Quality in Design - Medical Device Design: Innovation from Concept to Market - page 149

5 CONTROL 1 DEFINE

4 IMPROVE 2 MEASURE

3 ANALYZE

Figure 7.14
The 6 σ DMAIC wheel.

There are several building blocks that we can consider in terms of building quality into our
design process. The first is DMAIC - Define Measure Analyze Improve Control. I do not intend
to replicate a 6 σ handbook, but it is safe to say that the terms are pretty obvious. If we are to
undertake any quality-based activity, first we should “Define.” One could argue that our PDS
process has done this. The second activity in the cycle is “Measure,” after all if we have defined
important issues we should measure our actual outputs to see if we have met them (remember
HoQ). The third activity is “Analyze” - just measuring data is not enough - we need to examine
the data robustly to determine what has been achieved and what is lacking. The fourth process
is “Improve,” which follows directly from analysis (DoE and FMEA fit in here). The last is
“Control” - it is pointless suggesting all of the former if we never know if they have been done,
or if they have been done correctly (remember how much time we spent on procedures).

This is not to say I have given you a Six Sigma design manual, but if you ever attend a
Six Sigma course you will understand many of the tools they present to you. Also, it means
that you should appreciate that you are well on the way to having a quality product. We shall
be meeting some more 6 σ tools later.

7.8 Summary
In this chapter we have examined some of the numerous tools used to ensure quality in
design. Explicitly we have examined Design of Experiments, House of Quality, FMEA,
Ishikawa diagrams, and more. It is important that you put these into practice, as too often
people do them “after the event” to try and prove to an auditor that they have a quality design
process. This stands out like a sore thumb; it will not fool anyone. More importantly you are
actually fooling yourself. The tools are there to help you to develop a robust design that meets