SafetyInsights.org

This surveyed 108 Swedish investigators (funnily enough) about accident investigation practices and investigator beliefs. Investigators were from a range of industries.

There’s a lot of findings, so I won’t cover all of them.

Results

Investigators detailed the time spent on different stages of an investigation. These were approximately planning at 8%, data collection at 30%, analysis 25%, report writing 24% and despite how important the investigators said that generating recommendations is, it was only given 12% of total time.

Investigators that emphasised a search for systemic factors were predominately from the patient safety group.

Concern about blame was present from some responses, e.g. people who were uncertain on the purpose of investigations, noting “‘people are worried that there will be blame and guilt attributed to the ones that did wrong’ or (that people) ‘believe that we are looking for the guilty ones’” (p861).

Responses also emphasised a ‘sometimes neglected positive function of event investigations’, being not necessarily the specific investigation findings, but also the process of engaging with people to understand their perspectives and surface insights from different parties.

Regarding data sources for investigations, data from safety analysis and audits wasn’t used to much higher extent. Thus, these investigators “do not make use of the potential data sources that in many cases exist” (p864).

For perceived causes of accidents, factors related to people, error, wrong behaviour etc. was the most common category (23% of responses). 20% of responses emphasised the organisational context (work conditions, weak standards). 21% recognises a combination of people and environment/organisation. Again, the patient safety group was more likely to address systemic issues rather than individual.

Interestingly, technical factors didn’t seem to be a strongly identified causal category amongst the groups, which according to the authors is presumably due to a mindset where technical weaknesses are seen as a symptom arising from people and non-technical factors and where technology “usually perform as expected”. [** Note: There’s plenty of accident reports which show that technology didn’t really operate as expected…]

Of further interest is that while considerable attention was given to circumstances where apparently “good rules” were breached, little attention was given to when people performed workarounds of apparently “bad rules”.

When asked who participates in the process of formulating recommendations:

• investigators were nominated in 44% of cases;
• whereas the “appointed experts” were only nominated in 15% of cases;
• those ordering the investigation + investigators in 12% of cases;
• those who were interviewed and investigators in 9% of cases;
• three or more parties who participate formulate recommendations (20%)

Indeed, the most common category was where “the investigators themselves suggest recommendations with no or little involvement by other parties in the recommendation phase” (p865). [** Note: You can also imagine the socio-political and power structures involved in investigations, where whoever gets to draw the line in the sand gets to determine the ‘appropriate’ recommendations.]

Regarding how recommendations should be formulated, some responses indicated things like simple or concrete, which the authors suggest could be a “possible bias for striving for oversimplified (and less effective) solutions rather than addressing some of the more complicated issues [of systems and culture]” (p867).

Further, few in the sample reported the use of systematic methods and tools to support finding effective recommendations.

Authors: Carl Rollenhagen, Joakim Westerlund, Jonas Lundberg, Erik Hollnagel, 2010, Safety Science.

Shout me a coffee

Study link: https://doi.org/10.1016/j.ssci.2010.04.001

Link to the LinkedIn post: https://www.linkedin.com/pulse/context-habits-accident-investigation-practices-study-ben-hutchinson