(176c) Improving Human Factors Review in PHA and LOPA

Human Reliability practitioners utilize a variety of tools in their work that could improve the facilitation of PHA-LOPA related to identifying and evaluating scenarios with a significant human factors component. These tools are derived from human factors engineering and cognitive psychology and include, (1) task analysis, (2) error-likely-situations, (3) human error rates, (4) systematic biases, and (5) System 1 (fast) and System 2 (slow) type thinking.

Human error is not random, although the absent minded slips we all experience seem to come out of nowhere. Instead, human error is often predictable based on situations created external or internal to the mind. Human error is part of the human condition (part of being a human) and as such cannot be eliminated.

For example, a task performed at high frequency (e.g., daily or weekly) develops a highly-skilled operator with an expectation of a low error probability for that task. However, as the operator’ skill increases, their reliance on procedures decreases, leaving them open to memory lapses caused by internal or external distractions. The fact that a skilled operator becomes less dependent on procedures is not a conscious decision. It is part of the human condition. Forcing a skilled operator to read the procedure while performing the task they are skilled at, is like asking you to think about what your feet are doing as you walk down a flight of stairs. In both cases a loss of adroitness will occur.

A large portion of this paper will be to describe with practical examples the five tools mentioned above. Task analysis is a talk-through and walk-through exercise of a task (typically focusing on one or two critical steps of a procedure) that is used to identify error likely situations (ELS). Quantitative human error rates can be attached to the ELS depending on if the error is associated with skill, rule, or knowledge based performance. Systematic biases produced by System 1 (fast) type thinking cause judgment and diagnosis errors related to response to abnormal situations.

Having a working knowledge of these five tools will improve a PHA-LOPA facilitator’s awareness and ability to better evaluate human error related scenarios and safeguards. In addition the facilitator will feel confident about recommending the need for a more detailed follow-up study such as an HRA (Human Reliability Analysis).