Statistical Design of Biomedical Devices and Systems

The aim of this article is to explore the use of statistical design in biomedical engineering applications. These techniques are not extensively used in this discipline possibly due to the current content of most curriculum. They may however be very valuable in optimizing the performance of biomedical devices and systems. We first review the two major methods of statistical design, namely the Taguchi Robust Design and the Response Surface Methodology. Examples of applications are given from other disciplines and in particular electronic engineering. We then explore statistical design experiences in biomedical engineering. Examples from imaging equipment design are described. It is shown that lesion detectability can be increased significantly and therefore misses be reduced. The effectiveness and efficiency advantages of these methods with respect to classical experimentation methods for system optimization