In Reply: We agree with all these authors that it would be useful to calculate an overall accuracy of ultrasound screening for fetal Down syndrome when structural abnormalities and ultrasonographic markers are used together. Unfortunately, studies that reported a composite ultrasound score (n = 18) reported statistically inconsistent results, and thus a reliable summary estimate could not be generated. The mean sensitivity (68%) and mean false-positive rate (8%) are not reliable, and we did not include them in the expected outcome data. It is meaningless to use these values to generate positive and negative LRs as the authors of these letters suggest. Although it is likely that a normal ultrasound result demonstrating no structural abnormalities or markers can reduce the liklihood of Down syndrome, the amount of this reduction is not known, and Dr Nyberg's estimate of a negative LR of 0.34 is not supported by our data.
Smith-Bindman R, Goldberg JD. Ultrasound Markers of Fetal Down Syndrome—Reply. JAMA. 2001;285(22):2856-2858. doi:10.1001/jama.285.22.2856