Page 110 - Safety Risk Management for Medical Devices

Risk Analysis Techniques 89

• People get stuck in long discussions on some items, which causes some partici-
pants’ attention to drift off.
• People get tired, creativity ceases, and generic vague answers are put into the
analysis.
• The product design process is already finished. There is no opportunity to
make any changes. Findings of high significance will cause great pains and costs
for the business. Therefore there is reluctance to do any deep analysis in the
fears of finding something significant.
• Granularity in rating scales is too high. When estimating a rating, a lower gran-
ularity scale is better than a higher granularity scale. A high granularity scale,
e.g., a 10-point scale, could lead into unnecessary long debates in choosing
between, e.g., a 6 versus a 7 rating because the differentiation between adjacent
ratings may not be so clear.

Some tips to help smooth the FMEA process and ensure success are as follows:

• First and foremost, use a skilled facilitator for the FMEA. Section 12.4.1
describes some of the responsibilities of an FMEA facilitator. Most FMEA par-
ticipants participate in FMEA work only occasionally. Therefore they become
rusty on the mechanics of the analysis. With coaching and guidance from the
facilitator, most participants climb the learning/remembering curve quickly.
• Where possible, try to reuse existing FMEAs to accelerate the work.
• Keep the duration of the sessions to less than 3 hours. Long sessions lead to
fatigue and reduced quality of work.
• Refresh participants on the ground rules, definitions of Failure Mode, Local and
End Effects, and be vigilant to ensure the entries in the FMEA are properly worded.
• If necessary, give a quick overview of failure theory, and definitions of Severity,
Occurrence, and Detectability.
• Make the definitions of the rankings for Severity, Occurrence, and
Detectability easily accessible, e.g., by printing them on posters and posting
them on walls.
• Make sure the agenda and objective of the meeting is clear, state it at the
beginning of the meeting and post it on a wall.
• Prework—participant should become familiar with the design under analysis,
before coming to the meeting. The FMEA session time is precious and should
not be used for explaining the basic understanding of the subject of analysis.
• Have physical samples, models, drawings in the room that participants touch
and use as discussion tools. Simply touching a physical sample is a great thought
stimulator. Also, it is much easier to convey thoughts and ideas about Failure
Modes using models/drawings.