In the previous post Some Fun with Human Factors we looked at a maze with a mouse trying to get to the cheese to illustrate how confusing it can be navigate the many user interfaces one encounters in a typical hospital workday. Now onto to another cheese analogy in Human Factors: James Reason’s Model of Accident Causation.

more…

Reason uses Swiss cheese to illustrate the need to investigate all levels of an organization when studying accidents. Working backwards from the time of the accident outward, Reason’s model first looks at unsafe acts by the employees involved in the accident (Wiegmann & Shappell, 2003). This level is known as the active failures level as it is usually evident which employee actions contribute to the mishap. The next three levels out in the model address latent failures or those conditions or actions that are not as evident in their effect on the accident (Wiegmann & Shappell, 2003). Level two is known as preconditions for unsafe acts. Crew Resource management is a key factor in this. Things such as poor communication systems, fatigue, experience, unclear or unfollowed policies and teamwork are examined to see if these issues facilitated or failed to prevent the unsafe act (Wiegmann & Shappell, 2003). Further out but just as influential is unsafe supervision. Here we look at supervisor presence or absence, decisions, delegations etc. As mentioned in the post Pat Tillman and Fratricide, according to the military’s reports on incidents of friendly fire, leaders and leadership issues have been identified as one of the top two contributers to incidents of this type, which result in disastrous injuries and death in soldiers. The final layer of human failure involves organizational influences. There may be organization decisions or mandates that in essence “set employees up to fail” (Norris, 2009).

Reason’s theory is that all these layers have the potential of contributing to the accident..but also to help prevent accidents.
One of the limitations in Reason’s model is in identifying what makes the holes in the Swiss cheese. One option to help identify what human failures contribute to these holes is to understand the Human Factors Analysis and Classification system (HFACS). This system is based on the levels in Reason’s model. This will be discussed in detail in a future posting.

In addition to the HFACS, another approach to manage the “holes” involves a hierarchy of interventions developed by Hale and Glendon (Norris, 2009). These interventions also correlate with the levels described in Reason’s theory. The idea is to design for safety as far as out from the accident as possible. The four levels are (Norris, 2009):

Eliminating the hazard or forcing function should always be done whenever possible. This is the most effective way to create safety (Norris, 2009). An example is buying enteral products with tips that will not connect to any Intravenous products. When forced function is not possible, creating barriers should be the next step (Norris, 2009). Limiting access to certain drugs or having an IV team that performs all IV functions to reduce central line infections is an example of these levels of interventions. The next level if neither of the other two are feasible is to mitigate the consequences of a failure. For example, one would have an extravasations policy for administering and monitoring the IV administration of vesicant medications but the antidote for these medications should also be available and ready to use. Another example would be designing a playground with a soft flooring as described in a comment by MBW in response to the BP Oil spill post on this site. The least effective method for preventing accidents is educating the front line people (Norris, 2009). This is the closest level to the accident or the patient. Any failure in this level would directly lead to an accident. And education does not necessarily counteract human factors. Training is important but should not be considered to be the primary intervention for safety (Norris, 2009).