The safety profession does have a bad reputation in some cases. One of the general perceptions is that safety work takes a lot resources and time without giving proper return on investment. While safety professionals will have a much more nuanced (and sometimes naïve and idealistic) view on the matter one has to admit that if you want to do a good job this can be demanding and time consuming. For example when one has to do a risk assessment. These strive optimally to be as complete as possible as possible. After that it’s an almost just as big job to get it documented in a proper, usable and understandable way.

Of course there are quite a few risk assessments (and other ‘safety’ activities) that are purely done as an administrative exercise as window dressing or to satisfy some requirement from regulations or internal procedures. Those should cut out immediately and stop unnecessary waste of time and energy. But that’s not what we talk about here. The question is if we can increase effectiveness majorly for the risk assessments that have to be done by working smartly and creating the possibility to relocate resources to actual improvement.

There is little literature on the subject so far (although I’m sure that many companies are practicing some form of recycling and reuse), but my friend Rune Winther has done some work in this direction to work towards a systematic method. Results so far are encouraging. For a series of similar assessments there was indicated an 80% reduction in use of resources from the first to the third assessment.

Winther’s paper describes that there is often a significant overlap of hazards between similar projects and the number of hazard and failure modes that are unique to a specific project may in fact represent a minor part of the hard and failure modes. There is a serious possibility that time is ‘wasted’ on identifying hazards and failure modes that are already well known and often adequately mitigated. Many safety professionals can probably confirm this finding from their own experience. The aim should be to use as little as possible effort on the identification process, while still being complete. Systematic reuse is one way to solve this.

However, one must watch out that reuse and recycling of material not just ends up in a simple copy and paste job. I remember all too well an episode sometime in the early 2000s when we were contacted by the Health and Safety Inspectorate that had visited a contractor and checked a couple of safety plans and other required documentation. They wondered why working on the rail tracks (which the project they had visited was all about) at one place included the painting of a railway bridge about 200 kilometers away.

Luckily there was no real harm done (if I recall well the inspector actually thought this was funny), but of course it does send a wrong message, it doesn’t really give an impression of safety management being taken seriously (the safety plans had after all been approved by our project manager) and worse: there are other ‘copy and paste errors’ thinkable that might even have led to dangerous, maybe even life-threatening situations. Another danger that lurks in cases like these is of course the possibility of systematically copying errors which may spread an isolated problem to other situations.

Also, if one wants to reuse information it’s important to check if the situation or system that has been assessed previously indeed is similar to the one at hand. One must one watch out that one doesn’t overlook things that are unique in the current situation when compared to previous assessments. Important keywords are the context (which may be quite different from situation to situation, even if they seem similar) and interfaces and interactions (for example as a consequence of the integration of a sub-system into the system as a whole).

To tackle these and other problems, Winther argues that reuse should be based on hazards defined on a subsystem level, because these hazards will be more generic than hazards defined on a system level. It’s important to use clear generic subsystem definitions because these are an aid to evaluate the relevance and completeness of generic lists of hazards and failure modes. Good definitions help to identify the differences between a generic case and each specific case.

As often with research, more is needed, but as many of us will assume from their own experience there is a good potential here to improve and do things in a smarter way. It’s important to go forward systematically and with caution, to accept some trial and error and to do a good deal of tinkering. The results may be very rewarding indeed. I for one am looking forward to hear about your experiences!


A Pragmatic Approach To The Reuse Of Qualitative Risk And Reliability Analyses - Experiences From Analyses Of Railway Traction Substations (Rune Winther, 2015, presented at this year’s ESREL conference).

Get the full article through Researchgate


Also posted on Linkedin