Fertility and assisted reproduction: The costs to the NHS of multiple births after IVF treatment in the UK
Professor W. L. Ledger, The Jessop Wing, Centre for Reproductive Medicine and Fertility, Sheffield Teaching Hospitals Trust, University of Sheffield, Tree Root Walk, Sheffield, UK.
Objectives To determine the cost to the NHS resulting from multiple pregnancies arising from IVF treatment in the UK, and to compare those costs with the cost to the NHS due to singleton pregnancies resulting from IVF treatment.
Design A modelling study using data from published literature and cost data from national sources in the public domain, calculating direct costs from the diagnosis of a clinical pregnancy until the end of the first year after birth.
Setting Academic Unit of Reproductive and Developmental Medicine.
Population Theoretic core modelling study using data from published literature.
Methods The analysis was based on the total annual number of births resulting from an IVF treatment in the UK. Main outcome measures total direct costs to the NHS per IVF singleton, twin or triplet family.
Main outcome measures Cost of singleton, twin and triplet IVF pregnancies in the UK.
Results Total direct costs to the NHS per IVF twin or triplet family (maternal + infant costs) are substantially higher than per IVF singleton family (singleton: £3313; twin: £9122; and triplet: £32,354). Multiple pregnancies after IVF are associated with 56% of the direct cost of IVF pregnancies, although they represent less than 1/3 of the total annual number of maternities in the UK.
Conclusions Multiple pregnancies after IVF are associated with high direct costs to the NHS. Redirection of money saved by implementation of a mandatory ‘two embryo transfer’ policy into increased provision of IVF treatment could double the number of NHS-funded IVF treatment cycles at no extra cost. Further savings could be made if a selective ‘single embryo transfer’ policy were to be adopted.
There were 6309 live births (8106 live babies) after IVF treatment in the UK between 01.04.2000 and 31.03.2001.1 Although the largest single proportion were singleton births (n= 4621/73%), the proportion of multiple births was high (total multiple births, n= 1688/27%; twin births, n= 1579/ 25%; and triplet births, n= 109/2%). This is in agreement with other reports that identify a substantial increase in multiple births after IVF compared with spontaneous conception (up to 27% vs 1%).2,3 The risk of multiple birth is directly related to the number of embryos transferred (odds ratio three vs two embryo transfer: 1.6; 95% CI 1.5–1.8).4 The rise in the number of multiple births in the UK is largely due to increasing IVF treatment.
There is a worldwide impetus to decrease the numbers of embryos transferred after IVF, although there are substantial differences in practice between countries. From March 2004, the UK Human Fertilisation and Embryology Authority (HFEA) has advocated a universal maximum of two embryos per transfer after IVF5 (reference: Code of Practice 6th edition HFEA, January 2004), regardless of the procedure used. In addition, the National Institute for Clinical Excellence (NICE) in the UK guideline ‘Fertility: Assessment and Treatment for people with Fertility Problems’,6 published in February 2004, provides recommendations for good practice based on the best available evidence of clinical and cost effectiveness. The guideline includes the recommendation that ‘no more than two embryos should be transferred during any one cycle of IVF treatment (220.127.116.11). This guideline should form the basis for contracts between UK Primary Care Trusts, who act as NHS budget holders, and NHS and private providers of fertility services. The NICE recommendation (18.104.22.168) is that couples in which the woman is aged 23–39 years at the time of treatment … should be offered up to three stimulated cycles of treatment’.
In response to the NICE Guideline, the United Kingdom Government made a clear commitment to improve NHS funding of IVF treatment for infertile couples. The Minister of Health, Mr John Reid stated in February 2004: ‘By April next year (i.e. April 2005), I want all PCTs, including those who at present provide no IVF to offer one full cycle of IVF to all those eligible. In the longer term I would expect the NHS to make progress towards full implementation of the NICE guidance’. Despite this clear guidance, progress towards meeting the Minister's direction is at best patchy across the country. The challenge to those wishing to improve NHS funding for IVF treatment is to demonstrate that sufficient financial savings can be made from improvements in clinical practice in this area to make an increase in number of cycles funded ‘cost-neutral’.
Currently, only approximately 25% of IVF cycles of treatment in the UK are funded by NHS, but this percentage may well rise if the NICE guidelines are fully implemented. In contrast to the low level of NHS support for infertility treatment, almost all costs associated with pregnancy (either normal or with complications) and neonatal care (either of a healthy newborn, or of any costs resulting from neonatal complication) are borne by the NHS. Even though awareness exists about the burden of multiple births imposed on the individuals, health service and society,7 to our knowledge there have been no published studies specifically addressing the costs to the NHS imposed by multiple births resulting from IVF treatment.
We therefore set out to determine the cost to the NHS in the UK resulting from multiple pregnancies (twins and triplets) arising from IVF treatment, and to compare those costs with the cost to the NHS due to singleton pregnancies resulting from IVF treatment.
This was a modelling study using data from published sources (i.e. published literature and other data available in the public domain). Where only outdated data were available, the group extrapolated the data from other countries to the current situation in the UK. A decision tree model was developed by the group to predict outcomes and serve as a basis for the calculation of associated costs. The outputs of the model are as follows:
- • The probability of a singleton, twin or triplet surviving to at least one year
- • The estimated cost of a singleton, twin and triplet at one year.
The analysis was based on the total annual number of births resulting from an IVF treatment in the UK, using the Provisional National Data Statistics (between 01.04.2000 and 31.03.2001).1 Base case analysis was performed to estimate the cost to the NHS per singleton, twin or triplet pregnancy resulting in a live newborn infant(s) surviving up to year one. The total annual cost to the NHS of the births resulting from the IVF treatment was than calculated for the total number of singleton, twin and triplet births in the observed period. The model was developed from an NHS perspective and used NHS costings based on information for 2002.
A decision tree was constructed using DATA 4.0 (Tree-age Software Inc, Williamstown, MA, USA).
The model begins with a singleton, twin or triplet pregnancy. The costs of the infertility treatment (drugs and procedures) leading to pregnancy are not included in the model because this cost is common to all arms. Only women with a clinical pregnancy (defined as the presence of a fetal heartbeat on early pregnancy ultrasound) stay in the model, while all women with no clinical pregnancy leave the model without incurring any cost. Clinical pregnancies are assumed to result in either late miscarriage, stillbirth or live birth, with pregnancies resulting in stillbirth leaving the model and incurring certain costs. A baby can receive routine care, when no additional costs of neonatal care are incurred, or is admitted to a neonatal unit where the baby may receive one of the three levels of neonatal care, as defined in the UK, each of which incur cost. Babies who do not survive to weeks one or four, or year one, leave the model. However, these outcomes are associated with different costs. The calculation of the total costs and outcomes associated with the model was based on cost and outcomes incurred by the mother, and those incurred by the baby.
The decision tree model and detailed description of the costs and outcomes included in the model, as well as sensitivity analysis, are in Appendix S1 (available online as supplementary material).
There are substantial differences between the total costs to the NHS per IVF family of a singleton, twins or triplets, as well as in the expected maternal and neonatal costs associated with IVF singletons, twins and triplets (1Table 1).
Table 1. Summary of costs to the NHS of singleton, twin and triplet births resulting from IVF infertility treatment in the UK
|Estimated maternal cost||£3122||£6058||£11,534|
|Estimated neonatal cost (per family)||£191||£3064||£20,820|
|Total cost per family||£3313||£9122||£32,354|
|Number of IVF births in UK (1)*||4621||1579||109|
|Cost of IVF births to the NHS||£15,309,373||£14,403,636||£3,526,586|
The total estimated ratio of costs for a singleton versus twins mother was 1:1.94, and for a singleton versus triplets mother 1:3.96. In absolute monetary terms, and according to 2002 NHS costings, the total estimated maternal cost to the NHS of a singleton IVF pregnancy was £3122, of a twin IVF pregnancy £6058 and a triplet IVF pregnancy £11,534.
The ratio of neonatal singleton versus twins costs was 1:16.04, and of singleton versus triplets was 1: 109. In absolute monetary terms and according to 2002 NHS costings, the total expected neonatal cost to the NHS of a singleton IVF baby is £191, of a set of IVF twins £3064 and a set of IVF triplets £20,820.
The ratio per singleton versus twins family was 1:2.75, and of a singleton versus triplets family was 1:9.76. That means that a family of IVF triplets during an 18-month period (from a diagnosis of a clinical pregnancy until up to one year after birth) would incur almost 10 times higher costs to the NHS than a family of an IVF singleton.
The total estimated annual costs to the NHS of all IVF babies up to year 1 were £33,239,595. Although the IVF singletons represented 73% of the total number of live births, they incurred only 46% of the total costs. By contrast, IVF triplets represented only 2% of the total number of the live births, but they incurred 10.6% of the total costs. IVF twins represented 25% of the total number and incurred 43% of the costs. Thus, 54% of the cost to the NHS was incurred by multiple births resulting from IVF, although they represent <1/3 of the total annual number of live births resulting from IVF (27%). In addition, the cost of IVF twins (£14,403,636) was almost the same as costs of IVF singletons (£15,309,373) for 2.9 times more births. Triplets account for only 2% of all IVF births, yet they incur 10.6% of the total cost.
Although there are some differences in absolute values of costs, the incremental ratio of costs per family of a singleton versus twins versus triplets is present in the results of all sensitivity analyses (2Table 2).
Table 2. Sensitivity analysis results
|a. Estimated costs assuming length of stay in neonatal units is 3 days for all babies instead of 3 days for singletone, 7 for twins and 18 for triplets|
|Total per baby||£191||£712||£1273|
|Total per family||£191||£1424||£3819|
|b. Estimated costs assuming length of stay in neonatal units is 18 days for all babies instead of 3 days for singletons, 7 for twins and 18 for triplets|
|Total per baby||£888||£3787||£6940|
|Total per family||£888||£7574||£20,820|
|c. Costs incurred at end of weeks 1 and 4, rather than beginning of weeks 1 and 4|
|Total per baby||£191||£1505||£6673|
|Total per family||£191||£3010||£20,019|
|d. Routine care cost|
|Total per baby||£1251||£2943||£7642|
|Total per family||£1251||£5886||£22,926|
None of the changes had any appreciable effect on the results compared with the base case (for detailed calculations see Appendix S1 available online as supplementary material).
At the time of writing, this is the first study to model the cost to the NHS caused by multiple pregnancies (twins and triplets) resulting from IVF treatment, and to compare these costs with those of singleton pregnancies due to IVF. Although the NHS funds only a relatively small proportion of IVF treatments in the UK (approximately 25%), the majority of costs associated with pregnancy (either normal or with complications) and neonatal care (either of a healthy term baby, or of any costs resulting from neonatal complication) are borne by the NHS.
This was a modelling study based on a decision tree analysis, using published literature data and other data from public sources, and NHS costings wherever possible. The study was based on the data from the latest HFEA Provisional National Statistics1 providing the number of live singleton, twin and triplet births between 01.04.2000 and 31.03.2001. The results demonstrate that the total costs to the NHS of twin and triplet IVF pregnancies are associated with substantially increased direct costs per family when compared with singleton IVF pregnancies (total singleton family cost: £3313; total twin family cost: £9122; and total triplet family cost: £32,354). In addition, there was an imbalance in the total costs to the NHS between IVF singletons and multiples: although the IVF twins and triplets represent <1/3 of the total annual number of live births resulting from IVF in the UK (27%), they incur 54% of the costs to the NHS.
As calculations were based on short term costs only (up to one year post delivery), they most probably represent an under-estimate of the total costs to be borne by the NHS over a longer period. A study involving 9065 live born infants after IVF in Sweden between 1984 and 1997 showed increased risk for hospitalisation until six years of age (OR: 1.8) compared with non-IVF infants.8 Even the term live born IVF infants had an increased risk of hospitalisation (OR: 1.3).8
The increased neonatal costs of multiple births result from a much greater risk of very preterm birth (or very low birthweight; VLBW) and thus the need for high levels of dependency after birth. As gestation and birthweight decrease, the risk increases for sequelae associated with high life-long ongoing costs, such as cerebral palsy, developmental and learning problems, sensory deficits and respiratory illnesses. For example, triplets are associated with 300% increase in the relative risk of handicapping conditions compared with singletons, and 650% increase in the rate of cerebral palsy compared with singletons (26.6 vs 1.6).9 The costly diagnostic and treatment procedures associated with suspected disability usually start when an infant fails to attain key developmental milestones, such as independent standing and walking. These developmental milestones usually occur from the 10th month of life onwards, thus the probability of starting a diagnostic and treatment procedure within the time frame of our study was low. VLBW babies also display other serious medical comorbidities (e.g. vision and hearing disorders, or seizures). Diagnosis and treatment of each of these comorbidities will inevitably increase the burden to the NHS. However, we were unable to identify reliable estimates of the likelihood and cost of such complications for IVF multiples.
There were several limitations in our study. As in any other modelling study, our results were based on available data in published literature and not only on national UK statistics. In addition, some of the published results are from single centres or based on small clinical samples. In order to control for any bias that could have been introduced by using these data, clinical experts on the study group evaluated each variable from the NHS standard practices perspective. In addition, limited data were available on the quantities of routine maternal care, hence for antenatal clinic and outpatient visits, and postnatal home visits, a simple base case was assumed whereby the number of visits for twin and triplet sets was assumed to be the same for singleton mothers. In reality, twin and triplet mothers would expect to require more visits than those of a singleton, therefore the calculated costs of twin and triplet care will be an under-estimate of the true cost. Another problem was related to the simplistic approach to utilisation of neonatal care. Although the calculated estimates are in line with published estimates of risk requiring complex neonatal care, our data should be treated with caution, as definitions of the different levels of care are not identical, especially when comparing studies from different countries. In addition, the sample sizes10 on which our calculations were based may be too small to calculate estimated proportion rates with certainty. However, sensitivity analysis based on varying lengths of stay and costs in neonatal units made very little difference to the results compared with base case, thus confirming the robustness of results.
Based on the trend in other European countries to limit the number of transferred embryos to one or two, we further explored the effect of elective two embryo transfer on NHS costs. The base case model assumed a zero probability of getting more fetuses than implanted embryos (i.e. dual embryo transfer cannot result in triplets). While we recognise the possibility of post-embryo transfer monozygotic twinning, this is a rare event and the impact on the outcome of the model was likely to be low. Thus, we assumed that two embryo transfer would result in singleton or twin pregnancies only, and we performed further calculations using the available data in public domain on numbers of live births after IVF treatment in the UK (n= 6309 live births between 01.04.2000 and 31.03.2001).1 Elimination of triplet pregnancies would offer the possibility of reallocation of the NHS funds currently used for IVF triplets (total costs of IVF triplets: £3,526,586) to the funding of additional IVF cycles. Using the HFEA cost of £1771 per cycle as per NICE Fertility Guideline6 means that the NHS would be able to fund an additional 1991 cycles per year, or fund three cycles of treatment for an additional 663 couples per year. Compared with the HFEA 2000–2001 data, that would represent an increase of 7.8% in the number of cycles started.
If both twin and triplet pregnancies are eliminated by elective single embryo transfer, an additional 10,124 cycles could be funded by NHS. Assuming three cycles per couple funding by NHS, this equates to additional 3374 couples treated per year through NHS funding by means of savings from twin and triplet multiple births.
Another assumption is that elective two embryo transfer will also result in a reduction in the number of twin pregnancies. If twin deliveries are reduced by 1/3 and triplets are eliminated through elective two embryo transfer, the cumulative savings to the NHS would be £8,327,798, thus allowing funding of additional 4702 cycles per year, or full three-cycle treatment of additional 1567 couples per year. That would represent an additional increase of 20.1% in the number of cycles in the UK compared with the HFEA 2000–2001 data. Thus, elective twin embryo transfer may potentially increase by 18.6% the total number of cycles performed over one year in the UK, and funded by the private and NHS sectors.
Based on the estimate that only 25% of all cycles in the UK are NHS funded, the implications are even greater if these increases are applied to the NHS sector only, allowing the increase of funded cycles by 74.4%, enabling an increase of three quarters in the number of couples to whom the NHS can offer a three-cycle treatment. It is also pertinent that a number of European countries, including Sweden and Belgium, have moved to single embryo transfer for many IVF patients, with concomitant reduction in twins.11 The Swedish National Board of Health and Welfare has determined in 2003 that more than one embryo can only be transferred following IVF in ‘only exceptional cases’,12 with a target of single embryo transfer in at least 50% of all IVF treatments. While a similar approach may be a ‘step too far’ in current UK practice, demonstration from Sweden that frequent single embryo transfer is not associated with a large reduction in pregnancy rate, particularly if frozen embryo transfer pregnancies are included, should lead in time to adoption of a similar policy in UK.
In conclusion, the results demonstrate that multiple IVF pregnancies are associated with high total costs to the NHS in the UK. The recommendation to permit in the UK only elective twin embryo transfer in IVF treatment and consequent reallocation of NHS funds may allow for a significant increase in the total number of IVF cycles performed over one year. Increased NHS funding should be accompanied by increased regulation of IVF practice in UK, particularly in areas such as embryo transfer policy in which the consequences of irresponsible practice result in large costs to the NHS as a whole.
All authors were involved into reviewing of the evidence and model development. EW developed the original model. CT and EW calculated the costs. WL drafted the manuscript, and all other authors played a part in reviewing and editing it. The guarantor WL accepts full responsibility for the conduct of the study, had access to the data and controlled the decision to publish.
WLL has supervised and undertaken research funded by Organon Laboratories Ltd and Serono Pharmaceuticals Ltd and has been reimbursed for conference attendance and lectures by these organisations and Ferring Pharmaceuticals Ltd. DA: none declared. NM: none declared. At the time of the study, CT was employed as health economics manager by Organon Laboratories Ltd, the manufacturer of Puregon, Orgalutan and Pregnyl. EW was employed as health economist by Organon Laboratories Ltd, the manufacturer of Puregon, Orgalutan and Pregnyl.