Objectives: To develop and apply a quality control system in a Down syndrome screening study using nuchal translucency as an interventional marker.
Methods: In a prospective Down syndrome screening study fetal nuchal translucency thickness was measured in 9236 of the 10 045 examined pregnancies. For quality assessment two models were introduced: firstly, image-scoring evaluation of the nuchal translucency thickness measurements and secondly, establishment of the distributions of nuchal translucency multiples of the median over time and the influence of intervention.
Results: The observer variability in the image-scoring evaluation was high with a kappa value of 0.48 in the overall validation. A revised model showed better interobserver agreement with a kappa value of 0.58; however, comparing the individual criteria the differences were still unsatisfactory, i.e. we found highly significant differences in the criteria "position of the fetus" (P = 0.0026) and "magnification of the image" (P = 0.0001). Regarding the distributions of the nuchal translucency multiples of the median, the median stabilized after a short learning phase representing the practical part of the sonographer's certification to nuchal translucency screening. In groups of medians of 50 nuchal translucency multiples of the median the intergroup standard deviation decreased from 0.100-0.060 after the learning phase to 0.046 after intervention.
Conclusions: When well-trained certified examiners perform nuchal translucency screening, continuous evaluation of the distribution of the nuchal translucency multiples of the median seems to be a good method to assess the quality for a center and may also be used to identify individual examiners deviating from the mean performance. The image-scoring methods we introduced cannot be recommended for quality control in a nuchal translucency screening program.