Thursday, July 19, 2012

From task analysis to use error analysis: Designing out the human error in medical devices

Most organizations have a good process in place to ensure that the products they are designing meet a high degree of reliability and can withstand technical breakdowns. However, equally important and sometimes overlooked aspect is the breakdown that occurs as a result of the human operator. As a human factors scientist, I have been trained to never criticize the human operator. To err is human - because of our limited cognitive processing capacity. As a result, it is important to design out the human error, when designing systems. Here I describe a technique that can be employed to reduce human error in medical device interaction. Of course, this can be employed in any domain.  

Start with a task analysis of the new or proposed system. A plethora of ways to do a task analysis exists. I typically employ a hierarchical approach, wherein a high-level procedure is broken down into associated activities, functions, tasks, and steps. Of course, it is up to the researcher to determine the level of granularity associated with this breakdown. I typically employ this level of analysis (i.e., procedure- >activity->function->task->step) because this helps in identifying primary operating functions (frequent or safety-critical functions), which needs to be identified to comply with EN 62366 (a usability engineering standard that medical device manufacturers should follow).

The task analysis helps to systematically identify and document the environment in which the task is conducted, the operators involved, the information required to perform the task, and the sources of information.

Once the task analysis is complete, the use error analysis is conducted.
  • A failure mode and effects analysis is employed.
  • This approach starts with the task analysis and is conducted at the function level of the task analysis.
  • For each function in the task analysis, the user errors (e.g., errors of omission, errors of commission) and the associated hazards are identified.
  • The use errors are generated by analytically determining what can go wrong from the user’s end when interacting with the system. The list of user errors that are observed during user testing is then added to this list.
  • Based on the probability of occurrence of user errors and hazards, hazard estimates are computed. This is done by working with a cross-functional team.
  • Hazard estimates that exceed a certain value (this criterion is very much dependent on your domain) need to be mitigated through effective design solutions and training strategies.  

Usability testing should include scenarios that evaluate the frequent and safety-critical functions. These scenarios should help determine whether user errors that were identified in the analysis detailed above have been mitigated to a satisfactory level.

Photo credit Mk2010 via Wikimedia Commons.

1 comment:

  1. Although errors could not be totally avoided they should be minimized. Taking necessary precautions will help you have a more accurate result.