How Do You Actually Implement the NRC’s Safety Culture Policy Statement?

As we all know the NRC issued a policy statement on safety culture almost two years ago, presumably to set expectations for the industry to maintain strong safety cultures.  The policy statement is long on fuzzy things like traits and values but unfortunately short on specific “what do we do’s”.  The biggest hint may be in the invocation to accord safety issues the priority warranted by their significance.  Sounds right in theory but how does this actually get operationalized?

We are going to suggest a specific approach to apply safety culture policy in day-to-day nuclear management decision making.  As we’ve argued many times, focusing on specific decisions moves safety from the realm of theory to practice.  In general we think there may be significant advantages for organizations to do more to highlight, document and measure decisions involving safety considerations.  The benefit will be insight and emphasis on the “how and why” decisions are made and whether they achieve the safety balance envisioned by the NRC - and more importantly by the organization’s leadership.

We start with a numeric scale for what we call “decision quality”.  In our minds quality means how well a decision balances the priority accorded safety in light of the significance of the issue being addressed; in other words how well the decision does what the policy statement asks.  Conceptually this implies that an optimal decision achieves just the right balance for safety and that other decisions could under or over-shoot the optimal balance.  Can there be too much priority for safety?  Sure - remember the goal is to perform the nuclear mission safely, not to just pursue safety itself.  Here is the scale:

The scale sets a 0 value as the measure of an optimal balance of safety - meaning that it meets the expectation of the policy statement to give the priority warranted by safety significance.  Increasing positive values are associated with decisions that accord extra weight to safety; increasing negative values accord too little.  Use of a quantitative scale is the first step in being able to grade, track and provide feedback on decisions on a consistent basis.  When coupled with discussion of how significance was assessed and what the appropriate safety response needs to be, it provides many opportunities for a check and adjust process and organizational learning.

This leads to the next question which is: how should significance be determined?  There is of course NRC guidance via the significance determination process (SDP), including the red, yellow, white and green rainbow of significance levels, and this is our starting point.  The SDP include both qualitative (e.g. significant reduction in safety margin) and quantitative criteria (e.g. values of delta CDF and delta LERF).  While qualitative criteria may seem to some as lacking specificity, we’re fine with their use and in any event they are endemic in safety regulation.  We’re actually not that fond of the quantitative criteria since they are inherently hardware centric and do not encompass the complexity of the overall “system” that ultimately determines safety.  To provide quantification our approach is to again create a scale that correlates numeric values with the qualitative criteria.  Such “anchored scales” are a common and effective tool in decision analysis.  In addition we feel that the significance determinations need to be supplemented with an assessment of their uncertainty. 

It doesn’t take reviewing many event reports to see that judgments about safety significance are not always clear cut or unambiguous.  This variability in the adjudged significance can be the enabling mechanism for safety to not receive the appropriate priority - not because the priority doesn’t match the significance but because the significance has been discounted to justify a lower priority.  The catalysts can be as simple as overly optimistic thinking, normalization of deviation, complacency, or failing to ensure that the burden is on showing that something is safe versus showing that it is unsafe.  Our approach is to explicitly address the uncertainty of safety significance by introducing a second quantitative scale for this purpose.  When used together a judgment regarding significance would include both a nominal value (per SDM) and an uncertainty value.  These scales are illustrated below:

For decision making purposes the three scales would operate together to help arrive at appropriate decisions.  The significance scale would provide a nominal risk value.  If there was a little uncertainty in the assessed significance then the objective would be to make a decision that scores approximately “0” on the balance scale.  If there was greater uncertainty in the assessed significance the objective would be to select a decision option that scored higher on the balance scale; essentially giving safety higher priority to accommodate the potentially greater significance.  Decision options that rated negative balance values would avoided.

We see much of the value in this approach to be the focus on making the decision formation process more explicit, transparent and measurable.  Over time this structure provides greater opportunities for the organization to understand decisions and learn from the process not just the outcomes.  We also believe it may provide the basis for inferring and trending the safety culture within an organization.

In an upcoming post we’ll apply these decision scales to a specific plant situation to see how they might work in practice.