Objective Frequent interruptions are assumed to have a negative effect on healthcare clinicians’ working memory that could result in risk for errors and hence threatening patient safety. The aim of this study was to explore interruptions occurring during common activities of clinicians working in emergency departments.
Method Totally 18 clinicians, licensed practical nurses, registered nurses and medical doctors, at two Swedish emergency departments were observed during clinical work for 2 h each. A semistructured interview was conducted directly after the observation to explore their perceptions of interruptions. Data were analysed using non-parametric statistics, and by quantitative and qualitative content analysis.
Results The interruption rate was 5.1 interruptions per hour. Most often the clinicians were exposed to interruptions during activities involving information exchange. Calculated as percentages of categorised performed activities, preparation of medication was the most interrupted activity (28.6%). Face-to-face interaction with a colleague was the most common way to be interrupted (51%). Most common places for interruptions to occur were the nurses’ and doctors’ stations (68%). Medical doctors were the profession interrupted most often and were more often recipients of interruptions induced by others than causing self-interruptions. Most (87%) of the interrupted activities were resumed. Clinicians often did not regard interruptions negatively. Negative perceptions were more likely when the interruptions were considered unnecessary or when they disturbed the work processes.
Conclusions Clinicians were exposed to interruptions most often during information exchange. Relative to its occurrence, preparation of medication was the most common activity to be interrupted, which might increase risk for errors. Interruptions seemed to be perceived as something negative when related to disturbed work processes.
- Emergency department
- Patient safety
Statistics from Altmetric.com
Frequent interruptions are assumed to have negative effects on the clinicians’ working memory (the ability to remember future intentions), resulting in risk for errors (mistakes made in the process of care that result or have the potential to result in harm to patients) and hence exposing patients to increased risks.1–5 Interruptions can be initiated by external (observable event) or internal (unobservable thought processes in an individual) sources.3 Whereas most studies have described the types, numbers and duration of interruptions,6–9 mostly as experienced by physicians, studies on the relationship between interruptions and errors are few, as well as studies investigating the cognitive effects and outcomes of interruptions.7 ,8 However, a recent study found significant positive associations between interruptions and medication errors.10
To date, studies have reported that interruptions increase the risk for errors,7 ,8 but comparison between studies is difficult because the terms and definitions used for the concept of interruption varies. This far, research on interruptions in healthcare seems to discuss interruptions as a negative phenomenon, although some interruptions may be necessary for safety and high-quality care11 (eg, when providing information on a patient's deteriorating condition). Interruptions are complex phenomena, including characteristics of the primary tasks that are being interrupted and the nature of the interruptions themselves.7–9
The emergency department (ED) is a complex and dynamic high-risk arena within healthcare, especially prone to errors.4 ,12 ,13 In recent years, research on errors occurring in the ED has increased considerably.14–18 Specific situations in the work environment, such as overcrowding, multitasking (managing multiple tasks simultaneously) and interruptions have been studied.1–4 19–21 The communication load on ED staff is considered to be high;22 interruptions that occur face to face or from technical devices (eg, pagers and telephones) are among the most commonly identified interruptions confronting the staff in the ED.3–5 ,20
As clinicians’ experiences of interruptions are rather unexplored, knowledge about interruptions occurring during ED clinicians’ work processes can provide a basis for identifying situations where interruptions may compromise patient safety. The aim of this study was to explore interruptions occurring during common activities of clinicians working in EDs.
Study setting and participants
This is a descriptive observational and interview study with a qualitative and quantitative approach that was conducted during 2009 in two Swedish EDs for adults. One ED was located at a large urban university hospital and the other at a medium-sized county hospital. The two hospitals were purposely selected by the research group to achieve variation in the context and to make transferability possible based on variation in characteristics of patient flow (72 000 vs 49 000/year), type of ED (level one trauma centre vs county ED), and staffing (ED physicians vs non-ED physicians). Non-ED physicians, especially common in Sweden at EDs located in county hospitals, are on duty in the ED at evenings, nights and weekends and are, in addition, often responsible for inpatients at wards. At both EDs, there were about 90 clinicians on duty over a 24-h period, of which a sample of 18 clinicians, 9 from each hospital, was selected by purposeful sampling: 6 licensed practical nurses (LPNs), 6 registered nurses (RNs) and 6 medical doctors (MDs). Variation in gender and length of work experience in ED care was sought. The work experience in ED care among all clinicians varied from 6 months to 30 years. Based on the length of work experience, the participants were distributed into three groups: junior, medium-experienced and senior clinicians. One clinician from each profession at the two EDs was selected from each group to achieve variation in clinical experience. Less than 6 months of work experience from ED care was considered as an exclusion criterion. The 18 clinicians were invited and all agreed to participate. The reason for including these three groups of clinicians was to get a comprehensive perspective of the interruptions occurring in the ED. The MDs are responsible for medical diagnoses and treatment, the RNs are responsible for nursing care, prescribed medical-technical tasks and preparation and administration of medications. The RNs as well as the MDs have a university degree, LPNs have a high-school degree and are assistants to MDs and RNs.
Data collection was based on direct observations of clinicians’ work in the ED and subsequent individual interviews with the observed clinicians. The direct observation phase of the study comprised 36 h (18 clinicians×2 h) and covered different points in time of the day and night (from 8:00 to 3:00) and different weekdays. Due to practical reasons the observations needed to be conducted during three consecutive days at each hospital, Monday to Wednesday and Tuesday to Thursday respectively. Variation was achieved as Mondays were known to be characterised by high workload while the other weekdays were assumed to represent normal or low workload. At both EDs, the observations were conducted on representative days for the ED environment, covering patients with different symptoms of varying urgency. The observations and interviews were conducted concurrently by two of the researchers (LMB and ASK), both RNs with extensive clinical experience from ED care. The two researchers worked concurrently as a pair during data collection and observed the same clinician in order to maximise the capture of conducted activities and interruptions (figure 1). The researchers wore private clothes during the observations to indicate that they were not on duty. During the observations, each clinician was followed unobtrusively (shadowed) in their work. The researchers documented the continuous work process on a minute-to-minute basis as well as all received interruptions and self-interruptions (figure 1) that the clinicians were exposed to.
No predefined categories were used to identify the clinicians’ activities during the observations and the researchers used their own words to describe their observations. The data recorded in the observation framework were divided into six categories: (1) type of conducted activity (eg, reading a patient's chart), (2) duration of the activity in minutes, (3) location and person involved in the activity, (4) occurrence of interruptions in each activity, (5) whether the observed clinician was a recipient of an interruption or caused self-interruptions and (6) whether the previously observed activity was resumed after the interruption. Because no previous data collection protocol existed for the specific purpose of the study, a framework was developed based on previous studies within the research field.1–3 Four pilot observations were conducted to test the framework, resulting in minor changes.
Directly after the observations, a short (approximately 15 min) semistructured interview was held with each observed clinician by the researchers; one researcher conducted all interviews at one ED and the other researcher at the other ED. The interviews took place in a secluded location in the ED and focused on clinicians’ perceptions of interruptions.
Approval from the research ethical review board was obtained (2009/1413-31/4). The managers of the EDs gave their permission and each participating clinician gave written informed consent. Information to patients about the study was posted in the ED waiting rooms. Because some of the observed activities were patient-related, the observers sometimes decided to refrain from following the clinicians’ into examining rooms to protect patient integrity. In other situations, if the patients were involved in the observed activities, the patients were asked for approval.
In a previously published study conducted by our research team the types and frequencies of activities and multitasking performed by clinicians in two Swedish EDs have been reported based on the same data collection procedure.23 Fifteen categories of activities conducted by the clinicians were identified and 12 of these were exposed to interruptions during the observations. These 12 categories of activities were; information exchange (every time clinicians are asking for or giving information through interaction with each other, for example, discussing a patient, talking on the phone), preparation of medication (prepare all medication), documentation (eg, signing or dictating a patient's chart), patient/family-nurse/doctor interaction (all situations where clinicians interact with patients or relatives, for example, communicating with or examining a patient, bedside nursing), preparation of medical-technical tasks, administration (eg, time spent working at the computer, faxing documents), patient data analysis (eg, reviewing ECGs), transportation (eg, when the clinicians are moving around or transporting something from one place to another), organisational planning, information seeking (seeking information without interacting with another clinician, such as reading a patients chart), maintenance (eg, cleaning a stretcher), break (eg, communication between clinicians not involving work-related activities).
Based on the observation data, frequencies of how often interruptions occurred in each activity was counted and the locations in the EDs where these interruptions occurred were identified. Further, the type of clinician involved in the interruption, whether an activity was resumed after the interruption or not, and whether the observed clinician was involved in self-interruptions or was the recipient of an interruption was registered. The observed clinicians sometimes interrupted non-observed clinicians, but these events were not counted. χ² analysis was used to analyse differences in being recipient of or causing self-interruptions for each category of clinicians.
Content analysis was used for analysing the data generated through the interviews.24 ,25 The interviews were audio recorded and transcribed verbatim and read through several times to get a sense of the whole. The text was divided into textual units, condensed and organised into subcategories and categories. The analysis resulted in three separate areas. The first area contained data about the locations in the ED where interruptions occurred, the second area related to the clinicians’ need to interrupt someone else during the observation period and the third area comprised clinicians’ perceptions of interruptions. Examples of the analysis of the third area are shown in table 1. The data were analysed primarily by two of the authors (LMB and ASK). Consensus on the categories was reached through repeated discussions in the research group until agreement was reached.
Types and frequencies of interruptions—observation data
Of the 1882 activities being observed, 184 (10%) were interrupted, which yields a mean rate of 5.1 interruptions per hour. Interruptions were observed during 12 of the previously identified 15 activities (table 2). The most commonly occurring interruptions were observed during information exchange (20%). The activity that was most exposed to interruptions in relative terms was preparation of medication (29%), followed by documentation (27%), patient/family-nurse/doctor interaction (19%) and preparation of medical-technical tasks (19%).
There were 73 interruptions identified for MDs, 50 for RNs and 61 for LPNs. Collectively the clinicians were more often recipients of interruptions (63%) than causing self-interruptions (37%) (p<0.001). MDs and RNs were more often recipients of interruptions (MDs 78%, RNs 66%) than causing self-interruptions (MDs 22%, RNs 34%) (MDs p<0,001, RNs p<0.05). The MDs were primarily recipients of interruptions while documenting patient care. LPNs to a higher extent (59%) caused self-interruptions (p<0.004), than MDs (22%) and RNs (34%). The LPNs self-interruptions mostly took place when they were involved in the activity patient/family-nurse/doctor interaction (eg, needing to leave the patient alone in the assessment room for a short time because of missing equipment).
Among all clinicians, the most common way to be interrupted was by face-to-face interaction with a colleague (51%). Most interrupted activities, 161 (87.5%), were observed to be resumed shortly after the interruption. Of the remaining 23 (12.5%) interrupted activities, 9 were prematurely terminated before attending to another task, 1 was handed over to a colleague and the final handling of the 13 interrupted activities was not observed during the remaining observation session.
The locations at the ED where most of the interruptions occurred were at the doctors’ (19%) and nurses’ stations (49%). Interruptions occurred less often in assessment rooms (18%), corridors (8%), utility rooms (2%), waiting rooms (2%), triage areas (1%) and medication rooms (1%).
The clinicians’ perceptions of interruptions—interview data
Some clinicians’ used the terms interruption and disturbance (figure 1) as representations of the same concept while others distinguished between the two as separate concepts. Further, regardless of the terms used (interrupted or disturbed), data from the interviews showed that the concept of interruption was not always limited to a negative feeling, as some interruptions were perceived as expected and necessary in order to pursue the work process. As figure 2 illustrates, whether an interruption was perceived as something negative or not was related to whether the interruption disturbed the work process or not.
The clinicians’ perceptions of interruptions were categorised as either an undisturbed work process or a disturbed work process (figure 1). The most common reasons for clinicians not to perceive an interruption as something negative (undisturbed work process) were related to their own expectations of the work environment or whether the interruption rate was perceived as low. The clinicians’ perceptions in the subcategory expectations of the work environment often consisted of thoughts about being interrupted as a natural part of their professional role. An example was: ‘That's the job at the ED, it is always decisions and you are always interrupted, it comes something between’ (10: senior MD). This perception was most often expressed by the MDs. Other perceptions in the subcategory expectations of the work environment were that interruptions were commonplace events. An example was:’ It's so commonplace, this is happening every day, so I don't think of it as an interruption’ (2: senior LPN). If the interruption occurred with low frequency (low interruption rate), it was not considered as something negative for the ongoing work.
The primary reason for perceiving an interruption as something negative (disturbed work process) was clinicians’ exposure to frequent communicative interruptions, mostly by colleagues, pagers or phones. Examples given were: ‘There are other days when there are significantly more calls on the pager than today and then it (disturbance) gets obvious. There are usually many more pagings than there were today’ (1: senior MD). Other reasons for perceiving an interruption as something negative were if the interruption was perceived as irrelevant, ‘When you get disturbed by, what I think is an irrelevant interruption, it is often from wards. Pagings of a non-urgent character that can wait’ (7: junior MD). Further, having to wait for someone else to get on with the tasks at hand, ‘I get disturbed by having to wait for the person who will help me make a decision. Yes, one gets disturbed by waiting’ (16: junior MD). Furthermore, missing equipment was another reason for being negatively influenced by the interruption, ‘It was poorly refilled in the boxes that I brought in when we would insert the catheter, so then I had to go out and get more equipment, it is disturbing because you have to interrupt what you are doing in there’ (14: junior LPN).
The interview data supported the observation data regarding in which locations of the ED the clinicians were exposed to interruptions. The respondents in most cases felt interrupted in the nurses’ and doctors’ stations and also in stations with a heavy workload, and in the triage area. Further, the question concerning whether the clinicians had to interrupt someone else to be able to perform their assignments showed that there was uncertainty among the clinicians if they had interrupted someone else. The uncertainty was related to that the clinicians did not remember or did not reflect upon if they interrupted someone else, as it was perceived as a commonplace event.
The main finding from the observations in this study was an overall interruption rate of 5.1 interruptions per hour. Preparation of medication, although rarely observed, was the most interrupted activity in relative terms. The interruptions were most common in face-to-face interactions with a colleague. While the MDs most often experienced interruptions during documentation activities, the RNs were mostly exposed to interruptions during information exchange activities. In contrast, the LPNs were most commonly interrupted during patient/family-nurse/doctor interactions and more often involved in self-interruptions than the other professions. Most interrupted activities (87.5%) were resumed after the interruption.
The mean interruption rate of 5.1 per hour for all the clinicians in this study was low compared with findings in other studies reporting interruption frequency twice as many for doctors and nurses in the ED.2 ,7 ,10 ,21 One explanation might be that the workload at the ED was relatively low during the data collection period in our study. Further, there are several factors that need to be considered when comparing the interruption rates in various contexts, such as the type of ED (ie, urban vs rural), the length of observation time, and the included sample of clinicians. In addition, different terms and definitions of interruptions (eg, distractions, break-in-tasks and disturbances) have been used in the literature, making it difficult to compare studies.7–9 However, our finding that commonly occurring interruptions are face-to-face interactions is consistent with other studies,2 ,3 ,26 as well as the previously reported finding that communication traffic is high in the ED.22 ,27 ,28 Since it is well known that medication errors frequently occur in the ED29 and that interruption during preparation of medication increase the risk for errors,10 it is of concern that preparation of medication in relative terms was the most commonly interrupted activity (28.6%), although seldom observed in this study. This result is similar to that of another study, where interruptions during medication-related activities amounted to 27.5% of all interruptions.26 Thus, the most commonly interrupted activity in relative terms was preparation of medication, the highest number of interruptions occurred during information exchange. It can be assumed that clinicians who are interrupted during these activities are at risk of making errors and ultimately jeopardising patient safety.
Most of the interrupted activities were resumed after an interruption (n=161, 87.5%), which is in accordance with previous findings.3 ,9 ,20 ,21 Of the remaining 23 interrupted activities, 10 were terminated or handed over to a colleague while the final handling of 13 was not observed during the remaining observation session. This finding might support the notion that frequent interruptions1–3 ,5 and high communication load have a negative impact on the working memory of clinicians.22 ,27 ,28
The main finding from the interviews was that the clinicians did not always perceive interruptions as something negative. Research on interruptions in healthcare has sometimes assumed interruptions to be negative and unsafe.7 ,9 In our study however, an interruption seemed to be perceived as something negative only when it was related to a disturbed work process. However, interruptions were not perceived as negative when they were related to an undisturbed work process that is, when the interruption rate was low or when the clinician expected interruptions to be an inherent feature of their professional role or the work context. Even if the interruptions were not always perceived as negative, interruptions may affect the clinician's working memory and thereby potentially compromise patient safety.
In the interviews, the respondents expressed that waiting was perceived as something disturbing (table 1). Having to wait disturbs or delays clinicians’ work processes, that is, a disturbed work process, even though the interruption is not face to face and not obvious to the observers. Other researchers have previously presented waiting time using different concepts, for example, impediments/delays,11 which causes obstructions in workflow and a role-category that causes delays in workflow.2
The finding that the clinicians perceived interruptions as commonplace events and therefore did not remember if they had to interrupt someone else raises the question about how to manage and decrease unnecessary interruptions, if the clinicians are not always aware of when they occur.
Some limitations need to be kept in mind when interpreting the results. The interviews with the clinicians were conducted immediately after the observation period. The clinicians were still on duty and might have felt time pressure by needing to leave their assignments and thus may have wanted to terminate the interview as fast as possible. The time factor might have had a negative effect on the richness of data from the interviews. However, our intention was to minimise the time between the observations and the interviews so that the clinicians could reflect upon their perceptions of interruptions that had occurred during their 2 h of observation.
The researchers that conducted the data collection (LMB and ASK) were RNs working at one of the participating EDs and consequently there is a risk that preunderstanding influenced the observations as well as the interpretation of the data. On the other hand, the fact that one of the researchers was familiar with the work context at each ED made it easier to capture details of conducted activities and interruptions during the observations, contributing to deeper knowledge of the subject of study. The presence of the two observers was a way of minimising the potential bias effect while at the same time increasing the possibility to capture the complex work processes.
Credibility in the data collection was ensured by concurrent data collection by two researchers. It was also assured by the use of a structured data collection framework to ensure that all observed activities were captured. On the other hand, the preunderstanding of the work context, although it can be seen as a limitation, made it possible to capture the complex work process during the observations. The credibility of the data is strengthened by the variation in observation periods, participants’ varying professions, gender and level of ED working experience. Using observations and interviews in the data collection also strengthened credibility and dependability. This triangulation made it possible to look at the same situations from two perspectives, which provided a deeper understanding of the phenomena. The fact that independent analyses were made by four of the researchers also strengthens the credibility of the data. Because the observations took place at a large urban university hospital and at a medium-sized county hospital, transferability of the findings to other similar Swedish EDs is possible.
In the interviews the clinicians also talked about their perceptions of interruptions in general, and not only in relation to the observation period. Thus, some of the obtained data provided more general information about the perceptions of interruptions in the work process, which were included in the analysis to get a broad perspective of the phenomenon. The data from the interviews were thus a combination of the clinicians’ perceptions of interruptions that occurred during the observations and their general perceptions of this phenomenon.
The concepts, definitions and categorisations used in the literature for interruptions vary. It would be useful with standardised definitions of these concepts. This study revealed that the concept of interruption was hard to define and distinguish from other similar and related terms. In this study the participants to some extent used the concept interruption synonymously with disturbance. It was, however, evident during the analysis that, even when the participants had used the concepts synonymously, both concepts could be perceived as something negative and as not being negative, depending on whether the work processes were disturbed or not. Therefore, during the analysis, efforts were made to assure that the inherent meaning of the respondents’ answers were categorised correctly, rather than focusing on the concept used. Based on our findings, and the lack of standardised defined concepts in the literature regarding interruptions, we suggest that future studies about interruption focus on exploring these concepts.
Ten per cent of all observed activities conducted by the ED clinicians were interrupted. Most interruptions occurred during information exchange at the nurses’ and doctors’ stations. Preparation of medication was the activity most exposed to interruptions in relative terms, and there is a previously established causality between interruptions and medication errors. The clinicians seemed to perceive an interruption as something negative only when it was related to a disturbed work process. Situations where clinicians actually perceive an interruption as something negative could be considered a patient safety risk. The results of this study provide a basis for further studies on interruptions during activities performed by ED clinicians.
The authors thank the Department of Emergency Medicine at the Karolinska University Hospital Solna, the Department of Emergency Medicine, Falun Hospital, AFA Insurance and the Centre for Clinical Research Dalarna for providing financial support.
LMB and ASK have contributed equally.
Contributors LMB and ASK: contributed equally to this article, and share first authorship. Design of the study: all authors. Data collection: LMB and ASK. Data analysis: all authors. Manuscript writing: all authors. Financing: all authors. Supervision: KEG, AE, JÖ and JF.
Funding This study was funded by the AFA Insurance (grant number 110076) and Centre for Clinical Research, Dalarna (grant number CKFUU-73061).
Competing interests None.
Ethics approval Research Ethics Committee Stockholm 2009/1413-31/4.
Provenance and peer review Not commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.