There is a shift taking place in safety practice, a BIG shift, significantly influenced by the collective term Resilience Engineering and the work of Eric Hollnagel.
Resilience Engineering, in this context, takes in a whole range of safety science ideas and combines them to improve an organisation.
Importantly resilient systems continue operations in the face of incidents, stress or pressures that are applied to them.
Resilience is a mode of preparation, assuming that trouble is always going to happen, anticipating the unexpected, exerting proactive control, eliminating unwanted variability, and emphasising more helpful variability. As the system stretches, it can undergo some stresses and then return to its former shape or potentially even improve and get better. That's really what we're talking about when we say resilience!
The critical theme under this approach is that we maintain performance despite variability and uncertainty. Thus, when things go wrong, deviate or surprises strike, we can cope with that, and more importantly, we can be successful.
We want to increase the organisation's ability to make correct adjustments, vary in helpful ways, and make sure that we successfully achieve our objectives.
There are four critical principles we need to understand as we design and implement our systems.
Systems are complex under this model. Work is generally underspecified, wherein we don't quite know everything that's going on. Designers can't possibly anticipate, assume or develop systems that are perfect.
Safety is about amplifying variability that leads to success and minimising variability that leads to failure. There's less focus on component failures - the little parts of the Machine that break down - and the deviations or violations across different levels of the organisation.
Monitoring and revisiting risk models with a systems view is crucial so that the organisation as a whole understands how everything functions together, as opposed to an individual operation.
Proactive use of resources aids in responding to and anticipating adverse events. One of the problems we have with traditional approaches to safety is that often, the number of errors to correct drops to zero. Processes may start to be uncontrollable, where we don't know exactly what's happening and have no data coming in (such as no injuries, no incidents and no near misses). Do we, therefore, assume that everything's safe? No. Henceforth, safety is about the existence of measures that pre-empt failure, thus enabling an organisation to fail successfully.
Rules and procedures can manage interactions between people and variability. High uncertainty requires the flexible use of rules. Rules (and rules about rules) shouldn't have one right way but instead, provide guidance around the level of flexibility permitted. For example, process rules guide how we know what rules to apply in different situations, giving people autonomy and decision-making power to be flexible when things go wrong.
Rules shouldn't be thrown away, but they should be used sparingly and effectively. For example, a new starter may very much benefit from a prescriptive step-by-step task list of how to do the job, while a more experienced expert could conversely benefit from more flexibility, and capacity to adapt and adjust within that rule framework.