We must accept human error as inevitable – and design around that fact.         Donald Berwick

In the archive section, the principles are implied in the texts rather than clearly stated. Here are gathered those which can be drawn explicitly or “between the lines”, from the goals to some concepts proposed to get there.


The ultimate goal of safety is to save lives, i.e. avoiding accidents “before they occur”. ILO 2003 figures estimate that around 2 millions people die annually from work-related diseases. Taking into account lost working and production time, training, retraining and all expenses, the losses are about 4% of the world gross domestic product (GDP). But figures are abstract, so more concretely, already this case in 2020… Definitely worth focus and efforts.


To reach that goal, a group considered needs to share a sincere concern about safety and accept a policy. In other words, the group will have a Safety Culture. According to the European Union Agency for Railways (ERA): Safety culture refers to the interaction between the requirements of the Safety Management System, how people make sense of them, based on their attitudes, values and beliefs, and what they actually do, as seen in decisions and behaviours.


The core of the culture is that safety becomes a value. A priority can change, a value is stronger than rules, people live with it. At first they use to “know what safety is and do not need to learn it again”, so they must get persuaded that safety is not always intuitive and is more than they believe. Fortunately, as individuals and international organizations have reflected and written for more than 50 years, a comprehensive litterature is available for the homework.


The first task of a safety representative will be to shift the members’ mindset from the belief they already know safety, through the awareness they still have to learn, to a detailed knowledge and accepting a new value. Now they can change their behaviour. Instead of trying to get compliant (having to do this) and follow the rules, they are convinced (wanting to do it), committed in the common effort and willing to promote the policy, for an optimal efficiency.


The difference between fault and error (= mistake) must be understood: a fault is a conscious legal infringement, while there is no harmful intention in an error. To punish for mistakes prevents future reports: deeply counter-productive. Anonymity promotes reports instead. An interesting concept is normal error: there is no perfect safety and there are always new unpredictable situations, so a few adverse events remain with the best possible safety, the “normal errors”.


Now let us find out what sets apart accidents and incidents or even “close calls”: luck. Explaining human losses (the so-called “90% of accidents from human mistakes”…) only by destiny or bad luck does not incent to change anything. On the contrary, seing all the many near-misses as “lucky disasters” leads to admit that relying on luck is unacceptable, because you never know how much you have left. Hence a policy to improve the overall system and organization design..


Finally, once the members are sincerely convinced by the policy relevance, they are caring for each other, whatever their rank. The mutual respect makes possible to talk and listen without judging, only focusing on the system improvement. The small daily events, near-misses, accidents waiting to happen can be studied and solved or at least mitigated, instead of only reacting after someone died. Proaction is a strong path towards prevention and saving lives.