- Количество слайдов: 13
The Multi-Disciplinary Development of Collaborative Grids: The social shaping of a Grid for Healthcare By Avgousta Kyriakidou and Dr. Will Venters The London School of Economics and Political Science a. [email protected] ac. uk / w. [email protected] ac. uk This research was undertaken as part of Pegasus EPSRC: Grant No: EP/D 049954/1 www. pegasus. lse. ac. uk
Introduction – The Grid Storing, distributing and processing the large amounts of data in a distributed manner • Computers, storage devices and sensors connected by very fast networks – and Middleware to “virtualise” these resources. • “Overturn strategic operating assumptions, alter industrial economics, upset markets, and pose daunting challenges to every user and vendor” (Carr 2005)… “will provide the electronic foundation for a global society in business, government, research, science and education” (Berman et al 2003) Virtual Organisations: • Reflects collaborative nature of science (Chompalov et al. , 2002) • Allows management of Grid resources by distributed groups • Defines what is shared and rules for sharing (Foster et al. , 2001) Health. Care. Grid (pseudonym) Grid Mammography Analysis • Collaborative image analysis without the clinicians having to co-locate; distributed resources sharing; security, standardization and compliance.
Aim of the study • There must be a greater appreciation of the social and political context of Grid infrastructures’ development and use (Orzech 2003). • Developing and implementing such complex infrastructure requires collaborative working among a range of disciplinary groups and organizations (Hanseth & Monteiro 1998). • Reflects debates in Management Studies literature on the need for research relevance advocating industry involvement in order to bridge “Relevance gap” and suggested through trans-disciplinary research collaborations (Starkey & Madon 2001) and interest in universityindustry relationships (Inzelt 2004)(Also Porac et al, Mitev & Venters).
Social Construction of technology SCOT as an approach to help us look at how technology is constructed in society and through its use (Howcroft et al. , 2004) We consider Grids as embedded in social systems and that their development is determined by social factors and by the meanings attributed to it by different groups. We explore Healthscan. Grid’s development through SCOT as a means of understanding how the interpretations and interactions among different groups influence and eventually shape their final attitude towards the prototype (Orlikowski, 1992). Relevant Social Groups & Technological frames Inconsistencies (incongruence) in the technological frames between the relevant social groups within a collaboration can create problems in developing, implementing and using technology, as well as a breakdown in communication, lack of participation by its users, etc (Orlikowski & Gash, 1994).
The Healthscan. Grid project (cont. . ) Aimed to investigate the feasibility of developing a European database of mammograms using a data grid to support collaboration and communication among clinicians across the EU in medical diagnosis. Collaboration of 5 different institutions: • Hospital. UK, Hospital. Italia: end-users, providing the information and medical data to go into the Healthscan. Grid, as well as the requirements and feedback to the developers • University. West. UK: responsible for gathering comprehensive requirements from users and for the development of the database • Medical. XYZ (a small spin-off SME company) (UK): adapt their already existing mammogram software to meet the Healthscan. Grid’s requirements • PAL the world’s largest particle-physics accelerator laboratory (Switzerland): provide the Grid together with the Grid expertise, and also leading the project as project-coordinator EU Funded for three years. Concluded with successful prototype delivered to Spanish private company
Problems emerged during development • 5 institutions: people from different disciplines and with conflicting priorities and expectations • “Each institution applies for funding… People are meeting together for a common purpose: the funding…It is just business… • Poor teamwork and leadership: Individuals felt they contributed more than they received from the project. • The PAL group developed the Grid but didn’t take medical computing needs into account e. g. security and confidentiality. • Hospital. Italia couldn’t “talk” to its own software; Italia Clinicians had to travel to Hospital. UK to see their images, as well as Hospital. UK’s images
Analysis and findings Three relevant social groups identified: 1. Academic-Developers’ group: two institutions; particle physics group from PAL and the team of academics and Ph. D. students from University. West. UK. 2. Commercial group: staff at Medical. XYZ, a small start-up company interested in business, growing as a company and making money. 3. Health Care Users’ group: end-users of the project; little computing competency or technical skills and interests had nothing to do with commercial work.
Analysis – Congruence or Incongruence? Initiation and impact: • Academic -Developers’ group: Improvement in clinical diagnosis; help society in general; opportunity for hospitals to change their structure, upgrade their technologies in order to be able to facilitate collaboration through the Grid. If successful, it could be adapted and used for other applications in medicine. • Commercial group: Impact in terms of the changes it could offer in the status, structure, the technical expertise of the company as well as in the way they did business. • Healthcare. Users’ group: Impact in terms of providing solutions to the problems surrounding breast cancer today, as well as to the improvement in clinical diagnosis and treatment.
Technological frames – Congruence or Incongruence? (cont. . ) Functionalities and Capabilities: • Academic-Developers’ group: • Epidemiological studies; as a second opinion doctor; comparisons of similar cases. • No formal implementation and training plan found to be necessary. Highlighted the need for security, patient consent and confidentiality to be considered. • Commercial group: • Opportunity to develop better algorithms for their own projects. • No clear expectations about daily usage. Knew, and only cared that, the project could help them commercially. Argued that hospitals might not wish to share their data. • Healthcare. Users’ group: • Focus on teaching and real-time usage to collaborate during surgery. • Little experience with technologies, expected formal implementation and training plan to be provided. • Expected Grid for epidemiology studies, as a second opinion doctor, as a teaching aid, as well as to be used in some way within surgery. • Concern about security and confidentiality.
Technological frames – Congruence or Incongruence? (cont. . ) Expectations regarding the final output: • Academic -Developers’ group: • It was just a prototype, so left aspects e. g. user-friendliness and automated aside. • Particle physicists’ team provide knowledge from physics to develop a prototype, that was useful and had potential to become a commercial product. • Academics’ team: a successful prototype which would provide them with a large number of Ph. Ds and publications upon which their personal success is judged. • Commercial group: • Did not expect benefits from the prototype itself, but from acquiring technical expertise throughout the development. • Wanted to be visible as technical leaders in the market field, a goal they achieved. • Healthcare. Users’ group: • Expected a system that could improve clinical diagnosis, project cancer risk accurately, be used in a real-time basis and as a teaching tool. • Expected it to be user-friendly and automated. • When the prototype was delivered to them and put into use they found it extremely hard to cope with.
Closure and Stabilization The groups knew that if they did not deliver what they had promised and what the EU was paying for, they would not get the money they reached a form of closure and stabilization, because in a way they had to, rather than common consensus. They put aside their expectations and understandings concerning the Healthscan. Grid, and they all worked together in order to manage to deliver. “People just stopped talking to each other… when this happened, [the funding representative] was shouting at them. He said ‘that’s it pack up the project and give me back the money’ That’s when they got scared and started being nice to each others. They had to, because they had promised the EU to deliver. So, they needed someone to shout at them. ”
Conclusions 1. While collaborative multi-disciplinary projects may be advantageous in order to bridge the ‘relevance gap’ (Starkey & Madan, 2001) by ensuring that Grids reflect the needs of users within various sectors; we should not forget that such disciplines have different reasons for participating in such projects. 2. While producing relevant Grids is important, those involved in leading Grid development (including funding bodies) should appreciate more the interpretation of the project made by its participants. 3. Socio-technical problems may be linked through the requirements gathering to the varying interpretations of the project by the relevant social groups involved.
Conclusions (cont. . ) 1. If Grid’s are indeed to support global collaborative working practice then we should consider how conflicting interpretations by the relevant social groups towards the technology and the project become embedded within Grids’ “Virtual Organisations” rather than considering Grids as a transparent infrastructure. 2. “Virtual Organisations” of Grids are more than technical mapping of the anticipated Grid users collaboration 3. We suggest that they are a socially constructed formula which represents the negotiation of the relevant social group’s collaboration in the Grid’s construction – and are thus central to the stabilization of the technology in use, and hence to the project’s success 4. If projects simply ‘mechanistically pool’ the expertise of the participant disciplines then such stability of Grids may prove difficult to achieve.