RAT, I agree with the direction of your argument (#170), but also consider that whilst the ‘Agency-level’ defines the requirements, the implementation is delegated to national authorities, who in turn can delegate to operators. This increases the distance between intent and action enabling misbelief and assumptions to erode safety arguments. In order to understand safety, regulators need to get out of the office and use double loop learning.
Note that beliefs and assumptions are the primary aspects which should be considered in double-loop learning.
See
systems thinking and
learning and OODA including ‘5 whys’, all aspects of awareness and decision making.
A systematic approach to safety would consider many aspects other than the sharp end operations and the human entity. A new approach would not profess a solution; instead provide opportunity via a wider view to consider small interventions which might enable improvement whilst not disturbing what has already been achieved. We need to encourage thoughts about safety from different viewpoints.
One example is that the human is a resource for improving safety, not a hazard to avoid, be constraining by SOPs, or continually checked for consistency (humans are naturally inconsistent).
We should review the safety attitudes used in discussions of accidents – often human error, a hazard, blame, train; but for a few major success, Sioux City, Baghdad, The Hudson, the human is a hero, the provider of safety, yet the underlying behaviours are the same for success and failure. It’s only the outcome of an event which differs, which we often use to erroneously judge performance opposed to looking at wider is issues of managing surprise, workload, and the environment (see
ASAGA study as an example).