Statistical modelling is used to predict screening detection and false-positive rates for policies using different marker combinations and screening modalities. Parameters are derived wherever possible from meta-analyses. Screening with four serum markers yields a similar detection rate in the first trimester as it does in the second, and both have a similar detection rate to that of ultrasound nuchal translucency screening. Ultrasound anomaly screening has a low detection rate. The integration of the different screening modalities can yield very high detection rates. Some methods of integration are preferable to others. The concurrent use of first trimester serum markers and ultrasound nuchal translucency is the most practical. The combination of three modalities with nondisclosure of the intermediate results is more efficient, but the gain in detection is too small to warrant the wait for up to 4 weeks for results. The sequential use of modalities with intermediate disclosure is more practical but will generate higher false-positive rates.