internat ional journal of med ical informat ics 7 8 ( 2 0 0 9 ) 629637
journa l homepage: www. int l .e lsev ierhea l th .com/ journa ls / i jmi
Synchronous communication facilitates interruptiveworkclinic
Ashley EShing LBruce Ha Healthcareb Saint Vincc Phelps Me
a r t i c
10 April 200
CorrespoTel.: +1 212
E-mail a1386-5056/$doi:10.1016/ow for attending physicians and nurses inal settings
dwardsa,, Leslie-Anne Fitzpatricka, Sara Augustinea, Alex Trzebuckia,ai Chenga, Candice Presseaua, Cynthia Mersmannb,eckmanc, Stan Kachnowskia
Innovation and Technology Lab, New York, NY, United Statesents Catholic Medical Center, New York, NY, United Statesmorial Hospital Center, Westchester, NY, United States
l e i n f o
a b s t r a c t
Study objective: Inter-clinician communication accounts formore 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 workow 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 24h 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
2024.67min. 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 workow. 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
0.982.24min 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 workow anal-
ysis may inform the development of communication devices to enhance inter-clinician
communication by reducing interruptions or deferring interruptions to more appropriate
2009 Elsevier Ireland Ltd. All rights reserved.
nding author at: Healthcare Innovation and Technology Lab, 3960 Broadway, Room 410, New York, NY 10032, United States.543 0100; fax: +1 212 543 0108.ddress: email@example.com (A. Edwards). see front matter 2009 Elsevier Ireland Ltd. All rights reserved.j.ijmedinf.2009.04.006
630 internat ional journal of med ical informat ics 7 8 ( 2 0 0 9 ) 629637
Communicmon and imthe healthmay occurmunicationvoicemail,channels; huse of syncto use syndled with iinterruptiofar-reachinpatient carInterruptioby disruptidenced inHuman: Butively convwith signion patientHealth carebreakdown
In a coenvironmecuted efcifor studyincians and nstudies prthe deliverow patterthe clinicadevelopingtion technoworkers [11clear deni
Observain order tobehaviors aterns, andAdditionalldations forsolution we
A non-partional stud2008 at on235-bed nostudy involhospital sitone internato the medthe greaterhospital ha
seven attending physicians (6 emergency room physiciansand 1 internal medicine physician) and two emergency room
rsesrovidhisy acy voled.
bservn Oudyarchb (Hationerswhasksriodstudwasatchetrucearcn teg liteelivey, clia c
oremn evceiveuniced cthe sile talso a
t reqnfore. Eveientas pasyncs) veunicd alobserved n
g. 1 stheationamonghealth careproviders is themost com-portant exchange of clinical information within
care system . Information and data exchangesalong both synchronous (e.g. face-to-face com-, telephone conversations) and asynchronous (e.g.e-mail, patient charts, notes, page) communicationowever, past research has noted a bias towards thehronous channels . As a result of the tendencychronous channels, the clinical workspace is rid-nterruptive and multitasking workows. Frequentns andmultitasking in a clinical environment haveg and detrimental implications on the quality ofe [5,6] and the nancial cost to institutions .ns and multitasking contribute to medical errorng short-term memory processes [7,8]. This is evi-the Institute of Medicine (IOM) report, To Err isilding a Safer Health System. This report effec-eyed that communication failures were associatedcant medical errors . Moreover, in a 2002 reportsafety, the Joint Commission on Accreditation ofOrganizations (JCAHO) found that communicationwas a major cause of sentinel events .
mmunication-rich workspace such as a medicalnt, it is important for communication to be exe-ently and accurately. There is a documented needgandunderstanding communicationamongphysi-urses in the clinical setting. In situ, observational
ovide insight into the signicant time gaps iny of care as well as common data and work-ns. Insights into communication processes withinl workow processes are crucial to successfullymobile, asynchronous information communica-logies (ICTs) that are readily adopted by healthcare]. Furthermore, workow analysis helps to createtions of user needs and ICT requirements [12,13].tional studies were conducted at two hospital sitesgain a better understanding of communicationmong health care providers, routine workow pat-the use of ICTs within the clinical workspace.y, at the conclusion of data collection, recommen-the development of an effective health care ICTre made.
dy design and setting
ticipatory, qualitative time-and-motion observa-y was conducted over a 5-day period in Februarye 489-bed tertiary care teaching hospital and onet-for-prot acute care hospital. The observationalved primarily the emergency department of eache. At the 235-bed not-for-prot acute care hospital,l medicine physician was shadowed and followedicine department. The study sites were located inNew York Metropolitan area, however, the largerd signicantly higher patient inux. A total of
(ER) nucare pfrom tsistencentirelobtain
The onicatio. Stby reseogy Laobservobservrecorddaily ttion pewhileTimestopwsemi-sthe resnicatioexistincare dactivitduringprovidactionand invoicemThe afnicationot recommobservduringingwhwhile aMultitow.
Anylimitedpatienious uperiodinefction wuse ofsationcommalogueAny oUn-codday. Fiduringparticipated in the study. Although other healthers were observed, those results were excludedpaper due to insufcient data and lack of con-ross institutions. Participation in the study wasuntary. Demographics of the participants were not
thods of measurement
ation methodology was based on the Commu-bservation Method established by Spencer et al.participants were shadowed for 24h at a timeers from the Healthcare Innovation and Technol-IT Lab). All researchers were trained in the same
methodology. The shadowing method allowedto unobtrusively follow health care providers toat, why, and how they performed their routinein the clinical setting . During each observa-
, inter-clinician communication events were timedy participants performed their routine activities.tracked in real-time using synchronized digitals. Communication events were recorded on atured, paper-based observation tool developed byh team at the HIT Lab. The denitions of commu-rms used in the study were consistent with therature on inter-clinician communication in healthry . Interruption was dened as any disruptivenician conversation, event, or alert that occurredommunication stream between two health carecommunication event was considered to be any
n in order to relay information to another clinician,ed conversation, telephone calls, pages, e-mails,noteswritten in a patient record, and textmessage.entioned actions were considered to be commu-ents even if the information was delayed or wasd by the intended recipient. Percentage time ination accounted for the percent of time that thelinician was involved with communication eventshadowingperiod. Lastly, multitasking (e.g. chart-lking) was dened as attending to at least one dutyttempting to communicate with another clinician.g was observed to greatly impact communication
uptions in communication ows, including but nottelephone calls, questions from other clinicians,uests, waiting for other team members, and var-seen events were logged during the observationnts that took place in order to compensate forcommunication were also noted. Particular atten-id to communication patterns associated with thehronous (e.g. telephone and face-to-face conver-rsus asynchronous (e.g. fax, paging, and e-mail)ation channels. Communication devices were cat-ng with usage frequencies and usage scenarios.able communication barriers were documented.otes were preliminarily coded at the end of eachummarizes the communicationmethodology usedon-site observational studies.
internat ional journal of med ical informat ics 7 8 ( 2 0 0 9 ) 629637 631
There wpersonnelmunicationwas kept toon observaabout the pentered in aMicrosoft
The timesevents wermunicationcarefully noby clinicianies of cliniccollected.
Before entefrom the inresearchersance PortaClinical Prawere takendential, andshadowed.
Seven attenover a periovides a sum
averal, paientandtienies (emuelecing tocumtershad
pletees inFig. 1 Summary of observational s
ere no structured interviews with health careduring the observational period and verbal com-between the observer and health care personnela minimum, except in the event that claricationtions was needed. No demographic informationarticipants was recorded. The resulting data wasMicrosoft Access database and later exported to
Excel for communication event frequency analysis.
associated with inter-clinician communicatione recorded. The personnel involved in the com-
OnOveralon pathistoryrect paactivit
Comety ofincludized dcompututionsa comincludevent, interruptions, and devices used were also
ted. All of the outcomemeasureswere categorizedtype (i.e., attending physician or nurse). Case stud-ian workows were constructed based on the data
ring hospital study sites, permission was obtainedstitutional review board (IRB) of each hospital. Allwere certied with the necessary Health Insur-
bility and Accountability Act (HIPAA) and Goodctice (GCP) requirements. Appropriate measuresto ensure that all data collected was secure, con-de-identied to protect the participantswhowere
ding physicians and two ER nurses were observedd of 1727.67 and 297min, respectively. Table 1 pro-mary of the distribution of communication events.
The systemfor health cthatwere sof communvariousdevin Table 2.
Most ofsynchronounication acmunicationtelephone cnication wesynchronouchannels, uwasmost cclinician tytions of syusage werewere equacase studysynchronoucommon clmethodology.
ge, there were 29 communication events per hour.rticipants spent 85.4% of their time (1728.33min)care (24.6% of time on direct patient care, e.g.physical examination, and 60.7% of time on indi-
t care, e.g. charting) and 14.6% on non-patient care.g. social greetings).nication was executed through the use of a vari-tronic and non-electronic methods and devices,elephones, face-to-face conversations, computer-entation systems, paper charts, fax machines,on wheels (COWs) and pagers. One of the insti-incorporated the Vocera communication system,proprietary voice-over-IP intercom system that
dividual intercom badges, into their workow.
operates over a wireless network and is designedare facilities. Interestingly, none of the clinicians
hadowed at that institution usedVocera as ameansicating to other clinicians. The frequency of use ofices at the twoobservational study sites is captured
the communication events (84.4%) owed alongs communication channels. Face-to-face commu-
counted for 70.0% (n=681) of the synchronous com-events, while 15.4% (n=152) of these events wereonversations. All forms of asynchronous commu-re utilized substantially less frequently than anys communication channel. Along asynchronousse of a computerized hospital EMR/patient chartommonly observed (see Table 2).When analyzed bype (i.e. attending physician and nurse), the propor-nchronous versus asynchronous communicationsimilar. Synchronous forms of communication
lly dominant within both clinician groups. Theshown in Fig. 2 demonstrates the dominance ofs communication patterns in the execution ofinical tasks. The case study details an attending ER
632 internat ional journal of med ical informat ics 7 8 ( 2 0 0 9 ) 629637
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 415Interruption time (min) 353.03 41.42 394.45
N multitasking events 142 28 170Multitasking time (min) 254.34 44.05 298.39
physicians effort to acquire a surgical consult while manag-ing standardworkload. The process lasted for a total of 56min.Of the 56min, 31min and 58 s were dedicated to communica-tion surrounding the specic patients 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 sytelephonemethods ofof labs andhospital EMstudy were
The predotributed toin which atime (353.0interruptio(2.24min)than they ievents (n=recipientsleagues. Asresidents wcians shad
nurses shadowed were the initiators of 41.5% (n=22) of theirinterruptions over a total of 25.15min. Like the attendingphysicians, nurses received more interruptions than theyinitiated31 interruptions received versus 22 interruptionsinitiated. The interruptions analysis for nurses is summarizedin Fig. 4.
on ource5 reus cto coreace thof coandcretarmes, haronoy for
Table 2 omm
AsynchronoComputeriPaper formPaper charWritten noOverhead pPagerPDAnchronous channels, including face-to-face andmethods. The only events utilizing asynchronouscommunication were those involving the ordering/or the entering of patient notes in a compute...