Using TURF to Understand the Functions of Interruptions Vickie Nguyen, MA, MS1,2, Nnaemeka Okafor, MD, MS1,3, Jiajie Zhang, PhD1,2, Amy Franklin, PhD1,2 1 The National Center for Cognitive Informatics and Decision Making in Healthcare; 2 School of Biomedical Informatics; 3Medical School, The University of Texas Health Science Center, Houston, TX Abstract Interruptions are an often lamented and frequently studied aspect of clinical practice. However, some interruptions, such as updates on patient care decisions and notifications of detrimental patient lab values, are in fact necessary to the work process. In this paper, we explore the interruptions as an emergent feature of communication in teams. Looking beyond the frequency of interruptions, we consider the source and intent of interruptions with the goal of discovering the functions served by such communications. Furthermore, in this study of an emergency department, we classify interruptions into those activities that support required work and those interruptions that create unnecessary breaks in workflow. The intent of our larger body of work is to develop health information technology systems that support team efforts including the functions currently served by interruptions. Introduction Interruptions Interruptions, or a break in task to execute an unplanned task initiated by an internal or external source resulting in the pause or termination of the original task1, are an often bemoaned and frequently studied aspect of clinical practice2-7. While unnecessary interruptions have the potential to lead to increased patient risk6-9, some interruptions such as updates on patient care decisions by clinicians and notifications of detrimental patient lab values or irregular imaging results, are in fact necessary to the work process1. In this paper, we explore interruptions as an emergent feature of communication in teams. We focus on emergency medicine as this complex environment requires team based concurrent management of multiple patients coping with limited resources in a life-critical and interruptionladen environment10. U.S. emergency department physicians experience interruptions, on average, 10 times per hour during a shift6, 8. These routine breaks in task are more frequent in nursing, where emergency department nurses are interrupted every 4.5 minutes6, 9. Such interruptions not only impact the clinical care by potentiating medication errors or tasks on the wrong on the wrong patient2, 4, 5, 11, and have also been correlated to lower patient satisfaction12. Interruptions are not always to the detriment of care. It has also been found that the act of interruptions can help increase shared knowledge amongst team members. With efficient communications between team members, new knowledge or work routines are exchanged and such conversations improve the team’s performance on completing tasks13. Other positive impacts on task performance, the emotional state, and social attribution of people, depending on their current cognitive load, have also been found 14. In addition to the above studies, much of the literature has focused on identifying the rate of interruptions, tasks that are interrupted, and the participants engaged in the interruption1, 8, 9, 15, 16. Understanding the reasoning for interruptions, or the function that interruptions serve, will allow us to create interventions, here in the form of a health information technology, that support required interruptions while mitigating unnecessary breaks in work. Inherent in our assumption is the need for this type of communication as part of ongoing clinical practice. TURF The TURF framework, a method of Work-Centered Design, provides a qualitative and quantitative means of developing useful and usable systems. TURF stands for Task, User, Representation, and Function (TURF) analyses17, 18 . TURF is intended to aid in the creation of tools that provide a more useful and satisfying experience for system users while promoting patient safety and quality. This is accomplished by understanding the needs of users while supporting the functions to be accomplished as part of overarching goals and the tasks that are required in those efforts. Additionally, the users and their preferred visualizations make up other components of well thought out systems.

917

The objective of this study is to understand the function of interruptions by describing their place in the workflow. Namely, are the interruptions serving the functions of clinical work or do they reflect issues within the system (e.g., the interface does not fulfill the needs of the clinician in providing patient care) or environment (e.g., clinicians failing to understand roles and responsibilities required for completing overarching patient care goals). Our findings will help to inform our design of a team support tool to help clinical teams manage interruptions in the emergency department. Domain versus Overhead Functions The term overhead has traditionally been used in the business organizational context to refer to expenses required in operating a company such as paying rent for the facilities to manage production. Overhead is not directly connected to the items produced. In such examples, these extraneous outlays are part of “operating costs.” Similarly, the term has been used in the engineering and computational domain to describe design and algorithm features that require additional features from the system in order to reach a goal such as expending additional energy to run a circuit or transmitting signals in addition to computing data. By contrast, in these contexts, domain functions are used to describe the elements directly required in the work. That is what is needed to create the product. Zhang, Butler, and colleagues (2007 & 2011) expand on the use of domain and overhead terms to describe required and extraneous functions used to operate systems17, 18, 19. For example, a well designed interface would only display the domain, or required, operations users must click through in order to execute a command. A poorly designed system would have additional overhead, or unnecessary, operations interspersed throughout the system which may generate unnecessary effort. Couched more tangibly, when domain and overhead functions were applied to an electronic health record (EHR) allergy entry workflow, researchers discovered 99 overhead, and thus, extraneous functions. This is in stark comparison to the 28 domain functions 20. This nearly 4:1 ratio of overhead to domain functions had the potential to dramatically impact the usability of this system. For this paper, we further expand on the use of domain and overhead functions by using these terms to describe the interruptions within a clinical context. An overhead interruption refers to non-clinically dependent communications that disturb existing work processes. Some examples of overhead interruptions include: tracking down clinicians, repeating tasks already completed by other clinicians, completing tasks unrelated to the realms of your responsibilities, and performing workarounds of poorly implemented systems. Domain interruptions are therefore, interruptions whose communication fits within the needs of work. Examples of domain-type interruptions are providing updates on patient care plans, interpreting electrocardiography (EKG) or lab results, or verifying patient disposition decisions. Overhead interruptions can delay or impede the patient care process as they are not a necessary part of the required work while domain interruptions clearly support the patient care process as they are a part of the required work. Differentiating between these two types of interruptions may help us further understand how to better manage communications. We seek to support domain interruptions while mitigating the cognitive and workload burdens created by overhead-type interruptions. Table 1 provides the definitions we used to classify interruptions by their function. Table 1. Definitions used to classify interruptions by function. Terms

Definitions

Interruption

“A break in the performance of a human activity initiated by a source internal or external to the recipient. This break results in the suspension of an initial task to perform an unplanned task which results in a break or termination of the primary task1.”

Overhead

Interruptions that are not clinically dependent and create disturbed work processes (e.g., Technology Workarounds). Often caused by the implementation of poorly designed systems or workflow methods17, 18, 19.

Domain

Interruptions necessary for the completion of clinical tasks and create undisturbed work processes (e.g., Interpreting EKGs)17, 18, 19.

Methods Data for this study was drawn from a larger project examining decision making and information foraging in the emergency department. The dataset includes ethnographic non-participant observations of emergency medicine work from multiple clinical perspectives at a Gulf Coast Trauma 1 Center Teaching Hospital. This hospital sees

918

about 6065 trauma cases per year and employs fifty-two attending physicians, fifty-three residents, and fifty nurses. The emergency department at this hospital uses an EHR system with clinical provider order entry functionalities, trackboard, and a traditional pager and phone communication procedure. Here, we shadowed emergency department attending physicians, residents, and nurses. From this dataset, a convenience sample of six attendings, five residents, and five nurses were included in this secondary analysis. Each participant was observed on two different days for a time period of four hours each during each participant’s start of shift. A total of 128 hours of observation are included in this study. Following our IRB approved procedure, participants gave written informed consent before observations began a. Two non-clinical graduate students shadowed the clinicians. Observations were collected using pen and paper, a digital wristwatch with the time in seconds, and an audio recorder. Information was collected using a field note taking procedure established in our previous ethnographic studies 21. The audio recording device was placed within a pocket on the participant’s person and its microphone clipped at the participant’s white coat or scrub shirt collar 21. Using a categorization framework created by Brixey et al. (2007) and Franklin et al. (2011), we identified the types of interruptions and task transition decision making (i.e., interruptions, opportunistic decisions, or protocol based actions) that occurred among the emergency department team during21, 22. Interruptions were noted from all types of communications ranging from patient updates to general social behaviors during the observed work hours. Observation notes were supplemented with the audio recording for clarification. The observers had a reliability of ĸ=.80 indicating strong consistency in data collection and the reliability of initial categorization. In this paper, we then further classified the interruptions from the above dataset. We identified whether the clinician observed was the recipient or initiator of the interruption16, and whether or not the interruption was an overhead (i.e., unnecessary communication for the completion of a task or maintaining team shared knowledge and awareness) or domain-type (i.e., necessary for completing task or maintain team needs) interruption17, 18, 19. We were able to capture a snapshot of the emergency department team from the attending, resident, and nurse perspective on the directionality and interruption type by a member of the emergency department team. Results Attending Perspective 910 interruptions occurred during the 48 hours of observations of the six emergency department attending physicians (M = 18.96 interruptions per hour). First, considering the directionality of the interruptions, that is creator or receiver, we found that attending physicians are more often interrupted (91%) rather than the source of interruptions (9%). 52% of interruptions were initiated by residents with only 13% initiated by nurses. When we consider the function, domain or overhead, served by these interruptions, we found 23% are overhead-type interruptions unnecessary for completing clinical tasks or maintaining teamwork needs. Common overhead-type interruptions that attending experienced were: being asked to help locate other clinicians, relaying information to other clinicians, and EHR problems (i.e., technical issues). Again, residents were the primary source of these overhead interruptions (35%) with a small portion contributed by consultants (10%) and nurses (14%). Resident Perspective For the resident physicians, a total of 697 interruptions occurred during the 40 hours of observations (M = 17.43 interruptions per hour). 26% of interruptions received by the residents were initiated by attendings while nurses initiated 21% of these breaks in task. As initiators of interruptions, residents caused (52%) interruptions for attendings and (20%) interruptions for nurses. For our functional analysis, we found 30% of the interruptions (n = 208) were classified as overhead-type interruptions. The overhead-type interruptions that residents experienced were: finding information for other clinicians, helping clinicians find patients, and helping other clinicians navigate the EHR. Overhead-type interruptions were initiated by attendings (24%), consultants (14%), other residents (15%), and nurses (12%).

a

All clinical team members provided consent for each day of observation, including non-shadowed members of the group.

919

Nurse Perspective Forty hours of nursing data was included in this secondary analysis. In that time frame, nurses experienced a total of 1053 interruptions (M = 26.33 interruptions per hour). Of the total interruptions, nurses initiated breaks in task on only 18% of the occasions observed. Residents contributed to 17% of the nurse-recipient interruptions. Considering the function of the interruptions, 40% of the total interruptions observed were classified as overheadtype interruptions. These most often included inquiries by other nurses (45%). Triangulation of Clinical Perspectives After categorizing the interruptions by directionality and type of function from each clinical role’s perspective, we performed a triangulation of the data comparing the occurrences of overhead and domain-type interruptions for a comprehensive analysis of the emergency department team’s interactions (Figure 1). By comparing the difference perspectives of interruptions, we can validate each angle and further understand how each clinical role communicates and interacts with one another to determine how to better manage interruptions in the emergency department environment.

Figure 1. Triangulation of interruption perspectives by clinical role to validate the clinical team perspective. Comparison of the different clinical role perspectives indicate an increasing number of domain interruptions from attending physician to nurse (Figure 2).

920

Figure 2. Percentage of domain and overhead types of interruptions experienced by role. Breaking down the interruption functions by initiators and recipients of interruptions along with role demonstrates that although everyone received more interruptions than they initiated, there are differences in clinical role and interruption role (Figure 3). Nurses initiated more domain interruptions when compared to physicians, and though comparable, residents initiated more overhead breaks in task.

Figure 3. Classification of the identified interruptions by function, domain/overhead, direction of communication, and clinical role. Discussion In 128 hours of shadowing, 2660 interruptions were observed. Across clinical roles, this is the equivalent of 21 interruptions per hour. 61.1% of these breaks in task were classified here as received domain interruptions (n = 1626).

921

These findings demonstrate the ubiquitous nature of interruptions in emergency care and suggest a need for the management of interruptions, especially interruptions unnecessary to clinical tasks. Falling in line with expectation, residents and nurses received the majority of the domain-type interruptions as they are subordinates in the team and these types of interruptions may be natural to their workflow as task directives. For nurses to initiate a majority of the domain-type interruptions suggests that some communication gaps may be bridged by nurses between the attending, residents, and other nurses. Residents were the main initiators of overhead-type interruptions. These types of interruptions may be used to manage responsibilities outside of clinical work, but related to duties of promoting standard social work behaviors and maintaining a team environment. In addition to the overhead interruptions, the domain interruptions also require support. The use of interruptions as a communication strategy is an emergent feature of clinical teams. Currently, breaking the process of someone’s effort may be necessary in some instances, but may be disruptive in other instances. We believe building from these initial classifications with additional effort, we will be able to design a team support tool to help better manage the necessary interruptions while mitigating the unnecessary interruptions for clinicians. These team support tools could be more informative dashboards or smart phone applications which could help clinicians obtain a better sense of when to push communications based on the overall workload of the emergency department. Improving our understanding of team user’s needs and interruption content and function may help in the interface design of systems in healthcare settings. Our application of the TURF framework in the design of a support tool managing the activity of interruptions can help provide further support in the importance of designing systems for clinicians. Patient safety and quality may improve through the enhancement of each clinician team’s efficiency, effectiveness, coordination, communication, and, more broadly, future acceptance and use of possibly any implemented technology, such as health information technologies. Limitations This study used a limited convenience sample of participants within the trauma portion of a major emergency department. Consequently, these results may not be generalizable to other clinical settings. We were also unable to follow patients through their care outcomes and were therefore unable to verify if the impact of overhead interruptions greatly hampered the patient care process. Despite these limitations, we were able to outline the emergency department workflow, clinician communication and interaction patterns, and the function of interruptions. Future Work The scope of this work focuses on one emergency department. We are working to replicate the study in two other emergency departments using different health information technology systems, workflow, and patient sizes. Additionally, we are creating and testing more informative dashboards and communication tools (e.g., smart phone applications) for clinicians to help better manage their workload and communications. We will compare the results of using these team support tools against the baseline findings of this study in determining how helpful the team support tools are to providing safer and faster patient care process outcomes. Conclusion In this paper, we explored interruptions between attendings, residents, and nurses in the emergency department environment. We looked beyond identifying the frequency of interruptions and considered the function of the interruptions. The interruptions we observed in the emergency department were classified into activities that supported required work and those that create unnecessary breaks in workflow. The results from this study will help inform our design of health information technology systems to support teams in better managing their communications. Acknowledgements We thank our participants and collaborators for facilitating and assisting us with this study. This research was supported by the Agency for Healthcare Research and Quality through a research grant (AHRQ R01HS021236-20) and a training grant (AHRQ T32HS017586).

922

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.

Berg LM, Källberg AS, Göransson KE, Östergren J, Florin J, Ehrenberg A. Interruptions in emergency department work: An observational and interview study. BMJ Qual Saf. 2013;22:656-663. Coiera E, Tombs V. Communication behaviours in a hospital setting: An observational study. BMJ. 1998;316:673-676. France DJ, Levin S, Hemphill R, Chen K, Rickard D, Makowski R, Jones I, Aronsky D. Emergency physicians’ behaviors and workload in the presence of an electronic whiteboard. Int J Med Inform. 2005;74:827-837. Institute of Medicine. Health professions education: A bridge to quality. Washington (DC): National Academies Press; 2003. Westbrook JI, Coiera E, Dunsmuir WTM, Brown BM, Kelk N, Paoloni R, Tran C. The impact of interruptions on clinical task completion. Qual Saf Health Care. 2010;19:284-289. Grundgeiger T, Sanderson P. Interruptions in healthcare: Theoretical views. Int J Med Inform. 2009;78:293307. Rivera AJ, Karsh BT. Interruptions and distractions in healthcare: Review and reappraisal. Qual Saf Health Care. 2010;19(4):304-312. Chisholm CD, Dornfeld AM, Nelson DR, Cordell WH. Work interrupted: A comparison of workplace interruptions in emergency departments and primary care office. Ann Emerg Med. 2011;38(2):146-151. Li SYW, Magrabi F, Coiera E. A systematic review of the psychological literature on interruption and its patient safety implications. J Am Med Inform Assoc. 2012;19:6-12. Crosskerry P, Cosby KS, Schenkel SM, Wears RL. Patient safety in emergency medicine. Philadelphia (PA): Lippincott Williams & Wilkins; 2009. Kalisch BJ, Aebersold M. Interruptions and multitasking in nursing care. Jt Comm J Qual Patient Saf. 2010;36(3):126-132. Jeanmonod R, Boyd M, Loewenthal M, Triner W. The nature of emergency department interruptions and their impact on patient satisfaction. Emerg Med J. 2010;27:376-379. Zellmer-Bruhn ME. Interruptive events and team knowledge acquisition. Manag Science. 2003;49(4):514-528. Adamczyk PD, Bailey BP. If not now, when?: The effects of interruption at different moments within task execution. CHI. 2004;6(1):271-278. Allard J, Wyatt, Bleakley A, Graham B. “Do you really need to ask me that now?”: A self-audit of interruptions to the ‘shop floor’ practice of a UK consultant emergency physician. Emerg Med J. 2012;29:872-876. Brixey JJ, Robinson DJ, Turley JP, Zhang J. The roles of MDs and RNs as initiators and recipients of interruptions in workflow. Int J Med Inform. 2010;79:109-115. Zhang J, Walji M. TURF: Toward a unified framework of EHR usability. J Biomed Inform. 2011;44(6):10561067. Zhang J, Butler K. UFuRT: A work-centered framework and process for design and evaluation of information systems. HCI Int Proc. 2007;1-5. Butler K, Zhang J, Esposito C, Bahrami A, Hebron R, Kieras D. Work-centered design: A case study of a mixed-initiative scheduler. ACM SIGCHI Int Proc. 2007;747-756. Harrington C, Wood R, Breuer J, Pinzon O, Howell R, Pednekar M, Zhu M, Zhang J. Using a unified usability framework to dramatically improve the usability of an EMR module. Am Med Inform Assoc. 2011;549-558. Franklin A, Liu Y, Li Z, Nguyen V, Johnson TR, Robinson D, Okafor N, King B, Patel VL, Zhang J. Opportunistic decision and complexity in emergency care. J Biomed Inform. 2011;44:469-476. Brixey JJ, Robinson DJ, Johnson CW, Johnson TR, Turley JP, Patel VL, Zhang J. Towards a hybrid method to categorize interruptions and activities in healthcare. Int J Med Inform. 2007;76:812-820.

923

Using TURF to understand the functions of interruptions.

Interruptions are an often lamented and frequently studied aspect of clinical practice. However, some interruptions, such as updates on patient care d...
244KB Sizes 0 Downloads 7 Views