Using DATA Pro (TreeAge Software Inc., Williamston, MA), we created a decision-analytic model to compare different prenatal screening strategies for fetal T21 detection in a general screening population. The screening strategies compared consisted of: (1) first-trimester combined screening (FTS), which included the measurement of serum markers pregnancy associate plasma Protein A (PAPP-A) and β-hCG as well as first-trimester ultrasound, including nuchal translucency (NT) measurement, (2) integrated screening (INT) which included FTS as well as Quad screening of serum markers (AFP, estriol, hCG, Inhibin A) and (3) NIPT with cfDNA analysis in which NIPT was performed first line in women 35 years and older or in those with a medical or family history to place them at increased risk, or performed as a second line test in those who had a positive conventional screening test. The general structure of the Markov model is shown in .
Simplified Markov model diagram showing FTS, INT and NIPT screening flow.
We searched MEDLINE from 1997 to 2012 for English-language literature using the terms Down syndrome, trisomy 21, prenatal screening, non-invasive prenatal diagnosis, non-invasive prenatal testing and cell-free DNA analysis. In addition, we reviewed abstracts from national meetings, data from Medicare and relevant data from Ariosa Diagnostics (San Jose, CA), Sequenom (San Diego, CA) and Verinata (Redwood City, CA), which represent companies marketing a non-invasive prenatal test.
For the analysis, we used a theoretical cohort of 4 000 000 pregnant women which represents the current estimated annual number of births in the US. The analysis is based on the entire cohort of women undergoing prenatal testing in the first trimester for each of the screening strategies with screening uptake rates as per . For each screening strategy, the first branch assigns probabilities for those that opt for screening versus those that decline screening. For those that proceed with screening, tests can result in true positives, false positives, true negatives and false negatives. Invasive testing following a screening test is possible for each of the screening outcomes although the rates of invasive testing are higher for those that test positive versus those that test negative (). Fetal loss from invasive testing complications are captured as well as fetal loss from spontaneous and elective termination of pregnancies. In the model, fetal T21 is considered diagnosed only if confirmed by invasive testing. The estimated prevalence of T21 at the time of screening (first trimester) was 1 in 530 for the entire population and then adjusted accordingly when segregating into high- and low-risk women.
All costs are represented in 2012 USD. Cost items, which are listed in , included those associated with screening tests, invasive testing, office visits and counseling, termination procedures and birth of children with T21. When possible, the Medicare 2012 Fee Schedule was used to estimate cost inputs. A range of cost values based on published literature were used for a sensitivity analysis. The cost for screening and invasive testing was based on the total cost, which included any expected payments by insurance as well as patient co-pays. Cost for performing a screening test was inclusive of the blood tests and imaging. For FTS and INT screening test, a physician office visit cost was also included since NT requires a certified ultrasonographer and referral from a general practitioner or Ob/Gyn to a specialist may be necessary. NIPT testing following a positive conventional screening test also incurred a physician office visit cost. The baseline cost for NIPT was $795 based on the lowest published list price (Harmony Prenatal Test, Ariosa Diagnostics) and varied widely in the sensitivity analysis. The baseline cost for Down syndrome was estimated based on direct medical costs as well as indirect costs. In the sensitivity analysis, the cost of Down syndrome was also evaluated based solely on direct medical costs for the first 5 years of life as this may be of interest from a payer perspective [22
]. Costs were adjusted with the medical component of the CPI and future costs discounted at 3%.
The primary outcomes of the analyses were total costs of each screening strategy, number of fetal T21 cases diagnosed and number of non-T21 fetal losses due to invasive procedures for each screening strategy. The one-way sensitivity analyses were performed on all cost and effectiveness variables over the ranges specified in and . A two-way sensitivity analysis was performed on NIPT costs, FTS and INT costs, Down syndrome costs, termination rates, as well as detection and false positive rates of different screening modalities.
Clinical and cost outputs for base case of each screening strategy for the US population.