Thursday, July 18, 2013

INPO: Traits of a Healthy Nuclear Safety Culture

The Institute of Nuclear Power Operations (INPO) has released a document* that aims at aligning their previous descriptions of safety culture (SC) with current NRC SC terminology.  The document describes the essential traits and attributes of a healthy** nuclear SC.  “[A] trait is defined as a pattern of thinking, feeling, and behaving such that safety is emphasized over competing priorities. . . . The attributes clarify the intent of the traits.” (p. 3)  While there is an effort to align with NRC, the document remains consistent with INPO policy, viz., SC is a primary leadership responsibility.  Leaders are expected to regularly reinforce SC, measure SC in their organization and communicate what constitutes a healthy SC.

There are ten traits organized into three categories.  Each trait has multiple attributes and each attribute has representative observable behaviors that are supposed to evidence the attribute's existence, scope and strength.  Many of the behaviors stress management's responsibilities.  The report has too much detail to summarize in this post so we'll concentrate on one of the key SC artifacts we have repeatedly emphasized on this blog: decision making.

Decision making (DM) is one of the ten traits.  DM has three attributes: a consistent process, conservative bias and single-point accountability.  Risk insights are incorporated as appropriate.  Observable behaviors include: the organization establishes a well-defined DM process; individuals demonstrate an understanding of the DM process; leaders seek inputs from different work groups or organizations; when previous decisions are called into question by new facts, leaders reevaluate these decisions; conservative assumptions are used when determining whether emergent or unscheduled work can be conducted safely; leaders take a conservative approach to DM, particularly when information is incomplete or conditions are unusual; managers take timely action to address degraded conditions; executives and senior managers reinforce the expectation that the reactor will be shut down when procedurally required, when the margin for safe operation has degraded unacceptably, or when the condition of the reactor is uncertain; individuals do not rationalize assumptions for the sake of completing a task; and the organization ensures that important nuclear safety decisions are made by the correct person at the lowest appropriate level. (pp. 19-20)  That's quite a mouthful but it's not all of the behaviors and some of the included ones have been shortened to fit.

In addition to the above, communicating, explaining, challenging and justifying individual decisions are mentioned throughout the document.  Finally, “Leaders demonstrate a commitment to safety in their decisions and behaviors.” (p. 15)

Our perspective

On the positive side, the INPO treatment of DM is much more comprehensive than what we've seen in the NRC Common Language Path Forward materials released to date.

But the DM example illustrates a major problem with this type of document: a lengthy laundry list of observable behaviors that can morph into de facto requirements.  Now INPO says “. . . this document is not intended to be used as a checklist. It is encouraged that this document be considered for inclusion and use in self-assessments, root cause analyses, and training content, as appropriate.” (p. 3)  But while the observable behaviors may be intended as representative or illustrative, in practice they are likely to become first expectations then requirements.  An overall tone of absolutism reinforces this possibility.

The same tone is evident in the discussion of DM's larger context.  For example, INPO asserts that SC is a board and corporate responsibility but explicit or implicit priorities from above can create constraints on plant management's DM flexibility.  INPO also says “Executives and senior managers ensure sufficient corporate resources are allocated to the nuclear organization for short- and long-term safe and reliable operation” (p. 15) but the top and bottom of the organization may not agree on what level of resources is “sufficient.”

Another problem is the lack of priorities or relative importance.  Are all the traits equally important?  How about the attributes?  And the observable behaviors?  Is it up to, say, a team of QA assessors to determine what they need to include or do they only look at what the boss says or do they try to evaluate everything even remotely related to the scope of their inquiry?

But our biggest difficulty is with this statement: “These traits and attributes, when embraced, will be reflected in the values, assumptions, behaviors, beliefs, and norms of an organization and its members.” (p. 3)  This is na├»ve absolutism at its worst.  While some members of an organization may incorporate new values, others may comply with the rules and exhibit the desired behavior based on other factors, e.g., fear, peer pressure, desire for recognition or power, or money.  And ultimately, who cares why they do it?  As Commissioner Apostolakis said during an NRC meeting when the proposed SC policy was being discussed: “[W]e really care about what people do and maybe not why they do it. . . .”  (See our Feb. 12, 2011 post.)

We could not say it better ourselves.

*  Institute of Nuclear Power Operations (INPO), “Traits of a Healthy Nuclear Safety Culture”  INPO 12-012, Rev. 1 (April 2013).  The report has two addenda.  One describes nuclear safety behaviors and actions that contribute to a healthy nuclear SC by organizational level and the other provides cross-references to other INPO documents, the NRC ROP cross-cutting area components and the IAEA SC characteristics.  Thanks to Madalina Tronea for making these documents available.

** INPO refers to SC “health” while the NRC refers to SC “strength.”


  1. I've had occasion in recent days to be reviewing Kahneman's "Thinking Fast and Slow." It occurred to me as the Myth of the Rational Economic agent was being taken under a wrecking ball to wonder how one might inquire: To what extent is the content of INPO 12-12 Rev. 1 reflective of the empirical results Tversky and Kahneman produced over a couple of decades.

    As decision-making is one of the major focus areas of this research, and as Lew has illuminated the sense of the INPO discussion on DM, one might expect the new standard (with all its elaborate detail, and sub-detail) to reflect the sense of this Nobel Prize winning research.

    Others might be able to find the correlation, but not I. Lew's characterization of "naive absolutism" must be applied to the entire document (and its two companion pieces) - they simply lack a realistic context in which to imagine this material being used in any fashion other than a checklist of superficial features of very complex performance management systems.

  2. Any list of safety culture traits that does not clearly espouse and strongly emphasize competence, integrity, compliance, and transparency is a cruel fraud.


Thanks for your comment. We read them all. The moderator will publish comments that are related to our content.