of such a termination, offers post-termination support as well. The nurse specialist also ensures that results from neonatal screening are given promptly to all other women who have been referred, by performing a home visit as soon as the result is available through the North Thames (West) Neonatal Screening Programme.
This antenatal screening programme is run concurrently with the North Thames (West) Neonatal Haemoglobinopathy Screening Programme, also based in the haematology department at the Central Middlesex Hospital. As such, it is able to share the expense of laboratory equipment and other capital costs
Cost information has been collected for both the laboratory and follow up components of the programme. The financial information used is for the year 1994/1995, and the laboratory process and follow up information are from the calendar year 1994, during which 2101 women booked at the Central Middlesex Hospital antenatal clinic. Both fixed and variable elements of laboratory costs have been determined and 46% of annual equivalent costs for high performance liquid chromatography equipment has been allocated to the antenatal programme, compared with only 4% of annual equivalent costs for the isoelectric focusing equipment. These figures reflect the overall use of equipment as shown in Table 1. Follow up cost information was collected by examining the antenatal caseload as a proportion of the total Brent Sickle Cell and Thalassaemia Centre activity, and includes fixed costs (support and overheads) and variable costs for salary which are dependent on the number of referrals (141 during the sample year).
Table 1. Use of equipment in antenatal and neonatal screening. Values are given as n (%). IEF = isoelectric focusing; HPLC = high performance liquid chromatography.
IEF tests performed
HPLC tests performed
Costs of current programme
The laboratory tests performed and subsequent progress through the process are shown in Table 2. Table 3 shows fixed and variable elements of laboratory costs during 1994. Table 4 shows the nurse specialist and other Brent Sickle Cell and Thalassaemia Centre costs for provision of follow up, including information, education and support. The salary component for the nurse specialist has been deemed variable because of the interchangeability of staff between the neonatal, antenatal and other sickle programmes (the number of hours dedicated to antenatal work by the nurse specialist varies according to demand).
Table 2. Tests and follow up in antenatal screening programme. IEF = isoelectric focusing; HPLC = high performance liquid chromatography; ZPP = zinc protoporphyrin assay; BSCTC = Brent Sickle Cell & Thalassaemia Centre; PND = prenatal diagnosis.
Women booking at antenatal clinic
Women with earlier result (not tested)
IEF and HPLC performed
IEF only performed
HPLC only performed
Women with abnormal haemoglobin (referred to BSCTC)
Women who attended counselling
Women whose partners were tested
Couples whose results indicate fetus at risk (PND offered)
Prenatal diagnosis accepted
Termination of Pregnancy
Table 3. Fixed and variable costs in screening programme. Values are given as n. IEF = isoelectric focusing; HPLC = high performance liquid chromatography; ZPP = zinc protoporphyrin assay.
1994/1995 prices (£ sterling)
*Including overall clinical direction, laboratory supervision, training and stationery/small replacements.
Information review/selection/worksheet preparation
IEF testing: women and partners
HPLC testing: women and partners
Table 4. Costs of services provided in follow up to screening. Values are given as n.
Follow up component
*Including supervision, overall clinical direction, training, computer equipment and stationery.
Counselling, liaison, information collection, case registration
Termination of pregnancy
Costs of detecting high risk pregnancies, providing genetic choice and averting cases
Overall, the cost of identification of at risk fetuses was £2455 per case including the follow up costs as shown in Table 5.
Table 5. Average cost of achieving programme outcomes. Values are given as n.
Average cost (£)
*Women and partner with abnormal haemoglobin.
Laboratory identification of abnormal
haemoglobin in woman
Laboratory identification of at-risk fetus*
Counselling of woman with abnormal
Prenatal diagnosis with amniocentesis
Termination of pregnancy
An important outcome of such a community programme is to allow genetic choice. In 1994, 12 women had proven at risk pregnancies (partner tested) with 37 pregnancies also at risk because the partners' haemoglobinopathy status was unknown. If the main objective is to provide genetic choice to parents, the cost per proven at risk pregnancy (12) is £2455, which is the costs of giving one couple a choice.
Cost-effectiveness of the Brent Sickle Cell and Thalassaemia Centre service
No detailed information relating to the NHS costs for either sickle cell disease or β thalassaemia major has been published. One estimate of hospital costs for sickle cell disease is £5000 per annum and those for β thalassaemia major are £81507. With optimal care, patients with β thalassaemia major are now living to over 40 years (assumed to be an average of 41 in the examples below) and the median survival reported for patients with sickle cell disease is 44 years8. Therefore, the present value of the flow of savings in health service costs (discounted at 6% as recommended by HM Treasury) per case of sickle cell disease is £77,000, and per case of β thalassaemia £123,000.
During the study year, the programme identified 12 pregnancies at risk of a major haemoglobinopathy because both partners had significant traits. As there is a one in four risk of a clinically significant haemoglobinopathy from a conception in which both partners carry a haemoglobinopathy trait, we can expect three of these to result in affected fetuses. The ratio of sickle cell disease to β thalassaemia major in the programme at the Brent Sickle Cell and Thalassaemia Centre is approximately 3:19,10. Therefore, 2.25 sickle cell disease cases (with total annual treatment costs of £11,250) and 0.75 β thalassaemia major cases (with treatment costing a total of £6112.50 per year) would have been detected if the parents had pursued prenatal diagnosis. If these pregnancies had been terminated, the net present values (discounted at 6%) for costs averted would have been £173,878 and £92,531 respectively.
The experience in this programme has been that less than one third of women with a fetus at risk of a haemoglobinopathy accept prenatal diagnosis, and these are mainly those with a fetus at risk of β thalassaemia. The result is that of all affected pregnancies, 10% of sickle cell disease and 95% of β thalassaemia have been terminated. This equates to 0.225 sickle cell diseases and 0.7125 β thalassaemia major cases per year. Total net present values for costs averted are £17,388 and £87,904 respectively. Since the likely financial savings (£105,292) exceed the programme cost of this service, universal screening and follow up leads overall to savings in the Brent Sickle Cell and Thalassaemia Centre service.
Financial savings from a programme cannot be the sole criterion for introducing a service. However, if a programme has other benefits and saves on resources, there can be no argument against it unless another programme saves more or produces even greater benefits.
Cost effectiveness, different prevalence of traits and different types of trait
Subtracting the annual programme costs (£42,629) from the financial savings, there is likely to be a saving of around £61,100 from a programme in an inner city area like Brent, with a high (7.5%) prevalence of haemoglobinopathy traits of which around three quarters are sickle cell traits. Table 6 shows the likely financial savings or costs at different combinations of trait prevalence and the proportions of these that are β thalas-saemia traits. At low levels of prevalence, of haemoglobinopathy traits, and where most of these are sickle cell traits, the financial savings are smaller than the cost of a universal programme. Financial savings are likely even at quite low (e.g. 1%) prevalence of traits if these are mainly for β thalassaemia. This suggests that any guidelines on screening policy should take account of the countries of origin of people from ethnic minorities as well as the number of people likely to carry a haemoglobinopathy trait.
Table 6. Table 6. Savings (costs) of universal screening (programme cost-financial savings) in 1994/1995 prices at different combinations of prevalence of traits and proportions of traits that are for β thalassaemia.
Prevalence of traits (£)
The programme at Central Middlesex Hospital shares resources between antenatal and neonatal programmes. This has major advantages in reducing the cost of maintaining access to the equipment and skills for an antenatal programme. This includes sharing of expertise and access to counselling services. Where cost-effectiveness of antenatal screening is being considered it is important to take into account the policy on neonatal screening and treatment.
We demonstrate that most antenatal screening programmes are likely to be self-financing and therefore clearly cost-effective, since the savings in service costs are greater than the costs of detection of an affected fetus and termination of pregnancy. Antenatal screening (especially when this is managed alongside a neonatal screening programme) is quite cheap, and may be considered cost-effective in terms of improved genetic choice. However, two significant uncertainties could change this conclusion. First, on the basis of experience in Brent, it is estimated that the choice to proceed with termination of pregnancy is likely in 95% of β thalassaemia major cases. If this proportion were lower, however, the financial savings would be lower and the programmes less cost-effective. Second, the costs of lifetime treatment of people with haemoglobinopathies have not been researched in detail.
Sensitivity analyses were carried out on these two estimates. On the basis of a high prevalence area, with traits carried by 7.5% of the population, and 25% of traits being P thalassaemia major, a programme is likely to ‘break even’ in financial terms even if termination occurs in only 50% of fetuses with thalassaemia. Financial savings are likely on this level of termination for 2.5% of the population with traits all of which are for thalassaemia, and only a small financial deficit is likely with 2.5% prevalence of traits in the community, half of which are for thalassaemia. Given that this calculation makes no allowance for the health and social benefits of screening, choice and better treatment, it is fair to conclude that the cost-effectiveness of antenatal screening is not very sensitive to the estimate of the numbers choosing termination of pregnancy. If the cost of lifetime treatment is overestimated by 50%, financial savings would occur only for areas where trait prevalence is above 5%, or where the proportion of thalassaemia traits is over 50% and the rate in the population is 2.5%. Again this suggests that the likely conclusions are not very sensitive to errors in the cost of treatment.
No attempt has been made in this paper to compare universal screening and selective screening. It is likely, however, that many of the benefits of universal screening can be achieved with an effective policy of selective screening. Nor has any attempt been made to estimate the benefits of genetic choice and better management of people affected with significant haemoglobinopathies. While further work is needed on costing care, likely decisions on termination of pregnancy, and the value of earlier knowledge of significant haemoglobinopathies, these results suggest that antenatal screening is likely to be considered cost-effective at least in areas with haemoglobinopathy traits at or above 2.5%, especially if a high proportion of these are for thalassaemia.
We thank our collaborators on this project, and especially Dr P. Greengross for help and advice. The funding for this project was provided by the NHS Research and Development, Health Technology Assessment Programme.