Article Text
Abstract
Objective To explore the nature of interruptions that occur during clinical practice in the emergency department (ED). We determined the frequency, duration and type of interruptions that occurred. We then determined the impact on patient satisfaction of those interruptions occurring at the bedside.
Methods This was a cohort study of ED physicians and physicians in training. Trained research associates were assigned to an individual provider during 4-hour blocks of time during day and evening shifts. The research associates recorded the activity that was interrupted, as well as the nature and the duration of the interruption. If the interruption occurred during the principal interaction with a patient, the patient's satisfaction score was recorded on a 10-point scale.
Results Physicians were commonly interrupted in all clinical activities, but most frequently during reviewing of data (53%) and charting (50%). Bedside interruptions occurred 26% of the time, and had a negative impact on patient satisfaction. The majority of interruptions (60%) were initiated by another healthcare provider (physician or nurse). Interruptions only rarely resulted in a physician changing tasks before completion.
Conclusion Interruptions occur commonly during all clinical activities in the ED, and are frequently generated by providers themselves. These have a negative impact on patient satisfaction. The direct impact on medical errors or on provider satisfaction has not been determined.
- Interruptions
- patient satisfaction
- communication
- multi-tasking
- quality assurance
- anaesthesia, general
- emergency care systems, emergency departments
Statistics from Altmetric.com
- Interruptions
- patient satisfaction
- communication
- multi-tasking
- quality assurance
- anaesthesia, general
- emergency care systems, emergency departments
Introduction
The emergency department (ED) is known for its hectic environment and variety of clinical responsibilities. This inherently predisposes providers in the ED to frequent interruptions.1–5 Any interruption that occurs in this environment has the ability to disrupt a physician's working memory and subsequently derail the clinician from his or her current task.6 This may result in a loss of efficiency, be detrimental to the task at hand, slow progress and decrease patients' satisfaction regarding the quality of care they receive.
In many other fields, it has been shown that numerous interruptions and distractions are linked to an increased amount of error.7 8 It has been well established that distractions during tasks that require concentration can be detrimental to the outcome of the task. In driving, distraction through cell (mobile) phone use has been shown to delay response times and increase crash risk.7 In pharmacy research, distraction is correlated with errors in filling prescriptions.8 Only recently, however, have these interruptions been studied in other settings.
Several studies have begun to examine the interruptions that occur in the medical field. Observational studies done in Australia and the USA found that medical providers with higher levels of responsibility tend to get interrupted more often.1 Additionally, these studies found that emergency physicians are interrupted more frequently and manage more patients simultaneously than providers in office settings.2
In order to better understand the effect of such interruptions, it is essential to know what types of interruption occur, how often they occur and how they affect the ED workflow. This project was designed to further examine these interruptions in a structured manner. The frequency, duration and type of interruption were explored and quantified. Additionally, we examined the impact of these interruptions on patient satisfaction.
Methods
A cohort study was performed at a 70 000 volume tertiary care centre involving ED physicians (attendings) and physicians in training (residents) to further characterise the type of interruptions they experience during a typical shift. This study was reviewed and approved by the institutional review board. Trained research associates observed ED physicians during a convenience sampling of 4-hour blocks during both day and evening shifts over the 2-month study period. Physicians were observed for a total of 132 hours during the study period. Research associates recorded the activity of the physician during observation periods and noted the nature and duration of any interruptions that took place during that activity. Associates also documented whether interruptions were stacked, which we defined as more than one interruption occurring without the provider able to return to the primary activity in between. We also attempted to determine the ability of the physician to smoothly transition back to the original activity after the interruption. This determination was subjectively made by the research assistant. If the interruption occurred at a patient's bedside, the patient was privately asked to rate the level of satisfaction with the encounter on a verbal 1-10 numerical rating scale, with 1 indicating that the patient was extremely dissatisfied and 10 indicating that the patient was extremely satisfied. In addition, patients were privately asked to rate the level of satisfaction with uninterrupted encounters.
We a priori defined ‘physician activities’ as direct patient interaction, professional discussion, phone conversation, documentation or charting and reviewing data. We likewise defined interrupting events as phone calls, critical lab results, interruptions by another provider, overhead pages, pre-hospital or ambulance (EMS) arrival, sick patients or interruptions by other patients and their family members. Physicians could engage in more than one activity at the same time (eg, documenting while talking on the phone).
The Mann-Whitney test was used to assess for differences in patient satisfaction comparing those interactions that were interrupted to those that were not. The two-tailed t test was used to assess if there was a difference in interruption length for those physician tasks that were derailed versus those that were resumed smoothly. Descriptive statistics were used for all other aspects of data analysis.
Results
Over the course of the study, we were able to document 249 professional discussions, 226 patient interactions, 204 episodes of documentation, 140 episodes of reviewing data and 108 phone calls. Eight simultaneous activities were documented: one episode of professional discussion while charting, one episode of talking on the phone while reviewing data, three episodes of charting while talking on the phone, two episodes of professional discussion and reviewing of data simultaneously and one episode of reviewing data while charting. We observed 304 attending physician activities and 534 resident activities. For 93 encounters, provider type was not listed.
Physicians in our ED were most often interrupted while charting or reviewing data, with interruptions occurring 53% and 50% of the time, respectively (109 and 75 interruption events). These interruptions most often comprised interruptions from other healthcare providers. During professional discussions, physicians were interrupted 35% of the time (86 occurrences). These were most commonly interruptions by other providers, followed by phone calls. Interestingly, phone calls as a primary activity were only interrupted 15% of the time (table 1, figure 1).
Physicians were interrupted during 26% of their interactions with patients, most commonly by phone calls (39 patient interactions interrupted by 53 phone calls), but also to a lesser extent by other providers (24 occurrences). Only rarely were physicians interrupted by lab results or overhead pages while interacting with a patient. We were able to record patient satisfaction for 120 bedside occurrences. Patients were significantly less satisfied with all interactions during which they were interrupted versus those without interruptions (8.99 CI 0.42 vs 9.55 CI 0.18, p=0.02). There was no difference in patient satisfaction with or without interruptions when the primary interaction was with an attending (p=0.18), but patients were less satisfied with resident interaction when they were interrupted (p=0.03).
During four patient interactions, six episodes of charting, eight episodes of reviewing data and 10 professional discussions, stacked interruptions occurred. In seven occurrences, there were more than five interruptions for a single activity. In 14 activities, the interruption resulted in the provider aborting the current activity and engaging in a new activity after the interruption. These interruptions were most commonly generated by another medical provider.
Interruptions tended to be brief, with an average length of 0.78 minutes. Phone interruptions averaged 1.10 minutes, ECGs or lab results averaged 0.44 minutes, EMS interruptions averaged 1 minute, page interruptions averaged 0.32 minutes, interruptions by other providers averaged 0.64 minutes, sick patients averaged 2.28 minutes and interruptions by other family members or patients averaged 0.78 minutes. Interruptions resulting in derailing of physicians' tasks averaged 1.67 minutes, while interruptions after which physicians had a smooth transition back to their previous task averaged 0.68 minutes (p=0.20).
Discussion
The ED has been recognised to be a high-risk environment for patient care errors because of the high cognitive workload, the necessity for multi-tasking and the frequency of interruptions.1–5 Other studies have shown that emergency physicians spend up to 80% of their time in communication with others in the ED, and that interpersonal communication is frequently interrupted.1–5 It has also been demonstrated that providers with higher levels of task responsibility are interrupted more frequently than providers with less responsibility. Further, recent studies have focused on the chain of communication and interruption as a source of error in the ED.5 9 Our study adds to this body of knowledge. We demonstrate that physicians are interrupted in about one-third of their professional discussions and in over one-quarter of their bedside encounters with patients.
Although phone calls commonly interrupted providers during patient care, providers were rarely interrupted while actively speaking on the phone. A small part of this phenomenon may be attributed to the fact that the provider was already on the phone, and therefore not subject to incoming calls during the conversation. There is also the possibility that, at the time of the event, the person initiating the interruption makes an assessment as to the appropriateness and acuity of the interruption relative to the perceived importance of the provider's current activities. When a provider is speaking on the phone, it is likely much more difficult to determine the nature or importance of the call compared to when the provider is engaged in face-to-face communication. Therefore, people who might otherwise interrupt a face-to-face communication may avoid interrupting a telephone communication because they cannot assess the nature of the interruption.
Many of our interruptions were generated by other providers. A total of 302 provider-initiated interruptions occurred, which accounted for 59% of all interruptions (figure 2). This did not include phone interruptions, as we did not determine from whom the phone call came; however, many of those were likely provider generated, as well. A study by Brixey et al10 showed that providers are more commonly recipients than initiators of interruptions; however, our study shows that the vast majority of interruptions are actually provider initiated. This is supported indirectly by Fairbanks et al's study,7 which demonstrated a high volume of interprovider communication. It makes intuitive sense that providers who interact often have greater opportunity to interrupt one another.
In addition, we found that physicians have their tasks interrupted with a higher frequency when those tasks are not communication-based. When reviewing data or documenting on a chart, physicians were interrupted more than half the time. This could be because these activities may not seem time-sensitive or critical to an individual considering interrupting the activity. Alternatively, these activities may be interrupted more often because providers spend large amounts of time engaged in these activities in a public area (eg, at the nurse's station) where they may be more likely to be interrupted. The impact of this is unknown; however, it is known that any interruption has the potential to dislodge working memory and derail tasks. Interruption also causes a lag-time in return to the previous task, wasting valuable time.6 In our study, physicians only rarely changed tasks after an interruption (3%), which has been shown to be the case in a study by Brixey et al.11 However, these determinations were made by an impartial observer and therefore represent only obvious changes in physical tasks, and do not address changes in mental tasks. It is certainly possible that an interruption while writing an order might result in not ordering the proper intervention, or an interruption while checking laboratory values might result in failure to recognise a critical value. These sequelae would not be recognised by an observer, but could have a large impact on patient care and safety. Likewise, the number of interruptions encountered in the ED may result in a large and as yet unquantified waste of time while providers resume tasks and try to regain focus. The true amount of recovery time as well as the direct impact on errors secondary to interruptions has yet to be elucidated, but there is a growing body of literature supporting a link between interruptions, multi-tasking and error in the ED,9 and this should be an area of further research.
Although intuitively it would seem that interruptions that derail a task would be longer in duration than ones that do not, this was not found in this study. This may be because the urgency, or need to address the interruption in an acute fashion is independent of the length of the interruption. For example, a critical call from a pre-hospital provider might be of short duration, but result in a physician changing tasks to prepare for the incoming patient, while an interruption from a patient asking for an explanation of a study might be time-consuming, but not result in task derailment. Further study is needed in this area.
Our study demonstrated that patient satisfaction is also adversely affected by interruptions at the bedside. It is not a surprise that a patient who has waited for a long period of time to be seen by a physician would be dissatisfied with having his primary clinical interaction interrupted. However, this decline in satisfaction seemed more prominent when the patient was interrupted while being seen by the resident. This may be because attending physicians have more experience dealing with disgruntled patients, and may have better interpersonal skills. To our knowledge, this is the first study that has examined patient satisfaction in the context of ED interruptions. Future studies will look at the impact of ED interruptions on provider satisfaction and stress.
Interruptions in the ED are an intrinsic part of the environment. However, there may be ways to reduce the number of interruptions or their impact. Most of our interruptions were generated by other providers. When there is a critical emergency, interruption is appropriate, but this is often not the case. This issue can be addressed with provider education and awareness training. Interruptions might also be limited by systems changes. ‘Critical lab abnormalities’ could potentially be more narrowly defined to actually reflect clinically relevant values to result in fewer interruptions, or senior level residents might be given some responsibility for reviewing abnormal tests or ECGs. It might be feasible in some institutions to have a single provider in a lower acuity area in the ED be responsible for taking all transfer phone calls, all critical labs and ECGs, and all pre-hospital calls to reduce the interruption burden on providers in higher acuity areas. Providers in high acuity areas could then avoid wearing personal paging devices or phones altogether, which would reduce their interruption burden. A private area in the ED could be designated as a physician workspace for charting and data review. This would help minimise interruptions from patients and family members and allow a more interruption-free zone for working. Finally, training could be instituted for resident physicians to assist them in interruption management and multi-tasking strategies, as interruptions are unlikely to be eliminated completely from an ED environment.
Limitations
There are several limitations to this study. This was a cohort study of residents and attendings, and did not address nurses' or mid-level providers' clinical practice. We also did not assess the effect that interruptions had on the attitudes or satisfaction of the providers involved, or attempt to determine errors resulting from interruptions. In addition, this was an observational study and relied upon the interpretation of clinical events by the research associate, which may not be a true representation of what providers themselves find disruptive. Finally, this study was performed at a single institution and may not be generalisable to EDs that use different modes of communication.
Conclusion
Interruptions occur commonly during communication-based and non-communication-based physician activities in the ED. These interruptions are usually initiated by a healthcare provider and are usually brief in duration. Interruptions only rarely result in the physician changing tasks without completion. Interruptions have a negative impact on patient satisfaction.
References
Footnotes
Presented at ACEP Scientific Assembly, October 2008, Chicago, Illinois.
Presented at NY ACEP Scientific Assembly, July, 2007, Lake George, New York.
Competing interests None.
Ethics approval This study was conducted with the approval of the institutional review board of the site where the study occurred.
Provenance and peer review Not commissioned; externally peer reviewed.