Normalization of Deviation and Diving

(This article was originally published in Quest magazine)

By Guy Shockey and Gareth Lock

On January 28, 1986, 73 seconds after launch, Space Shuttle Challenger (STS-51) exploded following the massive aerodynamic forces produced as the solid rocket boosters (SRBs) detached from the huge centerline fuel tank due to hot gases escaping from a failing seal. Unfortunately, this disaster was not a surprise to the engineering team from Thiokol, the manufacturer of the SRBs. In fact, they had predicted that the launch vehicle would be lost on the platform because the O-rings which sealed the joints in the SRBs would fail. The reason? They had never been tested in temperatures as low as forecast immediately prior to the launch. This prediction occurred during a telephone call with the NASA engineering team on the evening of 27 January but was dismissed by NASA as the Thiokol engineers did not have evidence of the O-rings failing in such low temperatures and NASA believed that the secondary O-ring would still maintain integrity if the primary failed. This is despite the backup never having been tested in this configuration and it was forbidden to use a system backup as a primary in a “Criticality 1” situation.

Of note, this wasn’t the first time the O-rings had failed. On STS-2, the O-rings had been damaged and this should have led to the programme being suspended as it was classified as “Criticality 1”. However, the engineering team completed a number of tests designed to see if the O-rings would fail: they didn’t. This validated a poor decision and the missions carried on. More damage occurred on future launches but each time they occurred, the time and cost pressures and an acceptance of increased meant that the missions carried on.

Diane Vaughan, in her book “The Challenger Launch Decision: Risky Technology, Culture and Deviance at NASA “coined the term “Normalization of Deviance” to describe the organizational behavior of NASA at the time. This can be simplistically described as resetting the ‘safe’ baseline based on deviations over time, deviations which did not end in failure and are therefore accepted as ‘safe’. The diagram below shows this drift towards failure. The difficulty is we don’t know when the failure will occur. When failure does occur, invariably we are only assessing the final delta and the not the delta from the ‘baseline’ to the current procedures.

In the 1990’s, a researcher named Rasmussen, still recognized as one of the key visionaries in the field of modern safety, wrote “there seems to be a natural migration toward the boundaries of acceptable performance in any active work organization” His views moved between accidents as being ‘normal’: ‘‘catastrophic system breakdown is a normal feature of systems which have self-organizing features and at the same time, depend on protection against rare combination of conditions which are individually effected by adaptation’’ and foreseeable ones: ‘‘in many cases, as judged after the facts, liabilities and losses could reasonably be anticipated, accidents were foreseeable and obviously preventable”.  This view was taken further by Amalberti et al in 2006 when examining violations in healthcare whereby he produced a model which examined the pressures faced by healthcare professionals which ‘encouraged’ behaviors which migrated performance and behavior towards failure. This model is shown below. Whilst based on healthcare professionals, the concepts are equally applicable to divers and the diving industry.

This migration has 3 phases: initial safe space of action, creation of borderline tolerated convictions of use (BTCUs) and normalization of deviance & tolerance of recklessness. The first phase is where the organization creates the rules, regulations and procedures for the operation based on the initial risk assessments. In diving, these could be considered the ‘rules’ by which we operate, albeit many of them are socially-acceptable norms as opposed to legally defined rules. In the context of GUE’s diving, these are the SOPs and Standards by which the GUE diving community operates by.

However, in the real world, these standards and procedures never survive contact with the reality of operations and humans adapt and overcome as a consequence, moving further from the original baseline and the interpreting the spirit of the rule differently commensurate with the pressures being faced by those at the front-line. This adaptability leads to the creation of BTCUs whereby drift occurs and is accepted as normative behavior in that community, primarily because nothing has gone wrong or if it did, then it wasn’t considered serious. For example, running low on gas, uncontrolled buoyant ascents or poor trim leading to silting out the conditions.

The final phase is the normalization of deviance whereby these deviations have been accepted as the new baseline and then more deviations stack on top of them. Again, this deviation continues to happen because of the lack of a feedback loop. This feedback loop can be in the form of formal or informal debriefs, or an adverse event happening which ends up with someone being scared, injured or killed. Indeed, research has shown that it often takes a significant negative outcome to occur before the baseline is recovered.

Incident reporting systems are poor at reflecting the reality of the drift which has taken place i.e. the establishment of BTCUs. The reason being is that they have become the socially accepted norm and therefore deviation is not considered an ‘at risk’ behavior or safety has been compromised.

To detect these deviations and the natural and systemic migration towards failure there is a need for a proactive approach in terms of monitoring and correction. GUE provides standards against which diving operations and training courses are conducted – this provides the baseline, the phase 1 in the model above. It also provides the ability to hold others accountable to a standard. Still, the proactive part of the equation is still missing and deviations will continue to build upon each other until eventually there is a significant event, such as the Challenger disaster, which causes everyone to take notice and look at how they can “reset the clock” as it were to the original GUE standards.

Without being aware of this entire body of knowledge and theory surrounding the concept of the normalization of deviation, during a busy last year of GUE instruction in a large community of GUE divers, it became obvious that drift was occurring in the greater Pacific Northwest GUE community. This was mostly observed in Tech 1 and Tech 2 classes where students arrived after several years had passed since their last GUE class and it didn’t seem to be instructor specific. It was also the case that students of the same instructor but now at tech 1 and tech 2 classes had occasionally adopted small but noticeable variations from SOP’s that were usually followed by the instructor’s comment of “I never taught you that”. Drift happens but when the sample group is small or when the group is spread out over time and distance, it becomes harder to recognize. In this case, there was a geographically confined group that was large enough and the time was short enough that drift became more noticeable.

These small but noticeable examples of normalization of deviation became significant enough that they began to affect the smooth transition of training between levels. It was a clear example of how new baselines are created. It also substantiated the view that unless something comes along that demonstrates why the original standard or baseline was important, then this new baseline may not even be noticed. The community is very active and there had not been any recognizable “incidents” that would have otherwise have shone a light on the negative consequences of drift.

The concept of building upon a solid foundation is one of the cornerstones of GUE training. While this can be understood, and is taught in entry level classes at the abstract level of the “idea”, this concept becomes much more relevant when the GUE student moves to upper level classes. It is at this point that the effect of these new normalized deviations observed became apparent. Small things which by themselves may not have been a big deal became impediments to further training. In several cases this required a “clock reset” before the training could move forward but, in some cases, it was more pronounced and would prove to be a safety concern.

Flash forward to a team of experience Tech 1 divers, doing their first experience dive of their Tech 2 class. The protocol for a bottom stage to back gas switch is built up from elements learned and practiced from the fundamentals level up through the Tech 1 level. The muscle memory of unclipping and clipping the long hose on the right shoulder d ring and the switch to back gas protocol are building blocks of prior classes. What is also emphasized during this time is the very important element of team positioning and team awareness and how this is a time for the team to ramp up their level of alertness. What is new at the Tech 2 level is the combination of these simple elements at depth. We of course practice this repeatedly in shallow depths before doing it for real, and even then, we gradually work our way deeper. In this case, it was the students first time breathing a bottom stage at 170’ (51m) and until you have done this, an appreciation for how quickly the bottom stage pressure drops is usually missed. What should have been a simple sequence of switching to back gas at a preselected pressure became much more dramatic when the bottom stage was breathed dry… Again, as you read this, the armchair quarterback will point out that it is a simple matter to switch to your necklace if you have an unexpected interruption in your stage gas supply and then with no stress, unclip and switch to your long hose. What this same observer is not accounting for is the mental stress that can be present on this first T2 experience dive where a student’s actions have very non-armchair consequences if done incorrectly.

It is here that the importance of why we stress team awareness and positioning during gas switches became clear. The team had managed to separate and, because they had done gas switches many times from back gas to deco gas, they were not only further apart from each other than they should have been but they were also not paying attention to their team mates gas switch. They had done this many times before and nothing had ever gone wrong. Suffice to say all ended well with the instructor about 1’ away from the student looking directly into their mask while things settled down but it certainly could have ended in a more dramatic fashion.

In retrospect, it was clear that in a non-class setting, the team positioning and awareness could have been critical to avoiding an incident. It was also clear that this was not how the students were taught in their previous level of training (by different instructors in this case) to monitor their team during a gas switch, nor had it not been addressed during the critical skills component of the Tech 2 class. However, over time, because nothing had happened when they gradually lowered their level of alertness during this phase of the dive, a new baseline had been created and there had been no event that acted to close the feedback loop and reinforce the original baseline. A new baseline had been created and it was now the new standard. Suffice to say, the debrief from that dive reasserted the importance of the original protocol and the new converts became the best advocates!

This one incident proved to be the tipping point for critically examining the level to which normalization of deviation had quietly snuck in the back door of the community and it was at this point that the concept of a “workshop” to address drift in our community started to form.  As GUE instructors we address this issue through regular requalification training and through a more informal network of an instructor email list where we regularly address common topics, etc. However, there was nothing at the time, short of taking another GUE class that was available for GUE divers who were not instructors. One community member observed that they would happily pay for a “workshop” where they could revisit the initial baseline and refresh their skills and procedures. At the same time, new decompression and gas planning protocols had been introduced to the GUE training syllabus and it seemed we might be able to accomplish multiple goals at the same time where we not only addressed drift but also brought community members up to speed with new protocols and standards.

From these ideas, the GUE Leading EDGE Performance Workshop in the Pacific Northwest was born. Several of us within the community discussed the feasibility of such a workshop and as we reached out to other members for interest, we received a very high level of support. As the idea blossomed, we realized that a three-day workshop would also provide recreational GUE divers the opportunity to work on upgrading their training level in the presence of an instructor.

Thus, this past March, the GUE affiliates of GUE Seattle and GUE BC worked together to host a workshop that was intended to reaffirm the baseline that is provided by GUE standards and protocols. Our three-day workshop had 36 attendees, eight instructors and six community members who were the volunteer “glue” that helped the workshop go off without a hitch. We organized the participants into similarly qualified teams, where we identified common goals and then assigned them to a GUE instructor for three days of field drills, SOP and protocol review and lots of diving. All 50 participants came together in the later afternoon for dinner and group presentations.

Our presenters covered the topics of decompression and gas planning, as well as Human Factors, Project Baseline and the future of photogrammetry as a means of energizing Project Baseline initiatives. We were usually too tired in the evenings for much else but it was very satisfying to watch 50 GUE members from Canada, the United Kingdom and the United States come together in the spirit of family, working together and sharing our common passion of diving. We accepted a donation to our non-profit GUE affiliates in lieu of a workshop fee while the instructors and volunteers donated their time. We raised $10,000 for use in purchasing off-axis lighting for our future photogrammetry projects to continue to build our GUE community.

We have effectively “re-set the clock” in our participating communities and we are much more aware of how easily deviation can be normalized. In our discussions, we also more fully appreciated just how important an effective debrief can be after a dive. To this end we developed and applied a better after dive debriefing structure that we plan to use to make us better team mates and safer divers.

Our workshop was very successful and accomplished all our goals. Following the debrief structure we even sent out a post workshop survey to find out what participants thought we could do better next time!  We will continue to stay alert for drift and look forward to our next workshop!

Then end.

Share this post

Share on facebook
Share on google
Share on twitter
Share on linkedin
Share on pinterest
Share on print
Share on email