i n t e r n a t i o n a l j o u r n a l o f m e d i c a l i n f o r m a t i c s 7 8 ( 2 0 0 9 ) 629–637

journa l homepage: www. int l .e lsev ierhea l th .com/ journa ls / i jmi

Synchronous communication facilitates interruptiveworkflow for attending physicians and nurses inclinical settings

Ashley Edwardsa,∗, Leslie-Anne Fitzpatricka, Sara Augustinea, Alex Trzebuckia,Shing Lai Chenga, Candice Presseaua, Cynthia Mersmannb,Bruce Heckmanc, Stan Kachnowskia

a Healthcare Innovation and Technology Lab, New York, NY, United Statesb Saint Vincent’s Catholic Medical Center, New York, NY, United Statesc Phelps Memorial Hospital Center, Westchester, NY, United States

a r t i c l e i n f o

Article history:

Received 14 November 2008

Received in revised form

10 April 2009

Accepted 16 April 2009

Keywords:

Communication

Medical informatics

Clinical informatics

Physician’s practice patterns

a b s t r a c t

Study objective: Inter-clinician communication accounts for more than half of all information

exchanges within the health care system. A non-participatory, qualitative time-and-motion

observational study was conducted in order to gain a better understanding of inter-clinician

communication behaviors, routine workflow patterns, and the use of information commu-

nication technologies (ICTs) within the clinical workspace.

Method: Over a 5-day period, seven attending physicians and two nurses were shadowed

for 2–4 h at a time. Inter-clinician communication events were tracked in real-time using

synchronized digital stopwatches. Observations were recorded on a paper-based, semi-

structured observation tool and later coded for analysis.

Results: Nine hundred and eighty-seven communication events were observed over

2024.67 min. Clinicians were observed to spend the majority of their time on patient care

(85.4% in this study) with about three-fourths of that time spent on indirect patient care (e.g.

charting). Clinicians were observed to prefer using synchronous communication modes,

which led to multitasking and created a highly interrupted workflow. Forty-two percent

(n = 415) of communication events were coded as interruptions and study participants were

seen multitasking 14.8% of the time. Though each interruption was short-lived (on average

0.98 ± 2.24 min for attending physicians), they occurred frequently. Both attending physi-

cians and nurses were the recipients of more interruptions than they initiated.

Conclusion: This study demonstrated that the clinical workspace is a highly interruptive

environment. Multiple interruptions in the communication processes between clinicians

consume time and have the potential to increase the risk of error. This workflow anal-

ysis may inform the development of communication devices to enhance inter-clinician

communication by reducing interruptions or deferring interruptions to more appropriate

times.

© 2009 Elsevier Ireland Ltd. All rights reserved.

∗ Corresponding author at: Healthcare Innovation and Technology Lab, 3960 Broadway, Room 410, New York, NY 10032, United States.Tel.: +1 212 543 0100; fax: +1 212 543 0108.

E-mail address: are27@hitlab.org (A. Edwards).1386-5056/$ – see front matter © 2009 Elsevier Ireland Ltd. All rights reserved.doi:10.1016/j.ijmedinf.2009.04.006

i c a l

630 i n t e r n a t i o n a l j o u r n a l o f m e d

1. Introduction

Communication among health care providers is the most com-mon and important exchange of clinical information withinthe health care system [1–4]. Information and data exchangesmay occur along both synchronous (e.g. face-to-face com-munication, telephone conversations) and asynchronous (e.g.voice mail, e-mail, patient charts, notes, page) communicationchannels; however, past research has noted a bias towards theuse of synchronous channels [3]. As a result of the tendencyto use synchronous channels, the clinical workspace is rid-dled with interruptive and multitasking workflows. Frequentinterruptions and multitasking in a clinical environment havefar-reaching and detrimental implications on the quality ofpatient care [5,6] and the financial cost to institutions [3].Interruptions and multitasking contribute to medical errorby disrupting short-term memory processes [7,8]. This is evi-denced in the Institute of Medicine (IOM) report, “To Err isHuman: Building a Safer Health System.” This report effec-tively conveyed that communication failures were associatedwith significant medical errors [9]. Moreover, in a 2002 reporton patient safety, the Joint Commission on Accreditation ofHealth care Organizations (JCAHO) found that communicationbreakdown was a major cause of sentinel events [10].

In a communication-rich workspace such as a medicalenvironment, it is important for communication to be exe-cuted efficiently and accurately. There is a documented needfor studying and understanding communication among physi-cians and nurses in the clinical setting. In situ, observationalstudies provide insight into the significant time gaps inthe delivery of care as well as common data and work-flow patterns. Insights into communication processes withinthe clinical workflow processes are crucial to successfullydeveloping mobile, asynchronous information communica-tion technologies (ICTs) that are readily adopted by healthcareworkers [11]. Furthermore, workflow analysis helps to createclear definitions of user needs and ICT requirements [12,13].

Observational studies were conducted at two hospital sitesin order to gain a better understanding of communicationbehaviors among health care providers, routine workflow pat-terns, and the use of ICTs within the clinical workspace.Additionally, at the conclusion of data collection, recommen-dations for the development of an effective health care ICTsolution were made.

2. Methods

2.1. Study design and setting

A non-participatory, qualitative time-and-motion observa-tional study was conducted over a 5-day period in February2008 at one 489-bed tertiary care teaching hospital and one235-bed not-for-profit acute care hospital. The observationalstudy involved primarily the emergency department of eachhospital site. At the 235-bed not-for-profit acute care hospital,

one internal medicine physician was shadowed and followedto the medicine department. The study sites were located inthe greater New York Metropolitan area, however, the largerhospital had significantly higher patient influx. A total of

i n f o r m a t i c s 7 8 ( 2 0 0 9 ) 629–637

seven attending physicians (6 emergency room physiciansand 1 internal medicine physician) and two emergency room(ER) nurses participated in the study. Although other healthcare providers were observed, those results were excludedfrom this paper due to insufficient data and lack of con-sistency across institutions. Participation in the study wasentirely voluntary. Demographics of the participants were notobtained.

2.2. Methods of measurement

The observation methodology was based on the Commu-nication Observation Method established by Spencer et al.[1]. Study participants were shadowed for 2–4 h at a timeby researchers from the Healthcare Innovation and Technol-ogy Lab (HIT Lab). All researchers were trained in the sameobservation methodology. The shadowing method allowedobservers to unobtrusively follow health care providers torecord what, why, and how they performed their routinedaily tasks in the clinical setting [5]. During each observa-tion period, inter-clinician communication events were timedwhile study participants performed their routine activities.Time was tracked in real-time using synchronized digitalstopwatches. Communication events were recorded on asemi-structured, paper-based observation tool developed bythe research team at the HIT Lab. The definitions of commu-nication terms used in the study were consistent with theexisting literature on inter-clinician communication in healthcare delivery [1]. “Interruption” was defined as any disruptiveactivity, clinician conversation, event, or alert that occurredduring a communication stream between two health careproviders. A “communication event” was considered to be anyaction taken in order to relay information to another clinician,and included conversation, telephone calls, pages, e-mails,voice mail, notes written in a patient record, and text message.The aforementioned actions were considered to be “commu-nication events” even if the information was delayed or wasnot received by the intended recipient. “Percentage time incommunication” accounted for the percent of time that theobserved clinician was involved with communication eventsduring the shadowing period. Lastly, “multitasking” (e.g. chart-ing while talking) was defined as attending to at least one dutywhile also attempting to communicate with another clinician.Multitasking was observed to greatly impact communicationflow.

Any disruptions in communication flows, including but notlimited to telephone calls, questions from other clinicians,patient requests, waiting for other team members, and var-ious unforeseen events were logged during the observationperiod. Events that took place in order to compensate forinefficient communication were also noted. Particular atten-tion was paid to communication patterns associated with theuse of synchronous (e.g. telephone and face-to-face conver-sations) versus asynchronous (e.g. fax, paging, and e-mail)communication channels. Communication devices were cat-alogued along with usage frequencies and usage scenarios.

Any observable communication barriers were documented.Un-coded notes were preliminarily coded at the end of eachday. Fig. 1 summarizes the communication methodology usedduring the on-site observational studies.

i n t e r n a t i o n a l j o u r n a l o f m e d i c a l i n f o r m a t i c s 7 8 ( 2 0 0 9 ) 629–637 631

rvati

pmwoaeM

2

Temcbic

2

BfraCwds

3

Sov

Fig. 1 – Summary of obse

There were no structured interviews with health careersonnel during the observational period and verbal com-unication between the observer and health care personnelas kept to a minimum, except in the event that clarificationn observations was needed. No demographic informationbout the participants was recorded. The resulting data wasntered in a Microsoft® Access database and later exported toicrosoft® Excel for communication event frequency analysis.

.3. Outcome measures

he times associated with inter-clinician communicationvents were recorded. The personnel involved in the com-unication event, interruptions, and devices used were also

arefully noted. All of the outcome measures were categorizedy clinician type (i.e., attending physician or nurse). Case stud-es of clinician workflows were constructed based on the dataollected.

.4. Ethical considerations

efore entering hospital study sites, permission was obtainedrom the institutional review board (IRB) of each hospital. Allesearchers were certified with the necessary Health Insur-nce Portability and Accountability Act (HIPAA) and Goodlinical Practice (GCP) requirements. Appropriate measuresere taken to ensure that all data collected was secure, confi-ential, and de-identified to protect the participants who werehadowed.

. Results

even attending physicians and two ER nurses were observedver a period of 1727.67 and 297 min, respectively. Table 1 pro-ides a summary of the distribution of communication events.

onal study methodology.

On average, there were 29 communication events per hour.Overall, participants spent 85.4% of their time (1728.33 min)on patient care (24.6% of time on direct patient care, e.g.history and physical examination, and 60.7% of time on indi-rect patient care, e.g. charting) and 14.6% on non-patient careactivities (e.g. social greetings).

Communication was executed through the use of a vari-ety of electronic and non-electronic methods and devices,including telephones, face-to-face conversations, computer-ized documentation systems, paper charts, fax machines,computers on wheels (COWs) and pagers. One of the insti-tutions had incorporated the Vocera communication system,a complete proprietary voice-over-IP intercom system thatincludes individual intercom “badges”, into their workflow.The system operates over a wireless network and is designedfor health care facilities. Interestingly, none of the cliniciansthat were shadowed at that institution used Vocera as a meansof communicating to other clinicians. The frequency of use ofvarious devices at the two observational study sites is capturedin Table 2.

Most of the communication events (84.4%) flowed alongsynchronous communication channels. Face-to-face commu-nication accounted for 70.0% (n = 681) of the synchronous com-munication events, while 15.4% (n = 152) of these events weretelephone conversations. All forms of asynchronous commu-nication were utilized substantially less frequently than anysynchronous communication channel. Along asynchronouschannels, use of a ‘computerized hospital EMR/patient chart’was most commonly observed (see Table 2). When analyzed byclinician type (i.e. attending physician and nurse), the propor-tions of synchronous versus asynchronous communicationusage were similar. Synchronous forms of communication

were equally dominant within both clinician groups. Thecase study shown in Fig. 2 demonstrates the dominance ofsynchronous communication patterns in the execution ofcommon clinical tasks. The case study details an attending ER

632 i n t e r n a t i o n a l j o u r n a l o f m e d i c a l i n f o r m a t i c s 7 8 ( 2 0 0 9 ) 629–637

Table 1 – Distribution of communication events, interruptions, and multitasking among study participants.

Attending physician Registered nurse Total

Observation time (min) 1727.67 297.00 2024.67

N communication events 874 113 987Communication time (min) 1595.25 203.47 1798.72Percentage time in communication 92.3% 68.5% 88.8%

N interruption events 362 53 415

Interruption time (min) 353.03

N multitasking events 142Multitasking time (min) 254.34

physician’s effort to acquire a surgical consult while manag-ing standard workload. The process lasted for a total of 56 min.Of the 56 min, 31 min and 58 s were dedicated to communica-tion surrounding the specific patient’s care over the durationof the task. Communication occurred between the attendingphysician and several key stakeholders in order to completethe surgical consult. Nearly all communication events werethrough synchronous channels, including ‘face-to-face’ and‘telephone’ methods. The only events utilizing asynchronousmethods of communication were those involving the orderingof labs and/or the entering of patient notes in a computerizedhospital EMR/patient chart. All interruptions within this casestudy were the result of synchronous communication.

3.1. Interruptions

The predominance of synchronous communication con-tributed to an interruptive workspace. During the 1727.67 minin which attending physicians were shadowed, 22% of thetime (353.03 min) was lost to interruption events (n = 362interruption events). On average, interruptions took 0.98 min(±2.24 min). Attending physicians received more interruptionsthan they initiated. They initiated 29.6% of their interruptionevents (n = 107 interruption events; 134.24 min) and were therecipients of 70.4% (n = 255) of interruptions from their col-leagues. As depicted in Fig. 3, nurses, other physicians, and

residents were the prime interrupters of the attending physi-cians shadowed.

Interruptions accounted for 20.4% (41.42 min) of nurses’communication time (n = 53 interruption events). The two

Table 2 – Categorical breakdown of synchronous and asynchron

Communication type

Att

Synchronous communicationFace-to-faceTelephone (mobile, cell, wired)

Asynchronous communicationComputerized hospital EMR/patient chartPaper formPaper chartWritten note on paperOverhead public address (PA) systemPagerPDA

41.42 394.45

28 17044.05 298.39

nurses shadowed were the initiators of 41.5% (n = 22) of theirinterruptions over a total of 25.15 min. Like the attendingphysicians, nurses received more interruptions than theyinitiated—31 interruptions received versus 22 interruptionsinitiated. The interruptions analysis for nurses is summarizedin Fig. 4.

3.2. Communication inefficiencies

Based on our observations, synchronous communication wasthe source of some inefficiency. The case study representedin Fig. 5 recounts one scenario in which reliance upon syn-chronous communication for obtaining consults resulted in afailure to complete a task because one of three consults couldnot be reached. The physician in this case study attemptedto make three consultations by landline telephone as the firstmode of contact. In the case of consult #3 the physician called,paged, and called a second time, finally leaving a message withthe secretary. Each of the three attempts at contact resultedin deferment to an asynchronous mode of communication.Perhaps, had the physician been able to rely on single, trustedasynchronous method of communication, he could have savedtime by forgoing the follow-up page and phone call.

3.3. Multitasking

Multitasking was defined as attending to many duties at thesame time while also communicating. The amount of timeduring which multitasking occurred was noteworthy. Whencompared to other means of transferring patient information

ous communication between clinicians.

Communicating with other clinicians

ending physician Registered nurse

68.4% 73.5%16.5% 7.1%

11.5% 5.3%2.9% 3.5%3.5% 1.8%3.2% 0.9%2.5% 0.0%0.2% 0.0%0.2% 0.0%

i n t e r n a t i o n a l j o u r n a l o f m e d i c a l i n f o r m a t i c s 7 8 ( 2 0 0 9 ) 629–637 633

Fig. 2 – Case study workflow of an ER attending physician requesting a surgical consult. Times delineated within boxesdetail the time expenditure of the event while the times between the boxes indicate elapsed time between communicationn terpr

bm

4

Ts

odes. Interruptions are coded as blue colored boxes. (For ineader is referred to the web version of the article.)

etween health care providers, the most time was spent onultitasking. This is shown in Fig. 6.

. Discussion

he results from the two observational study sites are con-istent with past research on inter-clinician communication

retation of the references to color in this figure legend, the

[3,7,14]. Similar to other studies, clinicians demonstrated apreference for the use of synchronous modes of communica-tion (e.g. face-to-face and telephone). It was observed within

this study, that synchronous communication did result insome inefficiency, however, further studies could supplementthese qualitative observations with quantitative findings.Moreover, the clinical workspaces at the two sites were highly

634 i n t e r n a t i o n a l j o u r n a l o f m e d i c a l i n f o r m a t i c s 7 8 ( 2 0 0 9 ) 629–637

by a

Fig. 3 – Frequency of interruptions initiated and received

interruptive and health care personnel were consistently mul-titasking and shifting their priorities in order to manage thevarious tasks that arose.

The high communication load that is common in clini-cal environments has proven to be demanding on humanworking memory [2]. The incessant use of synchronouscommunication in the workplace creates a highly inter-ruptive environment requiring a great deal of multitaskingthat is taxing on working memory. Although most inter-ruptions were observed to be short-lived (on average0.95 min/interruption ± 2.17 min), interruptions are undoubt-edly distracting and disrupt the thought process. Suchdistractions, though momentary, can have far-reaching impli-cations on the delivery of care, potentially leading to medicalerrors and adverse clinical outcomes. As shown in Figs. 3 and 4,both physicians and nurses received several interruptions,and received more interruptions than they initiated. Indeedthis suggests that these two groups should be targeted for the

evaluation of new ICT technology that could minimize inter-ruptions.

Multitasking during patient care delivery can also be detri-mental. Of the observed communication events, 17.2% were

Fig. 4 – Frequency of interruptions initiated and receiv

ttending physicians during the total observation period.

“multitasking events” in which the observed clinician was per-forming one task (e.g. documenting patient information) whilesimultaneously communicating with a colleague either face-to-face or on the phone. On several occasions, informationbeing exchanged during multitasking events had to be ver-ified at a later time. This observation is in accordance withpast research in which investigators showed multitasking tobe risky due to limited capacities of humans’ short-term mem-ory processes [15].

In addition to elucidating clinician communication trends,study observations also provided important information aboutclinician time expenditures. The majority of caretaker timewas spent on patient care (85.4% in this study), with nearlythree-fourths of that time spent on indirect patient care. Oftotal work time, 24.6% was spent on direct patient care (e.g.history and physical examination), 60.7% of time on indi-rect patient care (e.g. charting) and 14.6% on non-patientcare activities. Awareness of how clinician time is segmented

provides additional insight into how inter-clinician com-munication can be facilitated efficiently. It is likely thatthe observed work time segmentation is linked to clini-cian communication processes. For example, indirect patient

ed by nurses during the total observation period.

i n t e r n a t i o n a l j o u r n a l o f m e d i c a l i n f o r m a t i c s 7 8 ( 2 0 0 9 ) 629–637 635

Fig. 5 – Case study of an internal medicine physician requesting consultation from other clinicians. Consults #1 and #2fielded the physician’s call and discussed the case. Consult #3 did not respond to the first attempt at synchronouscommunication. In response to the lack of response from consult #3, the doctor left a voice mail message and thenproceeded to attempt an asynchronous channel to connect with the consult via an overhead page. This asynchronousm diffeT iona

ccmsc

isucvam

Fu

ethod failed and was followed by a second phone call to ahe attempt to contact consult #3 failed during the observat

are often involves relaying patient information from onelinician to another. Thus, streamlining inter-clinician com-unication could potentially shift work time distribution

uch that more time could be devoted to direct patientare.

Notwithstanding the concerns associated with multitask-ng and interruptions, it is important to recognize that ourtudy observations demonstrate a definite preference for these of synchronous communication. Synchronous means of

ommunication, particularly in the ER environment, may pro-ide a richer exchange of information. Thus properly designedsynchronous ICT should aim not to replace face-to-face com-unication, but rather mediate interruptions that are not

ig. 6 – Total time involved in relaying patient informationsing different modes of information transfer.

rent number where a message was left with a third party.l period.

urgent. Asynchronous ICTs can serve to divert face-to-facecommunication to more opportune times.

The use of asynchronous communication technologiescould provide significant cognitive benefits by reducing inter-ruption events. Ideally, the reduction of interruptions wouldlead to more efficient, less time-consuming communica-tion and consequently, lead to improvements in the qualityof care. Further, if an asynchronous communication devicewas to be designed specifically for the clinical setting,well received by clinicians, implemented, used to mediateface-to-face communication, and deemed dependable, inter-clinician communication could potentially be streamlined.Such an asynchronous device could help to postpone inter-ruptions to moments during the clinician workday that arenot centered on client care (14.6% of time). Some recentinnovations in communication technology such as reminderand escalation features [16–18] and context-aware architec-ture [19–21] could be incorporated into asynchronous devicesso that they reduce cognitive load and subsequent medicalerrors, thereby, improving the efficiency of health care deliv-ery.

4.1. Limitations

As a result of the small sample of study participants and

study sites, this study did not include early morning andlate-night shifts nor did it analyze when interruptions weremost likely to occur. Moreover, a limited range of health-care providers – only physicians and nurses – were shadowed.

636 i n t e r n a t i o n a l j o u r n a l o f m e d i c a l

Summary pointsExisting knowledge in the field of inter-clinician commu-nication

• Communication among health care providersaccounts for between 60 and 90% of all informa-tion transactions within the health care system. Mostof this communication is done synchronously.

• Frequent interruptions and multitasking in the clinicalenvironment have far-reaching and detrimental impli-cations on the quality of patient care and the financialcost to institutions.

• Interruptions and multitasking contribute to com-munication failures. Communication failures areassociated with significant medical errors.

Knowledge added to the understanding of inter-cliniciancommunication by this study

• Although the majority of caretaker’s time is spent onpatient care (85.4% in this study), nearly three-fourthsof that time is spent on indirect patient care. Of totalwork time, 24.6% is spent on direct patient care (e.g.history and physical examination), 60.7% of time onindirect patient care (e.g. charting) and 14.6% on non-patient care activities.

• The predominance of synchronous communicationcontributed to an interruptive workspace.

• About 95% of interruptions (in this study) wereobserved during synchronous modes of communica-

tion.

Since the study was situated in the emergency departmentof the hospital sites, which is a particularly chaotic envi-ronment, the transferability and generalizability of findingsare limited as well. Expanding the scope of this researchinvestigation by including different types of health care pro-fessionals in various clinical settings would allow for greaterexternal validity and thus a more complete profile of thecommunication processes that take place among health careproviders.

Due to the study design, observers limited conversationwith the clinicians. Thus, no demographic information orinformation about the experience of the participants wasobtained. Furthermore, the observers could not ascertain thequality of the information being communicated. For example,in terms of quality, an in-depth discussion with a clinician fora duration of 3 min is not equivalent to a cursory chart reviewthat also takes 3 min.

Inter-rater reliability (determined by using percentageagreements between the coders) was not established prior toconducting the study. Researchers were, however, given thesame training on executing the observations, which was con-

sidered a sufficient method of ensuring uniformity amongraters.

The 5-day study timeline was likely unable to over-come the Hawthorne effect as the observers’ presence

i n f o r m a t i c s 7 8 ( 2 0 0 9 ) 629–637

could have affected the behavior of clinicians being shad-owed.

5. Conclusion

Currently, the daily operations in the clinical workspaceappear to be heavily dependent upon the use of synchronouscommunication methods. Given the nature of the medicalenvironment and the legal and logistical needs for syn-chronous relaying of certain types of information, there isno doubt that synchronous communication will continue todominate in the clinical workspace. Notwithstanding this fact,the results of this study have demonstrated that the exis-tence of multiple interruptions is correlated with synchronouscommunication processes. It is possible that the developmentof information communication technologies that streamlineclinical workflow through the use of asynchronous communi-cation systems (e.g. instant text, e-mail, and voice messaging)could provide a solution by aggregating these interruptionsor diverting interruptions to more opportune times. Ideally,ICTs should aim to optimize medical workflows by reduc-ing the frequency of interruptions. More quantitative andqualitative studies on inter-clinician communication wouldprovide technology developers and innovators with moreinformation and insights so that they can respond to the chal-lenges met by health care providers. Consequently, furtherstudies examining communication patterns among healthcare professionals in various health care settings are recom-mended.

Acknowledgements

This project was supported by a grant received fromQualcomm® Enterprise Services. Special thanks to the studyparticipants who willingly volunteered to participate in thisstudy. The investigators would also like to thank the admin-istration of the two study sites for allowing them to use thehospitals’ space and resources during the observational studyperiod.

Contributions: Ashley Edwards was involved in the concep-tion of the study design. Additionally, Ms. Edwards composedthe final manuscript for publication. Ms. Edwards made criti-cal final revisions and submission to the International Journalof Medical Informatics.

Leslie-Anne Fitzpatrick was the research coordinator onthis study. She was actively involved in the day-to-day oper-ations of the study, including data collection, analysis, andinterpretation. Ms. Fitzpatrick composed a substantial portionof the manuscript and contributed significantly to the format-ting and final approval of the document.

Sara Augustine and Alex Trzebucki were study inves-tigators. They played integral roles in the conduct ofthe study, including managing IRB protocols and datacollection, coding, and entry. Ms. Augustine and Mr.

Trzebucki both had considerable input into drafting the arti-cle.

Angie Cheng and Candice Presseau were involved in thereview of the final manuscript. In addition to making edits

a l i n

tt

rTv

It

r

i n t e r n a t i o n a l j o u r n a l o f m e d i c

o text, Ms. Cheng assisted with formatting diagrams andables.

Cynthia Mersmann and Bruce Heckman were involved inevising the manuscript for important intellectual content.hey both served as the principal investigators for the obser-ational study at their respective institutions.

Stan Kachnowski is the Chair of Research at the Health carennovation and Technology Lab. He gave the final approval ofhe manuscript to be submitted.

e f e r e n c e s

[1] R. Spencer, E. Coiera, P. Logan, Variation in communicationloads on clinical staff in the emergency department, Annalsof Emergency Medicine 44 (3) (2004) 268–273.

[2] J. Parker, E. Coiera, Improving clinical communication: aview from psychology, Journal of the American MedicalInformatics Association 7 (5) (2000) 453–461.

[3] E. Coiera, V. Tombs, Communication behaviours in ahospital setting: an observational study, British MedicalJournal 316 (7132) (1998) 673–676.

[4] D.G. Covell, G.C. Uman, P.R. Manning, Information needs inoffice practice: are they being met? Annals of InternalMedicine 103 (4) (1985) 596–599.

[5] J.J. Brixey, et al., Interruptions in workflow for RNs in a LevelOne Trauma Center, in: AMIA Annual SymposiumProceedings/AMIA Symposium, AMIA Symposium, 2005, pp.86–90.

[6] J.J. Brixey, et al., Interruptions in a level one trauma center: acase study, International Journal of Medical Informatics 77(4) (2008) 235–241.

[7] A. Laxmisan, et al., The multitasking clinician:decision-making and cognitive demand during and after

team handoffs in emergency care, International Journal ofMedical Informatics 76 (11–12) (2007) 801–811.

[8] C.A. Vincent, R.L. Wears, Communication in the emergencydepartment: separating the signal from the noise, MedicalJournal of Australia 176 (9) (2002) 409–410.

f o r m a t i c s 7 8 ( 2 0 0 9 ) 629–637 637

[9] L.T. Kohn, J.M. Corrigan, M.S. Donaldson (Eds.), To Err isHuman: Building a Safer Health System, National AcademyPress, Washington, DC, 2000.

[10] Sentinel Event Alert, 2002 [cited 2008 May 1]. Availablefrom: http://www.jointcommission.org/SentinelEvents/SentinelEventAlert/sea 27.htm.

[11] D. Walsh, et al., A PDA-based point of care e-health solutionfor ambulatory care, Australasian Journal of InformationSystems 13 (1) (2005) 6.

[12] B. Kirwan, L.K. Ainsworth (Eds.), A Guide to Task Analysis,CRC Press, Boca Raton, FL, 1992.

[13] J.C. Hollingsworth, et al., How do physicians and nursesspend their time in the emergency department? Annals ofEmergency Medicine 31 (1) (1998) 87–91.

[14] M. Woloshynowych, et al., Communication patterns in a UKEmergency Department, Annals of Emergency Medicine 50(4) (2007) 407–413.

[15] G.A. Miller, The magical number seven, plus or minus two:some limits on our capacity for processing information,Psychological Review 101 (2) (1994) 343–352.

[16] P.R. Dexter, et al., A computerized reminder system toincrease the use of preventive care for hospitalized patients,New England Journal of Medicine 345 (13) (2001) 965–970.

[17] E.A. Balas, et al., Improving preventive care by promptingphysicians, Archives of Internal Medicine 160 (3) (2000)301–308.

[18] S.M. Austin, et al., Effect of physician reminders onpreventive care: meta-analysis of randomized clinical trials,in: Proceedings/the. . .Annual Symposium on ComputerApplication [sic] in Medical Care, Symposium on ComputerApplications in Medical Care, 1994, pp. 121–124.

[19] Y. Dahl, “You have a message here”: enhancinginterpersonal communication in a hospital ward withlocation-based virtual notes, Methods of Information inMedicine 45 (6) (2006) 602–609.

[20] J. Favela, et al., Integrating context-aware public displays

into a mobile hospital information system, IEEETransactions on Information Technology in Biomedicine 8(3) (2004) 279–286.

[21] M.A. Munoz, et al., Context-aware mobile communication inhospitals, Computer 36 (9) (2003) 38–46.