| Description |
Radiologists are frequently interrupted during the interpretation of volumetric medical images, such as chest CT scans. In prior research, we discovered that interruptions led to an increase in search time for the interrupted case (i.e., a transient cost), which was associated with impaired memory for which regions of the image had already been searched. In the current research, we tested the effectiveness of two interventions from the human-computer interaction literature at varying rates of interruption prevalence. In Experiment One, participants were given an auditory warning prior to the interruption on half of the interrupted cases. In Experiment Two, participants were shown a simple resumption cue that marked the most recent fixation and scrolling direction prior to the interruption. Overall, these interventions did not lead to a reduction in the time cost associated with interruptions. However, we found a large shift in overall performance as a function of interruption prevalence (i.e., a sustained cost). Greater interruption frequency tended to lead to faster search times and impaired nodule detection rates for the entire set of cases. This suggests that images interpreted in reading room environments with frequent interruptions may be approached less carefully. This finding represents an important paradigm shift from our current approach to understanding interruptions, which has largely focused on quantifying the immediate effects of interruptions. In addition, this research has important implications for which interventions may be successful in a given environment. |
