Predictive Human Error Analysis (PHEA)

Description

A taxonomy-driven walkthrough of a task analysis that predicts planning, action, checking, retrieval, communication and selection errors with their consequences and recovery.

PHEA is a close relative of SHERPA used heavily in chemical process safety. It uses a broader error taxonomy covering planning errors as well as execution, and emphasises performance-influencing factors such as time pressure, feedback quality and procedure clarity. Each task step is examined with the taxonomy, and analysts record error description, consequences, recovery points and error-reduction measures. Its strength is the explicit link between context (performance-influencing factors) and error likelihood, which helps teams see that the same interface can be safe in a calm setting and dangerous under interruption. In UX it can be run as a structured workshop with engineers and support staff, producing a shared error forecast.

Attributes

Loss type
Siloed
Stage
Design
Effort
Days

Steps

  1. Produce or reuse an HTA and agree the scenario conditions (time pressure, interruptions, user expertise).
  2. Walk each step with the PHEA taxonomy: planning, action, checking, retrieval, transmission and selection errors.
  3. List performance-influencing factors present for each step, such as poor feedback, similar options or fatigue.
  4. Record consequences and recovery opportunities, noting whether the interface makes recovery visible and easy.
  5. Rate likelihood qualitatively and prioritise steps with high consequence and low recovery.
  6. Specify error-reduction measures and assign owners; revisit after the next design iteration.

Sources

  • Embrey, Quantitative and qualitative prediction of human error in safety assessments (1992)
  • Kirwan, A Guide to Practical Human Reliability Assessment (1994)

Also known as

  • PHEA
  • Human error identification (HEI)

Categories

  • Accessibility
  • Human Factors & Error