coordinating flexible performance during everyday work: an ... et...

This article was downloaded by: [142.150.190.39] On: 31 July 2016, At: 10:23Publisher: Institute for Operations Research and the Management Sciences (INFORMS)INFORMS is located in Maryland, USA

Organization Science

Publication details, including instructions for authors and subscription information:http://pubsonline.informs.org

Coordinating Flexible Performance During Everyday Work:An Ethnomethodological Study of Handoff RoutinesCurtis LeBaron, Marlys K. Christianson, Lyndon Garrett, Roy Ilan

To cite this article:Curtis LeBaron, Marlys K. Christianson, Lyndon Garrett, Roy Ilan (2016) Coordinating Flexible Performance During EverydayWork: An Ethnomethodological Study of Handoff Routines. Organization Science 27(3):514-534. http://dx.doi.org/10.1287/orsc.2015.1043

Full terms and conditions of use: http://pubsonline.informs.org/page/terms-and-conditions

This article may be used only for the purposes of research, teaching, and/or private study. Commercial useor systematic downloading (by robots or other automatic processes) is prohibited without explicit Publisherapproval, unless otherwise noted. For more information, contact [email protected].

The Publisher does not warrant or guarantee the article’s accuracy, completeness, merchantability, fitnessfor a particular purpose, or non-infringement. Descriptions of, or references to, products or publications, orinclusion of an advertisement in this article, neither constitutes nor implies a guarantee, endorsement, orsupport of claims made of that product, publication, or service.

Copyright © 2016, INFORMS

Please scroll down for article—it is on subsequent pages

INFORMS is the largest professional society in the world for professionals in the fields of operations research, managementscience, and analytics.For more information on INFORMS, its publications, membership, or meetings visit http://www.informs.org

http://pubsonline.informs.org

http://dx.doi.org/10.1287/orsc.2015.1043

http://dx.doi.org/10.1287/orsc.2015.1043

http://pubsonline.informs.org/page/terms-and-conditions

http://www.informs.org

OrganizationScienceVol. 27, No. 3, May–June 2016, pp. 514–534ISSN 1047-7039 (print) � ISSN 1526-5455 (online) http://dx.doi.org/10.1287/orsc.2015.1043

© 2016 INFORMS

Coordinating Flexible Performance During Everyday Work:An Ethnomethodological Study of Handoff Routines

Curtis LeBaronMarriott School of Management, Brigham Young University, Provo, Utah 84602, [email protected]

Marlys K. ChristiansonRotman School of Management, University of Toronto, Toronto M5S 3E6, Ontario, [email protected]

Lyndon GarrettStephen M. Ross School of Business, University of Michigan, Ann Arbor, Michigan 48109, [email protected]

Roy IlanQueen’s University, Kingston K7L 3N6, Ontario, [email protected]

Our paper examines the challenge of coordinating flexible performance during everyday work. We draw on routinedynamics and ethnomethodology to examine how intensive care unit (ICU) physicians coordinate their actions—flexibly

yet intelligibly—as they handoff patients at change of shift. Through our analysis of interview and video data, we demonstratehow physicians use the sequential features of the handoff routine—i.e., the expected moves and their expected sequence—toadapt each performance of the routine to the unique needs of each patient. We show the need for ongoing coordinating despitea strongly shared ostensive pattern and we illustrate how participants use the sequential nature of the ostensive pattern of theroutine as a resource for flexible performance, to manage sequential variation and the sufficiency of moves at transitions. Ourfindings contribute to the routine dynamics and coordination literatures by providing a more nuanced understanding of howmutual intelligibility is achieved through coordinating, whereby participants create the conditions to move forward with acommon project.

Keywords : organizational routines; coordinating; ethnomethodologyHistory : Published online in Articles in Advance April 11, 2016.

IntroductionPeople working together in organizations often need totailor their performance of a shared task to the situationat hand. As work becomes increasingly complex anddynamic (D’Aveni 1994, Eisenhardt 1989, Weick et al.1999), everyday tasks are frequently adapted to theparticular circumstances that exist at the moment ofenactment. However, adapting the way that a task isusually performed can pose a coordination challengefor everyone involved. On the one hand, the collectiveperformance of the task must be flexible enough toaccommodate the necessary variety of the situation. Onthe other hand, since people are improvising to somedegree, they must also coordinate their actions so thattheir performance remains mutually intelligible—that is,to move forward with a shared task, they need to beable to make sense of what others are saying and doing.Okhuysen and Bechky (2009) highlight the importanceof shared understanding for coordinating. Research oncoordinating has provided significant insight into how theshared understanding that people bring into performanceinfluences coordinated action. However, we know muchless about how shared understanding is created andmaintained during flexible performance (Dionysiou and

Tsoukas 2013, Okhuysen and Bechky 2009). In addition,although previous research has identified the need forrepair during flexible performance (Turner and Rindova2012, Zbaracki and Bergen 2010), many questions remainabout how this repair takes place. Our paper tackles theissue of how people create, maintain, and repair mutualintelligibility during flexible performance.

To study the challenge of coordinating flexible per-formance, we draw upon two complementary researchtraditions: routine dynamics and ethnomethodology. Whenshared tasks occur as part of everyday work, we can thinkof these tasks as routines, defined as “repetitive, recog-nizable patterns of interdependent actions, carried outby multiple actors” (Feldman and Pentland 2003, p. 95).As people collectively perform a routine, interdepen-dent action unfolds sequentially (Feldman and Pentland2003, Pentland and Rueter 1994). The routine dynamicsliterature is centrally concerned with how the “same”routine can be flexibly performed—from this perspective,routines are not an invariant set of actions, but instead areconstructed by participants from a repertoire of possibleactions (Birnholtz et al. 2007, Cohen 2007, Feldman2000, Feldman and Pentland 2003, Howard-Grenville2005, Pentland et al. 2011, Pentland and Rueter 1994,

514

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

mailto:[email protected]




LeBaron et al.: Coordinating Flexible Performance During Everyday WorkOrganization Science 27(3), pp. 514–534, © 2016 INFORMS 515

Turner and Rindova 2012). The routine dynamics liter-ature provides theoretical handholds for understandingflexible performance by differentiating between the osten-sive aspects—the general and abstract patterns of theroutine—and the performative aspects of routines—thespecific actions taken by specific people at specific placesand times (Feldman and Pentland 2003). By examiningthe relationship between the ostensive and performativeaspects of the routine, we gain insight into how mutualintelligibility is created, maintained, and repaired duringflexible performance.

Additionally, we employ an approach to studyingcoordinating—ethnomethodology—that foregrounds therole of action in creating shared understanding. We arguethat to study coordinating flexible performance, we musttake seriously the obstacles that participants encounterduring performance. Not being able to read each other’sminds, people do not have a way of knowing what otherpeople are thinking, except through the actions that theymake available to each other during performance—that is,what they say and what they do. An ethnomethodologicalapproach (Garfinkel 1967, Heritage 1984, Rawls 2008)provides a way of observing and documenting the methodsthat people use to create, maintain, and repair mutualintelligibility. From an ethnomethodological perspective,mutual intelligibility is not something that people have (inthe sense of shared cognition), rather it is something thatpeople enact—for example, they display understandingto each other by marking prior actions as sufficient orinsufficient for moving forward with a common project.Because mutual intelligibility is something that peopleshow to each other during performance, the emergence ofthis displayed understanding during flexible performancebecomes accessible to researchers.

We study flexible performance at a moment when coor-dinating is especially important: handoffs at shift change,as people transfer information about and responsibilityfor a shared task. During handoffs, outgoing personnelhave a limited amount of time to transfer the necessaryinformation to the incoming personnel. At the same time,it is essential that the incoming personnel understand thehandoff so that they can take over the work. Handoffsoccur in a wide variety of dynamic and consequentialsettings, such as hospitals, nuclear power plants, airtraffic control, railroad dispatch, and shuttle missions(Mumaw et al. 2000, Patterson et al. 2004). Our studyexamines handoffs between intensive care unit (ICU)physicians at change of shift—a setting in which bothflexible performance and mutual intelligibility are vitallyimportant. Since each patient in the ICU has a uniqueconstellation of medical issues, no two handoffs can beperformed in exactly the same way; yet, breakdowns inmutual intelligibility during the performance of handoffscan have lethal consequences (Sutcliffe et al. 2004).

By interviewing ICU physicians, we found that theyhad strongly shared expectations about how handoff rou-tines should be enacted, including particular moves in a

particular sequence. By video recording and analyzing thesame physicians conducting handoffs, we saw how theyused these expectations to create, maintain, and repairthe mutual intelligibility of their flexible performance.We develop a process model to summarize the theorythat emerged from our findings. Our study makes atleast two contributions to research on the flexible per-formance of routines. First, in contrast to prior researchthat suggests that coordinating results from increasinglyshared understanding of a routine (Dionysiou and Tsoukas2013), we show how coordinating during flexible perfor-mance is ongoing, even when participants have stronglyshared expectations about the moves and sequence oftheir routine. Second, building on prior research thatshows that the ostensive aspect of the routine is a resourcefor performance (Feldman and Pentland 2003, Howard-Grenville 2005, Turner and Rindova 2012, Zbaracki andBergen 2010), we show how the sequential features of theroutine are a resource for flexible performance. We foundthat sequential variation of handoffs (e.g., rearrangingand/or skipping moves) was mutually intelligible, largelythrough its relationship to the physicians’ expectationsabout moves and sequence. We also found that physicians’shared expectations about sequence help them to negotiatemoves at transitions in sequence. Taken together, ourfindings contribute to a more nuanced understandingof how mutual intelligibility is achieved during flexibleperformance of everyday work.

Coordinating During the FlexiblePerformance of RoutinesOkhuysen and Bechky (2009) argue that routines facilitatecoordinating by establishing how interdependent tasks aretransferred from one person to another, by providing atemplate to structure and sequence tasks, and by creatinga shared understanding of the task at hand. In keepingwith past research that emphasizes the dynamic nature ofcoordinating (Faraj and Xiao 2006, Jarzabkowski et al.2012), we also focus on the “collective performanceaspect of coordination and emphasize the temporal unfold-ing and situated nature of coordinative action” (Farajand Xiao 2006, p. 1155). To explore issues related tocoordinating during the flexible performance of routines,we turn first to a discussion of routine dynamics and thenethnomethodology.

Routine DynamicsDrawing on Feldman and Pentland’s (2003) distinctionbetween the ostensive and performative aspects of theroutine helps us understand how the “same” routinecan be flexibly performed. Even though specific perfor-mances vary, any given performance of the routine takesplace against a “background of rules and expectations”(Feldman and Pentland 2003, p. 102). Although not fullyspecifying performance, the ostensive aspects of the rou-tine enable people to guide, account for, and refer to

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

LeBaron et al.: Coordinating Flexible Performance During Everyday Work516 Organization Science 27(3), pp. 514–534, © 2016 INFORMS

specific performances of a routine (Feldman and Pentland2003). Relatedly, the performative aspect of the routinecreates, maintains, and modifies the ostensive aspects ofthe routine (Feldman and Pentland 2003). As a result,the ostensive and performative aspects of the routinemutually constitute each other. Rerup and Feldman (2011)highlight that, over repeated performances, performativeactions generate ostensive patterns, which are instantiatedin future performances. Feldman and Pentland (2003)emphasize that the ostensive is not a “single, unifiedobject,” but instead “incorporates the subjective under-standings of diverse participants (Feldman and Pentland2003, p. 101). When people understand the routine indifferent ways, this makes coordinating more complicated.

Previous research has examined how people developshared understanding about the ostensive pattern of theroutine. For instance, Feldman and Rafaeli (2002) foundthat the repeated performances of routines built connec-tions between people that enabled them to develop sharedunderstandings about what to do in a particular instanceand why some actions were appropriate (or not). Morerecently, Turner and Rindova (2012) showed how wastecollection crews drew on artifacts and connections todevelop shared understanding that facilitated both con-sistent and flexible performances. These studies provideevidence that participants’ already established sharedunderstanding about the routine can be a resource forperformance.

Dionysiou and Tsoukas (2013) highlight that there isstill much to be learned about how shared understandingemerges during flexible performance—specifically, aroundhow people develop “joint situated understanding” and“align their actions.” Adopting a symbolic interactionistapproach, they draw on Mead’s (1934) concept of role-taking to develop theory about how this process takesplace. They argue that, through repeated performances of aroutine, participants develop increasingly shared schemataabout the ostensive aspect of the routine. They write,

During interaction (performative aspect, Figure 1, [P]),participants engage in role taking—that is, they take intoaccount fellow participants’ roles (their actual and potentialunderstandings, ideas, opinions, and actions) with respectto the joint activity in order to develop a joint, situatedunderstanding of the concrete situation at hand; identifyappropriate actions; and align their individual lines ofaction accordingly (Figure 1, arrow d; Blumer 1969, 2004;Weick 1995). (Dionysiou and Tsoukas 2013, p. 191)

They go on to argue that developing increasingly sharedschemata about one’s role is beneficial because participants“can then organize their conduct in accordance withmutually consistent behavioral expectations (Blumer 1969,Joas 1997, Miller 1973)” (Dionysiou and Tsoukas 2013,p. 192).

Dionysiou and Tsoukas (2013, p. 197) recognize thatparticipants’ “mutually established expectations” canbe misaligned, or that a particular performance of a

routine may need to be “nonprototypical,” and suggestthat these circumstances can lead to “a more ‘intelligent’engagement with the joint activity (Mead 1934, 1938;Weick 1979).” Yet how this “more intelligent engagement”is accomplished remains under specified. What is missingfrom their theory is an account of how participants aligntheir actions and develop joint situated understandingsas they engage in real-time adjustment and repair of theroutine, to ensure that the performance remains mutuallyintelligible despite being flexibly performed.

Early research on routine dynamics reminds us of theimportance of real-time adjustment and repair in theservice of mutual intelligibility during flexible perfor-mance. Feldman and Rafaeli (2002) likened the mutualadjustment that is required during flexible performance toa dance, where participants adapt and align their actionsto fit with their partner’s actions. Similarly, Pentland andcolleagues (Pentland 1992, Pentland and Rueter 1994)talked about “moves” as a “unitary act of the routine”made meaningful through interaction. Drawing on workby Goffman (1981) and Schegloff (1982), Pentland (1992,p. 530) writes that, “the accomplishment of a particularmove generally depends on its ratification (implicitly orexplicitly) by others.” Pentland (1992, p. 530) furtherargues that “moves cannot be isolated from the situa-tional particulars of the context and the sequence ofinteraction in which they occur.” Both metaphors (“dance”and “move”) point to the reciprocal nature of real-timeadjustment in which actions are simultaneously both aresponse and a stimulus during unfolding interaction.Although a few studies have identified the need for repairduring performance—for instance, when the truce hold-ing together a routine collapses (Zbaracki and Bergen2010) or when the expected sequence of the routine isdisrupted (Turner and Rindova 2012)—how repair takesplace during flexible performance is a largely unansweredquestion.

Our study takes on this question of how participantscoordinate their actions to create, maintain, and repairmutual intelligibility during flexible performance. Incontrast to Dionysiou and Tsoukas’ (2013) role-taking per-spective on coordinated action, we take an ethnomethod-ological approach that remains agnostic about the degreeof shared understanding people actually have (i.e., sharedcognition). Our focus is on the shared understandings thatthey display.

EthnomethodologyEthnomethodology is a useful complement to routinedynamics for studying flexible performance becausethis approach gives additional insight into how peoplecoordinate and make sense of their actions in the momentthat they perform them. Garfinkel (1967, p. 1) introducedthe notion of ethnomethods to describe the methods thatpeople use to “produce and manage settings of organized

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


everyday affairs.” Ethnomethodology takes seriously theknowledge and skills that people bring to their interactions,and it aims to uncover the practices that they use to createand maintain a shared social world—for instance, a classicethnomethod is the accounting practice, whereby peoplecreate narratives that account for the facts as presented.Garfinkel argued that we gain insight into how people“construct a mutually intelligible world” (Rawls 2006,p. 7) through a detailed study of the practices they use:both their situated actions—what they say and do in aparticular situation—and the sequence of those situatedactions. So one way to answer our research question—howdo participants coordinate their actions in a mutuallyintelligible manner during flexible performance?—is toconduct an ethnomethodological study of people who aredoing just that.

Ethnomethodology emphasizes that people use a ma-chinery of expectations as they choose and coordinatetheir actions (David and Sutton 2011, Sacks 1992, Sil-verman 1998). Ethnomethodology is concerned withhow participants construct the meaning of actions andevents, not only when actions are consistent with pre-vailing norms or rules, but also when they deviate fromthem. Interdependent action, Garfinkel argued, takes placeagainst a background of expectations that shape howparticipants coordinate and make sense of their every-day interactions. He described these expectations as the“‘seen but unnoticed,’ expected, background features ofeveryday scenes” (Garfinkel 1967, p. 36). Garfinkel wasinterested not only in how background expectations wereused by participants to create and maintain a “world incommon,” but also in how participants repaired breachesand violations of those expectations. Ethnomethodology isfamous for Garfinkel’s (1967) “breaching experiments” inwhich field researchers violate—and thereby bring to thesurface—the taken-for-granted expectations of participants.Garfinkel highlighted that breaches frequently occur aspart of everyday activities but are difficult to sustainbecause people are so skilled at repairing them.

An ethnomethodological approach also highlights howparticipants make sense of sequences of action. Garfinkelobserved that actions acquire their meaning in relationshipto prior and subsequent actions. On this point, Rawls(2008, p. 712) writes about ethnomethodology that, “eachnext thing done or said is taken in relation to the last(reflects back on the last), and this reflexive sequentialchain constitutes a basic order of sensemaking.” Thenotion of sequence that is central to ethnomethodologyhas influenced allied approaches, such as conversationanalysis (Heritage 1984, Sacks 1992, Silverman 2013).By applying an ethnomethodological approach to thestudy of coordinating flexible performance, we can bothidentify moves and their sequence, and also examinehow the sequential unfolding of actions is meaningfullyunderstood and displayed by participants.

MethodsResearch SiteCommunity Hospital (CH) is the flagship institutionof a healthcare system that includes 24 satellite sites,serving about 500,000 people. CH provides a full range ofstandard and specialty services, making it a destination forpatients who have complex and uncertain conditions thatcannot be treated elsewhere. Our research was conductedin this hospital’s ICU, which receives patients with a widevariety of problems (medical, surgical, trauma, and soforth). The ICU has 21 beds, a dozen attending physicians,and a full complement of nurses and hospital staff. Priorto our study, we obtained ethics approval and informedconsent from the 10 attending physicians that we recordedduring handoffs and interviews.

Our specific focus is ICU handoffs that occur at theshift changes of attending physicians. Within the contextof healthcare, handoffs occur frequently: by conservativeestimate, there are half a billion handoffs every yearin U.S. inpatient hospital settings (Cohen et al. 2012).Handoffs are an important and consequential routine inwhich misunderstandings could harm patients (Sutcliffeet al. 2004). At the CH ICU, attending physicians are oncall for a week at a time. Handoffs occur once each weekwhen pairs of outgoing and incoming physicians enter asmall meeting room and sit down to discuss each patientin their particular group. These attending physicians arepeers who have worked together for years, with a medianof 10 years of experience. Although there are growingcalls to establish industry standards for handoffs, thereis not yet a universally accepted or official form thathandoffs must take within the context of an ICU. Rather,the practice of handoffs is locally managed, left to thediscretion of the health professionals involved in theroutine, with physicians able to tailor the routine to eachpatient’s situation (Cohen and Hilligoss 2010, Hilligossand Cohen 2011, Solet et al. 2005).

Research Design and MethodsOur research project began with a general interest inorganizational routines and physician handoffs as a highlyconsequential activity in healthcare. We did not beginwith a specific research question, but rather with datacollection and analysis (see section “Data Analysis”),until a research question eventually emerged and enabledour theoretical contribution (Sacks 1992). We closelyexamined what people actually did, against a backdrop ofwhat they expected (ten Have 2004).

On one hand, our approach was conversation analytic(Atkinson and Heritage 1984). We recorded and analyzedthe conduct of physicians doing handoffs, for the purposeof locating the actions that they performed and explicatingthe sequential packaging of those actions (Pomerantz andFehr 1997). In this view, talk is a form of social action(Austin 1962) as words are a way of doing things, notjust meaning things, and sometimes the work that routines

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


accomplish takes place through unfolding sequencesof talk (Boden 1994). However, conversation analystsprefer the term “talk-in-interaction” when referring tothe object of their study (Drew and Heritage 1992),partly because talk is always a multimodal practice. Talkis always situated within a material environment thatincludes objects, artifacts, and tools; and talk is alwaysembodied by those who look, point, and orient theirbodies in ways that constitute social action (Streecket al. 2011). This conversation analytic treatment ofmultimodality is consistent with Orlikowski’s (2007)notion of sociomateriality, which emphasizes the “inherentinseparability” (Orlikowski and Scott 2008, p. 434) ofsocial and material aspects of organizational activity. Thus,along with talk as a focus of our analysis, we observedthe material and embodied aspects of handoffs capturedby our video recordings. In our analysis, we did notprivilege one modality over another: rather, we attendedto what the participants themselves treated as importantin the moment of performance, which is consistent withthe ethnomethodological emphasis on emic perspectives.Thus, talk became a focus of our analysis when physiciansmade it central to their performance.

On the other hand, we interviewed the physicians thatwe videotaped to learn more about the expectations thatthey carried into their handoff sessions. Interviews, incombination with close observation and analysis, arecomplementary strategies of ethnomethodology (ten Have2004). The expectations that physicians brought to hand-offs, after so many years of conducting them, provideda backdrop for coordinating their flexible performance.Because we recorded handoffs before conducting inter-views, our interview questions did not influence the par-ticipants’ recorded performance. In addition to interviewswith physicians, we talked to other medical informantswho helped us to understand medical terminology andother aspects of medical practice.

Data CollectionVideo recordings. Video recordings capture emergent

action processes as they unfold in time and space, throughorchestrations of discourse, bodies, and things (Streecket al. 2011). Video recordings allow researchers to repeat-edly observe the empirical details of select events andactivities within organizations. Analysts can continuallycompare their emerging theories to their original data,thereby extending and refining their coding as they go. Wetried to record naturally occurring interaction—i.e., behav-ior that would have occurred whether or not a camera wasin the room—so we always conformed our data collectionefforts to the practices of the physicians. We recordedonly when the physicians were ready to handoff, withoutchanging their schedule. We placed our camera in theirhandoff room, in a corner and out of the way, trying to notdisrupt their activity. We also left the room during handoffs,so that only our camera remained. If something surprising

happened, we adapted as the participants improvised;for example, once the handoff room was occupied, sophysicians stood at the nursing station, which is where weplaced our camera. Of course the physicians knew that wewere recording, which may have influenced their behavior,but our disruption was probably minimal. Researchers whouse video cameras have found that their “presence doesnot continue to pervade the action,” that “within a shorttime, the camera is ‘made at home,”’ and that “participantshave a job or work to do, and have to accomplish theirvarious responsibilities in routine and recognizable ways”(Heath et al. 2010, pp. 48–49). We continued to collectvideo data until we had captured all 10 of our attendingphysicians participating in the roles of both outgoing andincoming physician, for a total of 262 handoffs within 27handoff sessions.

Interviews. After concluding our recording of handoffs,we conducted semistructured interviews with 8 of the 10physicians that we recorded (one physician had retired andone was unavailable). The interviews were approximatelyone hour long, conducted by telephone, recorded withthe participants’ permission, and then transcribed. Ourquestions invited the physicians to talk about handoffs,including objectives, preparations, enactments, etc., mov-ing gradually from general inquiries to more probingquestions (see interview protocol, Appendix A). Someof the physicians followed up with us subsequently andinformally, offering additional observations and comments.

Informants. Throughout our process of data collection,a couple of the ICU physicians also acted as ongoinginformants. In addition to explaining medical conceptsand terminology, they answered our questions about theresearch context, such as the variety of patients andproblems, and the training and tenure of physicians. Occa-sionally, these informants also corroborated or correctedour interpretations of handoff interaction, helping us tounderstand the sometimes nuanced performance of thephysicians.

Data AnalysisIn an effort to identify, articulate, and then answer ourresearch question, we analyzed recordings of both thehandoffs and the interviews. Our data analysis had threephases: (1) an initial coding of action patterns and pack-aging during handoffs; (2) an analysis of themes in ourinterview recordings; and (3) further analysis of ourhandoff recordings, using the interviews as a backdropfor comparison and contrast.

Phase one2 Coding patterns and packaging of actionin handoffs. Our data analysis began with watching thevideo recordings of handoffs, carefully and repeatedly.Our aim was to identify recurring patterns of action inthe sequential organization of talk-in-interaction (Boden1994, Drew and Heritage 1992), and begin to explicatethe packaging of those actions. Occasionally a moment“jumped out” as obviously noteworthy, but usually the

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


significance of the physicians’ performance emergedslowly as our examination deepened. We made note ofrecurring action patterns and packaging, which we usedto create a set of codes; it was not our aim to create anexhaustive list of codes, but rather to capture recurringfeatures that seemed most prevalent and important duringthe interactive accomplishment of the handoff routine.

Phase two2 Analyzing themes of interviews. Conduct-ing interviews with the physicians gave us additionaland valuable information about what they thought theywere doing in handoffs. We asked physicians to tell usabout different aspects of their routine, including theirpurpose, preparation, performance, and post-performanceassessment (see Appendix A). Sooner than later, wesaw that the physicians had strongly shared expectationsabout what steps should be taken in handoffs and in whatorder—none of which was mandated by an official orstandard template for conducting handoffs. In seekingtheoretical handholds for our analysis, we turned to workon routines as recurrent patterns of action, comprisedof “moves” in a sequence (Pentland 1992, Pentland andRueter 1994). Notions of moves and sequences fromthis routine-dynamics perspective acted as sensitizingconcepts (Blumer 1954) for our analysis. When we ana-lyzed our interview data to identify recurring themes, andthen to organize our findings as collections of individualstatements grouped by theme, we were conscious ofmoves and sequence. We saw that the work of the handoffroutine was largely accomplished through talk. As weread through the transcripts of our interviews, we saw thatthe physicians had identified a similar set of topics thathandoffs needed to cover, and that they listed those topicsin a particular order. The physicians regarded topics assteps in their process. Thus, to address one of the requiredtopics was, in their description, to move in the directionof accomplishing a handoff.

Phase three2 Analyzing handoffs as sequential pat-terns of action. Following our analysis of interviews, wereturned to our video recordings of handoffs, continuingto use notions of moves and sequence as sensitizingconcepts (Blumer 1954) in our analysis of the video data.As recommended by conversation analysts (Atkinsonand Heritage 1984), we transcribed select moments ofour video data to highlight audible and visible details,small ways of talking and moving that are not especiallymeaningful in isolation but become meaningful throughtheir relationship to other behaviors within sequences ofaction. We observed that physicians were doing muchmore than they had described in interviews. They were notmerely talking about specified topics in a sequence—theywere continually managing the doing of that. Whetheror not a move in a sequence had been accomplishedwas something that the participants had to manage andnegotiate. We saw that participants in handoffs werecoordinating their progression from one move to the next,through audible and visible displays that a move had been

accomplished and that they were ready to move on. Thisobservation helped focus our analysis of video excerpts tolearn more about how participants were using sequentialfeatures of the handoff to coordinate their actions.

Thus, our data analysis proceeded incrementally. Ourinitial analysis and coding of the handoff recordingsinformed our subsequent interviewing, and then ourinterviewing enabled further analysis of the handoffrecordings. This three-phase process enabled our findingsabout coordinating flexible performance. At each phaseof our data analysis we engaged in member-checking, aprocess where we tested our observations with participantsto validate our emergent findings (Lincoln and Guba 1985).We worked to ensure that our findings were representativeof the phenomenon of interest (Lee 1999). And we wereable to triangulate our data (Jick 1979) by using videorecordings, interviews, and informant conversations.

FindingsOur findings demonstrate that the outgoing and incomingphysicians tailored each handoff to the patient’s uniqueconstellation of issues. Within the ICU, issues variedwidely from patient to patient; for example, the treatmentof a young but otherwise healthy patient following traumain a car accident differed significantly from the treatmentof an elderly patient suffering from a heart attack andmultisystem organ failure. Because outgoing physicianscould not possibly share everything they knew about eachpatient, they tried to handoff the information that was mostimportant, in a way that was mutually intelligible. Ourphysicians had a lot of autonomy in conducting handoffs,deciding what to say, when to say it, and how. Althoughmany aspects of ICU care are officially standardized,handoffs are usually not. Most hospitals have no setprotocol for handoffs and their conduct is left to thediscretion of physicians.

Although our physicians had no official protocol forconducting handoffs, they did have consistent expectationsabout how handoffs should be done. They were experts,not only about their patients’ medical care, but alsoabout their ways of working together. Over a period ofyears, the same 12 physicians had conducted thousandsof handoffs with each other: usually meeting as dyads,at approximately the same time each week, in the samelittle room, always talking about the particular set ofpatients at hand. The history of their handoffs was notlost to them. When two physicians sat down for yetanother handoff session, they brought expectations to thetable—expectations about what to do and how to do it—including the prospect that physicians would deviate fromexpectations in the service of a better handoff. Deviationsfrom expectations were useful, not detrimental, as long asthe physicians were coordinating their performance to bemutually intelligible.

We present our findings in two parts. In part one, weconsider the expectations that physicians brought to their

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


handoffs, as reported by them during interviews. Ourinterviews showed that physicians shared three interrelatedexpectations about their routine: (1) that handoffs wouldinclude particular moves; (2) that these moves wouldoccur in a particular sequence; and (3) that the movesand sequence of handoffs would be flexibly performed ortailored to the particular needs of patients.

In part two of our findings, we examine the flexibleperformance and real-time coordinating of physicians inhandoffs, as captured in video recordings. Although thephysicians did not explicitly talk about coordinating dur-ing our interviews, the video shows that there was ongoingcoordinating between outgoing and incoming physicians.Coordinating included displays of sufficiency—such asagreeing, nodding, continuing writing—which audibly orvisibly showed that the handoff was mutually intelligibleand that the participants were ready to continue. Coordi-nating also included displays of insufficiency—such asquestioning, repairing, stopping writing—which showedthat something was wrong and needed to be changedbefore the handoff could go on. Moreover, coordinat-ing was made especially salient during deviations fromexpected moves and sequence, as well as during transi-tions between moves in sequence. Thus, the expectationsthat physicians brought to the table provided a backdropfor the mutual intelligibility of their flexible performance.

Our presentation of findings proceeds in a tell-show-tellfashion (Golden-Biddle and Locke 2007): we explaina core idea to be depicted in the data that follow; thenwe show those data; and finally we tell more abstractlywhat the data show in relation to our ethnomethodologicalapproach and organizational routines in general.

What Physicians Expected About Moves andSequence in the Flexible Performance of HandoffsThrough interviews with ICU attending physicians (thesame physicians in our video recordings) we discoveredthe nature of their expectations about handoffs. Eventhough we interviewed each physician separately, theirdescriptions of handoffs and how they should be performedwere generally the same. Their shared expectations werenot officially mandated or standardized, but were ratherrooted in past experiences, such as their shared history ofcoordinating while flexibly performing handoffs. We beginby reporting the physicians’ expectations about what theirroutine included and how it should be conducted.

The moves in the routine. The physicians identifiedfive key moves that they expected in handoffs: patientidentification, past events, current issues, future plans, andfamily matters. Although they sometimes used differentterminology, all five moves were identified by all of thephysicians.

The sequence of moves. The physicians expected thatthe moves of handoffs would be conducted in a particularsequence. Consistently, the physicians said that they wouldstart by identifying the patient, then tell the patient’s

medical story (past events, current issues, future plans),and end with considerations of the patient’s family. Thethree moves at the core of handoffs—past events, currentissues, future plans—have an obvious chronologicalrelationship or progression. Physicians emphasized thatthis chronological sequence helped them to recognize whatwas happening during a handoff. For example, Physician Cexplained, “[I] extract, out of the head space of the guythat’s handing over to me, everything he or she knowsabout that patient in terms of past history, past events andcurrent happenings, particularly if I’m going to show upthe next day and manage that patient.” Physician C thentold about one handoff he found especially confusingbecause the outgoing physician was not following thesequence of moves. He said, “What’s the flow of what hadbeen going on? There’s no temporal relations, no storybeing told here, [no] chronological flow of what’s beenhappening. It has to be conveyed as 0 0 0where we were,where we are, where we’re going.” Overall, handoffs weremutually intelligible to physicians, largely through thechronological sequence at the heart of their routine.

The flexible performance of moves in sequence. Al-though the physicians expected five moves in a particularsequence, they also expected that handoffs would betailored to fit the patient’s situation. During our interviews,physicians said that their patients came in wide variety—both “demographic” and “pathological” (Physician B).Even though some patients recovered quickly, some were“not expected to survive” (Physician F), and others hadalready “been in the ICU for months” (Physician A).One minute, a physician might be caring for a youngman who had “gone into pulmonary edema, and [was]now on dialysis” (Physician H); and the next minute shemight be treating “an 86-year-old woman who came inwith a COPD exacerbation, triggered by an infection”(Physician F). Because of the wide variety of patients,the physicians had to be flexible in their performance ofhandoffs to match each patient’s situation.

Another reason for flexible performance in handoffswas that physicians could not possibly say everything.They needed “enough information to do a good job”(Physician E), but they wanted the outgoing physician to“act as a filter 0 0 0because you can’t give all the information”(Physician A). Because of the complexity and uniquenessof each patient’s issues, physicians were constantly dealingwith a figure-ground problem, the difficulty of seeing theforest for the trees, the risk of getting lost in details thatwere “[not] really important and they’re obscuring thepicture” (Physician D). On one hand, physicians wantedto see the trees: to “know the main clinical and socialissues” (Physician B). On the other hand, they wanted tosee the forest: a “sense of the big picture of what’s goingon with a patient” (Physician F), “the broader issue ofwhat are the overall goals of care” (Physician C), withoutgetting “bogged down with micro details” (Physician G).One physician summarized this way: “I just want to

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


know the pertinent pieces of information. I don’t want toknow every piece of information because the importantthings get lost in the mix” (Physician E). Thus, too muchinformation was as problematic as too little.

Physicians expected that their performance of handoffswould be flexible, that each move would be enacted in adifferent way, according to the needs of particular patients,and that the sequence of moves in a handoff might alsovary. By deviating from the first two expectations (i.e.,the particular moves and/or their sequence), physicianscould better achieve the third (i.e., tailor routines to thepatient’s situation). For example, physicians might omitinformation by skipping one or more of the moves thatphysicians said they expected. If a patient “doesn’t havevery complex or complicated problems 0 0 0 then it’s okayto omit some stuff that is not really crucial” (Physician C).Or as Physician E explained

Sometimes the way to start [a handoff] is, “When were youhere last?” Because we do pretty asymmetrical shifts, sosometimes you’re in the ICU fairly frequently, sometimesyou haven’t been there for a while. When you’re therefrequently, and the patients have been there for a while,you may know people very well. And then you couldtailor it to just bring them up to speed on what’s happenedsince [they] were there last.

In this way, physicians recognized that handoffs could beflexibly performed by changing the makeup of the movesthat they expected and/or the sequence of those moves.

In summary, physicians expected a particular pattern butalso flexibility in handoffs. On one hand, they expectedcertain moves and in a particular sequence, which func-tioned like a backbone for handoffs to stay the “same.”On the other hand, physicians insisted on flexibility: thathandoffs be tailored according to the patient’s situation.For example, information about patients might be omitted,depending on the patient. So our interviews revealed atension in what the physicians expected, with handoffsbeing both varied and the “same.” We now turn our atten-tion to enacted patterns of flexible performance capturedon video—what the physicians actually did rather thanwhat they said they did.

How Physicians Managed Sequence and NegotiatedMoves in the Flexible Performance of HandoffsDespite the physicians’ similar expectations about hand-offs, which suggested a strongly shared ostensive patternof their routine, their flexible performance relied on vari-ous forms of real-time coordinating. Our video analysisfocuses on nine excerpts (transcribed and annotated) thatshow the physicians coordinating about both the movesand the sequence of those moves in their routine.

First, we show how coordinating during handoffs wasongoing. At virtually every turn, the physicians signalledto each other regarding the sufficiency or insufficiencyof their unfolding routine. Coordinating was sometimesexplicit, as when an incoming physician explicitly askedfor more identification of a patient before moving on to

that patient’s medical history. By explicitly moderatingthe advancement of a handoff, the physicians indicatedthat their performance was insufficient and needed tobe improved before they were ready to move on. Coor-dinating was sometimes implicit, as when an incomingphysician remained silent and continued writing while theoutgoing transitioned from one move of their handoff(e.g., identification) to the next (e.g., past events). Bysilently proceeding with a handoff, physicians implicitlyshowed to each other that the prior performance wassufficient and that they were ready to move on.

Researchers with roots in ethnomethodology oftenexamine the openings of social and organizational activi-ties, because openings are usually rich with informationabout the participants and their ways of working together.When we examine the openings of our ICU handoffs,we see physicians doing patient identification—an initialmove that was made mutually intelligible through ongoingcoordinating. To illustrate, consider the opening of a hand-off that was conducted under rather chaotic conditions.When the physicians went to their handoff room, theydiscovered that it was already occupied, so they decidedto huddle at the nursing station (see Figure 1). They werestanding instead of sitting, using counter space instead ofa table, and talking over the noise of their surroundings.Despite the chaos, the opening identification of theirpatient went smoothly:1

Excerpt 1 (#1.002–0:00)1 Out: Okay, so umm bed number 2 mister Jack Jones,2 umm 78 years old guy, has been in the hospital3 for about a week now, came in most likely with4 [pneumonia] 0 0 05 ((Figure 1: outgoing watched incoming write))

This opening sequence is a textbook example of patientidentification: identifiers included the patient’s bed number(Line 1), name (Line 1), age (Line 2), gender (Line 2), anda differentiating fact about his admission to the hospital(Lines 2 and 3). What is less obvious—but somethingthat we emphasize—is the ongoing coordinating of thissequence. While the outgoing physician was talking, theincoming was writing it down, and the outgoing waswatching what the incoming was writing (see Figure 1).

Figure 1 (Color online)

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Within their chaotic environment, the physicians couldhave positioned themselves in a wide variety of ways, butthey adopted an embodied formation that enabled ongoingcoordinating, with the outgoing looking over the shoulderof the incoming physician to see what he was writing. Theoutgoing’s sequence of identifiers was not privately andmentally received by the incoming, but rather publicallyand materially registered through his writing. With eachidentifier, the outgoing spoke and the incoming wrote,thereby coordinating and coauthoring each part of theirunfolding sequence. Notice their seamless transition frompatient identification to the next move of the handoff:i.e., the patient’s past events (“[pneumonia]” Line 4).By jointly advancing to the next move, the physiciansshowed to each other—and to analysts who later watchedthe video—that the opening move (identification) wasaccomplished or sufficient, at least good enough to moveon. Although they made it look easy, the physicianswere actively coordinating at every turn: they positionedtheir bodies spatially so that they faced each other; theyattended to visible as well as verbal behavior; and theincoming physician did not object when the outgoingtransitioned to the next move in the sequence.

Sometimes coordinating was more pronounced. Some-times physicians spoke explicitly about the sufficiency orinsufficiency of their unfolding routine—in contrast to theincoming physician’s silent endorsements in Excerpt 1.For example, in the following excerpt the incoming physi-cian explicitly signaled that the opening move (patientidentification) was not sufficient:

Excerpt 2 (#3.015—7:26)1 Out: Next one is Tracy Adams ((Figure 2A: incoming looks2 down and takes notes)) she’s a woman who had a3 history of uh previous multiple MVCs [motor vehicle4 collisions] and so has a bit of a fusion and uh 0 0 05 In: ((Figure 2B: incoming stops writing and looks up))6 What bed is that?7 Out: Oh sorry, thirteen8 ((Figure 2C: incoming looks down and writes))9 In: How old is she?

10 Out: Thirty nine

At first, patient identification was going well. Theincoming was looking down and taking notes (see Fig-ure 2A) while the outgoing identified the patient: her


(A) (B) (C)

name (Line 1), gender (Line 2), and a differentiating factabout her motor vehicle collisions (Line 3). But then theoutgoing physician seemed to cross a line: he startedtalking about the patient’s fused vertebrae (Line 4), whichis more like past events than identification. Immediately,the incoming physician stopped writing and looked up(see Figure 2B), suspending the handoff activity andmarking the moment as problematic. Then the incomingexplicitly asked about the patient’s bed number (Line 6)and age (Line 8), which had not yet been identified.In this way, the incoming physician showed that theopening sequence was insufficient, that more identifi-cation was needed. The outgoing physician apologized(Line 7), acknowledging his mistake, and then answeredthe incoming’s questions (Lines 7 and 10). By returningto his writing (see Figure 2C), the incoming physiciantreated the outgoing’s response as sufficient and showedthat he was ready to move on.

Thus, the opening move in ICU handoffs (patient iden-tification) enabled the physicians to immediately “get onthe same page” about both their patient and their process.On one hand, patient identification provided information.The physicians clarified what patient they were talkingabout before delving into that patient’s story; and becausesome identifiers were medically relevant, the physicianstransferred important medical information, such as ageand gender, in the very act of identifying the patients. Onthe other hand, the move of patient identification primedthe physicians’ coordinating. Physicians had to immedi-ately show whether they were ready to proceed with ahandoff through their participation in the opening move.Identification unfolded sequentially (bed, name, gender,age, differentiating comment), which gave the participantsa series of opportunities to signal their involvement andreadiness to move on (or not).

Second, we show how sequence was a resource forphysicians in the flexible performance of their routine.Within a context of shared expectations, the physicianssometimes changed the sequence of moves in handoffs,in the service of a particular patient’s needs. Sometimesmoves were rearranged to foreground or emphasize crucialinformation; sometimes moves were entirely skipped toavoid redundant or superfluous information. We analyze

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.



(A) (B) (C)

video excerpts that demonstrate sequential variation, andwe highlight the coordinating work that made sequentialvariation mutually intelligible. Through explicit or implicitreference to their shared expectations, physicians bothdeviated from and at the same time instantiated theostensive pattern of their routine. Thus, sequence was anongoing resource because the difference between sequenceexpected and sequence enacted was made meaningful inthe moment of performance, giving physicians a betterpurchase on their patients’ situations.

To illustrate how the sequence of moves was mean-ingfully rearranged, consider the case of Mark Johnson,who was slowly dying. Mark was a 71-year-old manwho had been admitted to the ICU on four differentoccasions within the past month, and his care was com-plicated by pre-existing dementia that prevented himfrom communicating with his doctors and making gooddecisions. In his stead, decisions were being made byhis wife, a deeply religious person who was literallyhoping for a miracle and was refusing to “let go” of herdying husband. Increasingly, the wife was angry abouther husband and unhappy with his doctors. So when itcame time to handoff this particular patient, the outgoingphysician decided to foreground “family matters” as hisprimary concern. The interaction unfolded as follows:

Excerpt 3 (#3.013–5:25)1 Out: Next, in bed five is Mister Johnson2 In: Male, how old is he?3 Out: He’s seventy one4 In: Oh5 Out: And it’s really unfortunate, it’s uh6 ((Figure 3A: outgoing closes eyes, pinches nose))7 and maybe starting from the end8 ((Figure 3B: incoming stops writing and looks up))9 the wife, who’s the substitute decision maker for him10 ((Figure 3C: incoming looks down and resumes writing))11 is extremely angry 0 0 0

The opening sequence was familiar as the physiciansinteractively identified the patient by bed number (Line 1),gender (Line 1), name (Line 1), and age (Line 3). Butthen the outgoing physician did something unexpected:he relocated talk about the family from the end of theroutine, where it usually occurred, to the beginning of

the handoff, where it was foregrounded and emphasized.By rearranging the sequence of their handoff, the physi-cians demonstrated both the flexibility and sameness oftheir routine. On one hand, they were able to tailor theirperformance—in short order—to accommodate a difficultsituation. They shifted their focus from the patient tothe wife, who became a primary concern rather than anafterthought. On the other hand, the physicians sustainedthe usual sequence of their routine in the course of devi-ating from it. With the words “starting from the end”(Line 7), the outgoing physician explicitly explained thatand how the handoff sequence was being rearranged,which acknowledged and thereby instantiated the expecta-tions that they shared.

By coordinating their performance, the physiciansmanaged to move on with their handoff despite the rear-ranging of moves. First, the outgoing physician markedthe moment as unusual: rather than look across the tableor down at his notes as usual, the outgoing physicianmomentarily closed his eyes and pinched his nose, likesomeone concentrating or having a headache (see Fig-ure 3A). His preface “it’s really unfortunate” (Line 5)also put the incoming physician on notice that some-thing unusual was afoot. Second, the incoming physicianshowed that he was willing to play along. When the out-going said “starting from the end” (Line 7), the incomingmomentarily suspended his writing and looked up (seeFigure 3B), not immediately willing or able to continue.But soon he looked back down and resumed his writing(see Figure 3C), showing that he was ready to go on. Thus,sequence was a resource for the flexible performance ofroutines: whether physicians adhered to or deviated fromtheir shared expectations about handoffs, each perfor-mance was mutually intelligible through its relationship tothose shared expectations including sequence—explicitlyor implicitly referenced and thereby instantiated throughcoordinating.

In addition to rearranging moves in a sequence toemphasize information, the physicians also skipped movesto avoid unnecessary information. By coordinating theirskipping, physicians made their flexible performancemutually intelligible: they both attended to the needs ofparticular patients and sustained their shared expectations

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.



about (moves and sequence) in the course of deviatingfrom them. To illustrate, consider the case of GilbertFrank, a patient that the physicians knew well because hehad been in the ICU for almost a month. When it cametime for Frank’s handoff, his long and complicated historywas entirely skipped by the outgoing physician—but onlyafter the incoming physician recognized the patient. Withthe past events skipped, the handoff became efficient(only 40 seconds), unfolding as follows:

Excerpt 4 (#3.001–7:45)1 Out: Okay so Gilbert Frank, he’s in bed 9, um, he’s2 been here almost a month ((Figure 4: incoming3 starts nodding)) and uh4 In: I know him, yeah5 Out: Okay so major issue with him is weaning, and6 actually he’s making great progress in weaning7 In: Okay8 Out: Uh today he was going to 0 0 0

In the opening sequence, the outgoing physician identi-fied Frank by bed number (Line 1), name (Line 1), gender(Line 1), and the differentiating fact that he had been inthe ICU for almost a month (Line 2). At precisely thispoint—i.e., the transition point from identification to pastevents—the incoming physician showed that he recog-nized the patient. His coordinating was first visible as henodded in recognition (see Figure 4), and then audible ashe explicitly stated “I know him” (Line 4). By then skip-ping past events and going straight to current issues (“somajor issue 0 0 0 ” Line 5), the outgoing physician treatedthe past as unnecessary and the current as sufficient. By


(A) (B)

eventually talking about future plans (“today he was goingto” Line 8), the outgoing preserved the overall chronologyof the handoff, both deviating from and instantiating theparticipants’ shared expectations about their routine. Thus,their flexible performance of handoffs included skippinga move (e.g., past events), which enabled physiciansto focus on what they really needed to know, in theservice of a better handoff. Through coordinating, suchdeviations were made mutually intelligible in relation tothe participants’ shared expectations about moves andsequence. Again, sequence was a resource for flexibleperformance of their routine.

In the next excerpt, flexibility is even more pronouncedas the physicians skip two moves of their routine intailoring to the needs of a particular patient. Jane Stewartwas recovering quickly and would soon be dischargedfrom the ICU and transferred to a hospital ward, no longerin the care of ICU physicians. Under these circumstances,the physicians skipped both her past events and currentissues, going straight from patient identification to futureplans. It was unnecessary for the physicians to talk aboutStewart’s past and present, because she would no longerrequire ICU care and would soon be someone else’sresponsibility. The handoff unfolded as follows:

Excerpt 5 (#5.011–14:42)1 Out: Next door is Miss Jane Stewart2 ((Figure 5A: physicians on task, looking down))3 Do you remember her?4 In: I remember her5 Out: She’s up for transfer, she’s-6 In: Yes! ((Figure 5B: arms raised in celebration))

At the start of the handoff, both physicians lookeddown at their papers, with one of them reading and theother one writing, visibly on task (see Figure 5A). Theiropening identification sequence included bed number(Line 1), gender (Line 1), and name (Line 1), but then theoutgoing physician asked directly whether the incomingrecognized the patient: “Do you remember her?” (Line 3).As soon as the incoming confirmed recognition—i.e.,precisely at the transition point between identification andpast events—the outgoing physician skipped to futureplans and the patient’s imminent discharge from the ICU(“She’s up for transfer 0 0 0 ” Line 5). By skipping the

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


patient’s past events and current issues, the physiciansshowed that these were now unnecessary, no longer theirconcern. Instead of looking down at their papers, bothphysicians looked up at each other and raised their arms(see Figure 5B) in joint celebration about the patient’srecovery. Although their handoff was unusually brief andfocused (less than 30 seconds), it was sufficient for theirpurposes in this particular situation. The symmetry of theirvisible actions was a strong display of coordination—i.e.,both looking down and then up at each other, both handson papers and then up in the air. Thus, the sequentialvariation of routines can be extreme (e.g., skipping twomoves), as long as the participants are coordinating inrelation to their shared expectations about moves andsequence. We now turn to instances of repair, when onephysician deviates from shared expectations and the othertreats it as a breach.

Third, we show how the sufficiency of moves was negoti-ated in the flexible performance of handoffs. When thingsgo well, coordinating enables physicians to tailor their rou-tine to the needs of particular patients while also makingtheir performance mutually intelligible. But sometimes theflexible performance of routines miscarries. Sometimesphysicians, in the course of conducting a handoff, signalthat a move is insufficient and needs more work, or thata move is unexpected and needs to be changed. Coordi-nating became especially salient during transitions fromone move to the next, because one move was deemedsufficient as soon as the physicians moved on to the next.As long as a particular move was underway, its sufficiencywas pending: incoming physicians sometimes withheldquestions or complaints, giving the outgoing physicianan opportunity to fill gaps and fix errors; but as soonas outgoing physicians signaled that a transition to thenext move was imminent, incoming physicians respondedwith a flurry of questions or repairs. Our analysis showshow the sufficiency of moves was accomplished throughcoordinating.

Consider the case of Dale Elwin, a patient who had beenin the ICU for more than a month after a motor vehicleaccident (MVA). The patient’s problems were seriousand extensive, including a head injury (with intracranialbleeding), pulmonary effusion (fluid in the lungs), renal


(A) (B)

failure, and more. However, none of these problems werementioned by the outgoing physician—at least not initially.Rather, the outgoing physician skipped the patient’s pastevents to talk about current issues (possible bacterialinfection). Visibly confused, the incoming physicianinterrupted and asked about the patient’s past, therebyrepairing the sequence by returning to the move that wasskipped. The interaction unfolded as follows:

Excerpt 6 (#1.005–0:00)1 Out: First patient is Dale Elwin, and he’s a MVA accident,2 ((Figure 6A: incoming looks down and takes notes))3 uh, he was admitted on December 3, uh, he has, uh,4 been, this week the issues with him have been that5 ((Figure 6B: incoming looks up, stops taking notes))6 he has an ongoing infection and he’s being treated7 with Vancomycin, Clox 0 0 08 In: Okay just before we go any further, uh, can you just-9 just tell me his injuries? I think I was around when he

10 came in but, uh, I don’t recall him

Skipping the patient’s past events may be okay, aslong as the incoming physician recognizes the patient(e.g., see Excerpt 4). But that didn’t happen in this case.While the outgoing physician was identifying the patient(Lines 1 and 3), the incoming was taking notes (see Fig-ure 6A)—showing his participation in the handoff but nothis recognition of the patient. When the outgoing physicianskipped from patient identification to current issues (“thisweek the issues 0 0 0 ” Line 4), the incoming stopped writing(see Figure 6B) and eventually suspended the handoff(“before we go any further 0 0 0 ” Line 8) by requesting theinformation that was skipped (“tell me his injuries 0 0 0 ”Line 9). In this way, the sequence with skipping failedand had to be repaired. In this way, skipping the patient’spast events was treated as a breach of expectations: withthe words, “I don’t recall him” (Line 10) the incomingphysician indicated that the handoff was not intelligibleand that he needed the information that was skipped.

As noted above, transitions from one move in asequence to the next were an opportunity for physiciansto signal the sufficiency or insufficiency of their unfoldinghandoff. Coordinating is especially germane at transi-tions between moves, because displays of sufficiencyare both backward and forward looking in unfolding

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.



(A) (B)

sequences of moves. Moves are sufficient once they havebeen accomplished—at their end, not their middle—withparticipants ready to move on. To illustrate, considerthe handoff of Cindy Nelson, who was recovering wellfrom surgery despite losing a lot of blood: the outgoingphysician reported that she was awake, moving around,and quite comfortable. As soon as the outgoing turned totalk about the future—i.e., precisely at the transition pointbetween current issues and future plans—the incomingphysician interjected with a question about the patient’sbleeding. The interaction unfolded as follows:Excerpt 7 (#1.011–0:25)

1 Out: Blood pressure continues to be a problem even2 with her blood transfusions during surgery 0 0 03 ((45 seconds omitted: talk about infection))4 Out: She’s actually awake and moving around and, you5 know, she’s quite comfortable I think she’ll be6 ((Figure 7A: outgoing starts to turn the page))7 In: Why did she bleed so much?8 Out: Uh9 In: Was she on Coumadin or something?

10 Out: Uh yes I’m sorry I forgot to tell you that11 ((Figure 7B: outgoing stops turning the page))

When the outgoing physician identified blood pressureas one of the patient’s current issues (lines 1 and 2),he did not account for her excessive bleeding. So whenhe started to talk about future plans (“I think she’llbe 0 0 0 ” Line 5) and when he began to literally turnthe page on this patient (see Figure 7A), the incomingphysician interrupted and tried to repair the prior move.Through her questions about bleeding (Lines 7 and 9), theincoming physician showed that the prior move (currentissues) was insufficient and that she was not ready tomove on. Immediately, the outgoing physician apologized(“I’m sorry 0 0 0 ” Line 10), acknowledged his mistake(“I forgot 0 0 0 ” Line 10), and released the page that he wasturning (see Figure 7B). Eventually, the questions relatedto bleeding were answered, the move (current issues) wastreated as sufficient, and both physicians moved on to thenext move in the sequence (future plans).

Researchers with roots in ethnomethodology oftenexamine the closings of social and organizational activitiesbecause coordinating is especially conspicuous duringsuch transitions. When we examine the closings of ICUhandoffs, we see that the physicians didn’t just stop

talking. Rather, they incrementally marked their handoffsas sufficient through nuanced utterances such as “alright”or “okay” and through small gestures or shifts in bodyorientation. Sometimes closings appeared easy, as whenboth physicians showed that the handoff was sufficientand that they were ready to move on. Other closingsinvolved more work, as when one physician made a bidto close while the other resisted and pursued some otherline of action, such as more information.

To illustrate, consider the handoff of Tyron Dickson,who was dying. After a lengthy discussion of his pastevents and current issues, the patient’s future prospectsseemed bleak: key indicators showed that Dickson’sorgans were failing and the physicians admitted that theywere running out of options. Under these circumstances,closing the handoff was a delicate matter because thepatient’s medical problems remained unresolved. Practi-cally, closing the handoff closed off further discussion ofoptions. Their closing was coordinated and incrementallyaccomplished through both audible and visible displays,with each turn by one physician creating an opportunityfor the other to resist closing and pursue some new lineof action or medical care. Eventually, the physiciansinteractively showed that their handoff was sufficient andthat they were ready to move on to the next patient. Themoment unfolded as follows:Excerpt 8 (#1.017–3:40)

1 Out: His kidney function is declining2 In: His lactate is up?3 Out: Yeah …it’s seven and a half [so] it’s up4 ((Figure 8A: outgoing looks down at patient list))5 In: Yeah6 Out: Yeah7 ((Figure 8B: incoming looks down at patient list))8 Out: That’s about it9 In: Okay

10 Out: Okay11 ((Figure 8C: outgoing points at next patient listed))12 Out: Next is Carl Simmons, he’s in bed 4 0 0 0

Having already talked about family matters, which wasthe last move of their expected sequence, the physiciansrevisited a couple of critical topics in closing (Lines 1and 3). Notice the symmetry of actions that both accom-plished and signaled the physicians’ coordinating aboutclosing. They symmetrically agreed that their patient was

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.



(A) (B) (C)

failing, both saying “Yeah” in quick succession (Lines 5and 6). They visibly coordinated their attention to apatient list: first the outgoing physician looked down(see Figure 8A) and then the incoming (see Figure 8B).By looking toward their list, the physicians collaboratedin a silent bid to close; the list is where they alwayslooked before starting a new handoff because that’s wherethey found the next patient’s name. Their bid for closingbecame more pronounced when the outgoing physiciansaid “That’s about it” (Line 8). His words were a claimof sufficiency, an assertion that their activity was nearlycomplete, which made closing their handoff immediatelyrelevant. Symmetry in action was also demonstrated bythe final utterances of this handoff: their “Okay” sequence(Lines 9 and 10) was information poor but coordinationrich, a joint display that the handoff was sufficient andthat the physicians were ready to move on. Immediately,the outgoing physician leaned forward and put his penon the list and the next patient’s name (see Figure 8C),“Carl Simmons” (Line 12).

In our final excerpt, we highlight the challenge ofachieving sufficiency at the closing of some handoffs,especially when handoffs are performed in succession.The patient was Mr. Bronson who had been in a seriouscar accident: his injuries included fractured ribs, whichmade it difficult for him to breathe, leading to pneumoniaand then septic shock. After a long conversation aboutBronson’s past events, current issues, and future plans,the physicians began to close the handoff without talkingabout family matters, which is the last move in their


(A) (B)

expected sequence. When the outgoing physician movedon and started talking about the next patient, the incomingraised a question about Bronson’s family, which createdtemporary confusion as the physicians were talking abouttwo patients at the same time. Their coordinating anddisplays of sufficiency involved visible behaviors such aspage turns, transcribed as follows:

Excerpt 9 (#2.007–38:00)1 In: Feeding okay?2 ((Figure 9A: incoming turns the page))3 Out: Yep fine4 ((Figure 9B: outgoing turns the page))5 In: Good6 Out: Next is Mary Johnson7 In: And the family? Oh Mary is still here8 Out: For Mister Bronson?9 In: Yeah

10 Out: Yeah I haven’t seen a lot of his family to be honest with you11 In: So Mary is still here12 Out: Yeah so Mary is the next patient in bed 8

The closing of Bronson’s handoff is somewhat bungled.After asking about his “feeding” (Line 1), the incomingphysician turned a page—she literally turned her attentionto the next patient (see Figure 9A). In a symmetry ofaction, the outgoing physician answered the questionabout feeding (Line 3) and then he also turned his page tothe next patient (see Figure 9B). Following these strongdisplays of sufficiency, with coordinating both audibleand visible, the outgoing physician moved on to the nextpatient, “Mary Johnson” (Line 6). Then the trouble started:

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


the incoming physician asked about Bronson’s family(“And the family?” Line 7), within an utterance that alsoacknowledged the new patient (“Mary” Line 7). Thus, thephysicians were simultaneously talking about two differentpatients, while performing two different moves—i.e.,family matters and patient identification. The confusionwas interactively resolved as the physicians clarified whichpatient they were talking about (“Bronson?” Line 8), andthen they addressed his family matters (Line 10) beforejointly proceeding to the next patient (“Mary” Lines 11and 12). Thus, coordinating was especially salient at theclosing of handoffs because to negotiate the sufficiencyof the final move was to also negotiate the sufficiency ofthe entire sequence. And when routines are performedin succession, participants must coordinate to avoid anoverlap of closings and openings.

In summary, part two of our research findings expandsour ethnomethodological approach as we employ con-versation analytic methods to examine coordinating inthe flexible performance of routines captured in videorecordings. Our data show that flexible performanceunfolds within a context of ongoing coordination. Weexamine both audible and visible forms of coordinatingthat become especially salient when participants vary thesequence of moves within their routine, and when theymake transitions between moves of their sequence. In ourpresentation and analysis of data, we have deliberatelypreserved the meaningful entanglement of the partici-pants’ behaviors, especially the displays of sufficiency orinsufficiency that enable participants to get on with theirshared project of caring for patients. When a breach ofthe participants’ shared expectations occurs, a variety ofvisible and audible behaviors may be used to signal thatsomething is problematic. In the course of conducting

Figure 10 (Color online) Coordinating During the Flexible Performance of Routines

handoffs, physicians may signal that a move is insufficientand needs more work, or that a sequence is insufficientand needs to be changed.

Process SummaryOur findings, which are graphically represented in Fig-ure 10, show that the flexible performance of handoffroutines involves a confluence of expectations and circum-stances. Participants come to the routine with overlappingexpectations about the moves and sequence that willconstitute their performance, and their performance istailored to the particular circumstances that exist at themoment of enactment. Through ongoing coordinating, theparticipants reconcile the difference between what theyexpect and what they enact.

We have identified three forms of ongoing coordinatingin the flexible performance of routines. First, participantsproject future action, indicating when they will conformor deviate from expectations. This enables the otherparticipant to distinguish between deviations as mistakesneeding to be repaired, and intentional deviations inservice of the situation at hand. Second, participantsdisplay to each other the sufficiency or insufficiency ofthe unfolding performance. When one participant deviatesfrom expectations, the other can signal whether or notthat deviation is mutually intelligible, through variousaudible and visible displays. Third, when one participantsignals that something is wrong, showing that a deviationfrom expectations is not mutually intelligible, then theother participant can either account for the prior action orrepair it as a breach of expectations.

Within this context of ongoing coordinating, participantsfurther manage the flexible performance of their routine

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


in two distinct ways. First, they alter the sequence oftheir routine by rearranging or skipping select moves.Sequential variation enables participants to manage therelative importance of moves, according to their particularsituation: by bringing a move forward, the participantsmay emphasize its importance; and by skipping a movealtogether, participants may show that it’s irrelevant orsuperfluous. In this way, sequence is a resource for action:whether the participants deviate from expectations or not,the sequence enacted is meaningful through its relationshipto the sequence expected.

Second, the participants alter the constitution of movesas they negotiate their sufficiency at points of transition.While a move is underway or in the process of being per-formed, a participant may withhold questions and suspendrepair because answers and fixes may be forthcoming.But as soon as one participant indicates that a transitionis imminent, the other participant may resist or protestthat the present move is not yet sufficient. Coordinatingis especially salient during transitions between moves,because a particular move is made “sufficient” when theparticipants move on to the next.

Thus, moves and sequence have a reflexive relationshipduring the flexible performance of routines. Althougheach move of the handoff makes the next move relevant,the sequential unfolding of the routine makes thosemoves sufficient. To signal that a particular move wassufficient, participants move on to the next; and by movingon to the next move, participants show that the priormove is sufficient. The participants’ explicit and implicitcoordinating provided ongoing answers to the questions:Is our performance sufficient (or not)? Is our unfoldingroutine mutually intelligible (or not)? Are we ready tomove on (or not)?

DiscussionOur study examines how participants coordinate theiractions—flexibly yet intelligibly—during the performanceof routines (see Figure 10 for summary). Drawing onroutine dynamics and ethnomethodology to analyze inter-view and video data, we uncover both the expectationsthat participants bring with them to the performanceof the routine and how participants use those expecta-tions during performance. Specifically, we show howICU physicians use the sequential features of the hand-off routine—the expected moves of the routine and theexpected sequence in which those moves occur—as theyadapt their performance of the routine to address theparticular circumstances of each patient. Our analysisreveals the ongoing work of coordinating that is requiredto adjust the performance of the routine to maintainmutual intelligibility, especially as the participants managesequential variation in the routine and negotiate moves atpoints of transition. Our findings make two important con-tributions to the literature: (1) they identify the need forongoing coordinating despite a strongly shared ostensive

pattern and (2) they illustrate how participants use thesequential nature of the ostensive pattern of the routine asa resource for flexible performance. We elaborate thesecontributions below.

Ongoing Coordinating in Flexible PerformanceOur study complements and extends work by Dionysiouand Tsoukas (2013), who theorized that a role-takingperspective and repeated performances of a routine shouldlead participants to develop increasingly shared schemata,which, in turn, should lead to a reduced need for coordi-nation during performance. In our context, both of thesefactors—repeated performance and role-taking—werepresent. On average, the ICU physicians in our studyworked at the CH ICU for 10 years, and, in that time, theyperformed thousands of handoffs. Both participants in thehandoff routine were ICU physicians, which facilitatedtaking the perspective of the other participant. Moreover,they literally took on each other’s roles: the incomingphysician for one handoff was the outgoing physician forthe next handoff. Our interview data demonstrated thatICU physicians had largely shared expectations about thehandoff, both in terms of what the routine involved (theexpected moves in the routine and the expected sequenceof those moves) and how the routine would be conducted(tailored to the need of each patient). Yet, despite thisstrongly shared ostensive pattern, we found that the ICUphysicians engaged in significant ongoing work of coordi-nating to develop the “joint situated understandings” and“aligned actions” described by Dionysiou and Tsoukas(2013). Our study reveals that whatever shared schematawere in play, they did not preclude the need for ongoingcoordinating during flexible performance.

Our finding that ongoing coordinating occurred evenwhen participants had a strongly shared ostensive alsoraises questions about the relationship between cognitionand action in the broader coordination literature. Muchof the coordination literature related to interdependentwork, not just in the performance of routines but also inthe teams and group performance literature, highlightsthe importance of shared understanding (Okhuysen andBechky 2009). For many coordination researchers, theunderlying assumption is that if people who are engagedin interdependent work understand a situation in a sharedenough way, then coordinated action follows. This isparticularly true in the teams and group literature, wherethere is a large body of research linking shared cognitionin teams (for instance, shared schema, shared mentalmodels, team mental models, and team situation models)and more effective team performance (Cannon-Bowersand Salas 2001, DeChurch and Mesmer-Magnus 2010,Marks et al. 2000, Mathieu et al. 2000, Mohammed et al.2010, Rico et al. 2008). However, our study illustratesthat shared understanding by itself was insufficient, andredirects attention to the role of action for coordinatinginterdependent work.

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


With respect to action, Dionysiou and Tsoukas (2013)highlight the need for further examination of how partici-pants interrelate their individual lines of action during thecollective performance of routines. Our study shows howparticipants calibrated their performance to create, main-tain, and repair mutual intelligibility during the flexibleperformance of routines. Our ethnomethodological lensfocuses attention on the emergent, situated, and embodiednature of ongoing coordinating. We show how ongoingcoordinating was made audible and visible through mul-timodal displays, such as when participants physicallyco-oriented toward each other, engaged with physicalartifacts, and used talk, gaze, and gesture to calibrate theirperformance. At first glance, the performative actionsthat our ethnomethodological approach identifies—forinstance, stopping writing, looking up, asking a question,or turning a page—may seem inconsequential, but ourfindings illustrate that the way in which participantscoordinate just these types of actions is what enabled theaccomplishment of mutual intelligibility during flexibleperformance.

By unpacking the ways in which ongoing coordinatingwas accomplished, we contribute to a more nuancedunderstanding of the relationship between repair and theflexible performance of routines. Previous studies haveshown that people engage in repair during the performanceof routines when truces collapse (Zbaracki and Bergen2010) or when the expected sequence of the routine isdisrupted (Turner and Rindova 2012). Our study goesfurther and illustrates that breaches and repairs are partof the ongoing coordinating that occurs in the everydayperformance of routines. In contrast to Turner and Rindova(2012), we found that repair was sometimes requiredeven when the sequence of the routine was unfolding asexpected (for instance, the problematic transition frompatient ID to past events in Excerpt 2, or the problematictransition from past events to current issues in Excerpt 6).Conversely, we showed other examples where no repairwas necessary, even when the enacted sequence of theroutine varied significantly from the expected sequence(for instance, skipping from patient ID to future planswithout discussing past events and current issues inExcept 5). Thus, we demonstrate that lapses in mutualintelligibility (not necessarily deviations from expectedsequence) are occasions for repair.

In contrast to previous research about repair whenthe performance of the routine breaks down completely(Zbaracki and Bergen 2010), our findings show howparticipants are able to engage in ongoing repair as partof flexible performance. When the mutual intelligibility ofhandoffs decreased, physicians gave indications to eachother that something was wrong, and each indication wasan opportunity to make adjustments while still movingforward. An early and subtle indication that somethingwas wrong was when incoming physicians stopped writingand looked up (e.g., Excerpts 2, 3, and 6). Because such

behavior was silent, it did not stop people from talkingand continuing with the performance of their routine,but it did signal that the performance had become lessintelligible and legible. A more pronounced cue for repairwas to ask a question (e.g., Excerpts 2, 3, 5, 6, 7, 8, and 9),which explicitly signaled that something was wrong orthat more information was needed. Some questions wereespecially constructive, a form of scaffolding for thecompletion of a move underway. For example, considerthe incoming physician’s performance during patientidentification in Excerpt 2: he stopped writing and lookedup before asking a couple of questions (“What bed isthat?” and “How old is she?”) that told the outgoingprecisely what to say and when, in this way pushing theirperformance forward. Some questions were relativelyinterruptive, such as the incoming physician’s behavior inExcerpt 6: he stopped writing and looked up before askingthe outgoing to suspend forward progress (“before we goany further”) and repair a previous move (“Can you justtell me his injuries?”), which was a step backward but,at the same time, enabled the performance to continue.Thus, we show that ongoing coordinating provides forongoing repair, which is unfolding and even escalating, inthe service of moving forward with a common project.

Sequence as a Resource in Flexible PerformanceOur findings show that the physicians’ shared expectationsabout handoffs were a resource for their enactment of theroutine. Consider sequence specifically: the sequentialfeatures of their routine’s ostensive pattern—i.e., expectedmoves in an expected sequence—enabled the mutualintelligibility of their flexible performance. The mostobvious demonstration of sequence as a resource waswhen the physicians enacted exactly what they expected.In Excerpt 1, for example, the physicians accomplished thefirst move of their routine (patient identification) beforemaking a transition to the next move (past events). Thissequence was consistent with participants’ expectationsand the handoff proceeded without a hitch, illustrating thata sequential performance may be especially intelligiblewhen it unfolds as expected. However, sequence is aresource even when participants deviate from expectations.In Excerpt 6, the outgoing physician tried to skip a movein the handoff sequence, going from patient identificationto current issues rather than past events. Instead of justgoing along, the incoming physician protested and askedthe outgoing physician to talk about past events—a requestthat was mutually intelligible in relation to the physicians’shared expectations about sequence. A striking example isExcerpt 3, when the outgoing physician began the handoffby explicitly “starting from the end” of their expectedsequence, which both deviated from and instantiatedthe routine’s ostensive pattern. Thus, we found thatthe expected sequence was important not only whenit was enacted but also when it was not. Feldman andPentland (2003) explain that the ostensive aspects of

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


the routine helps participants to guide, account for, andrefer to the performative aspect, which, in turn, helps tocreate, maintain, and modify the ostensive aspects. Ourperformance data reveal that when participants deviatedfrom the expected sequence, they also instantiated it,either explicitly or implicitly.

Sequence is also a resource for negotiating movesin the flexible performance of routines. Our findingsshow how participants’ strongly shared expectations aboutthe sequence of their routine helped them to negotiatethe sufficiency of moves at point of transition. In oursetting, participants expected that routines would unfoldin a linear fashion as participants completed one movebefore going onto the next: handoffs were the story ofthe patient, progressing from past events to current issuesto future plans. Displays of sufficiency and insufficiency,as well as bids for repair, became especially salient atmoments of transition between moves, as participantssignaled their readiness to move on (or not). Displays ofsufficiency were often in the form of silent assent, butalso included audible displays. For instance, in Excerpt 4,the outgoing physician was able to transition from patientID to past events to current issues without interruption,as the incoming nodded, acknowledged (“I know him,yeah”), and responded with the word “okay.” Displays ofinsufficiency indicated that more work was needed or thatsomething had to be repaired before the handoff couldadvance to the next move. In Excerpt 7, the incomingphysician did not ask her question—“Why did she bleedso much?”—until the moment of transition, when itbecame clear that the outgoing physician was movingfrom past events to current issues without explaining thebleeding mentioned earlier. Thus, the participants’ sharedexpectations about the sequence of their routine helpedthem to know whether and when a move was sufficient orneeded to be repaired. Our findings complement Pentland’s(1992) notion of moves by showing how participantsnegotiate the accomplishment of moves in relation totheir unfolding sequence. By foregrounding the transitionsbetween moves in flexible performance, we show how the“building blocks” of routines may be contested by theparticipants. Put another way, a move in a routine is notdone until the participants jointly display that it is done.

In summary, our ethnomethodological approach empha-sizes the participants’ ongoing role in managing thesufficiency of moves, their sequential variation, and repairduring the flexible performance of routines. Regardingmutual intelligibility as something that participants per-form and display helps to deepen our appreciation ofroutines as effortful accomplishments: participants notonly assemble a sequence of moves from a repertoire ofpossible moves (Pentland and Rueter 1994), but they alsoengage in real-time adjustment to make the enactment ofthose moves mutually intelligible. Ethnomethodology alsoallows us to observe how the participants are making thesequence of the routine (ostensive pattern) relevant to each

other during performance (Becker 2005, Pentland et al.2010). Moreover, our attention to the sufficiency of moveswithin handoffs is part of a larger conversation about thesufficiency of action within social activity. Sufficiency ofaction, presented here as an ethnomethodological phe-nomenon, enriches how we think about coordinating, notjust in the context of routines. It is important to highlightthat sufficiency of action is an emic concept; that is,sufficiency of action is not something for the researcherto determine, but is something that participants display toeach other in the moment of enactment. By foreground-ing issues of sufficiency, we gain insight into anotherway of coordinating—not through shared cognition butthrough interrelated actions, whereby participants createthe conditions to move forward with a common project.

Boundary ConditionsAlthough we found that our participants engaged in signif-icant work of coordinating as they performed the handoff,other contexts may require even more coordinating. Thehandoff routine was fairly straightforward: there wereonly two participants, colocated, engaged in performancethat lasted for minutes. In contrast to other routines,which may be distributed across multiple actors in multi-ple locations, both of our participants performed in theentire routine. In addition, the ICU physicians usuallyconducted handoffs face-to-face, which gave them accessto both visible and audible behaviors. Our physicianswere expert practitioners and, since both were ICU physi-cians, some of the issues related to coordination acrossoccupational boundaries are not likely to apply in ourstudy (Bechky 2003). Also, our physicians were highlymotivated to work together in advancing their commonproject of patient care, whereas participants in other rou-tines may have different or even conflicting motivations(Zbaracki and Bergen 2010), which could complicatecoordination.

ConclusionOur ethnomethodological approach to the study of organi-zational routines brings coordinating to the foreground.Through coordinating, routines can be both flexibly per-formed and mutually intelligible: against a backdrop ofshared expectations about how their routine is enacted,participants tailor the moves and sequence of their per-formance to fit the situation at hand. Our study makestwo interrelated contributions to research on the flexibleperformance of routines. First, we show how coordinatingis ongoing, even when the participants have a stronglyshared ostensive pattern. Continually, the participantsdisplay the sufficiency or insufficiency of their unfoldingroutine, through various audible and visible behaviors thateither move their common project forward or suspend itsprogress if something needs to change. Second, we showhow sequence is a resource in the flexible performance ofroutines. The participants’ shared expectations about the

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


sequence of moves enables their performance to be mutu-ally intelligible: whether they enact shared expectationsor deviate from them, they perform with regard to theirshared expectations and thereby instantiate their sharedostensive pattern with each performance. Moreover, sharedexpectations about sequence help participants to recognizewhen transitions are underway, which creates opportunitiesto signal the sufficiency or insufficiency of moves, throughespecially salient forms of coordinating such as questionsand repair. Thus, moves and sequence have a reflexiverelationship during the flexible performance of routines.Each move makes relevant the next in the sequence, andthe sequence enacted renders each move sufficient.

AcknowledgmentsWe dedicate this paper to our dear friend and colleague MichaelCohen (March 22, 1945–February 2, 2013), a founding memberof our handoffs research team, whose ideas and passion continueto inspire our collaboration. The authors thank our editor,Martha Feldman, for all of her support and guidance throughoutthe process. The authors also thank four anonymous reviewersat Organization Science and Michelle Barton, Katy DeCelles,Emily Heaphy, Brian Hilligoss, Sarah Kaplan, Claus Rerup,JP Stephens, and Kathie Sutcliffe for helpful comments onearlier drafts. The authors are grateful for comments fromparticipants at events where previous versions of this paper werepresented, including the 4th International Symposium on ProcessOrganization Studies (2012), Academy of Management AnnualMeeting (2012), University of Sydney Business School (2013),29th EGOS Colloquium (2013), 20th Annual OrganizationScience Winter Conference (2014), Laguna Conference (2014),6th International Symposium on Process Organization Studies(2014), and the Wharton EOI Workshop (2015). Curtis LeBaronand Marlys K. Christianson share lead authorship.

Appendix A. Semistructured Interview Protocol1. When you participate in an ICU handoff, what are yourobjectives?

Are your objectives different if you are the incoming versusthe outgoing participant?

2. How do you get ready to do a handoff?What materials do you bring into the handoff room?When you are the outgoing physician, do you use anydocuments or notes?When you are the incoming physician, do you use anydocuments or notes?

3. Could you please walk us through a typical handoff?4. Handoffs are often complicated with a lot of variety, so whathappens in a handoff when it goes well?

Can you think of a recent example of a handoff that wentwell?

5. What happens in a handoff when it doesn’t go as well?Can you think of a recent example of a handoff thatdidn’t go so well?

6. How do you organize information during a handoff?

Endnotes1We use pseudonyms for patient names in all the excerpts.

References

Atkinson M, Heritage J (1984) Structures of Social Action: Studies inConversation Analysis (Cambridge University Press, Cambridge,UK).

Austin JL (1962) How to Do Things with Words, 2nd ed. (HarvardUniversity Press, Cambridge, MA).

Bechky BA (2003) Sharing meaning across occupational communities:The transformation of understanding on a production floor. Organ.Sci. 14(3):312–330.

Becker MC (2005) The concept of routines: Some clarifications.Cambridge J. Econom. 29(2):249–262.

Birnholtz JP, Cohen MD, Hoch SV (2007) Organizational charac-ter: On the regeneration of Camp Poplar Grove. Organ. Sci.18(2):315–332.

Blumer H (1954) What is wrong with social theory? Amer. Sociol. Rev.19(1):3–10.

Blumer H (1969) Symbolic Interactionism: Perspective and Method(Prentice Hall, Englewood Cliffs, NJ).

Blumer H (2004) George Herbert Mead and Human Conduct (AltaMira Press, Oxford, UK).

Boden D (1994) The Business of Talk: Organizations in Action (Polity,Cambridge, UK).

Cannon-Bowers JA, Salas E (2001) Reflections on shared cognition.J. Organ. Behav. 22(2):195–202.

Cohen MD (2007) Reading Dewey: Reflections on the study of routine.Organ. Stud. 28(5):773–786.

Cohen MD, Hilligoss B (2010) The published literature on handoffs inhospitals: Deficiencies identified in an extensive review. QualitySafety Health Care 19(6):493–497.

Cohen MD, Ilan R, Garrett L, LeBaron C, Christianson MK (2012)The earlier the longer: Disproportionate time allocated to patientsdiscussed early in attending physician handoff sessions. ArchivesInternal Medicine 172(22):1762–1764.

D’Aveni RA (1994) Hypercompetition: Managing the Dynamics ofStrategic Maneuvering (Free Press, New York).

David M, Sutton C (2011) Social Research: An Introduction (SagePublications, London).

DeChurch LA, Mesmer-Magnus JR (2010) The cognitive underpin-nings of effective teamwork: A meta-analysis. J. Appl. Psych.95(1):32–53.

Dionysiou DD, Tsoukas H (2013) Understanding the (re)creation ofroutines from within: A symbolic interactions perspective. Acad.Management Rev. 38(2):181–205.

Drew P, Heritage J (1992) Talk at Work: Interaction in InstitutionalSettings (Cambridge University Press, Cambridge, UK).

Eisenhardt KM (1989) Making fast strategic decisions in high-velocityenvironments. Acad. Management J. 32(3):543–576.

Faraj S, Xiao Y (2006) Coordination in fast-response organizations.Management Sci. 52(8):1155–1169.

Feldman MS (2000) Organizational routines as a source of continuouschange. Organ. Sci. 11(6):611–629.

Feldman MS, Pentland BT (2003) Reconceptualizing organizationalroutines as a source of flexibility and change. Admin. Sci. Quart.48(1):94–118.

Feldman MS, Rafaeli A (2002) Organizational routines as sources of con-nections and understandings. J. Management Stud. 39(3):309–331.

Garfinkel H (1967) Studies in Ethnomethodology (Prentice-Hall, Engle-wood Cliffs, NJ).

Goffman E (1981) Forms of Talk (University of Philadelphia Press,Philadelphia).

Golden-Biddle K, Locke K (2007) Composing Qualitative Research(Sage Publications, Thousand Oaks, CA).

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Heath C, Hindmarsh J, Luff P (2010) Video in Qualitative Research:Analyzing Social Interaction in Everyday Life (Sage, London).

Heritage J (1984) Garfinkel and Ethnomethodology (Polity Press,Oxford, UK).

Hilligoss B, Cohen MD (2011) Hospital handoffs as multifunctionalsituated routines: Implications for researchers and administrators.Blair JD, Fottler MD, eds. Bienniel Review of Health CareManagement, Advances in Health Care Management, Vol. 11(Emerald Group Publishing Limited, Bingley, UK), 91–132.

Howard-Grenville JA (2005) The persistence of flexible organizationalroutines: The role of agency and organizational context. Organ.Sci. 16(6):618–636.

Jarzabkowski PA, Lê JK, Feldman MS (2012) Toward a theory ofcoordinating: Creating coordinating mechanisms in practice. Organ.Sci. 23(4):907–927.

Jick TD (1979) Mixing qualitative and quantitative methods: Triangula-tion in action. Admin. Sci. Quart. 24(4):602–611.

Joas H (1997) G.H. Mead: A Contemporary Re-examination of HisThought (MIT Press, Boston).

Lee TW (1999) Using Qualitative Methods in Organizational Research(Sage Publications, Thousand Oaks, CA).

Lincoln YS, Guba EG (1985) Naturalistic Inquiry (Sage Publications,Newbury Park, CA).

Marks MA, Zaccaro SJ, Mathieu JE (2000) Performance implications ofleader briefings and team-interaction training for team adaptationto novel environments. J. Appl. Psych. 85(6):971–986.

Mathieu JE, Heffner TS, Goodwin GF, Salas E, Cannon-Bowers JA(2000) The influence of shared mental models on team processand performance. J. Appl. Psych. 85(2):273–283.

Mead GH (1934) Mind, Self, and Society from the Standpoint of aSocial Behaviorist (University of Chicago Press, Chicago).

Mead GH (1938) The Philosophy oí the Act (University of ChicagoPress, Chicago).

Miller DL (1973) George Herbert Mead: Self, Language, and theWorld (University of Chicago Press, Chicago).

Mohammed S, Ferzandi L, Hamilton K (2010) Metaphor no more: A15-year review of the team mental model construct. J. Management36(4):876–910.

Mumaw R, Roth E, Vicente K, Burns C (2000) There is more tomonitoring a nuclear power plant than meets the eye. HumanFactors 42(1):36–55.

Okhuysen GA, Bechky BA (2009) Coordination in organizations: Anintegrative perspective. Acad. Management Ann. 3(1):463–502.

Orlikowski WJ (2007) Sociomaterial practices: Exploring technology atwork. Organ. Stud. 28(9):1435–1448.

Orlikowski W, Scott SV (2008) Sociomateriality: Challenging the sepa-ration of technology, work and organization. Acad. ManagementAnn. 2(1):433–474.

Patterson ES, Roth EM, Woods DD, Chow R, Gomes JO (2004)Handoff strategies in settings with high consequences for failure:Lessons for health care operations. Internat. J. Quality HealthCare 16(2):125–132.

Pentland B (1992) Organizing moves in software support hot lines.Admin. Sci. Quart. 37(4):527–548.

Pentland BT, Rueter HH (1994) Organizational routines as grammarsof action. Admin. Sci. Quart. 39(3):484–510.

Pentland BT, Hærem T, Hillison D (2010) Comparing organizational rou-tines as recurrent patterns of action. Organ. Stud. 31(7):917–940.

Pentland BT, Hærem T, Hillison D (2011) The (n)ever-changingworld: Stability and change in organizational routines. Organ. Sci.22(6):1369–1383.

Pomerantz A, Fehr BJ (1997) Conversation analysis: An approach tothe study of social action as sense making practices. van DijkTA, ed. Discourse as Social Interaction—Discourse Studies: AMultidisciplinary Introduction, Vol. 2 (Sage, London), 64–91.

Rawls AW (2006) Respecifying the study of social order—Garfinkel’stransition from theoretical conceptualization to practices in details.Garfinkel H, ed. Seeing Sociologically: The Routine Grounds ofSocial Action (Paradigm Publishers, Boulder, CO), 1–98.

Rawls AW (2008) Harold Garfinkel, ethomethodology and workplacestudies. Organ. Stud. 29(5):701–732.

Rerup C, Feldman MS (2011) Routines as a source of change inorganizational schemata: The role of trial-and-error learning. Acad.Management J. 54(3):577–610.

Rico R, Sanchez-Manzanares M, Gil F, Gibson C (2008) Team implicitcoordination processes: A team knowledge-based approach. Acad.Management Rev. 33(1):163–184.

Sacks H (1992) Lectures on Conversations (Blackwell, Oxford, UK).Schegloff E (1982) Discourse as an interactional achievement: Some

uses of “uh-huh” and other things that come between sentences.Tannen D, ed. Georgetown University Round Table on Languageand Linguistics 1981 (Georgetown University Press, Washington,DC), 71–93.

Silverman D (1998) Harvey Sacks: Social Science and ConversationAnalysis (Polity Press, Cambridge, UK).

Silverman D, ed. (2013) Instances or sequences? A Very Short,Fairly Interesting and Reasonably Cheap Book About QualitativeResearch (Sage Publications, Thousand Oaks, CA), 56–85.

Solet DJ, Norvell JM, Rutan GH, Frankel RM (2005) Lost in translation:Challenges and opportunities in physician-to-physician communi-cation during patient handoffs. Acad. Medicine 80(12):1094–1099.

Streeck J, Goodwin C, LeBaron C (2011) Embodied Interaction:Language and Body in the Material World (Cambridge UniversityPress, Cambridge, England).

Sutcliffe KM, Lewton E, Rosenthal MM (2004) Communication failures:An insidious contributor to medical mishaps. Acad. Medicine79(2):186–194.

ten Have P (2004) Understanding Qualitative Research and Eth-nomethodology (Sage Publications, London).

Turner SF, Rindova V (2012) A balancing act: How organizationspursue consistency in routine functioning in the face of ongoingchange. Organ. Sci. 23(1):24–46.

Weick KE (1979) The Social Psychology of Organizing, 2nd ed.(McGraw-Hill, New York).

Weick KE (1995) Sensemaking in Organizations (Sage Publications,Thousand Oaks, CA).

Weick KE, Sutcliffe KM, Obstfeld D (1999) Organizing for high relia-bility: Processes of collective mindfulness. Sutton RI, Staw BM,eds. Research in Organizational Behavior (JAI Press, Greenwich,CT), 81–123.

Zbaracki MJ, Bergen M (2010) When truces collapse: A longitudinalstudy of price-adjustment routines. Organ. Sci. 21(5):955–972.

Curtis LeBaron is associate professor and Warren JonesFellow of Organizational Leadership and Strategy at the MarriottSchool of Management. He received his Ph.D. from the Univer-sity of Texas at Austin. He conducts video-based research thatemphasizes the embodied and material aspects of organizationalwork and activity: topics include identity, knowledge work, andstrategy as practice.

Marlys K. Christianson is assistant professor at the RotmanSchool of Management. She received her Ph.D. from the Uni-versity of Michigan. Her research interests include senesmaking,the coordination of complex and interdependent work, and

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


resilient organizing, particularly as it relates to error detectionand correction.

Lyndon Garrett is a doctoral candidate of managementand organizations at the Stephen M. Ross School of Business,University of Michigan. His research interests include relationalprocesses of interpersonal and group bonding, high qualityconnections, thriving, and meaningful work, drawing primarilyon qualitative methods.

Roy Ilan is associate professor of medicine at Queen’sUniversity, and practices critical care and internal medicine atKingston General Hospital, Kingston, Ontario, Canada. Hisresearch has focused on patient safety and quality improvement.He has studied adherence to various evidence-based prac-tices, patient safety reporting systems, handoffs in healthcare,goals of care communication, and utilization of information inthe ICU.

Dow

nloa

ded

from

info

rms.

org

by [

142.

150.

190.

39]

on 3

1 Ju

ly 2

016,

at 1

0:23

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

coordinating flexible performance during everyday work: an ... et...

Documents