Diagnostic performance and costs of contingent screening models for trisomy 21 incorporating non-invasive prenatal testing

Background: Contingent screening for trisomy 21 using non-invasive prenatal testing has the potential to reduce invasive diagnostic testing and increase the detection of trisomy 21. Aim: To describe the diagnostic and economic performance of prenatal screening models for trisomy 21 that use non-invasive prenatal testing as a contingent screen across a range of combined first trimester screening risk cut-offs from a public health system perspective. Methods: Using a hypothetical cohort of 300 000 pregnancies, we modelled the outcomes of 25 contingent non-invasive prenatal testing screening models and compared these to conventional screening, offering women with a high-risk (1 > 300) combined first trimester screening result an invasive test. The 25 models used a range of risk cut-offs. High-risk women were offered invasive testing. Intermediate-risk women were offered non-invasive prenatal testing. We report the cost of each model, detection rate, costs per diagnosis, invasive tests per diagnosis and the number of fetal losses per diagnosis. Results: The cost per prenatal diagnosis of trisomy 21 using the conventional model was $51 876 compared to the contingent models which varied from $49 309-66 686. The number of diagnoses and cost per diagnosis increased as the intermediate-risk threshold was lowered. Results were sensitive to trisomy 21 incidence, uptake of testing and cost of non-invasive prenatal testing. Conclusion: Contingent non-invasive prenatal testing models using more sensitive combined first trimester screening risk cut-offs than conventional screening improved the detection rate of trisomy 21, reduced procedure-related fetal loss and could potentially be provided at a lower cost per diagnosis than conventional screening.