Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
Feasibility of prenatal screening strategies--factors to consider
- Sandra Farrell, Tianhua Huang, Nanette Okun, Nathalie LePage (1 April 2009)
-
Cost of Integrated, sequential and contingent prenatal screening for Down’s syndrome
- Jonathan P. Bestwick, Nicholas J. Wald and Alicja R. Rudnicka (2 April 2009)
-
Cost effectiveness of contingent and integrated screenings for prenatal Down syndrome risk evaluation
- Jean Gekas, Audrey Durant, David Gradus van den Berg, Geneviève Gagné, Emmanuel Bujold, Jean-Claude Forest, Daniel Reinharz and François Rousseau (7 May 2009)
-
Computer simulation modelling for comparison of different strategies in prenatal screening for Down's syndrome
- Jean Gekas, Audrey Durant, David Gradus van den Berg, Geneviève Gagné, Emmanuel Bujold, Jean-Claude Forest, Daniel Reinharz and François Rousseau (7 May 2009)
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Sandra Farrell, Medical Chief, Clinical Geneticist, Genetics, The Credit Valley Hospital, 2200 Eglinton Avenue West, Mississauga, ON L5N 2N1 Canada, Tianhua Huang, Nanette Okun, Nathalie LePage
Send response to journal:
|
We have read the article "Comparison of the different strategies in prenatal screening for Down's syndrome: cost effectiveness analysis of computer simulation"1 with great interest. The authors applied statistical modeling to data collected as part of the SURUSS study2 to compare screening performance and cost effectiveness of integrated screening, contingent screening and sequential screening. Their results indicate that to achieve the same screening performance, contingent screening is the most cost effective if one considers global costs and the cost effectiveness plus incremental cost effectiveness ratios. They concluded that contingent screening, with a first trimester risk cut-off of 1 in 9, is the preferable screening strategy. While mathematical modeling indicates contingent screening is a promising option, we are concerned about the logistics of implementing this type of screening. Compared with conventional first or second trimester screening or with integrated prenatal screening, contingent screening inevitably would result in additional counseling of patients and increased administrative costs, creating an increase in the workload not only for genetic counselors but also for the primary care providers ordering the screen. For example, in the article by Gekas and colleagues, 1 the risk cut-off levels are set to enable 78% of the women to receive a screening result in the first trimester, but this means 22% of the screened population will require further discussion about the option of completing the screening process by having the second part of the screening. Furthermore, as recognized in the article, a significant amount of education would be needed for health care providers and pregnant women to accept the idea that a woman with first trimester risk as high as 1 in 10 should wait for the result of the second trimester before considering a diagnostic test. We strongly suspect many of these women will demand a diagnostic test before the second trimester, which would alter the cost effectiveness of contingent screening. One also can imagine that a woman with a risk of 1 in 1500 might not be enthusiastic about having a second trimester test. If some of these women skip the second trimester test, the detection rate (DR) of Down syndrome would be expected to drop. In the modeling, the risk cut-offs were decided based on a population with a median maternal age of 27 years. To obtain the same false positive rate (FPR) and to ensure the same proportion of women would receive results in the first trimester, the risk cut-offs would need to be adjusted if the screening population has a different maternal age distribution. For example, in Ontario, the median age of women screened using integrated prenatal screening is 31.6 years. Using the risk cut-offs of 1 in 9 and 1 in 2000 respectively for the first and second trimester cut-offs, only 71% of the Ontario screened population would receive a final first trimester result (using parameters from the SURUSS study). Hence, the screening performance and cost effectiveness of contingent screening could be different from those estimated by Gekas and colleagues if the median age differs. The calculation of the cost effectiveness ratio could be misleading since some of the additional cases of Down syndrome diagnosed by contingent screening by first trimester screening would be lost before the time of the second trimester component due to the higher spontaneous fetal loss rate see in affected pregnancies. It appears biased to include these cases in the denominator of the cost effectiveness ratio calculations. Integrated screening was introduced in Ontario in November, 1999, using first trimester nuchal translucency and maternal serum pregnancy- associated plasma protein A (PAPP-A) and the second trimester markers, alpha-fetoprotein (AFP), unconjugated estriol (uE3) and total human chorionic gonadotrophin (hCG), in combination with maternal age. As of February, 2009, over 260,000 women have been screened in this manner. Results of routine population screening population show the detection rate is 88%, with a FPR of 3.0%. Of women who started integrated screening, 94- 95% completed second part of the screening. These clearly shows women are prepared to wait until the second trimester result is available to complete screening, and that they comply with screening. It should be noted that the 5-6% "drop out" also includes women who had a spontaneous miscarriage between the first and second trimester samples.3 4 In 2008, 64% of our screening tests were integrated screening. Since the introduction of integrated screening, we have observed a steady increase of the screening uptake, from just below 50% of all pregnancies in 2000 to approximately 65% in 2008. Our experience indicates integrated prenatal screening is well accepted by pregnant women and their health care providers in Ontario. The Ontario experience shows that integrated screening can be implemented in a large population based program, covering a wide geographic area. Furthermore, it shows that integrated prenatal screening has performed as expected. In Ontario, when NT measurement is not available, the integrated screening protocol is modified to include PAPP-A in the first trimester and AFP, uE3, hCG plus dimeric inhibin in the second trimester. This serum screening option offers a relatively similar DR, with a slightly greater FPR, allowing for effective screening even if the NT measurement is not available, which occurs in the more rural regions where there might not be a sonographer trained to undertake NT measurements. With the rapid evolution of prenatal screening, new markers and screening protocols will continually emerge and be assessed. The choice of screening modalities has become a balancing act between early reassurance and additional information (for disorders other than aneuploidy and open neural tube defects), cost effectiveness and feasibility. As the authors indicate, their study did not consider the logistical problems that might occur in the reality of clinical practice, and we feel that the practicality of contingent screening should be carefully evaluated before being introduced as a large population based screening program. Sandra Farrell, Medical Chief, Clinical Geneticist 1 Tianhua Huang, Research Scientist 2 Nanette Okun Associate Professor, Obstetrician,3 Nathalie LePage, Clinical Biochemist4 1. Genetics, The Credit Valley Hospital, Toronto, Ontario, Canada 2. Genetics Program, North York General Hospital, Toronto, Ontario, Canada 3. Department of Obstetrics and Gynaecology, Mount Sinai Hospital, University of Toronto, Toronto, Ontario, Canada 4. The Division of Biochemistry, Children¡¯s Hospital of Eastern Ontario, Ottawa, Ontario, Canada References: 1. Gekas J, Gagn¨¦ G, Bujold E, Douillard D, Forest JC, Reinharz D et al. Comparison of different strategies in prenatal screening for Down's syndrome: cost effectiveness analysis of computer simulation. BMJ 2009;338:b138 2. Wald NJ, Rodeck C, Hackshaw AK, Walters J, Chitty L, Mackinson AM. First and second trimester antenatal screening for Down's syndrome: the results of the Serum, Urine and Ultrasound Screening Study (SURUSS). J Med Screen 2003;10:56-104. 3. Okun N, Summers AM, Hoffman B, Huang T, Winsor E, Chitayat D et al. Prospective Experience with Integrated Prenatal Screening and First Trimester Combined Screening for Trisomy 21 in a Large Canadian Urban Center. Prenat Diagn 2008;28:987-92 4. Summers AM, Huang T, Farrell SA, LePage N. Integrated Prenatal Screening for Down syndrome-The Ontario experiences. 14th International Conference on Prenatal Diagnosis and Therapy, Abstract 7-3, June 1-4, 2008, Vancouver, Canada. Competing interests: None declared |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Jonathan P. Bestwick, Academic Fellow in Medical Statistics Wolfson Institute of Preventive Medicine. EC1M 6BQ, Nicholas J. Wald and Alicja R. Rudnicka
Send response to journal:
|
According to Gekas et al (1), contingent screening for Down’s syndrome is about 31% less expensive than Integrated screening if alphafetoprotein (AFP) testing for neural tube defects (NTD’s) is not done and sequential screening is about 8% less expensive. In a similar analysis we reported that contingent screening without AFP testing for NTD’s was 15% less expensive than Integrated screening and sequential screening was 21% more expensive than Integrated screening (2). Given that the modelled population Gekas et al. used in their cost analysis was the same as the one we used apart from the maternal age distribution, the results in the two analyses should not be materially different. There are a number of reasons for the differing costs. These are set out in detail on the Wolfson Institute of Preventive Medicine website (http://www.wolfson.qmul.ac.uk/integratedcosts) In summary, (i) the cost effectiveness ratio (cost per Down’s syndrome pregnancy diagnosed) for sequential and contingent screening is calculated including Down’s syndrome pregnancies diagnosed in the first trimester that would have miscarried before second trimester testing: for a fair comparison with the Integrated test these should not be included, (ii) in their modelling, an incorrect false-positive rate is used for the Integrated test (too high), relating to 10 weeks gestation instead of 11 weeks as used for the other tests (iii) screening performance and risk cut-off estimates used for contingent and sequential screening include free β-human chorionic gonadotrophin (hCG) in the first trimester test but based on their definition of the tests seem to exclude it in the cost estimates, (iv) the 55% Down’s syndrome spontaneous fetal loss rate between the first trimester and term is high (higher than two estimates cited: 43% (3) and 30% (4)). These points contribute to the overestimation of the cost of Integrated screening relative to the costs of contingent and Sequential screening. Other issues which indicate potential problems in the modelling by Gekas et al. are on the previously mentioned website. For example with the simplest screening strategy (amniocentesis for women 35 years or older; 14.6% of women in their population) a total cost of screening 100,000 women is given as $C4.15 million but the cost of amniocentesis alone for women in their population comes to $C6.57 million (14,600×amniocentesis uptake rate [90%]×cost of amniocentesis[$C500]). The table shows the estimates in Table 3 of Gekas et al, revised on the basis of our previous unit cost and screening parameters (2,5). The table shows that programmes based on contingent screening (1 in 9 first trimester test risk cut-off) are 15% less expensive than programmes based on Integrated screening and programmes based on sequential screening are 15% more expensive. Per Down’s syndrome pregnancy diagnosed, Integrated screening costs about £2,300 more than contingent screening (much less than the $C12,000, about £6,750 at the current 2009 exchange rate, estimated by Gekas et al.). If AFP tests for open neural tube defects are performed in women that do not proceed to second trimester testing, there is no cost advantage of contingent screening programmes. Relative to programmes based on the Integrated test, in programmes based on contingent or sequential screening, up to 27 Down’s syndrome pregnancies detected in the first trimester would be unnecessarily terminated because they would have miscarried before the early second trimester. Table Cost effectiveness analysis of screening 100,000 women (including 305 Down’s syndrome pregnancies at the first trimester of pregnancy, 226 at the second trimester) for an overall 90% detection rate of affected pregnancies viable in the second trimester.
†Number of Down’s syndrome
pregnancies diagnosed that are viable in the second trimester of pregnancy, numbers
in parentheses are affected pregnancies diagnosed in the first trimester that
would have miscarried before second trimester testing. *Compared to amniocentesis
for women ≥35 The Integrated test for all women avoids the confusion of giving women two risk estimates in the same pregnancy, has a similar cost to contingent screening (the difference being much lower than estimated by Gekas et al.), has the highest detection rate, the lowest false-positive rate, the lowest number of diagnostic procedure-related unaffected fetal losses and the lowest number of unnecessary pregnancy terminations (2). On this basis, screening using the Integrated test for all women, not contingent screening, is the method of choice. JP Bestwick*, NJ Wald*, AR Rudnicka† *Queen
Mary University of London, Barts & The London School of Medicine and
Dentistry, Wolfson Institute of Preventive Medicine, Charterhouse Square,
London, EC1M 6BQ, UK †Department of Community Health Sciences, References 1 Gekas J, Gagne G, Bujold E, Douillard D, Forest J, Reinharz D, Rousseaeu F. Comparison of different strategies in prenatal screening for Down’s syndrome: cost effectives analysis of computer simulation. BMJ; 2009 2 Sequential and contingent prenatal screening for Down syndrome. Prenat Diagn 2006; 26(9):769-777 3 Morris JK, 4 Snijders R. Fetal Loss in Down Syndrome Pregnancies. Prenat Diagn 1999; 19:1180 5 Wald NJ, Rodeck C, Hackshaw AK, Walters J, Chitty L, Mackinson AM. First and second trimester antenatal screening for Down’s syndrome: the results of the Serum, Urine and Ultrasound Study (SURUSS). J Med Screen 2003; 10:56-104 Competing interests: NJ Wald holds a patent for the Integrated test. He is also a Director of Logical Medical Systems Ltd, which produces software for the interpretation of Down’s syndrome screening tests. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Jean Gekas, Medical Geneticist Unité de Diagnostic Prénatal, CHUQ, Québec city, G1V 4G2, Canada, Audrey Durant, David Gradus van den Berg, Geneviève Gagné, Emmanuel Bujold, Jean-Claude Forest, Daniel Reinharz and François Rousseau
Send response to journal:
|
We read with great interest the commentaries of Sandra Farrell and colleagues on our article "Comparison of the different strategies in prenatal screening for Down's syndrome: cost effectiveness analysis of computer simulation" (1). We previously discussed the facts that our results were based on mathematical modelling and that logistics of implementing this type of screening may affect the purported benefits of the contingent screening strategy (1). While the risks and benefits of different strategies have been partially reported on different populations (2-4), a detailed analysis of the cost effectiveness (CE) of these options was lacking. We believe that approaches that are most cost effective in one setting are likely to also show good performance in different settings that are nevertheless comparable such as a public health care system. With respect to maternal age, the demographic characteristics of the simulated population are comparable with many other western countries. Indeed, the mean maternal age and the proportion of women aged 35 years or older were comparable to the SURUSS (3) and the FASTER trial’s (5) populations which represented women in the United Kingdom and the United States. Moreover, the effect of maternal age distribution have been reported to be limited and unlikely to be large enough to influence DS screening policy decisions (6). However, we agree that the results may have been different for populations with a very different maternal age distribution, this could be simulated indeed. Doctor Farrell and colleagues estimated that including in cost effectiveness ratios DS diagnosed during the first trimester by contingent screening might bias the results because of the spontaneous losses before the second trimester. In the reported data (1), we estimated the cost per DS detected in 3 screening strategies (integrated test, sequential and contingent screenings). Since, the cost effectiveness ratios represent the cost per Down’s syndrome pregnancy diagnosed (7), we therefore think our results to be meaningful. Nevertheless, we agree with Sandra Farrell and colleagues, as stated in Figure 5 (1), that contingent screening is associated with more unnecessary terminations than integrated screening. 1. Gekas J, Gagne G, Bujold E, Douillard D, Forest JC, Reinharz D, et al. Comparison of different strategies in prenatal screening for Down's syndrome: cost effectiveness analysis of computer simulation. Bmj 2009;338:b138. 2. Ball RH, Caughey AB, Malone FD, Nyberg DA, Comstock CH, Saade GR, et al. First- and second-trimester evaluation of risk for Down syndrome. Obstet Gynecol 2007;110(1):10-7. 3. Wald NJ, Rodeck C, Hackshaw AK, Walters J, Chitty L, Mackinson AM. First and second trimester antenatal screening for Down's syndrome: the results of the Serum, Urine and Ultrasound Screening Study (SURUSS). J Med Screen 2003;10(2):56-104. 4. Wald NJ, Rudnicka AR, Bestwick JP. Sequential and contingent prenatal screening for Down syndrome. Prenat Diagn 2006;26(9):769-77. 5. Malone FD, Canick JA, Ball RH, Nyberg DA, Comstock CH, Bukowski R, et al. First-trimester or second-trimester screening, or both, for Down's syndrome. N Engl J Med 2005;353(19):2001-11. 6. Cuckle HS. Effect of maternal age curve on the predicted detection rate in maternal serum screening for Down syndrome. Prenat Diagn 1998;18(11):1127-30. 7. Caughey AB. Cost-effectiveness analysis of prenatal diagnosis: methodological issues and concerns. Gynecol Obstet Invest 2005;60(1):11-8. Competing interests: None declared |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Jean Gekas, Medical Geneticist Unité de Diagnostic Prénatal, CHUQ, Québec city, G1V 4G2, Canada, Audrey Durant, David Gradus van den Berg, Geneviève Gagné, Emmanuel Bujold, Jean-Claude Forest, Daniel Reinharz and François Rousseau
Send response to journal:
|
We would like to thank Bestwick J.P. and colleagues for their meticulous lecture of our simulation model and results presented in the article "Comparison of the different strategies in prenatal screening for Down's syndrome: cost effectiveness analysis of computer simulation" (1). Indeed, in our simulations, contingent and sequential screenings included human chorionic gonadotrophin (hCG) at the first trimester and the cost effectiveness analysis presented in Table 4 (1) includes the cost for the three markers (NT, hCG and PAPP-A) in the first trimester but hCG has been omitted in the definitions of screening procedures in Table 1 (1). They also suggested that the modelling presented should used a different false-positive rate for the integrated test, that would be more representative of the 11 weeks’ gestation than the 10 weeks’. While, the false positive rate available for a 90% detection rate was from data collected only at 10 completed weeks (2), data from the SURRUS trial for 85% detection rate showed no difference between 10 and 11 completed weeks: 1.2% false- positive rate in both situations. Using the data provided by Bestwick J.P. and colleagues (2.11% false positive rate for a 90% detection rate, risk cut- off of 1 in 198) that show a better efficiency of the integrated test for 11 weeks’ gestation, we performed a new simulation and cost effectiveness analysis for the integrated test: we obtained respectively, 3.261 millions ( Canadian dollars) for global costs, 36,089 (Canadian dollars) as the cost effectiveness ratio and 90.35 for effectiveness (DS detected). Theses new results are more accurate but change only slightly the relative rank of the integrated test in the cost effectiveness analysis presented previously in Table 4 (1). Bestwick J.P. and colleagues challenged the parameters used in our model, specifically regarding the spontaneous foetal loss rate between the first trimester and term which they estimated higher in our simulation than in the two cited references [43% (3) and 30% (4)]. However, the numbers of DS pregnancies shown in Table 3 (1) were reflecting all DS cases in pregnancies registered for the population at each pregnancy period in 2001. The rates of DS pregnancy losses seemed to be more important in our model than the two cited [43% (3) and 30% (4)] because Table 3 shows all DS pregnancies that will stop either spontaneously or due to voluntary pregnancy termination (for other reasons than a positive DS screen) in first and second trimester observed in this population in 2001. Theses DS pregnancies and their evolutions are also simulated in our model because some of these DS pregnancies were accessible to prenatal diagnosis at the first but not at the second trimester. We believe that our model is representative of the real situation of the studied population (number of pregnancies, affected or normal, available for screening at first and second trimesters or only at first trimester but not at second trimester). Again, although we believe our results are likely to be applicable to other public health care systems similar to Canada, cost-effectiveness studies using the exact parameters of other specific target populations may provide a more detailed estimate of costs and efficiency and, thus, may not find exactly the same values as those we presented. An overestimation of the cost of integrated screening relative to the costs of contingent and sequential screening could only be obtained by underestimating natural foetal losses for DS pregnancies between the first and second trimester testing. We applied stratified loss rates by gestational age as reported by Snijders (4-7) and as in Ball’s model (6) (7% spontaneous loss rate between 10 and 14 weeks). In Wald’s study (8), higher spontaneous loss rates were applied from other studies (3 9), which likely underestimated the efficiency of first trimester screening as previously noticed (7). The number of amniocenteses is defined from the number of pregnancies that are still in evolution at the time of amniocentesis in 35 years or older women when the maternal age alone (35 years or older) screening strategy is considered in our simulation (1). There is a need to consider the foetal losses of spontaneous or voluntary terminated pregnancies before this time (at first or early second trimester) but also the consent rates to the diagnosis procedure. Because it takes into account these considerations, the number of amniocenteses considered in our modelling is close to the real situation in our population and lower than the estimated number obtained by the analysis of Bestwick J.P. and colleagues of our simulation model. Our study confirms (8) that the integrated screening approach results in a very low number of procedure-related euploid miscarriages and unnecessary terminations because it allows for a diagnostic test only in the second trimester. But, if an integrated test was universally applied, no DS pregnancy would be detected and no patients could be reassured in the first trimester. This could be a disadvantage given that some studies suggest that women prefer an early diagnosis (10 11). 1. Gekas J, Gagne G, Bujold E, Douillard D, Forest JC, Reinharz D, et al. Comparison of different strategies in prenatal screening for Down's syndrome: cost effectiveness analysis of computer simulation. Bmj 2009;338:b138. 2. Wald NJ, Rodeck C, Hackshaw AK, Walters J, Chitty L, Mackinson AM. First and second trimester antenatal screening for Down's syndrome: the results of the Serum, Urine and Ultrasound Screening Study (SURUSS). J Med Screen 2003;10(2):56-104. 3. Morris JK, Wald NJ, Watt HC. Fetal loss in Down syndrome pregnancies. Prenat Diagn 1999;19(2):142-5. 4. Snijders R. Fetal loss in Down syndrome pregnancies. Prenat Diagn 1999;19(12):1180. 5. ACOG. First-trimester screening for fetal aneuploidy. The American college of Obstetricians and Gynecologists: Opinion 296. Obstet Gynecol 2004;104(1):215-7. 6. Ball RH, Caughey AB, Malone FD, Nyberg DA, Comstock CH, Saade GR, et al. First- and second-trimester evaluation of risk for Down syndrome. Obstet Gynecol 2007;110(1):10-7. 7. Snijders RJ, Sundberg K, Holzgreve W, Henry G, Nicolaides KH. Maternal age- and gestation-specific risk for trisomy 21. Ultrasound Obstet Gynecol 1999;13(3):167-70. 8. Wald NJ, Rudnicka AR, Bestwick JP. Sequential and contingent prenatal screening for Down syndrome. Prenat Diagn 2006;26(9):769-77. 9. Wald NJ, Watt HC, Hackshaw AK. Integrated screening for Down's syndrome on the basis of tests performed during the first and second trimesters. N Engl J Med 1999;341(7):461-7. 10. Mulvey S, Zachariah R, McIlwaine K, Wallace EM. Do women prefer to have screening tests for Down syndrome that have the lowest screen- positive rate or the highest detection rate? Prenat Diagn 2003;23(10):828- 32. 11. Spencer K, Aitken D. Factors affecting women's preference for type of prenatal screening test for chromosomal anomalies. Ultrasound Obstet Gynecol 2004;24(7):735-9. Competing interests: None declared |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||