Organizations are constantly striving to optimize performance and improve systems. However, this is challenging when the performance is conducted by humans rather than machines. People don’t operate at a constant production rate, and some jobs may be more complex or stressful than others, which not only affects that job’s completion time but also completion times of subsequent jobs. An important question is: How much does emotional state matter in service operations settings?
Examples include police officers and paramedics who routinely respond to calls with limited information. The calls may involve routine issues, or they may encounter a critical incident—a situation that is sufficiently disturbing to challenge their coping mechanisms. For EMTs, this could involve a patient dying at the scene from a heart attack. How does that incident affect their ability to complete future ambulance calls during their shift?
Past studies on the effect of emotional state on job completion have almost exclusively used surveys. A research center would send out a survey to surgeons asking how many times in their career someone died on the operating table and how they reacted afterward. These are subjective, requiring doctors to describe their performance and feelings long after the event occurred.
To get more quantitative answers on the extent of stressful events on operational task performance, Hessam Bavafa of the University of Wisconsin-Madison and I conducted a study in collaboration with the London Ambulance Service. We used data on every ambulance call—including driving, patient pickup at the scene, and patient handover at the hospital—undertaken by the service in 2011. The dataset contained information on 902,002 ambulance activations, including timestamps, crew information, patient characteristics, and the receiving hospital.
We identified critical incidents as an ambulance activation where the patient has a high probability of dying at the scene. Our sample included 8,404 critical incidents, with at least one critical incident occurring as part of 5.5 percent of shifts. This setting was ideal for our analysis because the probability of any given paramedic crew being assigned a critical incident was effectively random.
We measured completion time of an ambulance activation and the following sub-processes. The completion time is defined as the duration from the ambulance dispatch to an incident until the ambulance is available again. Shortening completion times is an important goal for the ambulance service, and the sub-process times of handover and ambulance prep are official key performance indicators for paramedics.
Focusing on the short-term impact of critical incidents within a given shift, we found that encountering a critical incident significantly impacted subsequent completion times. Specifically, crews spent from 2.6 percent to 7.5 percent longer to complete the remaining ambulance activations of the shift.
The effect diminished as the paramedics conducted more ambulance activations following the critical incident. However, the magnitude of the tapering-off was small, so the overall impact lasted throughout the shift.
In addition, the study showed that the effect was not uniform across sub-processes of ambulance activation. There was a difference in the effect on the crew at the site where they must use creativity and insight, versus at the hospital where there are structures in place. We saw a much larger impact on work at the scene as opposed to at the hospital.
Finally, we observed that while more experience mitigates the impact of prior critical incidents on performance, the effect was outweighed by the fact that paramedic crews of older age are more susceptible to the detrimental effects of critical incidents. While experience is beneficial, the lowest impact of critical incidents was on younger workers.
These findings show that the impact of critical or traumatic events is not only an emotional matter, but it also has immediate short-term implications for organizational performance. Managers should consider how they assign jobs to teams with recent experience with critical incidents. Those workers may be better off assigned to lower priority tasks, where the impact of diminished operational performance is minimized.
The bottom line is that despite training and experience, emotional state matters.
Jónas Oddur Jónasson is a professor at the MIT Sloan School of Management. He is a co-author of Recovering from critical incidents: Evidence from paramedic performance, which is forthcoming in the INFORMS journal Manufacturing and Service Operations Management.