This explored the concept of Latent Error Detection (LED) within junior and senior naval air engineers from the Royal Navy helicopter squadrons.
LED describes situations where potential system failures are successfully recovered by individuals via system-triggered detections of hidden errors. E.g. An individual forgets to replace an oil cap or a bolt and then an hour later suddenly remembers that they forgot to replace the oil cap.
Previous research has found that apparently spontaneous self-detection of latent errors post-task completion were influenced by time, location and other system cues. It was recognised that retrospective recall is susceptible to memory decay.
Although this uses the term error, the paper is strongly aligned to human performance and systems research where understanding how system factors influence human performance variability is critical. Further, they note that firms should leverage LED intervention techniques that “deliberately engage with system cues across the entire sociotechnical environment and full range of normal workplace behaviours”.
A combination of interviews and work diaries were used to track examples of how latent errors were detected and corrected post-activity.
Results
The majority of LED events were linked with simple tasks, where perhaps the routine nature of everyday tasks infer low cognitive arousal.
Authors speculate that perhaps most LEDs in this study reflect simple tasks rather than complex ones because safety system approaches tend to defend more critical complex tasks with more controls and checks compared to simple tasks.
For time factors, most LED events occurred within 2 hours of the error event. This 2 hour period is said to reflect a ‘golden window’ in which most LEDs occur and a prime opportunity to detect these instances of variability.
For environment, LEDs are thought to be more successful when they take place in the same or similar environment to where the error occurred. In this study, most LEDs occurred at work in the same or similar environment (eg an error in the hangar was detected in the aircraft maintenance office). This was expected as most LEDs occurred within 2 hours while engineers were still working.
Participants were asked if they intentionally reviewed tasks for errors or if the error came up for apparently spontaneous reasons. Intentional searches of past tasks were found to be 3 times quicker than spontaneous mechanisms for detecting latent error.
In most cases, error recall was associated with a cue. Authors said this “demonstrates the importance of system cues distributed across environments to trigger recall”.
For cues triggering recall, participants noted physical work-related cues, eg sounds, aircraft paperwork, toolboxes, equipment. When not at work, people recalled physical items like keys or indirect perceptions of work-related cues, like thinking about paperwork and then recalling that an error was made.
For LEDs not associated with a physical trigger – it’s reasoned that people may not have been aware of the trigger or it could indicate an autonomous mental scheme that recalls errors. Based on other research, the authors note that in some cases the general environment could be a trigger rather than specific cues.
Moreover, simply being immersed in a similar environment may trigger LED or “recreating associations [in their minds]”.
Other research indicated that written word cues are more likely to trigger past memories, and thus LED, than picture cues.
Also notably, only a few of the LEDs were associated with false alarms – suggesting that latest errors are usually identified correctly by operators.
Moving forward, it’s said that firms should “consider the ‘strength’ of existing cues as an intervention to help enhance an error-detecting environment” (p111). They argue that if physical objects, in this case aviation equipment or aircraft paperwork strongly influence accurate LED, then this may offer other avenues of engagement.
For example, “specific word cues strategically placed alongside data entry areas in aircraft paperwork”. Further, aircraft paperwork is said to be completed in a separate location to where the maintenance is done, thus placing paperwork nearer to where work is conducted may improve LED conditions.
Further, displaying physical objects associated with common errors could enhance cue recognition, like a oil dipstick, padlock or even a scaled model of the aircraft.
They reason that LED may be most effective if conducted within the 2hr period and alone, to avoid distractions; preferably in the vicinity of where the task was executed. They suggest that interventions are especially important for everyday habitual tasks carried out alone.
Finally, they suggest that LED may “[compliment] Safety II scenarios” but also likely to support any other safety strategies in efforts to develop resilient potentials.
Authors: Saward, J. R., & Stanton, N. A. (2017). Applied Ergonomics, 59, 104-113.
Study link: https://doi.org/10.1016/j.apergo.2016.08.016
LinkedIn post: https://www.linkedin.com/pulse/latent-error-detection-golden-two-hours-ben-hutchinson
SafetyInsights.org 