Outcome Measure

NTS detects cases of trisomy 21 earlier in pregnancy. This finding would allow some women to choose to have an earlier, less physically and perhaps emotionally traumatic termination. However, for others, there would be an increase in unnecessary anxiety, invasive procedures and terminations of affected fetuses that would have otherwise miscarried before a second trimester serum screening was done. There are several possible measurable consequences of a screening program (5). The two outcome measures chosen here—trisomy 21 cases detected and live trisomy 21 births prevented—represent proxies for the benefits of prenatal screening.

Some studies, such as Vintzileos et al. (9), have gone further and calculated the lifetime extra cost of care for infants born with trisomy 21. Although there are no studies of these costs for Australia, and it is not possible to take overseas results directly, it is clear that these costs are likely to be substantial. Life expectancy with trisomy 21 is now in excess of 50 years and is usually associated with major health problems. However, including only the cost of care as a negative social value of a child with trisomy 21 provides a distorted picture of the net benefits of screening. The extra cost of care needs to be balanced by the value of the benefits of having a child with trisomy 21. There are no reliable estimates of the money value of a life with or without trisomy 21, and there are conceptual difficulties in the calculation not least because of the potential for a replacement pregnancy. For these reasons, the perspective of the analysis has been constrained and the full net benefits of screening have not been calculated.

Data

There were approximately 260,000 births per year in Australia in 2001. It is unknown what the prevalence of trisomy 21 would be without any screening program. Data from South Australia (1) suggests that the prevalence was 1.1 per 1,000 live births in 1982. Given the upward trend in maternal age since 1982, the model assumed an underlying prevalence of 1.6 per 1,000 live births (10). This rate would suggest 416 trisomy 21 live births each year.

There is a high spontaneous rate of miscarriage of trisomy 21. Between the first trimester and term, the rate is reported to be 40–48 percent (8), and between the second trimester and term, 23–30 percent (2). These data are not precise and do not accord exactly with the dates of NTS and serum screening. For the purposes of the primary modeled analysis, the assumption was a 40 percent rate of spontaneous miscarriage of trisomy 21 between the time when NTS would be performed in the first trimester and term. The rate of fetal loss between the time when a second trimester serum screen would be done and term was assumed to be 23 percent.

A systematic review of the literature (MSAC report 2002 Table 21 (4)) suggested that the detection rate for triple-marker serum screening in the second trimester (alpha-fetoprotein, total human chorionic gonadotropin, inhibin A) was no more than 64 percent with a 5 percent false-positive rate. CVS or amniocentesis is offered to those who have a risk in excess of 1:250 in second trimester and 1:300 in first trimester either determined through screening or because of the risk associated with maternal age. Amniocentesis has been reported to have a procedure-related fetal loss rate of 0.9 percent (95 percent confidence interval [CI]=1.2 to 0.6 percent) in the second trimester of pregnancy (10). The current uptake of amniocentesis among women with a positive screen has been reported as 85 per cent (6). The vast majority of women whose fetuses are diagnosed as having trisomy 21 seek elective terminations of the pregnancy, and 90 percent was assumed in this analysis.

Yates et al. (11) estimated in 1991 and 1992 that 97 percent of pregnant women had at least one ultrasound scan and that 46 percent had two or more. NTS may substitute for the current use of ultrasound in the first trimester, with considerable implications for the cost-effectiveness of ultrasound screening. If we take the current Medicare fee for genetic defect screening ultrasound at 17–22 weeks ($100) as indicative of the likely cost of a first trimester NTS, then the extra cost of NTS over a current ultrasound examination at less than 16 weeks ($60–70) would be only $30–40. The primary analysis assumes that there will be no substitution for current first trimester ultrasound, but the sensitivity analysis considers an 80 percent substitution in the first trimester at a marginal cost of $40. This approach is in contrast to that of Gilbert et al. (3) who, based on the cost structure of the United Kingdom, assume a very low marginal cost for NTS.

Evidence from Snijders et al. (7) suggests that, with a false-positive rate of 5 percent, the sensitivity of NTS for trisomy 21 was 77 percent (95 percent CI=72 percent, 82 percent). A combination of NTS, double-serum markers, and age were assumed to raise the detection rate to 86.4 percent. Women who screen positive would be offered CVS or amniocentesis at a fee of $683 and $782.90, respectively. The uptake of diagnostic testing after a positive screen was assumed to be 85 percent with a 100 percent detection rate and a procedure-related rate of fetal loss of 0.9 percent, irrespective of the test. The cost of a termination is $1,016.

An additional cost of counseling was added for those who have a positive screen and for those who decline further diagnosis while the cost of an additional ultrasound scan was added for those who went on to have a diagnostic test.