Cost of providing cell-free DNA screening for Down syndrome in Finland using different strategies

cfDNA strategies

Three strategies are compared:

1. ‘Primary’ screening, entirely replacing the Combined test by routine cfDNA.

2. ‘Secondary’ screening, whereby all women with Down syndrome risk based on the Combined test ≥1 in 250 are offered cfDNA.

3. ‘Contingent’ screening, as with Secondary screening but using a lower cut-off risk whereby 10–30% of women are offered cfDNA.

Additional and marginal cost

Three different measures of cost are estimated for each strategy:

1. Additional cost for 10,000 pregnancies compared with the current paradigm of the Combined test. This represents the ‘new’ money needed to maintain the new policy.

2. ‘Marginal’ cost of avoiding a Down syndrome birth which occurs in a pregnancy that would have been false-negative using the Combined test.

3. Marginal cost of preventing the iatrogenic loss of a non-Down syndrome birth which occurs in a pregnancy that would have been false-positive using the Combined test.

Factors contributing to cost

Table 1 lists all the factors and the values which have been assigned to them in the analysis. The detailed derivation of these values are as follows:

Cost computations

The methods used to compute: proportion of Down syndrome births avoided; proportion of unaffected pregnancies undergoing CVS; cost per 10,000 pregnancies; cost per Down syndrome birth avoided; additional cost for 10,000 pregnancies compared with the Combined test; marginal cost per Down syndrome birth avoided; and marginal cost per fetal loss prevented are shown in the Supplementary Material.

Financial benefits

In 11 economic studies of cfDNA screening benefit was derived from the cost of continuing a Down syndrome pregnancy [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23]. One only included costs until age 18 [22].and the remainder considered lifetime costs which ranged from $583,000 to $1,497,000. Two studies discounted the cost to account for the reduced value of today’s currency in the future when the expenses will be incurred, yielding estimates of $677,000 and $857,000 [16, 21]. In the absence of specifically Finnish data on these costs, the average of these two studies (converted at $1.00 = €0.85) was used as a guide amounting to €902,000.

The benefit of preventing an iatrogenic fetal loss can be derived from the lost productivity of an adult’s working life. In a recently published study from the UK this was estimated to be ₤213,000 which, converted at ₤1.00 = €1.10, amounts to €234,000 [24]. Another comparator is the upper cost limit set by the UK National Institute of Health and Care Excellence (NICE) when judging the value of potential new interventions. The limit is ₤30,000 per Quality Adjusted Life Year (QALY) saved by the intervention – a higher limit of ₤100,000, for treatment of the elderly or in the case of rare conditions [25]. A newborn child has 25 QALYs so the limit amounts to €825,000 after discounting at 3% per annum. Neither the lost productivity nor the QALY comparator is entirely satisfactory though since in a proportion of cases an iatrogenic fetal loss will be replaced by a subsequent viable pregnancy.

Ethics

According to the Finnish Law (https://finlex.fi/fi/laki/ajantasa/1999/19990488), the cost-analyses in medicine that uses data which may not be personalized in any way are exempted from the Ethical approval and Research permit in local setting.

Down syndrome births avoided and CVS performed in unaffected pregnancies (Table 2)

Primary cfDNA screening increases avoidance of Down syndrome births by 15% compared with the Combined test and by 30% assuming enhanced uptake. It also reduces CVS being performed on unaffected pregnancies by 33–37 fold. Secondary cfDNA screening results in a slight reduction in Down syndrome births avoided except for enhanced uptake where there is a 12% increase. The reduction in CVS performed is 800–910 fold. Contingent screening increases avoidance of Down syndrome births by 6–12% or 20–27% with enhanced uptake; reduction in CVS is 110–370 fold.

Mean cost per pregnancy and per Down syndrome birth avoided (Table 2)

For all assumptions about uptake, the mean cost per pregnancy is higher for Primary cfDNA screening compared with the Combined test and lower for Secondary cfDNA screening. For Contingent cfDNA screening selecting 10% or 20% mean cost is also lower but selecting 30% it is higher than for Primary screening. The mean cost per Down syndrome birth avoided is higher for Primary cfDNA than Secondary cfDNA and intermediate for Contingent cfDNA strategies with 10% or 20% selected.

Additional cost for 10,000 pregnancies (Table 3)

Replacing the Combined test by Primary cfDNA screening will require additional funds amounting to €129–253,000, depending on uptake assumptions. If instead Secondary or Contingent cfDNA screening with 10% selected is adopted there will be a reduction in costs (€214–337,000 and €58–181,000 respectively) but for Contingent screening with 20% or 30% selected additional funds will be needed (€66–190,000 and €177–301,000 respectively).

Marginal costs (Table 3)

For each additional Down syndrome birth avoided when replacing the Combined test by Primary cfDNA screening the cost will be €43,000–€95,000 depending on uptake assumptions. Each iatrogenic fetal loss of an unaffected pregnancy prevented by the change will cost €164,000–€365,000.

Secondary cfDNA and Contingent cfDNA screening with 10% selected have no marginal costs as they are cost saving. Contingent cfDNA with 20% selected has a lower marginal cost per Down syndrome birth avoided (€32,000–€56,000) than Primary cfDNA screening and a lower marginal cost per iatrogenic loss prevented (€82,000–€268,000). Selecting 30% increases avoidance at a higher marginal cost (€72,000–€107,000) except under the assumption of enhanced uptake when it is lower than for Primary cfDNA (€79,000 compared with €95,000) but the marginal cost of fetal reduction is high (€221,000–€425,000).

Comparing marginal costs and benefits

For Primary and Contingent strategies under all uptake scenarios, the marginal cost per Down syndrome birth avoided (€32–107,000) is considerably less than the discounted lifetime medical and indirect cost incurred by a Down syndrome birth of €902,000. Under the most realistic scenario of enhanced uptake the marginal cost per loss prevented (€268–425,000), due to the very low cfDNA false-positive rate is higher than the lost productivity of an adults working life of €234,000, but at most about half of the upper cost limit per QALY set by NICE when judging the value of potential new interventions of €825,000.

Termination of pregnancy

The cost of ToP is higher than the cost of treatment following a spontaneous abortion; $874 and $552 in one economic evaluation of cfDNA screening, a difference of $322 or about €270 [10].When a cfDNA testing strategy leads to the avoidance of additional Down syndrome births, there will be an extra expense due to the termination of affected pregnancies which would have otherwise miscarried. However, this does not materially alter the marginal costs; for example, even when the additional births are 10% it would only increase by €270 × 10%/0.54 or €50.

Screening for common aneuploidies other than Down syndrome

Both the Combined test and cfDNA screening can be used to detect Edwards and Patau syndromes. The birth prevalence of Edwards and Patau syndromes are considerably lower than Down syndrome and they have a very high infant mortality rate. One economic study of cfDNA screening included all three aneuploidies and Turner syndrome (‘Monosomy X’) but found that the results were dominated by Down syndrome. The authors showed that the financial benefit of the other aneuploidies was offset by the additional cost of screening, and did not alter the overall cost-benefit comparisons between the Combined test and Primary cfDNA screening.

Primary cfDNA: retaining the ultrasound scan

The first trimester scan visit has many functions: confirm fetal viability; more accurately estimate gestational age than menstrual dates; identify multiple gestations and determine chorionicity; and examine and locate the placenta. In Finland, it is also an early ‘anomaly’ scan used to identify fetal structural anomalies. Moreover, using a very high NT cut-off level can select women at high risk of uncommon aneuploidies. Two studies have quantified this potential [26, 27]. The California Prenatal Screening Program found that using an 3.5 mm NT cut-off would select a group including 2.6% with uncommon aneuploidies detectable by classical karyotyping [26]. A national study in the Netherlands used a lower 95th centile cut-off but identified a group including 5.4% with uncommon aneuploidies detectable by karyotyping, submicroscopic copy number variants detectable by micro-array or single gene disorders detectable by sequencing [27]. Under these circumstances some centers would want to retain the NT scan after introducing Primary cfDNA screening. Table 4 shows how this would substantially increase the additional and marginal costs. The required additional funds would rise to €992,000 for 10,000 pregnancies. The marginal cost would still be much less than the lifetime medical and indirect cost incurred by a Down syndrome birth but the marginal cost per unaffected iatrogenic fetal loss prevented will be extremely high. The additional funds might be justified when taking account of savings resulting from the detection of serious structural abnormalities but this would require an economic analysis specifically focused on them.

Primary cfDNA: retaining maternal serum PAPP-A

Some centers may want to retain maternal serum PAPP-A as part of a screening program for pre-eclampsia (PE) [28]. However, it can be argued that the cost of this marker needs to be financially justified in a separate cost-benefit analysis rather than factoring them into an analysis of Down syndrome screening. For example, there is a published cost-benefit study of PE screening in Lebanon using first trimester mean arterial pressure measurement and maternal serum placental growth factor determination [29]. The study showed that the financial savings were more than sufficient to pay for screening.

Secondary cfDNA: choice of invasive prenatal diagnosis

At FMC the current Secondary cfDNA strategy allows individual choice between cfDNA and invasive prenatal diagnosis and one-fifth choose invasive testing (Vedran Stefanovic, unpublished results). Such an uptake would reduce the estimated cost savings in Table 3 by 19–30%. In 2017 Denmark introduced a similar Secondary cfDNA strategy and 80% choose invasive testing [30] which may be explained by their use of a micro-array rather than QF-PCR thus detecting submicroscopic copy number variants and counseling lower procedure-related miscarriage rates (negligible for CVS and 0.3% for amniocentesis) than quoted in Finland (0.5–1% for both procedures). This will have reduced the savings further. About 2% of cfDNA tests do not yield a result and such ‘no-calls’ were not included in the meta-analysis used to assess performance in this paper [1]. Consequently, the practical performance will be poorer than in the meta-analysis depending on how no-calls are treated. Referring all of them for invasive prenatal diagnosis, as some medical bodies currently suggest [31] will increase the detection rate but also considerably increase the false-positive rate. Taking no further action will reduce detection and false-positive rates whilst repeat testing will resolve about two-thirds of those due to low fetal fraction [32, 33].This has not been taken into account in the current analysis but this can be readily done. For example, with Primary cfDNA and 100% uptake the marginal detection rate is €43,000 (Table 3). Invasive prenatal diagnosis, assuming a 2% no-call rate, will substantially increase this to €102,000; no further action reduces it to €38,000 and repeat testing, assuming all no-calls are due to low fetal fraction and two-thirds are resolved, increases it to €46,000. Other categories of results may also have been excluded from some studies entered into [34] the meta-analysis, for example cases of confined placental mosaicism. This will have reduced the false-positive rate compared to clinical practice but this is likely to be a small effect.

In the 12 years since the discovery that maternal plasma cfDNA can be used to detect Down syndrome there has not been widespread replacement of conventional screening by cfDNA. A recent survey of European countries, Australia and the USA shows very uneven utilization hilst some have a form of Secondary or Contingent cfDNA screening, few provide even this minimal provision by public health funds. Only two countries, Belgium and the Netherlands, offer Primary cfDNA screening and whilst in Belgium the cost is fully reimbursed by insurance, in the Netherlands the pregnant woman pays for some of the cost.

The uptake of conventional screening in the Netherlands has been low, with a rate of abot 30% in 2007–2013 [35].There are many reasons for this, but a major factor is likely to be an unfavourable attitude towards screening in the local midwife community. Other potential factors could be a relatively positive attitude towards raising a child with Down syndrome, a negative view of selective pregnancy termination and the costs of screening. Screening in the Netherlands was never free of charge in contrast to Finland where there is no charge. The charges in The Netherlands are approximately €170 for combined screening and €175 for cfDNA screening. With the introduction of cfDNA screening, the uptake has increased to 46% [35] which is still much lower rate than in Finland. Furthermore, it should be noted that women in the Netherlands are not routinely screened for structural fetal abnormalities because the first‐trimester ultrasound is not part of the current Dutch prenatal screening program. Hence, the Dutch experience is that the introduction of Primary cfDNA screening did not lead to a major increase in uptake. Also, regional and maternal age variations were observed. Authors highlight the importance of tailoring counseling to the diverse needs of pregnant women and a centralized approach ensuring access, quality and continuous monitoring. High‐quality counseling for aneuploidy screening is imperative to ensure that women are free to make decisions in line with their personal values [35].

In Belgium, the uptake of first trimester screening is about 80% [36]. The introduction and reimbursement of Secondary cfDNA has substantially reduced procedure-related miscarriages without increasing the short-term costs. Almost full reimbursement of Primary cfDNA (for three common trisomies) has been available since 2017. All eight Belgian genetic centers perform a cfDNA test based on whole-genome sequencing, allowing not only for the detection of the three common trisomies, but also of rare autosomal trisomies (RATs), maternal subchromosomal aberrations, and, in some cases, large, noncryptic fetal aberrations throughout the genome [36].

Finland has a higher screening uptake rate than the Netherlands and Belgium and standardized first trimester fetal scan criteria, particularly nuchal translucency measurement, Also, invasive fetal procedures in Finland are concentrated to five university hospitals, while in Belgium many small centers still perform invasive procedures that may influence the procedure-related miscarriage risk [37].

Norway currently restricts the Combined test to women aged over 38 and recent changes in the law allow cfDNA testing in this age group. However, a survey of pregnant women found that 78% thought such testing should be offered as a public health service to all [38].

In Finland, for large centers like FMC, the practical choices are limited and the associated costs are shown in Figure 1. cfDNA secondary screening implementation in 2015 had a dramatic impact in decreasing the rate of invasive procedures (Figure 2).The first trimester scan would be retained for general obstetric care, as an early anomaly scan and to select women for further testing on the basis of a very large NT. Hence the Primary cfDNA strategy would be relatively costly. A version of the Secondary cfDNA strategy is already in place and is cost saving. Replacing this by a Contingent cfDNA strategy selecting 10% for testing would improve performance slightly and would also be cost saving. Contingent cfDNA selecting 20% or 30% for testing would involve additional costs but would be cheaper than Primary cfDNA and would increase the number of Down syndrome births avoided.

Figure 1:

Different screening strategies: impact on Down syndrome births avoided and associated costs.

Down syndrome births avoided (−) and, compared with the Combined test, additional cost for 10,000 pregnancies (■), marginal cost per Down syndrome birth prevented (●) and per fetal loss avoided (▲), when there is a cost, according to cfDNA screening strategy. Combined test uptake, 93% Screen, 95% CVS; cfDNA, 100% Screen, 95% CVS; both, 80% ToP.

Figure 2:

Invasive procedures (CVS + amniocentesis). Dept. of OB/GYN, Fetomaternal Medical Center Helsinki, Finland 2013–2018 before and after introduction of cfDNA and micoarray (aCGH).

While Contingent cfDNA screening would be the most cost-effective strategy with the current prices, one has to take into account that using a lower cut-off risk (≥1 in 1,000) compared to the present one (≥1 in 250) would increase the proportion of screen positive tests from 5% to 20–30%. As a consequence more women would experience anxiety and conflicting decisions that require increased resources for counseling.

Due to the high uptake of the first trimester screening in Finland, we speculate that it would be possible to find a cfDNA laboratory provider with a relatively low price due to the high volume of tests from Primary cfDNA screening. We are not aware if the public health funds would be able to cover the price. If the cost would be only partially reimbursed, the Finnish Law should be changed: currently the combined screening and second trimester screening are provided completely free of charge for all women. In the current Finnish organization where FMC operates under a hospital budget, not as a part of public health, short-term health savings result from either not screening at all or selecting the cheapest option. This in turn results in higher downstream long-term health care costs, including offspring lifetime costs and the burden to families which are not valued when the cost effectiveness threshold is to be decided at the hospital level for budgetary purposes. The costs should therefore be explored across such borders and across health care providers to maximize the value for the money. This paper brings transparency to the controversy surrounding the cost-effectiveness estimates in the Finnish setting and shows that Contingent cfDNA screening is an underused tool despite its cost-effectiveness. One of the strengths of this study is that we included the screening uptake rates as well as the costs and uptake rates of invasive diagnostic which were the most uncertain variables in the most of the publications on the similar cost analyses [39].