We have learned several lessons from these studies.

1. A high proportion of the interruptions actually experienced by physicians are relevant to a temporarily suspended task.
2. This is especially true for more senior physicians, whose caseload, and thus number of suspended tasks, is higher than for more junior physicians.
3. Relevant interruptions can increase overall efficiency in completing a demanding task, in part by reducing fruitless actions.
4. Relevant interruptions lead to better integration of the various sources of information, thus leading to higher scores on a comprehension test.
5. The substantial benefits of relevant interruptions may come at a cost to local performance. That is, whatever is in working memory at the time of an interruption may be forgotten. This is consistent with the traditional findings on this topic, since this constitutes an irrelevant interruption.
6. Although people with low working memory capacity may suffer the greatest impairments from irrelevant interruptions, they stand to benefit most from timely and relevant ones.

Several researchers in the area have argued strongly for communication training in Emergency Departments^i,ii. There certainly is room for improvement, given that communication errors appear to be responsible for nearly 40% of medical errors in critical care unitsⁱⁱⁱ and are twice as likely to lead to patient death as is inadequate clinical skill^iv. But our findings suggest that interruptions should not and should not be eliminated, or even drastically reduced. Instead, policy and training should target the ways that we can optimize the timing, nature, and management of interruptions so that they support ongoing and suspended goals and allow demands on working memory to be distributed across the team.