I would like to congratulate the authors of a recent paper in this journal on the feasibility of an image scoring method for quality control of fetal biometry in the second trimester1. They have made a valuable contribution to the literature in an area of work that is too often neglected, and have carefully addressed the areas of intra- and interobserver reproducibility. I would also like to draw their attention to the earlier work of myself and my colleagues2, where we demonstrated the value of using such scoring as a feedback mechanism for training and coaching. We showed improvement in measurement quality over four cycles of audit, in both second and third trimesters, and demonstrated the value of the process in a wider trial in six centers3. We also found that comparing gestational ages from biparietal diameter and femoral length is a useful quantitative adjunct to image review.
In our work we showed that ultrasound practitioners show improvement in two stages. Given feedback on quality criteria not met, they first demonstrate improved recognition of the relevant features. They then begin to further develop the technical skills required to meet the quality criteria. In order to measure baseline performance and demonstrate progress against these two distinct stages, it is important to record the practitioner's own assessment of image acceptability; in our studies we have asked them simply to record whether the image was acceptable or not (prospectively they make this assessment using the predefined quality criteria).
Since this work was published we have carried out a larger scale study, as yet unpublished, including a large number of centers. In addition to head and abdomen measurements, we have reviewed femoral length and crown–rump-length images. We have also studied reviewer reproducibility; if a small number of reviewers discuss and agree the interpretation of quality criteria and review a number of images together, reproducibility is good, but cross-auditing, where each image is independently reviewed by two observers, is valuable. The approach is feasible on a large scale.
I urge all centers undertaking fetal biometry to adopt these quality-control procedures and to ensure that they record the practitioner's own assessment to facilitate the two-stage improvement we have demonstrated. We have shown that the process is effective in generating improvements with a single experienced reviewer. However, where scan numbers are high, for example where the process is run on a multicenter basis, multiple reviewers will be required and both intra- and interobserver reproducibility become critical elements. In this case cross-auditing and dual review where there are a significant number of cases of disagreement are helpful.