Designing Collaborative Workspaces for Particular Complex Work Settings Yingjing(Jane) Li A thesis submitted for the degree of Doctor of Philosophy in Computing Sciences Faculty of Engineering and Information Technology University of Technology, Sydney February 2016
218
Embed
Designing Collaborative Workspaces for Particular Complex ... · Designing Collaborative Workspaces for Particular Complex Work Settings Yingjing(Jane) Li A thesis submitted for the
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Designing Collaborative Workspaces for Particular Complex Work Settings
Yingjing(Jane) Li A thesis submitted for the degree of Doctor of Philosophy in Computing Sciences Faculty of Engineering and Information Technology University of Technology, Sydney February 2016
i
Certificate of Authorship/Originality
I certify that the work in this thesis has not previously been submitted for a degree nor has it been submitted as part of requirements for a degree except as fully acknowledged within the text.
I also certify that the thesis has been written by me. Any help that I have received in my research work and the preparation of the thesis itself has been acknowledged. In addition, I certify that all information sources and literature used are indicated in the thesis.
Signature of Candidate:
Date:
ii
Acknowledgements This research work has been a long journey. The completion of this work would not
have been possible without the support of many people whom I would like to thank
sincerely.
To Professor Toni Robertson, my supervisor, thank you for your guidance, inspiration,
input and encouragement. I appreciate your constructive advices which helped me to
establish the overall direction of this research and your suggestions which inspired and
challenged me to explore my research in depth. You constantly helped me by pointing
out the weaknesses in my work and by guiding me to clarify my ideas. I have been
fortunate to have worked with you since my early years of research at CSIRO and
throughout the journey of my PhD. Your guidance and encouragement have helped me
grow, both personally and professionally.
To Adjunct Associate Professor Duncan Stevenson, my advisor and former CSIRO
scientist, thank you for your mentoring and continuous support. I am grateful that you
shared your extensive research knowledge and experience with me and helped me better
understand the results of my explorations. I appreciate your detailed feedback which has
greatly improved the robustness of my thesis.
To Dr Christian Mueller-Tomfeld, Dr Kenton O’Hara, Susan Hansen and Dr Alex Hyatt,
former CSIRO scientists, and to Dr Tim Mansfield, former NICTA researcher, whom I
worked closely with in the studies that are presented in this thesis, thank you for all your
invaluable contributions to the field studies, data analysis and writing of papers. Your
knowledge and insights have greatly helped shape my research.
To my CSIRO managers and colleagues, Dr Laurie Wilson, Dr Jesper Kjeldskov, Dr
Leila Alem, Dr Jill Freyne, Dr Dimitrios Georgakopoulos, Dr Sarah Dods, Dr Trevor
Bird, Dr David Hansen, Dr John Zic and Narelle Clark, thank you for giving me a
supportive environment to enable me to pursue my doctoral degree. Your understanding
and encouragement are gratefully acknowledged.
To the members and associates of the Interaction Design and Human Practice Lab at the
University of Technology, Sydney, thank you for your friendship and support. I am
grateful that I did my PhD in this wonderful academic environment. Special thanks to
iii
Dr Julia Prior, Dr Jeni Paay, Dr Lian Loke, Dr Tuck Leong, Dr Penny Hagen, Dean
Hargreaves, Jeannette Durick and Anita Gisch for your company during my PhD
journey.
Most of all, to my husband Kai, my two children and my parents, I would never have
gone this far without your love and support.
iv
Table of Contents
1 Introduction .................................................................................................................. 1 1.1 Background ............................................................................................................. 1 1.2 Research Context – Designing Collaboration Technologies .................................. 4 1.3 Challenges in Supporting Distributed Collaboration .............................................. 6 1.4 Research Questions and Objectives ........................................................................ 8 1.5 Research Approach and Three Case Studies ........................................................ 10 1.6 Summary of Findings and Contributions .............................................................. 11 1.7 Structure of the Thesis .......................................................................................... 13
2 Supporting Distributed Collaboration - Literature Review ................................... 15 2.1 Distributed Collaboration, Interaction and Design ............................................... 16
2.1.1 An Overview of Media Spaces and Related Research ............................... 17 2.1.2 Social Interaction ........................................................................................ 19 2.1.3 The Space of Media Space ......................................................................... 21 2.1.4 Communication .......................................................................................... 24 2.1.5 Asymmetries and Fractured Ecologies ....................................................... 26 2.1.6 Issues in Early Media Space ....................................................................... 28 2.1.7 Reflections and Challenges ........................................................................ 30 2.1.8 Multi-display Environment and Shared Digital Workspace for
Meetings and Information Sharing ............................................................. 32 2.2 Collaboration across Different Local Settings ...................................................... 35
2.2.1 Variations in Local Practices ...................................................................... 36 2.2.2 Common Information Spaces ..................................................................... 39 2.2.3 Flexibility and Design ................................................................................ 42 2.2.4 Configuring Technology, Practices and Resources .................................... 44
2.3 Summary ............................................................................................................... 48 3 The Research Process ................................................................................................. 50
3.1 Evolution of the Research ..................................................................................... 50 3.1.1 Original Research Interest .......................................................................... 51 3.1.2 Perspective .................................................................................................. 52 3.1.3 Getting Focused .......................................................................................... 52
3.2 Workplace Studies in a System Design Context .................................................. 55 3.2.1 Workplace Study in CSCW ........................................................................ 55
v
3.2.2 Challenges in Conducting the Studies ........................................................ 58 3.3 Research Process ................................................................................................... 63 3.4 Data Collection and Data Analysis ....................................................................... 64
3.4.1 Data Collection ........................................................................................... 65 3.4.2 Data Analysis ............................................................................................. 66 3.4.3 Synthesis ..................................................................................................... 67
3.5 Summary ............................................................................................................... 68 4 Case Study: Collaboration in Multidisciplinary Medical Team
Meetings ...................................................................................................................... 69 4.1 Multidisciplinary Team Meetings ......................................................................... 69 4.2 Related Work ........................................................................................................ 71 4.3 Research Context and Motivation ......................................................................... 74 4.4 Methods ................................................................................................................ 76 4.5 Meetings Within and Between the Hospitals ........................................................ 79 4.6 Local Meetings ..................................................................................................... 81
4.6.1 Physical Settings ......................................................................................... 81 4.6.2 Preparing Information ................................................................................ 84 4.6.3 Presenting Information ............................................................................... 86 4.6.4 Context of Different Information Practices ................................................ 87 4.6.5 Conversation and Case Discussion ............................................................. 88
4.8 Exploration of the Physical Setup ......................................................................... 94 4.9 Discussion ............................................................................................................. 95
4.9.1 Designing Physical Space........................................................................... 96 4.9.2 Supporting Information Sharing ................................................................. 98 4.9.3 Challenges of Integration ......................................................................... 100
4.10 Conclusion .......................................................................................................... 101 5 Case Study: Distributed Collaboration in Emergency Response on
Animal Disease .......................................................................................................... 103 5.1 Emergency Response on Animal Disease .......................................................... 103 5.2 Related Work ...................................................................................................... 106
vi
5.3 Research Context and Motivation ....................................................................... 108 5.4 Methods .............................................................................................................. 110 5.5 CCEAD Meetings ............................................................................................... 111
5.5.1 Physical settings ....................................................................................... 112 5.5.2 Decision Making ...................................................................................... 115 5.5.3 Information Sharing ................................................................................. 116 5.5.4 Background Work and Multi-tasking During Meetings ........................... 119
5.6 Exploration of the Physical Setup ....................................................................... 120 5.7 Discussion ........................................................................................................... 122
5.7.1 Coordination and Working Together ....................................................... 123 5.7.2 Supporting Information Sharing ............................................................... 124 5.7.3 Different Settings and Existing Mechanisms ........................................... 125
5.8 Conclusion .......................................................................................................... 127 6 Case Study: Distributed Scientific Collaboration across
Biocontainment Barriers ......................................................................................... 128 6.1 Distributed Scientific Collaboration ................................................................... 128 6.2 Related work ....................................................................................................... 129 6.3 Research Context and Motivation ....................................................................... 131 6.4 Research Methods ............................................................................................... 133 6.5 The work of AAHL ............................................................................................. 135
6.5.1 The Physical Setting ................................................................................. 136 6.5.2 Work Groups ............................................................................................ 137
6.6 Collaboration within AAHL ............................................................................... 138 6.6.1 Collaboration within and between Groups ............................................... 138 6.6.2 Collaboration Issues ................................................................................. 141 6.6.3 Information Sharing Issues ....................................................................... 143
6.7 Informing the Design .......................................................................................... 145 6.7.1 Scenario One: General Group Discussion ................................................ 146 6.7.2 Scenario Two: Real-time Scientific Data Discussion .............................. 149 6.7.3 Scenario Three: Collaborative PC4 Work ................................................ 150
6.8 Post-deployment User Study .............................................................................. 151 6.9 Discussion ........................................................................................................... 153
6.9.1 Supporting Coordinative Practices ........................................................... 154 6.9.2 Linking People with People, Information and Facilities .......................... 155 6.9.3 Designing for Different Local Requirements ........................................... 156
vii
6.10 Conclusion .......................................................................................................... 157 7 Issues and Design Guidelines................................................................................... 158
7.1 Understanding Distributed Collaboration in Context ......................................... 158 7.1.1 Summary of Findings ............................................................................... 159 7.1.2 Common Issues ........................................................................................ 160
7.2 Guidelines for Designing Collaborative Workspace .......................................... 163 7.2.1 Guideline 1: Focusing on Supporting Information Sharing ..................... 165 7.2.2 Guideline 2: Constructing Physical Space to Support Social
and Communicative Interactions .............................................................. 167 7.2.3 Guideline 3: Configuring Physical Space and Information
Space to Support Integrated Communication and Information Sharing ...................................................................................................... 169
7.2.4 Guideline 4: Configuring Collaborative Workspaces to Enable Local Practices ......................................................................................... 171
7.2.5 Guideline 5: Configuring Collaborative Workspace in the Broader Organizational and Procedural Arenas ....................................... 174
7.3 Summary ............................................................................................................. 175 8 Contributions and Conclusion ................................................................................ 176
8.1 Contributions ...................................................................................................... 176 8.2 Future Work ........................................................................................................ 179 8.3 In Conclusion ...................................................................................................... 181
Figure 1.1. A research prototype - Blended Interaction Space installation ...................... 3 Figure 2.1. The components of the literature review in this thesis ................................. 16 Figure 3.1. Getting focused ............................................................................................. 54 Figure 3.2. An overview of the research process ............................................................ 63 Figure 4.1. Site plan of the local meeting at hospital A .................................................. 83 Figure 4.2. Site plan of the local meeting at hospital B .................................................. 83 Figure 4.3. Preparation of patient histories ..................................................................... 84 Figure 4.4. Preparation of radiology images ................................................................... 85 Figure 4.5. Preparation of pathology images .................................................................. 86 Figure 4.6. Site plan of the room for distributed meeting at hospital A in
distributed meeting ....................................................................................... 90 Figure 4.7. Site plan of the room for distributed meeting at hospital B .......................... 91 Figure 4.8. Examples of clinicians’ design sketching ..................................................... 95 Figure 5.1. Membership of CCEAD ............................................................................. 105 Figure 5.2. A large meeting room ................................................................................. 112 Figure 5.3. The office used for the meeting .................................................................. 114 Figure 5.4. Information access and distribution ............................................................ 117 Figure 5.5. Room arrangement 1................................................................................... 121 Figure 5.6. Room arrangement 2................................................................................... 122 Figure 6.1. The physical layout of the work areas at AAHL ........................................ 136 Figure 6.4. Overview of the work groups and interactions between them ................... 139 Figure 6.5. Diagnostics collaboration between related work groups ............................ 139 Figure 6.6. Communication and information sharing between work groups ................ 142 Figure 6.7. Three design scenarios to drive the design of the collaboration
platform ...................................................................................................... 145 Figure 6.8. Shared workspace at the general office area in AAHL .............................. 147 Figure 6.9. Side view of the physical setting of the shared workspace ........................ 148 Figure 6.10. Collaboration between scientists in PC4 and scientists in the control
room in PC3 ............................................................................................... 151 Figure 7.1. Complex and interrelated issues in distributed collaboration across two
different local settings ................................................................................ 161 Figure 7.2. Mapping the five guidelines to the factors that shape distributed
collaboration in particular local settings .................................................... 165
ix
Figure 7.3. Integrating physical space and information space in a large meeting room environment ...................................................................................... 170
x
List of Tables
Table 4.1. Participants interviewed ................................................................................. 77 Table 4.2. Participants at hospital A ............................................................................... 81 Table 4.3. Participants at hospital B................................................................................ 82
xi
List of Abbreviations
AAHL Australian Animal Health Laboratory
BISi Blended Interaction Space installation
CCEAD Consultative Committee for Emergency Animal Disease
CSCW Computer Supported Cooperative Work
CSIRO Commonwealth Scientific and Industrial Research Organisation
CVO Chief Veterinary Officer
DAFF Department of Agriculture, Fishery and Forestry
DSTO Defence Science and Technology Organisation
EPR Electronic Patient Record
HCI Human Computer Interaction
HD High Definition
JCSCW Journal of Computer Supported Cooperative Work
LCD Liquid Crystal Display
LIMS Lab Information Management System
MDTM Multidisciplinary Medical Team Meeting
NICTA National ICT Australia
PACS Picture Archive and Communication Systems
PC3 Physical Containment Level 3
PC4 Physical Containment Level 4
ViCCU Virtual Critical Care Unit
xii
Abstract
This research explores how new collaboration technology can be designed to enhance
distributed collaboration in particular complex work settings. Collaboration in work
environments increasingly involves complex interactions between individuals and teams
working across geographical, institutional and professional boundaries. This research
addresses the challenges of supporting real-time communication and information
sharing between different teams and across variable local settings. These issues are
explored within the context of developing collaborative workspaces which integrate
sophisticated video conferencing and information sharing technologies in multi-display
environments. This research aims to understand the characteristics of interactions that a
collaborative workspace needs to support and how to design a collaborative workspace
for collaboration across different local settings without compromising the integrity of
local work practices.
The research issues were explored through three case studies in three work domains:
multidisciplinary medical team meetings in two hospitals, collaboration in a national
committee responsible for the emergency response to animal disease, and scientific
collaboration across containment barriers in a biosecurity laboratory. Workplace studies
were conducted in each of the studies. The case studies were research components of
design-oriented projects carried out by the Commonwealth Scientific and Industrial
Research Organisation (CSIRO) with aims to inform the design of a collaborative
workspace within each domain. The case studies are the empirical contributions of this
thesis.
This research has shown that a set of socio-technical factors relating variations in local
physical settings, information sharing practices and organizational contexts can
influence the dynamics of collaboration across different local settings. The results
highlight different kinds and levels of configuration work required in designing
collaborative workspaces. These include the careful integration of physical settings with
information sharing practices, the appropriate configuration of collaborative workspaces
to enable diversity of local practices and the configuration of collaborative workspaces
at an organizational level and in the context of coordinative practices. The results of the
study have contributed to the development and deployment of an integrated
xiii
collaboration platform in a scientific laboratory and have demonstrated that a generic
collaborative workspace can be extended by components developed in response to the
specific requirements of the work of the local setting. A set of design guidelines has
been developed that can be used to guide the design and development of collaborative
workspaces which provide coherent collaboration environments across different already
existing local settings while respecting the variations within local practices.
Introduction
1
1 Introduction
The research presented in this thesis explores collaborations between individuals
working at different locations and in complex work settings. It focuses on how to design
new collaboration technology to support these collaborations that can fit well with
particular work practices and organizational context. This focus is explored within the
research fields of computer supported cooperative work (CSCW) and human computer
interaction (HCI). Three case studies of distributed collaborations in three different
work settings form the empirical contributions of the research and theoretical
contributions are derived from the results of the three case studies.
This chapter introduces the thesis and describes its contributions to the fields of CSCW
and HCI. It includes a discussion of the background of the research, an overview of its
context and challenges, the research questions addressed in it, an introduction to the
three case studies and a summary of its findings and contributions.
1.1 Background
As a researcher working in a laboratory of the CSIRO (Commonwealth Scientific and
Industrial Research Organization), I have been involved in a number of projects in the
design and development of new technologies to enhance distributed collaboration. One
of the research foci of these projects has been the design of technologies to be deployed
in real-world work environments to support collaborative activities such as diagnosis,
planning, analysis and decision making. An example of early work that I was involved
in was the development of the Virtual Critical Care Unit (ViCCU) which was a
broadband telehealth system allowing a medical specialist to remotely supervise a
clinical team in an emergency department managing critically ill patients (Li et al. 2006,
Wilson et al. 2010). My initial role in this project was visiting hospitals and talking with
clinicians in order to understand their requirements. These understandings were then
used to generate technical specifications for technical design. After an iterative design
process the ViCCU system was successfully deployed and provided a link between the
emergency departments of a large public hospital and a small rural hospital. I
participated in the evaluation of the system in the emergency departments where I
Introduction
2
observed how the system was used. The design team was not familiar with CSCW and
HCI research and evaluation methods but was motivated to work closely with the
clinical users. The evaluation results showed that the factors contributing to the
effectiveness of the system were not just the high quality audio-video channels enabled
by a broadband network. Our understandings of the dynamic interaction patterns in the
clinical practices and the tailored design that benefited from field visits before and
during the design process had contributed to the successful integration of the system
into the work practices of the emergency department. These findings made it clear to the
technology designers that our design approach needed to shift from the traditional
emphasis on laboratory experiments to the real-world context in order to overcome both
technical and non-technical issues in designing and implementing new technologies.
The experiences from this project directed my attention to CSCW and HCI fields which
are design-oriented research fields that inform technology design by empirically derived
understandings of how people work within groups and organizations and the impacts of
technology on those processes.
I participated in the “Braccetto” project within the HxI Initiative, a research
collaboration between Australia’s leading publicly funded research organisations –
CSIRO, DSTO (Defence Science and Technology Organization) and NICTA (National
ICT Australia) from 2007 to 2009. One of the research outcomes of this project was the
demonstration of a collaborative workspace, a research prototype called Blended
Interaction Space installation (BISi) (Broughton et al. 2009, Paay et al. 2011, O’Hara et
al. 2011) (Figure 1.1). It was designed to facilitate collaboration within a small group of
distributed participants and as a walk-up-and-use system where the user does not have
to deal with camera and audio setups. It enabled distributed participants to share a range
of digital content on large displays, including four large displays and a tabletop display,
with simultaneous multi-user interactions. The configuration of the furniture and
displays provided a “blending” of the local physical workspace with the video
conferencing images of the other side so the two sites appeared as one workspace. It
enabled more complex collaborative activities than the normal interactions in a board-
meeting-room type of video conferencing. By creating a shared physical and digital
workspace, it aimed to support complex hands-on collaborative activities and facilitate a
being-together user experience in distributed collaborations.
Introduction
3
Figure 1.1. A research prototype - Blended Interaction Space installation
The work of HxI also included conducting field studies in different work domains in
order to provide high level design insights for the design of BISi. After the completion
of the HxI Initiative, the CSIRO research team decided to extend the design of this
generic prototype and deploy the technology to real-world environments. Studies of
distributed collaboration were carried out in a number of work domains in order to
inform specific designs. Based on the conversation with the managers of related
organizations and initial understandings of the collaboration requirements, some of the
work domains were identified as potentially suitable sites to proceed with field studies
and extension design of BISi.
Since the ViCCU project, I have participated in a range of field studies of distributed
collaboration in real-world work environments. These studies include:
Collaboration in the multidisciplinary medical team meetings in two hospitals in
Sydney. This study was done by the HxI team. I was one of the key members of
the team.
Distributed collaboration between members of a large-scale committee working
in managing the responses to emergency animal disease. This study was done by
CSIRO researchers. I was one of the key members of the study.
Distributed scientific collaborations within a national animal health laboratory
which has high bio-containment facilities. This was conducted by the CSIRO
researchers and I led the study.
Introduction
4
My involvement in these design-oriented investigations led to the research in this thesis.
The aim of the research is to address the challenges of particular complex work settings
and the design of technologies to support distributed collaboration within and across
these complex settings.
I had approvals from the University of Technology, Sydney and CSIRO ICT Centre to
be enrolled as a part-time PhD student and I had permission from CSIRO to use the
intellectual property from my work at CSIRO towards my degree.
1.2 Research Context – Designing Collaboration Technologies
There has been an interest for more than three decades, within the research fields of
CSCW and HCI, in developing systems that use audio-video communication
technologies to support real-time collaborations between people based in different
locations. A central and on-going focus of this stream of research has been the design
and development of systems to provide support for activities and interactions in real-
world settings. This research focuses on understanding people’s work practices and
experiences in order to design technologies that support the way people communicate
and interact in their everyday and working lives (Bannon & Schmidt 1991, Blomberg
1993, Schmidt 1998, Rogers et al. 2011, Kuutti & Bannon 2014). Studies in these
research fields, particularly CSCW, provide rich insights into how work is
collaboratively achieved as well as guidelines for designing technologies that support
that work (Blomberg & Karasti 2013, Fitzpatrick & Elingsen 2013, Bjorn et al. 2014).
Research in distributed collaboration explores technical design challenges (e.g. Tang &
Minneman 1990, Gaver 1992, Childers et al. 2000), recognizes issues of computer-
mediated communication (e.g. Bly et al. 1993, Heath et al. 1995, Olson & Olson 2000)
and, importantly, addresses the complex socio-technical issues that shape the
collaboration (Harrison 2009 and most of the studies referred to in this thesis). It is
widely accepted that a collaborative system has “the technological components which
determine how the system will behave and the social components which determine the
acceptable use and behavior” (Dourish 1993, p.125). Interaction in a technology-
enabled environment is, by its nature, a social activity and the collaborative systems
used to manage the connectivity of the work environments are always embedded within
social and organizational contexts (ibid). Researchers need to go beyond traditional
Introduction
5
cognitive and usability-based models to understand the relationship between people,
technology and practice.
The interrelations between work practice and the design of collaboration technology
have been widely explored in a wide range of work domains, such as healthcare
delivery (Heath & Luff 1996, Reddy et al. 2001, Kane & Groth 2013, Fitzpatrick &
Ellingsen 2013) and collaborative scientific work settings (Finholt & Olson 1997, Olson,
Zimmerman et al. 2008), that relate to the case studies in this thesis. Research in the
healthcare domain has highlighted the needs for a “contextual” understanding of the
actual work practice and a socio-technical approach in the design process (Zuiderent et
al. 2003, Kane et al. 2011). Similarly, research in scientific collaboration has called for
a shift from a focus on technical design and development to consider how technologies
and infrastructure can be embedded into the everyday working lives of scientists
(Jirotka et al. 2013). One key challenge for technology designers has been to understand
the practices that new technologies will support and the context in which they need to
be integrated (Kuutti & Bannon 2014).
There has been a range of collaboration technologies developed to support work-
focused interactions since the early 1990s when audio-video communications were used
to connect office environments (e.g. Dourish & Bly 1992, Gaver 1992, Bly et al. 1993).
These technologies include Access Grid which is an ensemble of resources such as large
displays and conferencing facilities for group-to-group collaboration (e.g. Childers et al.
2000, Corrie & Zimmerman 2009), high-definition and high fidelity telepresence video
conferencing system (e.g. Gorzynski et al. 2009) and the recent developments
integrating telepresence video conferencing with shared digital workspaces (e.g. Wigdor
et al. 2009, Stevenson 2011, Yamashita et al. 2011, Luff et al. 2013, and BISi described
in Section 1.1). The technology investigated in this thesis relates to this recent
technology development which supports high quality audio-video communication and
advanced information sharing and interaction in a multi-display environment. The term
“collaborative workspace” will be used to refer to this technology development.
Researchers have found that despite the increasing use of collaboration technologies and
the creation of prototypes, people are still not successfully embedding these systems
into everyday work life (e.g. Baecker et al. 2008, Luff et al. 2009, Tang 2009, Jirotka et
al. 2013). As Luff et al. (2009) commented, it has been a “highly intractable” problem
Introduction
6
to put a successful and usable collaborative system into an organizational setting (p.27).
Social, behavioural and cultural issues are still relevant (Baecker et al. 2008, Harrison
2009, Olson & Olson 2013, Bjorn et al. 2014). Interaction issues are still evident (e.g.
Harrsion 2009, Luff et al. 2013). Technology improvements are still being explored by
researchers, such as shared interaction with information by remote pointing and
annotation (e.g. Stevenson et al. 2010, Norris et al. 2013) and configuration of the
physical setup to create coherent collaboration environments (e.g. Ciolfi et al. 2008,
Buxton 2009, Paay et al. 2011). While introducing collaboration technology has been
considered as a way to improve work efficiency, getting it right has been considered an
There are tensions when researchers use ethnographic approaches for studying
distributed collaboration, particularly in synchronous and real-time distributed
collaborations. Ethnography within one work site or a confined environment, such as a
control room, has relatively clearly visible tasks to observe. Scaling such observation to
spatially expanded settings and to processes distributed in real-time is much more
difficult (Hughes et al. 1994, Blomberg & Karasti 2013).
Millen (2000) points out that one of the valuable approaches to improve the efficiency
of field research is to have more than one researcher in the field at the same time. For
example multiple researchers in the field can split up and observe different activities or
groups; multiple views of the same events can be turned into a richer representation and
understanding of the situation. This collaborative observation and analysis strategy is
applicable to the study of distributed collaboration although it is expensive and not
always possible within a design project with limited resources.
I was fortunately able to work together with a group of researchers in the field studies.
For example, the MDTM study involved observations of real-time distributed meetings
in two hospitals so three to four observers split up at the two sites to observe the
meetings and met after the observation to debrief. Audio and video recordings of some
meetings at the two sites were collected simultaneously. I joined and synchronized the
two sites recordings using the Vegas computer software to allow the researchers to
review the distributed meetings during the data interpretation process.
3.2.2.4 Work Environment Constraints and Other Constraints
Conducting research in a complex work environment has particular constraints caused
by the specific work setting. Field work researchers not only require access to relevant
sites but also need acceptance from those who work in them and need to accept the
work regulations in these settings (Hughes et al. 1994).
The ethical issue of conducting a workplace study in a hospital was a particular concern.
During the MDTM study we worked closely with the key study participants to design
The Research Process
62
the study and obtained ethics approval from the two hospitals. Issues regarding patient
privacy and video recordings were addressed by anonymising the data. All the study
participants were provided with an information sheet explaining that their and their
patients’ identities would be protected and the data would be used only for the purpose
of the study.
In the CCEAD study I was not able to observe the actual CCEAD meeting in action.
This was due to the unpredictable nature of emergency disease outbreak and, more
specifically, due to the high level confidentiality of the meeting. However we were able
to visit the meeting rooms. Where possible we conducted interviews and focus group
meetings in these rooms to allow the participants to better explain the meeting process
and resources used in the meeting.
The study at AAHL required me to follow the strict biosecurity regulations at the
containment areas in that laboratory, such as wearing laboratory clothing to go in and
having a proper shower to exit. There were also particular constraints and procedures on
getting data collection materials, such as the camera, audio recorder and notes, out from
the containment areas. I was not able to go into the most secure containment area at all
because wearing a fully encapsulated suit is required for entering into that area and a
chemical shower is required when leaving.
The CSIRO working context had practical constraints on my effort in data collections in
addition to the project time frame constraint described in Section 3.2.2.1. As a full-time
CSIRO employee, I worked in multiple projects at the same time. My involvement in a
study could be reduced if I was allocated to a new project by the managers. This
happened in the AAHL study as I will describe in Chapter 6. The part-time nature of my
PhD study also limited my time in conducting the data analysis, particularly the
synthesis analysis across the three case studies, after the completions of the individual
projects. I managed to make progresses by working closely with other researchers
involved to publish the findings of individual studies (the permissions of using the
photographs in this thesis for publications were obtained when we published related
papers). I kept on reading recent related work and used updated literature review to help
reflect my findings.
The Research Process
63
3.3 Research Process
Conducting three case studies in this research was an iterative and reflective research
process and it needed to be considered as a whole process rather than individual study
followed by individual study. Figure 3.2 below outlines this process.
Figure 3.2. An overview of the research process
The research process in each of the case studies consisted of these components (Flick
2008, Thomas 2011):
Study design, which is proposed at the beginning of the study;
Data collection, which involves gathering data about work practice and
collaboration practice in individual domains;
Data analysis, which involves inductive analysis of field data;
Synthesis, which includes drawing the significance of individual findings and
merging together of three case study findings guided by theories in the literature.
The research in the AAHL study also included a stage of using the field study findings
to inform the design of a collaboration platform.
The Research Process
64
Although the data collection in the three case studies was done in different calendar
years, the research process was not a linear one of one case study following another. As
described in section 3.1, multiple case studies and an iterative approach allowed the
refinement of the research questions and the research objective during the collection of
data and analysis process. There was a constant reflective process between existing case
analysis, previous data analysis, theories and synthesis, as described below.
This reflective process was particularly important for the study designs. In each study
considerable time was spent before conducting the workplace study in the field to
ensure appropriate research design. During this period, issues identified in earlier
studies were taken into consideration, including informing the investigation focus and
selecting of data collection methods.
The reflective process was not only required for cross-case design and analysis but also
within each study itself. Conducting interpretive research often begins with issues
researchers can not know ahead of time and the research focus sharpens during
immersion in the field. Interpretive research requires a degree of openness at the
beginning and constant moving between field observation, literature and theory
(Altheide & Johnson 2011).
Frameworks and theories from literature were used for analyzing and interpreting
findings in this research. For example, prior work utilizing insights from ethnographic
research and concerning variations in local practices (e.g. Ellingsen & Monteiro 2006,
Schmidt et al. 2007, Balka et al. 2008) is one of the foundations that shaped my
research. Another is the interrelated social, spatial and communicative interactions in
media space (Harrison 2009) that suggested dimensions of analysis for my case studies.
A third is the configuration needs and the categories of configurability (Binder et al.
2004, Balka & Wagner 2006) that have guided the development of three levels of
configuration work for designing a collaborative workspace as presented in Chapter 7.
3.4 Data Collection and Data Analysis
One recognized strategy to ensure the research quality of interpretive study is to use
triangulation techniques - verifying facts and interpretations through multiple sources,
such as methods, data and investigators (Cho & Trent 2006). Qualitative researchers
usually deploy a range of interconnected interpretive methods to understand the
The Research Process
65
practices that are studied (Denzin & Lincoln 2011). Using a diversity of material allows
the researcher to approach the research questions from different angles and a broader
picture of work practices.
In the research presented in this thesis, multiple sources of information and methods
were triangulated to ensure the credibility of the qualitative research. I was the key
researcher in each study and worked through the whole workplace study process in
each, while other researchers participated in different studies and in some parts of data
collection and analysis of the studies.
3.4.1 Data Collection
A range of data collection methods was used throughout the research process that
included interviews, focus groups and direct observations. Interviews were useful to
determine participants’ perceptions and explanations of processes whereas direct
observations (including video recordings) helped to identify phenomena that
participants may not have reported (Denzin & Lincoln 2011). The aims of both
interviews and focus group meetings were to gather details about the collaboration
practices that participants had with other people; issues around their existing
collaboration and information sharing practices; their design ideas and expectations of
new systems. The following is a description of each of these data collection methods:
Interviews: Semi-structured interviews were conducted in each of the case studies. They
were conducted in situ in conjunction with the observations or site visits in the MDTM
and AAHL studies. Telephone interviews were used in the CCEAD study since the
CCEAD members were located in seven states and territories. For consistency semi-
structured interview schedules and basic scripts were prepared before the interviews so
that the same topics were covered with each participant. These interviews were audio
recorded and transcribed for later analysis.
Focus groups: A focus group was used in the CCEAD and AAHL studies. One large
focus group meeting with 10-20 participants was held at the beginning of each study. It
allowed diverse or sensitive issues to be raised by participants in a supportive
environment and in a social context by talking to others (Rogers et al. 2011). The focus
group discussions were audio recorded and analyzed in a similar way to the interviews.
The Research Process
66
Direct observation: Direct observations in the field have helped to capture details that
were not elicited from the interviews. They provided context for the tasks and important
information about why activities happen in particular ways (Rogers et al. 2011). In the
MDTM study observations of the meetings were carried out once a week over a three
month period by multiple researchers. Observations focused on different aspects of
interactions suggested by Jordan and Henderson (1995), including the structure of the
meeting, temporal organization of the meeting, turn taking, trouble and repair, and the
spatial organization of activities, artefacts and documents. Pointing and gesturing were
particularly examined. There were no opportunities to directly observe the collaborative
meetings in the CCEAD and AAHL studies, but during the site visits physical settings
and arrangements of artefacts were investigated and photographed. Notes were taken
during the observations. Video recording methods were used in the MDTM study.
3.4.2 Data Analysis
Data analysis methods adopted in each of the case studies included thematic data
coding, video reviewing and analysis with writing. The first two methods were common
across the case studies and the last one was used in the MDTM study only. The
qualitative analytic process is cyclical rather than linear (Saldana 2009). The first step is
to gain an overall impression of the data by looking for patterns (Randall et al. 2007).
Some recurring patterns or themes may have emerged during the data gathering process.
Since I worked closely with a couple of other researchers I was able to have discussions
with them after each observation and interview to identify the patterns that emerged. In
my three case studies, initial data were analyzed and discussed while data collection
continued. The following is a description of each of these analysis methods.
Data coding: Interview transcripts were coded in each of the studies. I was the
researcher working on the data coding. Coding data is a process of investigating the data
to identify codes and categories and of organizing and analysing these to identify
themes (Saldana 2009). Coding is an iterative process from data, to code, to category
and back to data. There were several iterations in the analysis process of each case
study. I first worked on the initial data coding based on the interview structure. A code
table was generated either using paper or using NVivo software. Key issues were
identified and grouped into categories relating to aspects of the particular local setting,
The Research Process
67
the practices of participants and information sharing issues. Based on the categories,
several sub-themes were identified. These themes were then closely examined by using
findings from other data collection methods and literature reviews. The data were also
reviewed by the other researchers involved to validate the categories and themes. A list
of core themes and sub-themes were identified by this iterative process. This inductive
approach allows research findings to emerge from the frequent, dominant or significant
themes inherent in raw data.
Video reviewing: The video recordings of the distributed MDMTs were synchronized
and annotated using Vegas software. Due to the limited number of video recording
sessions that we were able to obtain, formal video analysis was not involved in the
study. The videos were reviewed and discussed together by multiple researchers
drawing on the “interaction analysis” framework introduced by Jordan and Henderson
(1995). Video materials were helpful to understand the particular actions with regard to
the immediate context and a particular interactional environment in which they arose.
Analysis with writing: Initial findings of each case study were summarized into reports
to present to the design team. Insights were carefully written into research papers and
published following the in-depth analysis process and while conducting the next study.
During this paper writing process, theories from literature were used to further reflect on
and interpret findings. Importantly, this provided a concrete iterative feedback between
studies and was useful for the cross-case analysis.
3.4.3 Synthesis
The synthesis process included synthesis in an individual case study and synthesis
across case studies (see Figure 3.2).
At an individual case study level, this was done by using data triangulation, as described
above, and by constantly moving between field observation, literature and theory. Prior
to the cross-case synthesis process, each case was analysed as one entity (a within-case
analysis). These analyses involved detailed write-ups of each as a coherent entity to
cope with the diversity of data collected from each case. The results, therefore, included
in-depth empirical descriptions of work practices in the three cases.
The Research Process
68
At a multiple case study level, based on results from the individual cases I conducted
cross-case analysis for patterns which occurred across the cases. All cases were
workplace studies of collaboration in particular settings that shared some basic
characteristics such as variations of practices in the collaboration sites. Categories based
on literature were selected and combined with the themes emerging from the three cases
during the analysis. For example, the three dimensions of social, spatial and
communicative interactions in media space (Harrison et al. 2009) have been helpful
when I investigated different aspects of interactions and information sharing practices in
a particular collaboration setting; the design principles identified in media space
research (e.g. Luff et al. 2003, Buxton 2009, Paay et al. 2011) were referred to when I
considered the designs related to physical space and gesturing in supporting the access
to remote participants, their environments and digital resources; and the configurability
categories (e.g. Binder et al. 2004, Balka & Wagner 2006, Balka et al. 2008) have been
one of the key related works that inspired me to think about how to provide flexible
support for different forms of collaborative activities and organizational context. I used
these dimensions as a means to bring together results from the case studies and
construct design guidelines concerning designing collaborative workspace to be used
across different local settings.
The data collection and analysis processes in each of the three cases will be described in
more detail in the Method sections in Chapters 4, 5 and 6.
3.5 Summary
This chapter has focused on the purpose, approach and process of this research. The
evolution of the research questions after a reflective process was presented. Workplace
study in CSCW was reviewed along with a rationale for the methods adopted in this
research to address particular methodology challenges. To address the research
questions, three case studies were conducted in an iterative and reflective research
process which involved moving between study design, data analysis, literature and
theory. The data collection and data analysis of these three empirical studies led to the
final synthesis process from which design guidelines for collaboration technologies to
support different local settings were generated.
Collaboration in Multidisciplinary Medical Team Meetings
69
4 Case Study: Collaboration in Multidisciplinary Medical Team Meetings
This chapter describes the study of multidisciplinary team meetings (MDTMs) in two
groups of breast cancer care teams in two hospitals - a large public teaching hospital and
a smaller private hospital. Regular MDTMs were held at each hospital for the treatment
and on-going management of its breast cancer patients. Some clinicians at the public
hospital were responsible for the care of some patients at the smaller private hospital
and the MDTM discussion of the ongoing treatment of these patients was done via
video conferencing between the two hospitals. The focus of the study was to investigate
how the MDTM participants collaborated in their local meetings in each hospital as well
as in the video conferencing meetings between the hospitals. The study examined the
MDTMs from various perspectives, including organizational context, physical setup of
the meeting, collaboration technologies used, medical information sharing practice and
technology-mediated conversation in both local meetings and distributed meetings.
Implications of designing collaborative workspace to support distributed collaboration
in MDTMs were explored. I was particularly interested in understanding how local
variations shaped the particular interactions in the distributed meetings, how different
local arrangements and practices arose and how to address the collaboration issues
caused by these differences.
4.1 Multidisciplinary Team Meetings
Multidisciplinary care is an integrated team approach to healthcare in which relevant
healthcare professionals collaboratively develop a treatment plan for individual patients.
They are considered important for the treatment of cancer since there is increasing
evidence that multidisciplinary care improves cancer patient outcomes (Zorbas et al.
2003, Bain et al. 2013). Cancer care involves a range of services including screening,
diagnosis, treatment (surgery, chemotherapy and radiotherapy), rehabilitation and
supportive care. Multidisciplinary cancer team members commonly include surgeons,
radiologists, pathologists, medical oncologists, radiation oncologists, psychologists,
oncology nurses and social workers.
Collaboration in Multidisciplinary Medical Team Meetings
70
Regular multidisciplinary team meetings are an integral component of multidisciplinary
care (Marsh et al. 2008). A central theme of the meetings is for clinicians from different
disciplines to get together to review patient cases, establish diagnosis, and decide upon
the management of cancer patients. Patient case details, including patient records and
medical images are presented and discussed by the team in the meeting. A typical
procedure of patient case discussion (e.g. Kane & Luz 2006, Kane et al. 2013) begins
with the cancer patients’ details being read out to the team, followed by the presentation
of relevant medical images, such as radiology and pathology images. The clinician in
charge of each patient leads the discussion of the case and the proposed treatment.
Oncologists, surgeons, radiologists and pathologists then contribute their points of view
about treatment and management options. Nurses, psychologists and social workers
contribute other relevant information. Eventually a decision is made about future care
and treatment.
MDTMs play an important role in the care coordination between different hospital
departments or different health institutes and provide clinicians involved with
opportunities to discuss their recommendations with the rest of the team (Kane & Luz
2009b). The main outcome of the MDTM is to reach a decision on cancer patient
management. Besides the formal review activities, patient referrals for clinicians
working in different hospitals, recruiting cancer patients for clinical trials and related
research advances can be discussed in the meeting. A MDTM also provides an
education environment for medical students and junior doctors who attend the meeting.
In addition, the meeting serves a social function to support the development of collegial
relationships, particularly through the informal interactions before and after the
meetings.
Video conferencing has been increasingly used in MDTMs to facilitate communications
between team members based in different hospitals or cancer centres (e.g. Delaney et al.
2004, Kane & Luz 2006, 2013). Collaboration technologies have extended the
traditional MDTM, with local case discussions and other local activities involving
people in the same room, to enable collaboration with remote team members.
Collaboration in Multidisciplinary Medical Team Meetings
71
4.2 Related Work
Collaborations in MDTMs have been investigated by Kane and Luz (2006, 2008, 2009a,
2009b, 2011, 2013) and other CSCW and HCI researchers from various perspectives.
Kane, Groth & Randall (2011) summarize related research in a special issue of the
Journal of Behaviour and Information Technology where a collection of papers
investigating MDTMs are presented including one of my publications Li and Robertson
(2011). In this section I will describe three related themes in MDTM research: the
practices of communication and decision-making, comparison of co-located meeting
and video conferencing and the design of collaboration technologies to support
information sharing.
The importance of efficient communication and coordination in the practices of
MDTMs has been reported by Kane and Luz (e.g. Kane & Luz 2006, 2008, 2009a, b,
2013). They conducted a long-term ethnographic study in MDTMs for lung cancer
patients at a teaching hospital over several years. Some of their investigations include
the MDTM’s role in the diagnosis and treatment of patients, the practices in MDTMs
within the broader context of pre and post meeting activities and the social and
organisational contexts of the different participating hospitals. The practices of
assimilating information and decision making and how diagnosis is collectively
achieved have been also explored in detail (Kane & Luz 2009a, b). In providing
technological support for the MDTM, it needs to be recognized that the task of patient
management discussion through interaction and communication is one of the main
functions of the meeting (Kane & Luz 2009a). For any solution to be satisfactory,
account also needs to be taken of the organizational and social factors (ibid). Recently
Kane and Luz (2013) studied the changes, in particular the rhythms of discussion, when
introducing electronic patient record (EPR) and picture archive and communication
systems (PACS) (Kane & Luz 2013). They found that the number of patient cases in a
meeting has increased and there is less information and less time used in a patient case
review. They argue that although more information is potentially available at MDTMs
because of PACS and the EPR implementations, it is not as easy to access the electronic
information as it is to review a paper chart. However clinicians still value “being there”,
being co-located with their multidisciplinary team colleagues, and having the chance to
interact with one another. By comparing and contrasting MDTM practices in three large
Collaboration in Multidisciplinary Medical Team Meetings
72
teaching hospitals in Europe, Kane and Groth (2014) highlight the complex nature of
interpersonal collaboration at MDTMs and the importance of individual role
participation, communication and team cohesiveness.
The effects of the use of video conferencing on different aspects of the meetings and the
differences between co-located meetings and video conferencing meetings have been
explored by various researchers. Studies have shown that each case discussion takes
significantly longer in video conferencing meetings when compared with the local
meetings although the structure of the patient discussions remains the same (Kane &
Luz 2006, Groth et al. 2009). Delaney et al. (2004) report similar findings in their study
of MDTMs in an Australian hospital by showing that there is a slower and more formal
presentation and discussion of the case information when using video conferencing.
Kane and Luz (2006) argue that the formal style had less benefit for the core MDTM
members while this gives other participants opportunities for obtaining information and
making contributions. In their study of the effectiveness of MDTM, Kane and Luz
(2013) examine the timing of a MDTM in the context of using PACS and EPR. They
point out that video conferencing still takes more time per case, hinders interaction
dynamics (such as turn-taking) and causes difficulties in coordinating information
sharing.
Although video conferencing technologies have enabled relevant disciplines to
participate in discussion, the collaboration in MDTMs has not yet been achieved in a
satisfactory way (Kane & Luz 2006, Kane & Groth 2014). One of the improvements
needed is the use of more powerful systems and designs, such as pointing devices and
tools that allow for rapid manipulation of medical images (Kane & Luz 2009). Kane and
Luz (2008, 2009b, 2011) and Groth et al. (2008, 2009) have investigated the design of
collaboration technologies to support information presentation and sharing in MDTMs.
These investigations include the following three aspects:
Clinical information workspace. In their discussion of a “clinical information
workspace” for MDTMs, Groth et al. (2009) recommend the use of collaboration tools
for individual participants for the visualisation of, and interaction with, patient
information during meetings (see also Frykholm & Groth 2009). Groth and her
colleagues have developed a multi-display groupware system by incorporating multi-
user multi-model interaction techniques in personal handheld devices to support
Collaboration in Multidisciplinary Medical Team Meetings
73
MDTMs (Olwal et al. 2011). They evaluate the use of tablet devices for individual
meeting participants to view patient case details, which is a summary of a patient’s
electronic patient record (Frykholm et al. 2012). They found that using a summary
prepared before the meeting is more efficient than directly browsing through electronic
patient records but the use of the tablet distracts them from following the meeting
discussion. The drawing and annotation functions on the tablet devices and showing
these annotations on the shared large displays are very helpful for meeting participants
to discuss radiology images.
Text-based recording of the case discussion. With an aim to obtain a real-time text-
based recording of the meeting discussion as a communication record, Kane and Luz
(2008) investigate the use of visual display for text data in addition to the audio resource
to support co-located and teleconferencing discussion. They also look at using text-
based recording of case discussion as an electronic record for patient files (Kane et al.
2013). They point out that it is a challenge to develop a record that will not detract from
the synchronous collaboration between clinicians and that is easily processed.
Displays and pointing devices. Kane and Luz (2008, 2011) argue that collaboration can
be enhanced by the development of architecture spaces that are integrated with tools
supporting interaction and the exchange of information, such as large shared interactive
areas. Kane and Luz (2008, 2011) analysed the configuration of the video-displays and
offered a number of recommendations about how new configurations might facilitate
improved discussion and decision-making during video-mediated MDTMs. In their
discussion of an “ideal room” for a MDTM, Kane and Luz (2009b) suggest a range of
technologies that could equip the room to improve MDTM services. These include, for
example, suggestions about pointing technologies, the use of multiple screens for
concurrent display of medical information and ubiquitous devices to record the presence
of individuals.
The series of papers published by Kane and Luz have demonstrated that the MDTM is a
valuable setting for research into collaboration. Research themes around cooperation
and coordination, roles of participants, time, information sharing, space and place, and
technology employed are all relevant to the general CSCW issues. Exploring how the
teams work in MDTMs and the problems they encountered provide insights into
collaborative work that can be applied in other work settings. Although substantial
Collaboration in Multidisciplinary Medical Team Meetings
74
CSCW research has engaged with MDTM settings, Kane et al. (2013) point out that
some of the CSCW issues that arise when collaboration technology is employed at
MDTMs are still not fully understood.
4.3 Research Context and Motivation
The healthcare system in Australia is a complex mix of public and private systems.
Large tertiary teaching hospitals are funded and operated by state governments. Private
hospitals owned by private companies are usually small with no teaching functions.
Some specialists, such as surgeons, radiologists and pathologists, work in both the
public and private sectors and divide their time and their sources of income between the
two sectors.
There is a mixed practice and referral network for cancer patients in Australia. Cancer
patients can be referred by their general practitioners, who are a patient’s fist contact
with the health system, to specialists at public hospitals, private hospitals or private
practices. Specialists can refer their cancer patients to other discipline specialists for
investigation (e.g. radiology) or treatment (e.g. chemotherapy, radiotherapy). In public
hospitals, patients are treated by specialists nominated by the hospitals. Private patients
can choose their own specialists but need to pay for the service gaps which are not
covered by government funding. For example, a private patient can be referred by his or
her surgeon to a senior radiologist who works at a public hospital. Or a private patient in
a small private hospital can be managed by a surgeon at a large public hospital.
In Australia MDTMs have been held regularly in many hospitals as an integral part of
multidisciplinary care (Marsh 2008, Bain et al. 2013). The operations of MDTMs
present some unique features because of the mix of private and public health delivery
and shortage of specialists (Marsh 2008). The specialists work in different hospitals or
different sectors and are sometimes required to travel to attend case discussion sessions
at different hospitals. To minimize the need for travel, video conferencing has been
introduced to multidisciplinary cancer care settings as a solution to support case
discussion across sites (Delaney et al. 2004, NBCC 2005).
The MDTM study presented in this chapter was conducted in two hospitals in Sydney, a
large public teaching hospital (hospital A) and a smaller private hospital (hospital B). At
both hospitals the weekly meetings began with discussion of those cases in which all
Collaboration in Multidisciplinary Medical Team Meetings
75
patient data are held locally. These local meetings were followed by a distributed
meeting supported by video conferencing where the focus of discussions was those
cases where the patient data were distributed between the two hospitals. This was
because there was a mixed public and private service provision in the multidisciplinary
team between the two hospitals. Some of the surgeons at the public hospital A also
spent time working in hospital B to treat the private patients. If MDTMs involved the
discussions of these patient cases, that is, hospital B patients treated by a surgeon at
hospital A, the two meetings were joined via video conferencing for the discussion of
these patients. ISDN connection at the speed of 256kbit/s was used for the video
conferencing meetings.
In order to improve multidisciplinary team functioning and patient outcomes, the breast
cancer teams at the two hospitals proposed and implemented a number of quality
assurance strategies, such as the employment of a breast cancer nurse at each hospital
for the coordination of the meetings. The strategies also included technical improvement
of their video conferencing meetings between the two hospitals. This strategy led to the
collaboration between our design team and the breast cancer clinicians with an aim of
designing an enhanced collaboration platform to support the distributed MDTMs. Initial
conversation with a number of key clinicians in the two hospitals revealed audio and
video problems relating to video conferencing technology. The list was interesting to
the design team because the clinicians referred to each problem mostly as a technical
one, but further investigation revealed a more complex socio-technical issue. A field
study was then planned in order to have a better understanding of the team collaboration
over distance and to identify possible technical interventions to improve the
effectiveness of the meetings and team communication. However due to a number of
management and funding issues there was no system development activities after the
completion of the study.
There were a large number of meeting participants at hospital A and a smaller number
of participants at hospital B. The meetings at hospital A were held in a large multi-
purpose meeting room with a standard commercial video conferencing system while at
hospital B they were held in a smaller, dedicated room with integrated video
conferencing facilities.
Collaboration in Multidisciplinary Medical Team Meetings
76
It was clear from the early field work that there were differences in organizational
structures, physical setup of the video conferencing environment and the ways of
presenting medical images in the two hospitals. I was motivated to understand these
differences, to trace them to their various sources within their particular socio-technical
context and understand how these differences affected the meeting practices.
This research focus relates to a number of themes in the literature on the design of
coherent environments for distributed collaboration as described in Chapter 2. These
include the relationship between people, physical space and information space in
technology-mediated interaction (e.g. Bannon & Bodker 1997, Harrison 2009),
variations in local practices in healthcare (e.g. Schneider & Wagner 1993, Balka et al.
2008, Randell et al. 2011) and integrating technology across different settings (e.g.
Ellingsen & Monteiro 2006, Monteiro et al. 2013, Fitzpatrick & Ellingsen 2013). This
focus also extends related work in MDTM research by exploring the role played by
variations in local settings, particularly physical settings and information sharing
practices. Furthermore, studying the basic collaboration technology - video
conferencing - in use gave me a valuable understanding of the fundamental issues that
need to be addressed in designing collaborative workspace. This study and its results
directly impacted on my research questions and their necessary focus.
4.4 Methods
The three-month workplace study combined semi-structured interviews with
observations and video recordings of the meetings. A similar qualitative approach was
used by Kane and Luz (e.g. 2006, 2009a, 2009b, 2011) in their studies of MDTMs and
has been shown to be effective in understanding the mechanics in this environment. A
team of four researchers observed twelve meetings and interviewed eleven of the key
participants, six in hospital A and five in hospital B. The use of different methods,
conducted in parallel, was intended to enable a quick review items of interest from the
observations and, if needed, to provide evidence to interview participants of their
exhibited behaviours.
This study was conducted in a design-oriented context and in order to match the pace of
the design cycle, “rapid ethnography” techniques (Millen 2000) were used, such as
focusing on interviewing the key participants to quickly identify the basic requirements;
Collaboration in Multidisciplinary Medical Team Meetings
77
working with a “field guide” participant (in this case the breast cancer nurses); and
developing good relationship with participants. I coordinated the study schedule with
support from the hospital breast cancer nurses who suggested a list of interview
participants and helped to contact these participants. It was important to build up a good
relationship with the meeting participants and to obtain their trust as the meeting content
was sensitive and confidential. Researchers attended the weekly meetings and
communicated actively with meeting participants before and after the meetings and via
telephone and email at other times.
Observations focused on group behaviour, the group activities and interaction, in both
the local meeting and the distributed meetings, and the roles played by various
technologies and artefacts used in the meetings (Jordan & Henderson 1995). Members
of the research team split into two groups to attend both local meetings and swap around
from week to week so each member had an opportunity to observe both sites. A debrief
session for the observers was held after each meeting to share initial understandings and
perceptions of the meeting. The debrief session was helpful to maintain research focus,
reflect on findings and direct further stages of work.
Observations notes were taken by each of the researchers. These notes also included
interesting issues discovered during the informal talk with the clinicians before and after
the meetings. Audio and video recordings were collected in some of the meetings. The
video stream from the video-conferencing system was directly recorded, and in
addition, one or two video cameras were used for recording in each of the meeting
rooms. The recording equipment was set up and tested one hour before the meetings.
The positions of the room cameras were arranged to obtain the best views while
minimising the intrusiveness to the meetings.
Table 4.1. Participants interviewed
Role Hospital A Hospital B
Surgeon 1 1
Medical oncologist 0 1
Radiation oncologist 1 1
Pathologist 1 1
Radiologist 2 0
Breast Cancer nurse 1 1
Collaboration in Multidisciplinary Medical Team Meetings
78
Eleven meeting participants were interviewed as shown in Table 4.1. They were from
both hospitals and from each of the key clinical disciplines represented in the team. I
worked with another researcher on the semi-structured interviews. The interviews
focused on participants’ perspectives on their work in general and specifically their
comments on the meetings, the functioning of the group and how these might be better
supported. Interviewing medical staff in hospital is not straightforward. A number of
interview strategies were employed:
Participants in this study were busy clinicians and work under constant time
constraints. So the interview questions were prioritised to ensure that the most
important information, generally individual practices and routines, was captured
in even the shortest interview
Some of the participants might be on-call when we conducted interviews. For
example the interview with a pathologist was stopped several times since she
had to work on an urgent diagnosis request from a surgeon waiting at the
operating theatre. It required the balance between flexibility and consistency to
deal with work-related interruptions during the interviews
Interviews with the radiologists and pathologists were held in their departments.
This provided us with an opportunity to see their work environment and
appreciate the difficulties and complexity of their preparation work before the
meetings
Interviews were held immediately after the multidisciplinary team meetings if
possible because clinicians usually had some time available after the meetings
and had a fresher memory about their experience of the meetings
Snapshots taken from the video recordings of the meetings were found to be
helpful in prompting participants to discuss their experiences in the meetings
As part of the interviews, participants were encouraged to create draft layout
diagrams of the new collaborative system based on their expectations
Interview transcriptions and observation notes were read by all of the observers and
interesting issues were identified and discussed. I was responsible for the data coding
based on the interview structure and issues identified. Video editing software (Vegas)
Collaboration in Multidisciplinary Medical Team Meetings
79
was used to synchronise and annotate recordings of each side of the video-mediated
meetings. Videos were reviewed by the researchers during the various stages of
analysis.
The research team included Tim Mansfield from NICTA, Susan Hansen from CSIRO
and Toni Robertson from University of Technology, Sydney. Tim led the research team
and all team members contributed to the study design. I worked with Susan on
designing and conducting the semi-structured interviews. All members of the team were
involved in the observations although I was the only researcher who participated in all
the observations. I coordinated the field studies and led the data analysis.
4.5 Meetings Within and Between the Hospitals
The MDTMs at both hospitals started around 8am and lasted two to three hours every
Wednesday morning. The formal discussion of patient cases was firstly conducted
locally at each hospital. The timing of the shift to video conferencing meeting was
coordinated by the breast cancer nurses at each hospital. When the patient discussions
ended the video conferencing also ended. The meetings become local once more.
Participants went on to other meeting agendas such as new research in breast cancer and
management matters that were relevant to the group.
The weekly MDTMs at both hospitals followed the standard MDTM guidelines and
protocols that had been developed by the healthcare government authority (NBCC
2005). Both hospitals had implemented the MDTM program for breast cancer care for
eight years before the study while the video conferencing meeting between the two
hospitals was introduced two years before our study. Case discussions were chaired by
one of the senior specialists who ensured the patient cases were presented and discussed
by clinicians of different disciplines. The presentation of each patient case followed the
same order: clinical case summary, radiology findings and then pathology results. These
materials were prepared before the meetings and presented in the meeting by related
clinicians. Other members of the team such as surgeons, oncologist, nurses,
psychologists and social workers expressed their opinions. Discussions about the
treatment and management of a patient can require references to relevant evidence bases
and guidelines. The outcome of these discussions was agreed recommendations about
ongoing treatment and care. The chairperson summarised the recommendations at the
Collaboration in Multidisciplinary Medical Team Meetings
80
end of each patient case discussion. The breast cancer nurses recorded the
recommendations and coordinated with responsible clinicians to inform the patients at
the patients’ next visit to the clinicians. Our interview participants told us that an
effective meeting constituted the following aspects: all relevant patient issues are well
presented; everyone has the opportunity to provide input to the management of patients;
participants work as a team to make decisions and support each other; participants could
learn something about the meeting; and patients were actually effectively treated as a
result of the decision made in the meeting.
The presentations of the radiology and pathology images were central to the discussion
which led to the agreement on the diagnosis, the disease stage and treatment plan (e.g.
Kane & Luz 2006, 2008, Groth et al. 2009). The presentations structured the flow of the
discussion and references to the images made them the central focus of the discussion.
Radiology images included X-ray, CT scan and MRI, either on film or disk. Pathology
images were samples of affected body tissue stained on a glass slide and captured by
microscope. These two types of images needed to be transformed and displayed for the
participants to look at.
The video conferencing meeting was actually embedded in the two local settings. There
were a variety of differences between the two local settings, such as physical setup, size
of the team and the way of preparation and presentation of medical information.
Interaction problems emerged when the local practices of the two local settings were
joined. All interview participants indicated that they were satisfied with the presentation
and discussion of the local meetings but ten out of the eleven interview participants
expressed negative sentiments about the presentations and discussions in distributed
meetings. For example one of the medical oncologists commented:
“The meetings are pretty efficient until you go to the video link and then their
efficiency drops off dramatically and people disengage at that point”.
While the overall research goal was to suggest intervention in the distributed meetings,
it is essential to develop an understanding of the differences in the local meeting
practices and why various arrangements were developed in the particular ways. When
presenting the results of the study, I will first describe the processes of presentation and
discussion of patient cases in the two local meetings in section 4.6. Then I will present
Collaboration in Multidisciplinary Medical Team Meetings
81
the distributed meetings in section 4.7 to develop the discussion of how variations in
local practices affected the interactions in the distributed meetings.
4.6 Local Meetings
The findings will be structured into the categories of physical setup, preparing
information, presenting information, context of different information practices and
conversation and case discussion. These aspects of each local meeting are presented
“side by side” in order to highlight the different processes and practices between
hospital A and hospital B.
4.6.1 Physical Settings
There were twenty to thirty meeting participants at hospital A and ten to fifteen
participants at hospital B. Large public teaching hospitals typically have more
participants due to the MDTM’s role in education, while highly specialized groups, such
as the teams in private hospitals, have fewer people in attendance (Kane & Luz 2009b).
Table 4.2 and Table 4.3 list the participants at hospital A and B. These include “key”
participants who had a formal role in the meeting and “non-key” participants who had
no formal responsibility. There were more non-key participants in hospital A, such as
junior doctors and registrars who might transfer from different departments and medical
students who might be new to the meeting. At hospital A, a surgeon chaired the local
meetings. At hospital B, a surgeon and a medical oncologist took turns to chair the local
meetings each week.
Table 4.2. Participants at hospital A
Key participants Non-key participants two breast cancer surgeons one senior radiologist and one junior radiologist one pathologist one medical oncologist one radiation oncologist one breast cancer nurse
nurses psychologists social workers registrars junior doctors medical researchers medical students
Collaboration in Multidisciplinary Medical Team Meetings
82
Table 4.3. Participants at hospital B
Key participants Non-key participants One or two breast cancer surgeons one radiologist one pathologist one or two medical oncologists one radiation oncologist one breast cancer nurse
one project officer one psychologist one social worker one researcher occasionally one or two junior doctors or medical students
The MDTMs at hospital A were held within the radiation oncology department in a
large “lecture room” which was a common meeting room. This room was used
primarily for presentations so its furniture, console and projection facilities were
organised primarily for this purpose (Figure 4.1). A large projector screen was set up at
the front. The default seating arrangement was rows facing the projector screen. The
video conferencing system and its related artefacts were positioned in a way that could
be accommodated within this arrangement. A table was positioned across the front of
the room. The video conferencing system was set up on the left side of the table with a
75cm TV screen and a video camera. The conference microphone was situated on this
table.
The chairperson sat in the front of room and faced either towards the rows of chairs or
the front projection screen. The four or five key participants sat in the first row of the
audience. The remaining participants sat in the rows of seats at the back of the room and
faced towards the main projector screen at the front. A junior pathologist and radiologist
stood at the console of the room, facing the audience and sometimes turning to see the
projection screen at the front.
At hospital B, the MDTMs took place in a small dedicated meeting room where the six
or seven key participants sat around a semi-oval table with two large TV screens side by
side on the wall at one end (Figure 4.2). The table was shaped so key participants could
easily see their local colleagues as well as the TV screens. There was a light box
mounted on the wall adjacent to the presentation screens. Other participants sat around
the periphery of the room.
Collaboration in Multidisciplinary Medical Team Meetings
83
Figure 4.1. Site plan of the local meeting at hospital A
Figure 4.2. Site plan of the local meeting at hospital B
Collaboration in Multidisciplinary Medical Team Meetings
84
4.6.2 Preparing Information
Usually there were between five to ten cases discussed at each local meeting. The
number of cases for the video conferencing meeting varied each week but usually
around five cases needed to be reviewed. The preparation work involved radiologists,
pathologists and breast cancer nurses. There were commonalities and differences in the
preparation processes in the two hospitals.
Patient history. Surgeons initiated the list of patients to be discussed and informed the
breast cancer nurses. The nurses at each of the hospitals coordinated with their
respective surgeons and circulated the list to the relevant radiologists and pathologists
before the early afternoon on Friday. Since radiology and pathology examinations of the
patient might be performed at different hospitals, the nurses needed to assist in ordering
appropriate materials. The nurses also looked for patient records to generate patient
summaries. There was no electronic patient record system used in the two hospitals at
the time of the study. At hospital A each anonymized summary was put into a
PowerPoint presentation that was used in the meeting and at hospital B each summary
was presented in the form of a cover sheet to be distributed to meeting participants
(Figure 4.3).
Figure 4.3. Preparation of patient histories
Radiology image. At hospital A, radiology images were included in the PowerPoint
presentation. It took six to eight hours, sometimes after work hours, for a junior
radiologist to prepare medical radiology images before the meeting. Since a digital
medical image database - PACS was not used in either of the hospitals, films needed to
be located from physical storage, reports needed to found or faxed from where the
imaging was originally done. Then the radiologist analysed the films and reports, found
Collaboration in Multidisciplinary Medical Team Meetings
85
appropriate images for illustration, captured these images by digital camera, uploaded
the camera images to a computer, used Photoshop to produce the best quality images
and annotate the area of interest using a red circle etc, and put the images into a
PowerPoint presentation (Figure 4.4). At hospital B, radiologists were only involved in
the local case discussions and brought films with them to present directly on the light
box during the meeting.
Figure 4.4. Preparation of radiology images
Pathology image. Before the meeting, the pathologists located the slides and reports
required from various storage places and then looked for the best slides and areas of
interest for the discussion. At hospital A, a pathologist captured the pathology images
by a camera attached to a microscope, uploaded the images to a computer, and inserted
the images and a summary of the diagnosis into a PowerPoint presentation (Figure 4.5).
The total process took around two to three hours. At hospital B, the pathologists spent
around ten to twenty minutes per case to review the pathology slides and mark the area
of interest directly on the slides before the meetings, then brought the pathology slides
to the meetings and presented them directly from a digital microscope which was
integrated with the video conferencing system.
Collaboration in Multidisciplinary Medical Team Meetings
86
Figure 4.5. Preparation of pathology images
4.6.3 Presenting Information
At hospital A, the patient summary was presented by the patient’s responsible surgeon
sitting in a chair in the first row of the audience while the radiology results and
pathology results were presented by a junior radiologist and a pathologist who both
stood at the console. A PowerPoint file incorporated the patient summary with selected
radiology and pathology images. The areas of interest in the PowerPoint radiology
images were annotated before the meeting to assist the multidisciplinary audience to
follow the presentation. The computer used for the presentation was on the console and
the junior radiologist and pathologist were responsible for presenting and navigating
through the slides to support the case discussion. A senior radiologist sitting in the first
row often took over the junior radiologist’s discussion to add his interpretation. He used
a laser pointer on the projected display to make reference to the areas of interest to
support his discussion. The pathology images were presented by the pathologist at the
console who was generally the only pathologist at the meeting.
At hospital B, cases were presented by the surgeons responsible for the patient. The
radiologist used the light box to directly display the radiology films for each patient to
the rest of the participants. She stood beside the light box and pointed directly to parts
of the film during discussion. The pathologist used a digital microscope with a camera
attached to present selected pathology slides. These were displayed on one of the
screens at the end of the table. Areas of interest in the slide were highlighted by
Collaboration in Multidisciplinary Medical Team Meetings
87
physically moving the slides under the microscope so the parts under discussion were in
the centre of the display. This direct slide projection was used both during the case
presentation as well as when questions raised by the other participants in the local
meeting were discussed. When answering questions, the pathologist sometimes referred
to the printed pathology reports placed on the table and adjacent to the microscope.
Slide selection and presentation were then tailored to fit with the immediate needs of the
case discussion.
4.6.4 Context of Different Information Practices
The interview and observation analysis revealed a number of reasons why these specific
preparation practices had developed in each hospital. Firstly, hospital A was a public
teaching hospital where the senior radiologists and pathologists had support from junior
medical professionals. Although the preparation of the presentation was time
consuming, it did not incur any further financial cost. For a private hospital, there were
no teaching responsibilities and neither the senior nor the junior staff was salaried. So
there were no staff to facilitate the preparation of images and their integration with the
various patient summaries in hospital B. Secondly, the situation of a smaller number of
people in a smaller room at hospital B allowed the radiology images on the light box
and pathology images on the TV screen to be clearly visible to people in the room.
Within this setting, there was no need for large projection. However the size of the
meeting at hospital A meant that the images had to be transformed so they were able to
be viewed by presentation on the large projector screen. Finally, hospital A was a
teaching institution and the senior clinicians usually had academic roles in universities.
The materials used during the meeting served as an important resource for subsequent
teaching and research purposes.
Participants at hospital B preferred the way of moving slides around to look at areas of
interest during the meeting as it provided the opportunity of accessing images that were
not prepared previously and therefore better positioning the pathologist to address
unanticipated questions at the meeting. However the PowerPoint presentation method
was highly valued by hospital A interview participants who felt that it was a more
cohesive and straightforward way to present, especially in a meeting which might have
Collaboration in Multidisciplinary Medical Team Meetings
88
ten patients cases to be discussed in one hour. The preferences of participants at hospital
A are captured in the following interview comment made by a pathologist:
“It’s much better to be able to just show – even though it’s a lot more work for
us to do it like that it’s much better for the meeting… a few select areas
photographed and you can just demonstrate it straight away. It’s a much more
cohesive way to present a meeting.”
4.6.5 Conversation and Case Discussion
At hospital A, the overall atmosphere of the meetings was professional and cooperative.
People explored issues of interest in a relaxed way. They shared jokes, sometimes left
their seats to get drinks and lean backward or forward to gossip to the people
surrounding them. The discussion itself was formal and was mostly among the
surgeons, oncologists and radiologists who may from time to time turn to talk to the
psychologists, social workers or nurses who sat at the back rows and had less input to
the discussion. It was a large teaching hospital and there was a hierarchical structure
both within the hospital and also within each of the disciplines. Any individual’s place
in the hierarchy tended to be reflected in where they sat in the room. Medical students
and junior doctors and the majority of these people sat at the back observing and did not
participate in the discussion.
At hospital B, the communication in the local meeting was informal and free flowing.
There were fewer people who all sat around a semi-circular table with good eye contact
with each other. This seating enabled better management of interruptions and general
conversation flow. Organisational hierarchy was much less prevalent than in hospital A.
Participants appeared to be quite relaxed and willing to express uncertainties or
disagreements. For example, the pathologist at hospital B typically received more
questions from the team than the pathologist at hospital A. If necessary he or she looked
for the answers in the reports and moved the slides around on the microscope while
answering.
4.7 Distributed Meeting
Patient cases discussed in distributed meetings involved clinicians from both of the
hospitals. These patients were from hospital B and either the responsible surgeon was
Collaboration in Multidisciplinary Medical Team Meetings
89
from hospital A or the reviewing radiologist was the senior radiologist at hospital A.
The junior radiologist at hospital A was responsible for the preparation of selected
radiology images in PowerPoint including marking the areas of interest with red circles.
The pathologist at hospital B was responsible for the pathology images for the
distributed meetings. Similar to the way of presentation at local meetings, the
pathologist brought the original pathology glass slides to the meeting and presented
them directly from the digital microscope. The image from the digital microscope was
made available to hospital A via the video conferencing connection.
The two breast cancer nurses phoned each other to coordinate the video conferencing
connection after local meetings. When the two sites joined together, participants moved
to a different meeting setting. As mentioned earlier most interview participants reported
a drop in the level of discussion in the distributed meetings compared to the local
meetings and major technical and interaction problems.
I will first describe how both sites organized their physical setups and presenting
information in distributed meetings. Following this, the technological and interactional
problems in video-mediated conversations and interactions will be presented.
4.7.1 Physical Settings
Positions of participants at hospital A. During the distributed meetings, the chairperson
and the other key participants (one surgeon and two oncologists) moved from the rows
of seats to sit around the table in front of the TV screen and camera (Figure 4.6). The
other participants remained in the rows of seats at the back of the room and faced
towards the main projector screen at the front. Similar to the local meeting, the junior
radiologist stood at the console in the front right corner, facing both the participants at
the video conferencing table and those in the rows of seats.
The camera in the video conferencing unit pointed towards the table and the console but
not to the back of the room. The core team members could be seen by the remote
participants at hospital B. However those peripheral participants sitting in rows towards
the back of the room could not be seen by the people at hospital B. These peripheral
participants were also a long way from the microphones on the table at the front and it
was difficult for their voices to be heard at hospital B.
Collaboration in Multidisciplinary Medical Team Meetings
90
Figure 4.6. Site plan of the room for distributed meeting at hospital A in distributed meeting
Screens and information on the screens at hospital A. The video conferencing screen
showed video streams from hospital B which automatically switch between a view of
the participants at hospital B or the pathology images from hospital B. The project
screen showed either the local PowerPoint presentation or the video feed of the
participants at hospital B or the pathology images from hospital B. However, this switch
between views needed to be done manually by the junior radiologist standing at the
console.
The core members, seated at the table, either turned to the projection screen or the video
conference screen. The participants in the rows of seats watched the projection screen
only. They relied on the junior radiologist standing at the console to manually switch
between different views.
Positions of participants at hospital B. Participants were already seated facing the video
conferencing system in local meetings so there was no seating change (Figure 4.7). The
key participants sitting at the shaped table could easily see their local colleagues as well
Collaboration in Multidisciplinary Medical Team Meetings
91
as the display screens. The smaller size of the room, the smaller number of participants,
the appropriate sized displays and purpose-built video conferencing setup at hospital B
meant they could all be seen and heard by the remote participants at hospital A.
Screens and information on the screens at hospital B. The two display screens at
hospital B were well positioned so that everyone in the room could see what was
displayed on them, that is, they all watched the same material displayed at the same
time. However only one of the screens was used to display both the video feed of the
core participants at hospital A and the radiology or pathology images being discussed.
The remote room view reduced to a small picture-in-picture display when the medical
images were being displayed. The second screen always displayed the video feed from
their own camera – the view of themselves which was clearly not helpful. This display
configuration was set up by the company supplying the video conferencing system.
While the participants at hospital B were vocal in their criticism of this configuration
nobody in the team knew how to change it.
Figure 4.7. Site plan of the room for distributed meeting at hospital B
4.7.2 Presenting Information
Presentation. Both hospitals used the patient summary and images they would normally
use in local meetings. The prepared PowerPoint presentation of radiology images from
hospital A was used in the presentation and discussion along with the pathology slides
from hospital B. A typical discussion started with a surgeon (either hospital A or
Collaboration in Multidisciplinary Medical Team Meetings
92
hospital B) verbally reporting the case. In hospital A, the junior radiologist displayed
the prepared radiology PowerPoint slides while presenting the patient summary. The
pathologist at hospital B presented the pathology slides directly from the slides on the
microscope.
Focusing group attention. When the junior radiologist at the console presented the
PowerPoint slides he used the computer cursor as a pointer to the relevant area of
interest. The cursor movements were visible on the projection screen and the video
conferencing displays at hospital A and B. The senior radiologist sitting in the audience
provided further explanation and used a laser pointer to point at relevant parts of the
images displayed on the projection screen at the front of the room. At hospital B the
PowerPoint presentation was shown on the video conferencing screen. In order to make
the senior radiologist’s referencing gestures visible to the participants at hospital B, the
junior radiologist had to move the computer cursor over the image displayed on the
computer screen following the laser pointer gestures made by the senior radiologist on
the projection screen.
Similar to the practice in the local meeting, at hospital B the presenting pathologist
verbally indicated areas of interest on the images while moving the slide on the
microscope so these areas were in the centre of the display. No computer cursor was
used to facilitate this activity.
4.7.3 Technology-mediated Conversations
In comparison to the local meetings, distributed meetings appeared to be more formal.
Participants in general were perceived by other members to be less willing to ask
questions, discuss issues and resolve disagreements. Besides the inter-personal
relationships between the clinicians at these two different institutions, another reason for
this lack of engagement was the participants’ “gap” felt introduced by the setup of the
rooms and video conferencing facilities. These problems resulted in interruptions to the
flow and ease of the conversation. The following comment made by a surgeon
exemplifies the sentiments expressed by the participants:
“In the local meeting, I think people feel that they have the opportunity to
express their views…everyone’s views are listened to and valued. I think any
disagreements about management can usually be appropriately resolved without
Collaboration in Multidisciplinary Medical Team Meetings
93
any aggravation….people who attend our meeting actually get on well with each
other and get on in a cooperative manner that’s in the interests of the
patient…there is much less cross-interaction between the two groups than there
should be.”
As described before, the peripheral participants at hospital A sitting in the rows of seats,
including the senior radiologist, were outside the camera capture area. Their
participation in the discussion was compromised by the fact they needed to get up and
move to the table at the front of the room if they were to be seen and heard by those at
hospital B. The issue of lack of social and spatial awareness due to this limited visual
information was evident from the instances of unawareness of individual people’s
presence, such as asking “Is X there or not?”, “Is Y sitting in the background?”.
Interview participants (e.g. a breast cancer nurse) expressed their desire to see the whole
room of the remote site this way:
“It certainly does limit… The body language is not often seen on a video link,
particularly if not everyone in the room is able to be seen on the screen, and that
could be part of why one end doesn’t talk so much because people can’t see you
know somebody on the screen”
4.7.4 Synchronising Conversation and Image Sharing
As explained in section 4.7.1, at hospital A the key participants watched the video
conferencing screen where the views of hospital B participants and hospital B pathology
images switched automatically in response to the pathologist operation of switching
on/off the camera display from the microscope. However, for the peripheral participants
to see this change, it required the junior radiologist at the console to manually switch
between the room view of hospital B and pathology images to be displayed on the
projector screen. There were a number of occasions during the meeting we observed
that this manual switch did not occur because the radiologist forgot to do so. This meant
the various participants were not seeing the same things. It caused frustrations of the
peripheral participants but neither the key participants at hospital A nor those in hospital
B were aware of this breakdown.
Another related problem was extra work required for the junior radiologist to use the
computer cursor to trace the gestures of the laser pointer being used by the senior
Collaboration in Multidisciplinary Medical Team Meetings
94
radiologist at hospital A so they could be seen at hospital B. This synchronization relied
on the junior radiologist to have a quick understanding and response to the laser
activities in order to allow participants at hospital B to see the references to areas of
specific interest while listening to the senior radiologist.
Another source of frustration for participants at hospital A stemmed from the direct
display of pathology slides at hospital B. The pathologist moved the slides round
according to the ongoing discussion so that the area of interest was in the centre of the
display. This practice worked well for those participants at hospital B who experienced
this selection and movement of slides in real time. However the transmission time lag
between these actions at hospital B and the actual display of the images at hospital A
meant that the image rendering often lagged behind the actual talk. It led to the
decoupling between what was visible to participants at hospital A and the talk
accompanying it. This caused both frustration and eventually disengagement for the
participants at hospital A who all identified it as major problem in the distributed
meeting. A radiation oncologist expressed it:
“They relay down to us the pathology which they’re looking at down a
microscope… then they move the slide and it goes all out of focus and it takes
two or three seconds to come back into focus…”
In summary, there were regular breakdowns in the synchronization of different
subgroups of participants to follow the conversation and images. These were
compounded by the fact that this breakdown was not always visible to the other
subgroups for whom the synchronisation of conversation and images was not affected.
4.8 Exploration of the Physical Setup
Interview participants were encouraged during the interviews to think of solutions to the
problems. Two of the participants drew their designs of the physical setup based on
their expectations and experience. A medical oncologist from hospital B generated a
sketching of a new layout arrangement at hospital A (shown at the left of Figure 4.8)
with the rows of seats at the back of the room arranged in semi-circles around the main
table. Another example was drawn by the senior radiologist at hospital A who illustrated
an advanced system supporting the multidisciplinary team meeting he had seen in
Switzerland (shown at the right in Figure 4.8). It shows a ramp-seat room with a good
Collaboration in Multidisciplinary Medical Team Meetings
95
setup of microphones and displays, access to digital imaging database and a seamless
way of presenting in the meeting.
Figure 4.8. Examples of clinicians’ design sketching
Frustrations with the “technical” problems led to attempts by the team members to
improve the situation by changing the physical setting at hospital A. They tried to
arrange the rows of seats at the back of the room into semi-circles. They turned around
the video conferencing unit and the central table, so that the camera faced the central
table and the rest of the participants were sitting in semi-circular rows of seats. This
layout is similar to that of the left picture of Figure 4.8. With this new arrangement, the
participants in the rows of seats were able to participate more actively in the discussion
and did so. However because this meeting room was a shared common facility then
everything had to be returned to its original position after the meeting. Despite the
acknowledged improvements, the rearrangement of the space before the meeting did not
continue after these initial experiments.
4.9 Discussion
The study has shown that the MDTM collaboration between different settings using
collaboration technology is characterized by a range of asymmetries caused by
variations of local practices. Relating to the asymmetry of physical space were the
physical layout of the room, the size of the room, the arrangement of the camera and the
microphone and the size and position of the displays. The context of different teams in
different hospitals contributed to another set of asymmetries: the numbers of
participants (large vs. relatively small), the structure (hierarchical vs flat), the
presentation of medical images (static vs dynamic) and preferences about how the
images are displayed or transmitted to the others. The analysis presented here highlights
how these variations in local settings impact on the interactions in the MDTMs.
Collaboration in Multidisciplinary Medical Team Meetings
96
Importantly, in presenting the findings I have shown how these different local
arrangements and practices had developed in the two hospitals.
I will discuss the implications these particular practices and settings have for the design
of collaborative workspace, focusing on design of physical spaces in distributed
MDTMs and design to support the information sharing practices. Challenges of
integrating a collaborative workspace into a healthcare context and enabling particular
local practices will also be discussed.
4.9.1 Designing Physical Space
Social interaction. One of the issues that emerged from the study relates to social
interaction in media space which has some core aspects relating social cohesion of the
group and awareness of remote participants’ presence and the communication and task
status in distributed meetings (Aoki & Tang 2009). Kane and Luz (2006) and Groth et
al. (2009) have revealed issues in coordination and awareness in video-mediated
communication based on their analysis of vocalization patterns in MDTMs. The study
presented in this chapter showed how particular spatial arrangements affect the
dynamics of the interaction, participant’s awareness of who was speaking and the
coordination of the discussion. There were problems of visibility and identity of remote
participants and audibility of remote participants, especially for the hospital A
peripheral participants who were not visible to the participants at hospital B. These
problems were caused by the setups of the seating and the audio-video devices,
limitation of the audio-video quality and especially the multiple displays where
sometimes what was displayed was not necessarily the same for different participants.
The study showed that in distributed meetings, these problems affected the spontaneous
conversation and open discussion which are important features of an effective meeting.
One of the fundamental design challenges is to tackle the visual and perceptual issues
that are caused by the asymmetry and fracture ecology in video-mediated
communication for individuals (e.g. Gaver 1992, Luff et al. 2003). The integration of
audio-video devices in the physical space needs to be carefully designed to produce the
right interplay between infrastructures, artefacts, shared understanding and activities
(Binder et al. 2004, Buxton 2009). Effort to support synchronous communication is also
required to concentrate not only on ‘same time, different place’ issues, but also to better
Collaboration in Multidisciplinary Medical Team Meetings
97
support ‘same time, same place’ co-located communication in distributed meetings
(Kane et al. 2013).
Function-space relationship. The meeting space arrangements in MDTMs were affected
by the division of roles and activities of the different participants (Kane & Luz 2013). In
MDTMs, different roles can correspond to different locations in the room (Frykholm et
al. 2012). For example, radiologists and pathologists presenting images sat either at the
console or equivalent, or next to the devices used to display medical images, the person
chairing the meeting sat at the central table, the other key team members sat at the
centre of the camera view and the less involved participants sat in the peripheral areas.
The study has shown that the spatial arrangement of the participants could influence the
interaction patterns in distributed MDTMs. For example, the involvement of peripheral
participants was improved with seating re-arrangements in hospital A where peripheral
participants were more visible to the remote team and had better visual access to both
the remote and local participants. These arrangements could also ensure, by the design
of the physical setting that participants all look at the same thing at the same time
(Henderson 2009). Note that the point here is not to change the various social functions
of the various participants. It is instead to think about how space can be used to support
these functions particularly the interactions which enable them (Fitzpatrick et al. 1996,
Nova 2005, Luff et al. 2009).
This study has shown how the shared visual space influenced the interactions between
team members and the sharing of information, for example the problematic positioning
and control of the visual spaces in hospital A. An optimal arrangement with the right
size and position of the display could support participants’ sense of co-presence,
including their perceptions of non-verbal cues like gesture and gaze, and their
awareness of others’ reactions.
However I am not arguing for the requirement of a sophisticated telepresence setup in
specialised "board" meeting room applications. The purpose that defines an appropriate
physical space in distributed MDTMs is to support the construction of a common
information space among team members (Bonnon & Bodker 1997, Bossen 2002). The
job of the technology is to enable people to negotiate shared understandings across
differences in disciplines by mediating their capacities to see, talk and gesture with each
other (Robertson 1997, 2002, Buxton 2009). Any computer-mediated technology
Collaboration in Multidisciplinary Medical Team Meetings
98
introduced into MDTMs needs to reflect and respect the space-function relationship so
that appropriate information can be distributed at the appropriate locations within that
space.
Challenge of local constraints. The study has also shown that different room settings
were constrained by their organizational contexts. The facilities in hospital A evolved
from pre-existing common meeting rooms and did not necessarily provide an ideal
environment for distributed MDTMs. While the teams at hospital A noticed
improvements after reconfiguring their seating arrangements, there were tensions
between the demands of the local meeting and the distributed meeting as well as
tensions arising because the MDTMs were held in a multi-purpose room that was also
used by other members of the hospital. The issue here is not that optimal setups can
better support particular demands of distributed MDTMs but that what constitutes an
optimal setup needs be considered within the broader set of arrangements and evolved
work practices that define these spaces (Fitzpatrick et al. 1998, Dourish et al. 2006b).
The tensions between these arrangements introduce additional work and effort for
participants that can be barriers to the implementation of more optimal settings (Martin
et al. 2006, Ellingsen and Monteiro 2006).
4.9.2 Supporting Information Sharing
Challenges of variations in information practices. Different ways of presenting medical
images in distributed MDTMs had both advantages and disadvantages and there were
always tensions between traditional ways of viewing medical images and computer-
supported alternatives. For example pathology images could be projected directly from
slides or digitized and displayed using digital display and presentation technologies,
radiology images could be digitized for computer access or directly shown from a light
box. Manually digitizing medical images so they could be embedded in a PowerPoint
presentation was time consuming. However PowerPoint presentation was
straightforward and saved time searching for images during meetings. Both ways work
sufficiently well in their local settings that there was no local imperative to change.
However there were interaction and communication problems caused by different local
practices in the distributed meetings.
Collaboration in Multidisciplinary Medical Team Meetings
99
The medical image presentation and preparation practices in the two hospitals varied in
both format and content and this issue had its source within different health service
structures. There were differences in the availability of resources to support MDTM
preparation between public and private hospitals. The underlying issue is the particular
mix of public and private health service structures in Australia and the particular
relationships and practices that had developed to integrate the two systems. These
include very different funding practices and responsibilities. While participants at
hospital A considered their way ‘the better way’, participants at hospital B were quite
happy with their own solution. A public, teaching hospital might consider recording the
material presented in MDTMs for research and education purposes. Private hospitals do
not maintain the infrastructure that enables resources to be allocated to the preparation
of images for presentation in MDTMs. Moreover, there were no existing mechanisms at
hospital B to change this particular practice. Any technical solutions will need to be
sufficiently flexible that these very different requirements in the sharing of medical
images can be accommodated and integrated (Schmidt et al. 2007, Randell et al. 2011).
Interaction needs. The study highlights the importance of providing an "interaction
space" for the person chairing the meeting and the radiologists and pathologists who are
the “human mediators” in distributed MDTMs (Bannon & Bodker 1997). These key
participants not only assist in the production of content for the information spaces in
MDTMs, but also play the major role in interpreting the meaning of information for
those who might wish to use it (Kane & Luz 2013). As already discussed, their physical
position affects their interaction with other team members both co-located and
distributed. They need to position themselves to enable easy access to devices and
materials at the appropriate time (Buxton 2009, Luff et al. 2009). The study has also
shown that different meeting participants had different roles and their interaction needs
varied. Most of the non-key participants only needed to observe the information being
shared while key participants engaged in the interactions with information.
The study supports the recognition that participants need the support of technology to
synchronize their interactions with information across physically distributed sites, in
particular to support the pointing and other indexing gestures (Luff et al. 2009, Olwal et
al. 2011) as well as to solve the fractured ecologies problem in viewing images (Luff et
al. 2003, Zuiderent et al. 2003). In the design of technology to support distributed
Collaboration in Multidisciplinary Medical Team Meetings
100
collaboration these indexing actions need to be simultaneously available to those acting
and those perceiving the actions in the shared work space (Robertson 1997, Buxton
2009). Interaction problems resulted if these human mediators were not near the
computer displaying the images and/or if they were using a microscope or light box
which does not have a pointer that is represented in the shared workspace. Otherwise
some other participants needed to manually translate these actions to a form that can be
visible in the shared space, such as, in our studies, controlling the direction and zoom of
the camera capturing film images or using the computer cursor to follow the path of a
laser pointer.
4.9.3 Challenges of Integration
Kane and Luz (2009b) suggest an ideal room for distributed MDTMs – the room “will
have high speed wireless network, utilise ubiquitous devices to record the presence of
individuals, will maximize the visual display area potential of the space and have
enhanced audio support. The room would be equipped with enough visual display area
to allow comparison of several images at once, and allow for the simultaneous display
of the bronchoscopy image, microscope image and/or video taken in the operating
theatre. A record, or outcome, of the discussion will be available for review, as needed,
afterwards.” (p. 388). Multi-display environments and shared digital workspace
technologies and prototypes that seamlessly integrate information interaction in large
interactive displays and dedicated physical setups might potentially solve the problems
in supporting collaboration in distributed MDTMs. However the design of a
collaborative workspace to successfully support distributed MDTMs requires more than
the provision of elegant architectural solutions and advanced interactive display
technologies (Kane & Luz 2009b). People's behaviour and interactions, including their
information sharing practices in a complex healthcare environment, are different to
those that might have been possible in the controlled settings where these technologies
are often designed.
It is not just physical layout alone that can provide a solution to the coordination and
communication of MDTMs. It is the arrangement of the physical spaces that are suitable
to participants and particular practices. There are differences in the preparation before
the meeting, the way of running the material and coherent narrative in the meeting. The
Collaboration in Multidisciplinary Medical Team Meetings
101
interaction design for an integrated workspace will require new technologies that can
support a wide variety of existing local practices including, for example, the integration
of image viewing practices using a light box and the support of live presentation of
pathology images at hospital B. Again, the problems arising as these practices were
brought together may suggest particular solutions, such as facilitating digital image
systems or increasing network bandwidth to solve the image delay problems. However
it is important not to just consider these solutions in the context of the immediately
available local arrangements and any possible re-configurations. It is also important to
understand the complex social, technical and organizational factors that shaped how
these arrangements and configurations emerged and were experienced in the first place.
The ways in which these practices may evolve in the future will be influenced by
changes in technology and health policies. PACS can provide direct access to digital
images and solve the issue of the work involved in digitising images. There was a clear
interest from the radiologists at hospital A in introducing a PACS to the MDTM to
lighten the meeting preparation work and to support the research and teaching activities
of the hospital. However, a radiologist at hospital B expressed concern that a PACS
would introduce technical problems into a setting. Kane and Luz (2006) observed that
PACS would not resolve problems with maintaining image integrity and exchange due
to the lack of standards in different institutes. It also causes inefficiency because
searching for unprepared images from large amount of data in PACS during the meeting
is not straightforward (Kane & Luz 2013). Similar to PACS, the development of
electronic patient records poses special challenges for their use in MDTMs (Groth et al.
2009, Frykholm et al. 2012). Electronic support for case summaries and record keeping
of the meetings will need to be developed. These challenges still relate to the
relationship between the information space and the meeting practices in MDTMs as
well as configuration issues (Binder et al. 2004, Balka & Wanger, Balka et al. 2008)
that need to be supported in the design of collaboration technologies for MDTMs.
4.10 Conclusion
This study examined the MDTMs in the two hospitals and particularly focused on the
variations in local practices both in the local settings of each hospital and in the
distributed setting when the local meetings were linked. The results have shown how
Collaboration in Multidisciplinary Medical Team Meetings
102
key factors such as the physical setup of the meetings, the presentation of the patient
cases, the preparation of images used in patient discussion and the asymmetries of these
factors across the sites clearly influenced the dynamics of collaboration and experience
with the distributed MDTMs. Designing a collaborative workspace for MDTMs needs
to take account of how these arrangements and practices arise and the implications these
have for technical interventions.
The study has shown that careful configuration of physical space can potentially
improve the social interaction and communication within MDTMs. Design also needs to
consider appropriate integration of information sharing functions and interaction tools to
allow for effective patient information discussion which is the centre of the
collaboration. Importantly, this study has highlighted that designing collaborative
workspaces to support MDTM collaboration across variable local settings needs to
accommodate local constraints and to provide appropriate solutions to support the
configuration of available technologies and resources to enable local practices.
Distributed Collaboration in Emergency Response on Animal Disease
103
5 Case Study: Distributed Collaboration in Emergency Response on Animal Disease
In this chapter I present the study of collaboration in emergency animal disease
response which focuses on the work of high-level analysis and decision making by the
Australian Consultative Committee for Emergency Animal Disease (CCEAD). CCEAD
is a geographically distributed committee established to recommend action plans during
animal disease outbreaks at a national level. The study investigated the ways in which
the CCEAD members shared and analysed information together, with particular
emphasis on their teleconference meetings across multiple sites. The aim of the study
was to develop understandings of this work that could inform decisions about where
appropriate collaborative workspace intervention could be made to facilitate the
information sharing and decision-making performed by this committee. This study
allowed me to explore the collaborative practices of a highly distributed set of teams
across a large number of sites and the particular design challenges this entails.
5.1 Emergency Response on Animal Disease
Emergencies are critical situations that can cause damage to life and property and
require immediate response to minimize adverse consequences (DHA 1992). Examples
of emergencies and responses include firefighting (Jiang et al. 2004, Denef et al. 2008),
emergency medical care (Reddy et al. 2001, Kristensen et al. 2006), search and
evacuation, natural disaster and crisis (e.g., hurricane, nuclear disaster) (Palen & Liu
2007, Pipek et al. 2012). The main characteristics of emergency events are their
unforeseen occurrence and the need for coordination and immediate response
(Kristensen & Kyng 2009).
An emergency response activity usually involves several teams from different
organizations working cooperatively to eliminate or reduce the impact of the event
(Diniz et al. 2008). Management and response to emergencies requires fast and effective
action in collaboration between different teams in order to meet the dynamic challenges
in the situation (Kyng et al. 2006). These teams usually follow established procedures to
deal with emergencies contained in emergency plans. Factors such as speed of events,
Distributed Collaboration in Emergency Response on Animal Disease
104
number of teams involved, time available to make decisions, resource availability and
stress create challenges to team collaborations in complex situations (Pipek et al. 2012).
This case study was of a particular sector of emergency response, namely that of
emergency response for infectious animal disease outbreak. Diseases such as foot-and-
mouth disease, mad cow disease and Hendra virus are examples of emergency animal
diseases. The consequences of infectious disease outbreak within an animal population
are potentially enormous. In addition to the basic concerns about animal welfare, such
outbreaks can cause significant social disruption and environmental damage through
ecological imbalances introduced. The economic consequences of these outbreaks are
huge. The epidemics of foot-and-mouth disease in the UK in 2008 were estimated to
have cost in the order of several billion dollars. This is not simply the immediate costs
of the livestock or disease intervention programs but also the costs of ongoing
disruption to trade links that extend well beyond the disease outbreak itself.
Given the potential consequences of such emergency animal disease outbreaks, a strong
national biosecurity management infrastructure for emergent animal disease has been
developed by the Australian government. One component of the initiative has been the
establishment of the Consultative Committee on Emergency Animal Disease (CCEAD).
CCEAD is a technical and operational committee devoted to the operational
management of emergency animal disease. It is a coordinating body set up to provide a
technical link between the Commonwealth, States, Territories and affected industries for
decision-making during animal health emergencies.
Membership of CCEAD comprises the Commonwealth Chief Veterinary Officer (CVO)
in Canberra, state and territory CVOs, management representatives from the
Department of Agriculture, Fishery and Forestry (DAFF) in Canberra and the Director
of the Australian Animal Health Laboratory (AAHL) which is in Geelong (Figure 5.1).
In addition, relevant expertise is drawn in as necessary, such as diagnostics experts from
AAHL. According to the particular disease there will also be representatives from
affected industries. For example, in the case of equine influenza, representatives from
the horse-trading and horse-racing industries will be included. In the case where there is
also a threat to human health, the Chief Medical Officer may also be present as in the
case of the Hendra virus outbreak.
Distributed Collaboration in Emergency Response on Animal Disease
105
Figure 5.1. Membership of CCEAD
CCEAD is chaired by the Commonwealth CVO and coordinated by a couple of
dedicated secretaries working at DAFF in Canberra. The Commonwealth CVO and the
management team in Canberra lead the overall work on risk assessment, surveillance,
laboratory diagnostics, veterinary epidemiology and administration. State CVOs
maintain emergency animal disease monitoring and control at the state level based on
the epidemiological data analysis of the disease distribution. AAHL provides diagnostic
services based on its expertise and facilities to support evidence-based decision making
which is one of the major tasks of the committee. Industry representatives are involved
in decisions about the management of animals and trade. These representatives can help
provide an economic perspective on how particular interventions such as movement
restrictions or a culling policy would impact on the affected industry.
The committee is essentially a distributed set of representatives that collaboratively
analyse information and discuss the strategies for dealing with an emergency animal
disease outbreak. The committee represents various vested interests and its membership
includes core members as well as ad hoc members drawn in as necessary for a particular
disease or point during an outbreak. CCEAD members are geographically dispersed
across Australia. CCEAD meetings using teleconferencing are held throughout the
entire outbreak course of an emergency animal disease until the disease is demonstrably
eradicated.
Distributed Collaboration in Emergency Response on Animal Disease
106
5.2 Related Work
There is a growing interest in CSCW and HCI research into the role information
technology can play in supporting emergency response. Sociological, organizational and
technical issues have been identified as major obstacles that need to be addressed in the
development of effective emergency information and collaboration technologies (Manoj
& Baker 2007, Kristensen & Kyng 2009). Research efforts have sought to understand
the challenges of time-critical communication and coordination as well as the
collaboration between disparate groups in a variety of different emergency response
domains (e.g. Palen & Liu 2007, Chen et al. 2008, Pipek et al. 2012, 2014). To my
knowledge no study has been reported on the emergency response for infectious animal
disease in CSCW and HCI. Each of the different sectors under the broad umbrella of
emergency response has its own unique characteristics and socio-technical challenges.
The work of the CCEAD is not about in-moment coordination activities. Rather it deals
with higher-level strategic recommendations for intervention and involves distributed
meetings and information sharing across multiple organizations. Issues of collaboration
across different settings have been reviewed in Chapter 2. In the context of supporting
communication and information sharing in CCEAD, this section will outline three
related social-technical challenges: collaboration across multiple organizations,
common information space and collaboration technology development.
One of the socio-technical challenges has been supporting intra- and inter-
organizational collaborations involving different teams. Klann et al. (2008) and
Kristensen and Kyng (2009) explore the design of interaction technologies to support
communications in emergency response and highlight the need to understand the
organization and division of work as well as the characteristics of collaborations and the
ways in which they have unique implications for technological interventions. Pipek et
al. (2012) at a workshop of the CSCW conference argue that emergency response can
be considered as a “continuous social process of a network of interdependent actors and
organizations” (p7). Studies of how technology design may be undertaken to achieve
interoperability in emergency response suggest the importance of understanding the
loosely coupled structures and highly heterogeneous environments (e.g. Mendonca et al.
2007, Pipek et al. 2012, 2014). Olson et al. (2009) in their paper of “What still matters
about distance” illustrate three classes of distributed work typologies by using a team of
Distributed Collaboration in Emergency Response on Animal Disease
107
four: hub-to-spoke (3-1), hub-to-hub (2-2) and fully distributed (1-1-1-1). Collaboration
practices of different sized teams can be significantly different in several ways (Mark et
al. 2003). These include multiple actors, subgroup interaction, physical environment and
the display settings required (ibid). Most of the literature focuses on understanding the
collaboration in hub-to-hub configuration in which there is a critical mass represented at
the headquarters and remote hubs (Hinds & McGrath 2006). Koehne et al. (2012)
conducted a study to investigate how individual remote members develop strategies to
cope with the challenge of working alone with other remote members. They argue that
the same strategies and technical solutions that resulted from the symmetrical hub-to-
hub setting may not generalize to the other settings. There is a need to understand the
collaboration in the other types of configuration, especially the mechanics of how the
asymmetrically configured groups collaborate across distance in real time (Koehne et al.
2012, Olson & Olson 2013).
Another social-technical challenge relates to the issue of information sharing. For the
work of CCEAD, there is a particular challenge in constructing common information
space across large scale and distributed communities. Studies have shown that
coordination problems in emergency responses are generally caused by breakdowns in
information sharing and the communication process (Diniz et al. 2008, Ley et al. 2012).
Information is disseminated within and between different actors and organizations through
different channels and entities. Contextual information is generated from the
development of the events and actions carried out by the teams. The prompt distribution
and sharing of related information can play an important role to build shared
understandings between different teams and support the actions carried out by different
teams (Denef et al. 2008). Ley et al. (2012), in their study of improvisation work in
inter-organizational emergency response, found that there is a lack of awareness about
what information is available for the team involved since information is mostly
distributed. They outline implications for the design that focus on geographically
visualized data about information and collaboration resources for collaborative situation
assessment. Importantly, supporting the construction of common information space is
not just to provide access to information (Bannon & Bodker 1997). Harrald and
Jeffereson (2007) explore how to achieve common situation awareness for emergency
response teams in the information sharing process and highlight that the emphasis
should not only be on the data transformation process but also the sense-making
Distributed Collaboration in Emergency Response on Animal Disease
108
processes of the team. Supporting common information space for a large-scale team
also requires the development of custom components to increase flexibility for the end
users as reviewed in Chapter 2 (e.g. Rolland et al. 2006, Hjelle & Jarulitis 2008). Ley et
al. (2012) suggest solutions for individualization of information composition to allow
individual users to annotate and add new information for their personal needs when
sharing information in emergency response.
A range of technologies have been investigated in different types of emergency
responses. These include formal control rooms, mobile and wearable computing, large
public displays for group coordination and mobile incident command, distributed
collaborative systems, geospatial technology and more recently the social networking
sites and crowd sourcing approaches (Jiang et al. 2004, MacEachren et al. 2006,
Landgren & Nulden 2007, Heard et al. 2014, Ginige et al. 2014). Among these, there
are a number of explorations in designing collaboration technologies. Kyng et al. (2006)
in their study of emergency incident response investigate the challenges of designing an
interactive system for time-critical collaboration. The challenges they identified range
from equipment and communications to the professionals and information technology
supporting these. Ginige et al. (2014) demonstrate a spreadsheet-based collaborative
system built in mobile devices to support information sharing among disaster
responders. Heard et al. (2014) present a real time information gather and share system
which is a web-based visual collaborative environment and is designed to facilitate
teleconferencing over maps. They emphasize the needs to support flexible
customization of shared information according to roles of participants and integration of
a variety of data to be overlaid onto a map.
5.3 Research Context and Motivation
When there is a disease outbreak, the CCEAD convenes and members begin to meet
regularly. Depending on the nature of the disease, likely prognosis and economic
impact of a particular disease, the meetings happen more or less often. In the case of a
major outbreak such as equine influenza, the group initially would meet for several
hours every other day with this gradually slowing down to weekly and fortnightly
meetings as the disease begins to show signs of being under control. Because the
CCEAD consists of members located in different states, and because of the frequency
Distributed Collaboration in Emergency Response on Animal Disease
109
with which these meetings have to occur during an outbreak, it is not logistically
possible for all of the groups to get together in a co-located setting. Rather, it is a
necessity to conduct these meetings in a distributed manner. The CCEAD meeting can
be considered a hub-to-spoke setting and the Canberra site can be considered a hub site
given the chairing role by the Commonwealth CVO. A significant challenge faced by
the committee is for its members to collaborate, across multiple locations and groups,
representing different sectors and states, to share information in a timely manner and
make effective and appropriate evidence-based decisions.
An effective collaboration platform was identified by key members of the CCEAD as
one of the solutions to enhance the connectivity and information management in
emergency animal disease response. In order to understand the work practices of the
CCEAD and provide insight into what and how appropriate ICT interventions can be
made to facilitate their collaboration, I worked with a senior researcher in CSIRO to
conduct the workplace study in CCEAD and their distributed meetings.
It was interesting to me at the beginning of the fieldwork that such distributed meetings
were conducted with very basic collaboration and information technology. For
example, the core technology for communication was teleconferencing facilities which
supported audio connections only while a large number of documents were also
presented on paper with no basic capability for sharing these in real time. Fax machines
were a necessary part of the communication infrastructure. Given the apparent lack of
sophistication in these information and communication technologies, I was keen to
understand why they remained, how they supported work practices in positive ways and
how they hindered any communication, analysis and decision-making.
Findings from the MDTM study had showed differences in local practices and how
technical, organizational and procedural factors shaped the particular ways of
interactions in distributed collaborations. The understandings obtained from MDTM
study helped me to develop the investigation focus of the collaboration practice in
CCEAD, including particular ways information and communication artefacts were
configured and arranged during particular situations and why such arrangements came
about.
Distributed Collaboration in Emergency Response on Animal Disease
110
5.4 Methods
The study included a series of interviews with key members of the CCEAD and
supporting groups and was conducted over a period of five months. In total twelve
participants from four different sites were interviewed. These participants included the
Commonwealth CVO, two CCEAD secretaries, two state CVOs, three key members of
AAHL in Geelong and four members of the management team at the DAFF in
Canberra. Where possible these interviews were conducted in situ at the place of work
and lasted approximately one hour – though some participants were interviewed via
telephone. A larger focus group meeting involving 15 participants from the DAFF was
conducted in Canberra at the beginning of the study. The participants worked in various
jobs supporting the work of the committee, including core members of CCEAD, the
secretariat for the committee and the administrative managers. The purpose of the focus
group meeting was to understand collaboration practices relating to CCEAD and the
physical environment of the Canberra site. I worked closely with Kenton O’Hara, a
senior CSIRO researcher, in conducting the interviews and focus group meeting. I was
responsible for the data analysis and Kenton provided overall guidance on the study.
The interviews focused on the distributed collaboration practices of the CCEAD during
teleconferences but also in the surrounding work before and after particular meetings. In
building a picture of the work practices, efforts were made to ground the discussion in
the details of actual disease outbreaks and the major focus of the work for the
committee in recent years. These details provided richer descriptions of the work than
discussing the work purely in terms of abstract generalisations. The articulation of the
work also included particular ways information and communication artefacts were
configured and arranged during particular situations and why such arrangements came
about.
In situ interviews were combined with site visits to the DAFF in Canberra and the
AAHL in Victoria. The physical set-up and the arrangement of various artefacts in the
key rooms, where committee members participated in the meetings, were documented
through photographs.
Interviews were audio recorded and transcribed for later analysis drawing out key
themes relating to why work practices were organised in particular ways and the
Distributed Collaboration in Emergency Response on Animal Disease
111
implications this had for future ICT interventions in support of their distributed
collaboration. The focus group discussion and site visits were video recorded and
subject to similar analytic orientation.
It is worth mentioning again the difficulty in the study of collaboration in emergency
response in general and in this study. Since emergency response meetings cannot
always be anticipated, systematic research is difficult to plan for. The intention of this
study was to combine interview with observations of actual meeting practices of the
CCEAD. Unfortunately due to the unpredictable nature of emergency disease outbreak,
there was no opportunity to observe the committee meeting in action.
There were no design activities carried out for supporting CCEAD collaboration
immediately after this study. This was because the management of AAHL was keen to
introduce collaborative workspaces to support the collaborations within AAHL and the
CSIRO design team decided to focus on the AAHL work first. However, the CCEAD
collaboration was revisited two years later when a collaborative workspace was
developed and used in AAHL. Some of the CCEAD members at Canberra were
interested in setting up a similar collaborative workspace at the Canberra site to support
the CCEAD collaborations. A one-day workshop was conducted by me and a second
CSIRO researcher at DAFF to explore this potential development. Two CCEAD
members, three biosecurity service managers and two IT infrastructure managers of
DAFF participated in the workshop. Although this workshop was not planned as part of
the original study, the findings were related to the overall study of the collaboration in
emergency response work of CCEAD and will be presented in section 5.6.
5.5 CCEAD Meetings
At the time of the study, the members making up the group were distributed across at
least eight sites and sometimes more. There were differences in terms of numbers of
meeting participants at each site as well as participants’ responsibilities and
involvements. A unique feature presented in this large-scale committee was the
combination of different groups with different sizes and structures across multiple sites.
The hub site, which is the Canberra site, has a larger number of participants involved.
The other sites, namely the state CVOs and the AAHL, can be considered satellite sites
and were smaller in terms of the size of the physical spaces and number of participants
Distributed Collaboration in Emergency Response on Animal Disease
112
attending – usually somewhere between one and three participants per site. The physical
settings of the meeting rooms and findings relating to communication and information
sharing activities in the meetings will be described in this section.
5.5.1 Physical settings
The group at the Canberra site used a large meeting room, the Emergency Meeting
room, to accommodate around fifteen participants sitting in a big meeting table at the
centre of the room (Figure 5.2). The Commonwealth CVO, a couple of key management
representatives and one secretary sit close to the speaker phone which was in the middle
of the big table. They are the “active” meeting participants because they respond to the
agenda items. Most of the management team staff were not CCEAD members but had
indirect roles in CCEAD related activities, such as ministerial liaison and public
communication. They were observers of the meeting, taking notes and minutes for their
own records. There was a seat close to the speaker phone for participants to move to
when they needed to speak.
Figure 5.2. A large meeting room
The Emergency Meeting room was a common meeting room shared by different teams
in DAFF. Six to eight small tables formed the big meeting table in the middle of the
room. All the small tables had wheels, so the size and shape of the big table were
configurable. It was intended to keep this flexibility because the meeting room was used
Distributed Collaboration in Emergency Response on Animal Disease
113
as a meeting room for different emergency response meetings involving different
committees concerning different areas, such as food, plants. There were meeting
facilities, such as white boards and projector screens for general meetings. There were a
number of information resources available for co-located participants to access and
view necessary information: during an outbreak, the walls were used to display various
paper artefacts such as a range of different maps depicting geographical representations
of data relevant to a particular disease; computers on the side table allowed access to
related information if needed during the meetings. These information resources
provided useful common references during the meetings. However the shared visual
information for the co-located members was unavailable for view by remote
participants, as I will describe further in 5.5.3.
Three fax machines were placed on the side table for sending and receiving meeting
materials during the meeting. While fax could be considered an old fashioned
technology, it particularly related to the ways that real-time information could be shared
securely. Information security was of high priority because of the confidentiality of the
information. In managing disease outbreaks, the way that any information was
presented externally needed to be well managed. This was because of the economic
consequences of information being leaked in an uncontrolled manner. It was also to do
with political sensitivities in relation to particular intervention strategies and animal
welfare. For example it could create a sense of panic for individual farmers because of
the economic impact on themselves due to the intervention strategies such as movement
restriction or regional slaughter. These security concerns were also reflected in
restrictions on wireless Internet access within the DAFF building. The security concerns
could impact on information access practices and potential ICT solutions.
The participants of the state CVOs and AAHL sites typically attended the meetings
from the individual offices of core CCEAD members of their local sites. These offices
were fairly standard and had small meeting tables where CCEAD members sat with a
couple of other representatives as needed for particular meetings. Figure 5.3 shows the
office of the AAHL Director. The office had a small oval meeting table where a
telephone was placed. Participants had Internet access through personal desktop
computers, laptops or mobile phones that were brought to the meeting. The Personal
Assistant (PA) of the AAHL Director was in an office close by. While not participating
Distributed Collaboration in Emergency Response on Animal Disease
114
directly in the meetings, the PA assisted in gathering the large amount of meeting
information before the meeting and collecting incoming documents during the meeting.
The assistance from PAs was an important factor in the way that real-time information
was monitored, received and managed at the state CVO and AAHL sites.
Figure 5.3. The office used for the meeting
For the more ad hoc industry representatives participating in the meeting, information
and communication infrastructure was often more compromised. For example, during
the equine influenza outbreak a number of industry representatives would have to
connect from remote rural areas. Low bandwidth and unreliable networks in these
remote regions impacted on the ways that these representatives could participate. Some
of the industry participants had difficulty in receiving documents in a timely fashion
because of the poor network connectivity. It also created extra effort for the secretaries
at DAFF to organize the sharing of information with them. The key point here is that the
sites within this collaboration were operating under asymmetrical conditions in terms of
their information and communication access. This potentially undermined equitable
participation for some of the groups and industries. In this context, the design of a
platform to support collaboration across multiple different local settings is extremely
challenging. It includes the design complexity of each local setting, especially the large
group in the Emergency Meeting Room in Canberra, and also the linkage between the
asymmetries and multiple sites. One of the reasons why teleconferencing offered some
value (in spite of its limitations) was because it was a “lowest common denominator”
technology equally accessible by all participants.
Distributed Collaboration in Emergency Response on Animal Disease
115
5.5.2 Decision Making
The work of the committee during the meetings was essentially analysing large amounts
of information updated from multiple sources and debating appropriate courses of
action on the basis of this information. Representatives from the different groups and
sectors had different views and expectations. States might be differentially affected by
the disease outbreak and have different responsibilities for implementing programs and
interventions. Each state had certain vested interests in terms of economic consequences
of decisions. Likewise, industry representatives who wish to protect their industries
might have different perspectives on certain intervention strategies because of the
immediate effects on the ability of the industry to operate effectively. The issues being
discussed in the meetings were inherently controversial, as commented by one the
CVOs:
“It is almost impossible for people to not come to CCEAD with some invested
interest in terms of protecting their own position… So the whole response, right
through the equine influenza response, a lot of it was trying to get the right
balance between disease control and enabling the industry to operate effectively.
The biggest debates we had in CCEAD were about movement conditions and
what were acceptable…There were lots of different views… a lot of debate”.
There were also inherent uncertainties with some information, such as climatic
conditions that might impact the spread of airborne diseases. These uncertainties add to
the potential for debate and interpretation in terms of particular vested interests. As one
participant pointed out, the analysis was a process of “managing risks”. The decision
making was about getting a balance between the different perspectives.
While there was potential for conflict, the decision making was actually well managed
in the group according to the CVOs we interviewed. This was partly due to a good
working culture within the group. The good relationship was established through other
work that the CCEAD members did together such as meeting regularly when serving on
other animal health committees. Importantly, the decision making was also facilitated
by the provision of a national biosecurity strategy called AUSVETPLAN (AHA 2010).
This strategy outlined particular courses of action and guidelines during certain
scenario. It also outlined important cost sharing measures across the states so that
Distributed Collaboration in Emergency Response on Animal Disease
116
affected states did not have to bear the full costs of implementation so long as certain
criteria were followed by the affected states. It raised the concerns up to a shared
national level and effectively reduced potential economic conflict impacting on the
decision making process.
5.5.3 Information Sharing
Information was a central component in the CCEAD meetings in supporting the
evidence-based debate and decision-making. The meetings were document-centric.
Understanding these practices and the role of documents in them were important factors
in design decisions for potential ICT solutions for such meetings.
The document set used during a meeting was prepared by various members especially
the CVOs of the affected states. Typical types of documents included situation reports
and implementation reports from each of the affected jurisdictions, for example
epidemiological data, graphs and maps of new movement conditions between different
zones. There was a variety of computerized documents in different formats which
included Word, PDF, Excel sheets and PowerPoint presentations. Occasionally short
video clips of sick animals were sent for the members to watch before the meeting.
Figure 5.4 shows the process of information access and distribution. The documents to
be used for the meetings were collated by the CCEAD secretariat based in Canberra.
Due to confidentiality concerns, documents were not allowed to be sent in email
attachments. There was a common repository for the CCEAD members. Microsoft
SharePoint was used as a platform to support the document management and document
sharing across organizations. The access control of the documents was centralized at
DAFF in Canberra. The secretariat administered the database and established access
permissions to SharePoint for each participant. The key secretariat who worked closely
with the Commonwealth CVO acted as an information gate-keeper to maintain the
SharePoint and documents.
Distributed Collaboration in Emergency Response on Animal Disease
117
Figure 5.4. Information access and distribution
Getting documents onto the common information repository was a cumbersome
process. After the document owner logged into the password-protected SharePoint
website and submitted documents, he/she needed to send an email to inform the key
secretariat. The secretariat then sent an email to the list of contacts to notify them of the
updated documents. Since there were a large number of documents sent from different
groups at different times, this communication process increased the effort of each
CCEAD member and has been considered as a barrier to the requirement of near real-
time information sharing. Some of the members had to rely on their PAs to monitor the
updates. One of the interview participants commented that “the way things work in an
emergency situation is less than ideal.”
Relating to this issue, due to the quick changes in a disease situation and various data
requirements, some of the documents were sent very close to the beginning of the
meeting or even during the meeting. At times some meeting participants did not even
receive all the documents. This was because the responsibility for many of the
documents lay with the affected states. The time pressures were particularly acute for
these affected states since they were involved with the day-to-day management of the
Distributed Collaboration in Emergency Response on Animal Disease
118
outbreak. This issue was a source of frustration for meeting participants. It affected
their pre-meeting analysis of the contents but also could lead to communication
difficulties during a meeting since there was no common resource. The cumbersome
nature of the document distribution was a key issue to be addressed. There would be
great benefit gained from collaboration technologies to support real-time sharing of up-
to-date information during the meetings in order to maintain the flow of conversation.
Without any computational means for shared visualization of the documents during the
teleconference, all the participants brought printouts of the documents to the meeting.
Not having the common visual resource could lead to conversational inefficiency since
there was continual effort necessary to orient all the members to the appropriate places
in the document resources. Second, because of the last minute preparation of the
documents the circulated documents sometimes had errors in them. The presenters often
found areas that needed to be updated or corrected as they read through the documents
in the meetings. This was very frustrating for other members to find those places and
make same changes in the paper in front of them. The issue of shared visualisation
could be especially difficult for data in graphical format, such as maps and graphs, when
communicating by telephone. When we asked about the possible solutions to these
problems, most of the interview participants believed that apart from the improvement
in the coordination across multiple organizations, technology could play an important
role in mediating the information sharing and communication. A state CVO emphasised
that the support for interactive sharing of documents, maps and graphs was necessary:
“Need to have some kind of feature there, once the document was edited, you
could all see the edits being done…that document should then be able to send to
everybody…. The other thing is that it would be very good if we could all view
a map simultaneously, and someone could explain the map and we could see and
understand what he was explaining”.
In spite of these difficulties, the paper format of the documents also had a certain value
in the way that the analytic work and discussion were achieved. Paper documents were
important personal resources that were annotated both prior to and during the meeting in
support of particular conversational contributions the individual members wished to
make. Paper documents as personal resources also allowed each participant to view
different parts of the document set according to their individual needs while other
Distributed Collaboration in Emergency Response on Animal Disease
119
people were looking at different documents. Interview participants discussed the
importance of being able to see multiple documents at the same time in order to be able
to fully synthesise information from multiple sources and :
“Particularly when there are situation reports such as these, each state is talking
individually and often you want to look at a map, two maps side by side and
work out if they have the same bits on either side of the border”..
“It is not easy jumping around multiple documents on the laptop either, it saves
on paper but it is not necessarily faster”.
In addition, the paper documents supported co-located sharing practices at the
individual sites. For example, a participant might check some information detail to get
consensus from colleagues before talking to the remote teams. They looked over the
same paper and muted the telephone while doing this. These practices were an
important part of managing side-work during the meetings and facilitating the fluid
development of different conversational threads.
5.5.4 Background Work and Multi-tasking During Meetings
Because of the amount of information to be discussed and certain inefficiencies in
information sharing, the CCEAD meetings sometimes were very long. In some
situations it could last up to 3 or 4 hours. Although there was a structured agenda
prepared by the key secretariat at DAFF, the complexity of the disease, the cross
discipline knowledge required to understand the content and the tensions between
different interests and controversial opinions were all factors that could lead discussion
away from the agenda and make it difficult to coordinate the flow. This was one of the
common sources of frustration among the meeting participants because during the
disease outbreak there was a high workload that they had outside the meeting context.
Access to personal computing resources during the meeting was considered useful in
dealing with some of these concerns. Some of the meeting participants at AAHL and
state CVO offices brought laptops to the meetings or, on occasions, used the phone
while sitting at their desks. The laptops were not being used for reading the meeting
materials but rather for monitoring incoming information and accessing background
work or other materials. The use of teleconferencing (rather than video conferencing)
Distributed Collaboration in Emergency Response on Animal Disease
120
and the use of paper materials provided some of these participants with greater
flexibility in choosing where the meeting was held and greater flexibility to access their
personal computing resources. As well as monitoring incoming information, this access
provided a means by which some participants could manage their background workload,
in particular those participants whose key inputs were only required at certain points
during the meeting. These participants were able to peripherally monitor the
conversation and continue with other important background tasks through the use of
their personal computing resources. One of the values of teleconferencing was that it
reduced the visibility of this activity to other meeting participants, allowing people to
participate when necessary but still manage their other important work tasks.
5.6 Exploration of the Physical Setup
A workshop was conducted at DAFF after the CSIRO design team developed and
deployed a collaborative platform at AAHL to support the collaborations within AAHL
(detail of this development is described in Chapter 6). The aim of the workshop was to
explore the potential of integrating a collaborative workspace into the CCEAD meeting
for the Canberra group. This one-day workshop was conducted in the Emergency
Meeting room where the CCEAD meetings were held. The participants included the
deputy Australian CVO, one CCEAD secretary, three biosecurity service managers
from DAFF, two IT infrastructure managers from DAFF and one diagnostics manager
from AAHL. The collaborative workspace developed for the small group discussion
meetings (four participants) within AAHL was presented at the workshop and how to
extend this solution to fit CCEAD meeting was discussed.
One of design issues that the workshop participants pointed out was to support the
peripheral meeting participants of the large group at the Canberra site. Similar to the
setting of MDTMs at hospital A, there were two different types of people in the
CCEAD meeting at the Canberra site. As described before, one was the key participants
which included the Australian CVO and another two to three people. The other twelve
to fifteen people were observers or peripheral participants. The workshop participants
expressed their concern that those peripheral participants might feel excluded if they
were not able to see the information showing on the digital displays or had a narrow
field view of the video conferencing. Although there was a difference in terms of level
Distributed Collaboration in Emergency Response on Animal Disease
121
of participation between the key participants and non-key participants, it was important
to make sure that all the participants felt involved in the meeting. In addition, the design
solution would be providing all meeting participants with appropriate views of the
remote people and information discussed.
Possible locations of the screens and room layout were investigated during the
workshop. The tables with wheels in the meeting room were easy to move, combine and
recombine. This means that it would be possible to flexibly configure the existing tables
and chairs for the best seating arrangement and put them back to normal positions after
the CCEAD meeting. We experimented with different layouts during the workshop.
The tables and chairs were set up around the workspace displays to allow key staff to sit
at the main table close to the displays and observers to sit at the other tables around
them. As shown in Figure 5.5, one of the arrangements was that the tables of the
observers were arranged in an L-shape around the main table where key staff sat. The
workspace displays, mimicked by an electronic white board in the exercise, were placed
next to the column in the room. Another arrangement (Figure 5.6) was that the displays
were envisioned to be between the two electronic white boards with the main table in
front of it. Other tables were grouped around the main table for the observers.
Figure 5.5. Room arrangement 1
Distributed Collaboration in Emergency Response on Animal Disease
122
Figure 5.6. Room arrangement 2
The workshop participants pointed out that an additional display may need to be
installed at a position close to the observers for showing information being discussed
because the observers may not be able to see the details showed on the main workspace
displays. They felt that the video conferencing (the views of remote participants) may
not need to be visually replicated for the observers since the remote participants’ audio
was available in the room. They also pointed out that the Australian CVO needed to
chair the meeting and the secretary needed to taking notes so they were not able to
manage the device operation, such as opening files using the mouse of the platform
during the meeting. An alternative assistant responsible for controlling the information
flow on the displays would be required.
5.7 Discussion
Through this study I have been able to explore different aspects of the collaboration in a
geographically distributed committee which has responsibility to make high-level
strategic decisions when there is an outbreak of emergency animal diseases. These
aspects relate to the issue of coordination across loosely coupled teams in emergency
response (e.g. Pipek et al. 2012, 2014) and other distributed collaboration situations
where social and communicative connections are important (e.g. Kane & Luz 2006,
Olsen, Zimmerman et al. 2008, Jirakta et al. 2013, Fitzpatrick & Ellingsen 2013). It also
echoes the information sharing challenges identified in designing interactive systems for
Distributed Collaboration in Emergency Response on Animal Disease
123
a large-scale team (e.g. Rolland et al. 2006, Hjelle & Jarulitis 2008, Kristensen & Kyng
2009). Combined with what is known about emergency response in general and
collaboration challenges in inter-organizational collaboration, the findings from this
case study contribute to my overall exploration in supporting collaboration across
different local settings. It allows me to further explore the issues of asymmetries in
physical settings and differences in local information sharing practices that I have found
in the MDTM study.
5.7.1 Coordination and Working Together
Responding to animal disease emergencies required multiple decision makers who must
reason, communicate, and make decisions about complex systems and material
resources. The time pressures that the group was under to make decisions and the
enormous potential impact of these decisions influenced the way that the work was
done. The distributed nature of the collaboration was a pragmatic requirement for this
group of people given their large geographical separation and other ongoing work
commitments during emergency disease outbreaks. One of the coordination problems in
this distributed collaboration concerned inter-organizational mechanisms of document
sharing, particularly the difficulties arising from a rapidly evolving set of information.
We saw how breakdowns and communication inefficiencies arose out of cumbersome
methods for delivering up-to-the-minute documents. While these methods had their
particular reasons, such as the security concern, there was a need for efficient ICT
solutions which could better support the real-time sharing of documents and less
cumbersome but secure methods for document exchange.
The study findings also show how the collaboration involved individuals, groups and
organizations who had different backgrounds and interests working together.
Researchers (e.g. Olson, Hofer et al. 2008) have shown that there are two particular
challenges of coordinating across diversity and distance for collaboration: the greater
the diversity, the less common ground and trust, which together impede the
understanding of each other; and the larger the scale, the greater the coordination
overhead. One of the related issues for CCEAD collaboration is the division of work in
the inter-organizational coordination in emergency response. Definitions about
competencies and responsibilities are important for smooth interaction. Pre-negotiation
Distributed Collaboration in Emergency Response on Animal Disease
124
routines and improvisation work are also required in order to better coordinate the
decision making process. Although in CCEAD there was a common high level goal of
managing the disease outbreak, the different perspectives and vested interests of the
represented organisations can lead to intense debate over action plans that balance their
sometimes competing needs. While technology could play a role in facilitating
consensus building, solutions would not be simply technical. One of the important
features of allowing consensus building in these situations was the introduction of the
AUSVETPLAN. This actually helped mitigate some of the factors that would
otherwise have led to more partisan interpretation and representation of the data, for
example by removing some of the economic concerns underlying these vested interests.
5.7.2 Supporting Information Sharing
Gathering and analysing information, assessing the potential impacts of the animal
disease outbreak and executing related actions were the main activities of CCEAD.
These activities involved a complex process of constructing common information
spaces (Bannon & Bodker 1997, Reddy et al. 2001). The study identified the challenges
in aggregating information (as discussed in 5.7.1), supporting shared visualisation of
information and individualisation of information for different users. Specific
implications for designs to support shared visualisation and support individualisation of
viewing information resources will be described below.
This study has shown the difficulties arising from not having shared visual access to
information being discussed in the distributed meetings. Similar to the findings of the
MDTM study, analysing information with colleagues in the same room usually worked
well, especially if a large situation map was available on the wall for co-located
participants. One of the key design considerations has been enabling shared
visualization and interaction over the network for participants located at different sites.
There have been existing systems for document sharing that could be introduced.
However, given the complexity of the information and sense making, it is important that
new technologies need to support the concurrent visualisation of multiple documents at
the same time and not simply the sequential presentation of single documents. New
technologies will also need to support real-time shared interactions across different
sites, especially the references over maps and graphs. The design will need to address
Distributed Collaboration in Emergency Response on Animal Disease
125
the functionalities of audio-video communication, presenting information, visualizing
information and interacting with information as we have seen in other work in the
broader media space research (e.g. Boxton 2009, Luff et al. 2009, 2013).
Likewise, such shared visualisation in itself would not solve all the problems. The study
showed that individual participants had particular ways of being involved in the
discussion of information. It should be recognized that paper versions of the documents
would remain an important part of the work practice during such meetings, particularly
for the role paper resources play in concurrent visualisation across multiple information
sources. Similarly, the use of personal computing resources would also play an
important role in the ongoing management of work during the meetings, whether
immediately relevant to the conversation at hand or whether directed at tasks outside the
meeting agenda. Efforts could be made to link these personal resources to the shared
resources, such as through the real-time sharing and visualisation of up-to-date
information accessed via personal computing resources.
5.7.3 Different Settings and Existing Mechanisms
One of the challenges for the design of technology to support CCEAD collaboration is
the context of multiple teams of different sizes, multiple locations with different
physical settings, loosely coupled participants and a requirement for relationships
demanding a high level of trust. This context introduces a complex set of asymmetries
in the system that needed to be accounted for in any technology interventions proposed
for this group. Any system introduced needs to be able to work across these different
sites and within the context of different information, network and communication
resources.
Designing collaboration technology for a distributed meeting involving different teams
relies on understandings of how current technology and mechanisms are used in the
current meeting setting. One of the reasons why teleconferencing was used was because
of its status as a “lowest common denominator” technology that was equally available
to all parties, including co-located and distributed participants. Similar to the findings of
the MDTM, this study has shown that any technology introduced needs to take account
of disparities in information access that it introduces into the socio-technical system
which makes up the CCEAD and any consequential biases it introduces in terms of
Distributed Collaboration in Emergency Response on Animal Disease
126
different participants being able to participate equitably. The potential solution of a
collaborative workspace needs to provide participants with equal access to audio, video
and shared visualization of information so efficient participation in the functions of the
meeting can be achieved.
It is also important to maintain local differences since specific settings have their own
social and functional meanings that can be valuable in the distributed collaboration
practices of individuals and individual sites. Adding personalization of information
viewing functionality to the basic shared workspace is one of the considerations to
maintain individual preferences. Another consideration is the physical design of each
local setting, for example the hub site in Canberra which had a larger number of
participants. Similar to the physical setup re-arrangement exercise done in hospital A in
the MDTM study, the workshop described in this chapter has shown potential practical
solutions to support both the key participants and peripheral participants by allowing
both types of participants to have common views of information and to have better
visual communications with remote participants.
The extent to which new technologies can close the gap of the “space between” settings
and groups is a challenge when enabling a distributed large-scale group to engage in a
meaningful collaboration. Technology can serve to close the gap of the space between
groups, but it can also introduce its own gap (Mark et al. 2003). The uncertainty of the
value of participants seeing each other in this study is an example of potential problems
when introducing a video-based environment which provides better visibility of the
remote situation. The valuable aspect of multi-tasking work during a long meeting is a
consideration in any video-mediated solutions that could render this work more visible.
The abilities to mute or walk away, that we have seen in a teleconferencing meeting,
would also be limited.
A collaborative workspace itself will not solve the problems of information aggregation
and distribution process. Accessibility to information is a challenge in inter-
organizational collaboration. Existing security policies required a centralized procedure
which hindered sending and updating information to CCEAD members in a timely
manner. This impacted on the efficiency of the collaborative situation assessment
process for the distributed members of CCEAD. One of the fundamental design
requirements for supporting emergency response work for the inter-organizational level
Distributed Collaboration in Emergency Response on Animal Disease
127
is to provide a complementary information infrastructure, not a tool to replace existing
infrastructures (Ley et al. 2012). One of the solutions and possible future work (see
Chapter 8) can be maintaining the existing shared repository and integrating existing
authentication and authorization mechanisms into the design of a collaborative
workspace so that the access, read or write policies can be predefined and protected at
the collaborative workspace level.
5.8 Conclusion
In this chapter I have described an investigation of collaboration within the emergency
animal disease response committee in their decision-making and management meetings.
Issues around the coordination and information sharing in the collaboration involving
the large-scale multi-site multi-organizational committee have been presented and the
related design implications have been discussed. The importance of understanding the
existing social-technical context and mechanisms that affect the particular collaborative
practice in the committee has been highlighted. Similar to the findings of the MDTM
study, the specific implications for technological support for the emergency response in
the animal disease domain will have broader implications for other settings where
collaborative information synthesis is across different organizations with different local
settings.
.
Distributed Scientific Collaboration across Biocontainment Barriers
128
6 Case Study: Distributed Scientific Collaboration across Biocontainment Barriers
This chapter presents the study of distributed scientific collaboration within the
Australian Animal Health Laboratory (AAHL). Collaboration in this setting is
challenged by physical containment barriers which ensure the safe handling of animal
diseases. The aim of the study was to understand how the scientists communicate across
the barriers, particularly how they share information and collaborate on its analysis. The
study was part of a large design project which has delivered the outcome of developing
and deploying an integrated collaboration platform in AAHL. In this chapter I will first
briefly review related work in distributed scientific collaboration and describe the
research context of this study. Following a description of the research methods, the
work of AAHL and collaborations between the scientists at AAHL will be presented.
This chapter will discuss how these findings shaped the design of the integrated
collaboration platform that has been used in AAHL. This study allowed me to explore
distributed collaboration across different settings within a single laboratory.
6.1 Distributed Scientific Collaboration
Scientific work is collaborative in nature. Sharing a wide variety of data and
collaborating on its review and analysis are central to scientific research (Finholt &
Olson 1997). Research work can be shared among scientists in various ways. The tasks
usually are divisible and can be performed either sequentially or concurrently
(Sonnenwald 2007). For example, in the natural sciences, one scientist may develop
data samples and a second scientist may analyse the samples using specialized scientific
instrumentation (ibid). In the social sciences, scientists may jointly develop data
collection instruments, separately collect data using the instruments in different
geographical areas and then analyse the results together.
Scientific collaborations have increasingly involved research teams which are
distributed over distance. It is common today for scientists to work together with their
colleagues from different disciplines and institutes (Ackerman et al. 2013).
Interdisciplinary collaboration involves the coordination of scientific work and the
Distributed Scientific Collaboration across Biocontainment Barriers
129
integration of knowledge from different areas of expertises. The complexity of scientific
collaboration, the coordination work it relies on and the increasing need for scientists to
collaborate with other groups and over distance have made this a particularly rich
domain for investigation (Olson, Zimmerman et al. 2008).
Computer-supported systems have been developed to enable scientists to work with
each other, with a range of research facilities and with shared data repositories without