Swiss Medical Informatics - SMI 51 by Pascal Walliser

SMI 51

SGMI • SSIM • SSMI

SGMI SSIM SSMI

Schweizerische Gesellschaft für Medizinische Informatik Société suisse d‘informatique médicale Società svizzera d‘informatica medicale Swiss Society for Medical Informatics

Swiss Medical Informatics

Schwerpunkt / Thème principal: 18. Wissenschaftliche Jahresversammlung der SGMI Luzern, 26.6. - 28.6.2003 18ièmes journées annuelles de la SSIM Lucerne, le 26 au 28 juin 2003 • Panel: Health Informatics Education in Switzerland • Ausgewählte Beiträge der Jahresversammlung / sélection de contributions des journées annuelles Schwabe & Co. AG Verlag · Basel

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles Inhaltsverzeichnis / Table des matières

SGMI SSIM SSMI

Inhaltsverzeichnis / Table des matières 2

Editorial (Antoine Geissbühler)

Schwerpunkt / Thème principal: 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

SMI 2003; No 51

Panel: Health informatics training in Switzerland

Introducing Peer-to-Peer (P2P) in Healthcare (Claus Eikemeier)

Integration der elektronischen Kommunikation zwischen Spital und Arztpraxen ins elektronische Patientendossier (Stefan Hunziker, Cécil Cheah, Peter Steinmann, Guido Schüpfer)

Carte de santé: un premier pas vers le futur Réseau d‘informatique sanitaire dans une logique de changement culturel (Marzio Della Santa)

Die neu gewählten Vorstandsmitglieder / Les nouveaux élus du comité

Clinical study on the acceptance of a Web-based self-help guide for bulimia (Tony Lam, Isabelle Carrard, Maria Damoiseau Carola, Anne-Christine Volkart)

Bridging the gap between medical narratives and structured data in the computerized patient record (Christian Lovis, Alexander Lamb, Anne-Marie Rassinoux, Antoine Geissbuhler)

RAFT: un réseau de télémédecine en Afrique francophone (Ousmane Ly, Antoine Geissbühler)

Clinical Decision-Support Tool for Acute Appendicitis (Stéphane Meystre)

The global Image Management Strategy within the HUG (David Bandon, Christian Lovis, Jean-Paul Vallée, Antoine Geissbuhler, François Terrier)

Tips & Tricks - Instant Messaging, Dipl.-Ing. Claus Eikemeier - Trucs et astuces pour Office Prof. Christian Lovis

Events in Medical Informatics

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles Editorial

Editorial Antoine Geissbühler University of Geneva

SGMI SSIM SSMI

A picture is worth a thousand kilobytes Healthcare is literally swamped by images: new CT-scanners produce several hundreds of images for each exam, functional magnetic resonance produce thousands of raw images per study, not to mention endoscopy or telemedicine with their streams of video images. Gigabytes, terabytes of data must be moved between the technicians, radiologists and clinicians who want to access them in real time. This is obviously stressful for computer networks and storage devices, but the exponential growth in computer performance keeps technology able to cope with this challenge. Less obvious until recently but more profound are the challenges to the unexpandable human cognitive abilities. Inundated with multimedia signals, it becomes more and more difficult to find the appropriate information at the time it is needed for safe and efficient medical decision-making. Improving the situation implies a) a better understanding of the added value of each stakeholder in the image lifecycle and its contribution to the diagnostic and therapeutic processes, b) the enhancement of human-computer collaboration through better interfaces and richer knowledgedriven systems, and, above all, c) a better understanding and instrumentation of the organisational and cultural aspects of our complex healthcare environments.

At the architectural and organizational level, David Bandon details the global image management strategy and lessons learned in a real world large-scale implementation at Geneva University Hospitals, and explains how imaging is expanding beyond the field of radiology, thus creating new challenges. Stefan Hunziker describes the experience of integrating electronic communication between the hospital and private practitioners. Ousmane Ly proposes a multinational telemedicine network to foster collaboration in healthcare between developing countries of Northern and Western Africa. Claus Eikemeier explores how new communication and sharing paradigms, and in particular the Peer-to-Peer architecture, made famous and controversial by Napster in the music industry, can be used productively in healthcare. These contributions illustrate how medical informatics has become a key player to help shape the future of a safer and more efficient healthcare delivery system, and, in particular, in the field of medical imaging.

These concerns were at the centre of the debates during the conference jointly organized in Lucerne in June 2003 by the Swiss Societies for Radiology, for Radiooncology, and for Medical Informatics, under the title “Image Communication”. Selected papers presented at the conference, dealing with the health information aspects of these challenges, are published in this issue of Swiss Medical Informatics.

Correspondence: Prof. Dr. med. Antoine Geissbühler University of Geneva Department of Radiology and Medical Informatics 1204 Genève

At the human-machine interaction level, Christian Lovis shows how to improve the contents of the computerized medical record by bridging the gap between medical narratives and structured data. Stéphane Meystre proposes a new approach for developing a clinical decision-support tool for the diagnosis of acute appendicitis. Tony Lam describes a clinical study on the acceptance of a Web-based self-help guide.

SMI 2003; No 51

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

Panel: Health informatics training in Switzerland Certificat de formation continue universitaire en informatique médicale

Antoine Geissbühler Christian Lovis Faculté de médecine, Université de Genève

Depuis quelques années, la capacité à disposer d’informations médico-économiques est devenue un enjeu stratégique. Les difficultés inhérentes à l’implémentation d’outils informatiques au cœur des processus de soins ont alors été mises en évidence, ajoutant au domaine une dimension organisationnelle et cognitive. Récemment, le rôle des systèmes informatiques cliniques se redéfinit, évoluant d ’une fonction essentiellement descriptive à une fonction plus active, afin de mettre à disposition des médecins et soignants des outils pouvant mieux les assister. La mise en place de tels systèmes nécessite des compétences particulières: compréhension systémique des processus de soins, maîtrise conceptuelle des technologies impliquées, familiarité avec les enjeux infor-

mationnels et cognitifs et leurs traductions sous forme d’outils informatiques. Le but de ce programme est de fournir les outils nécessaires à la compréhension des aspects médicaux, informatiques, techniques, organisationnels et juridiques liés à la mise en oeuvre des technologies de l’information et de la communication dans la pratique médico-soignante, ainsi que pour l’évaluation et le pilotage de systèmes de santé. Ce programme s’adresse aux professionnels de la santé et aux professionnels des technologies de l’information désireux de développer ces compétences spécifiques. Pour en savoir plus: http://www.dim.hcuge.ch

Medizin-Technologie-Management, Nachdiplomstudium Medizin-InformatikManagement, geplantes Nachdiplomstudium Stelia Fuhrer Andreas Fuchs Fachhochschule Bern

Die Hochschule für Technik und Architektur der Fachhochschule Bern bietet eine grosse Auswahl von Nachdiplomstudien (NDS) und Nachdiplomkursen (NDK) an: Gut bekannt sind die Weiterbildungsmöglichkeiten der Software-Schule Schweiz (SWS) sowie der betriebswirtschaftlich (BWL) orientierten Fächer wie BWL/Unternehmensführung und BWL/Innovationsmanagement. Seit 1999 werden auch im Bereich Medizintechnologien ein NDS sowie Weiterbildungstagungen angeboten. Bereits der dritte Jahrgang besucht das NDS Medizin-Technologie-Management (MTM). Im Rahmen dieses Studiums werden die Grundlagen von Medizin und Biologie, der Medizinaltechnik und -technologien unterrichtet. Dazu noch wichtige Umfeldkenntnisse wie ganzheitliches Entwicklungsmanagement,

SMI 2003; No 51

integrierte Managementsysteme (Qualität, Umwelt, Arbeitssicherheit) oder Regulatory Affairs. Dieses Nachdiplomstudium dient Nachwuchskräften aus den verschiedenen Sparten der schweizerischen MedizinaltechnikIndustrie als Wissensquelle und zur Weiterqualifizierung. Der Kurs kann als Ganzes besucht werden – 600 Lektionen Unterricht sowie eine Diplomarbeit im Umfang von 200 Stunden – oder es können je nach Bedürfnis einzelne Semester oder Module besucht werden. Basierend auf der umfassenden Erfahrung der HTA Bern mit Nachdiplomstudien wird nun ein Studiengang zum Thema Medizin-Informatik entwickelt. Die Lerninhalte wurden in Zusammenarbeit von Frau Dr. Stelia Fuhrer vom Center for Medical Technology und Herr Dr. Arno

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

Schmidhauser von der Software-Schule Schweiz entwickelt, wobei der Autor die konkrete Ausführung des Curriculum übernahm. Inhaltlich orientiert man sich an den Empfehlungen der SGMI (Schweizerische Gesellschaft für Medizinische Informatik), und deren Präsidentin Frau Dr. Judith Wagner, der IMIA (International Medical Informatics Association), an Empfehlungen von Repräsentanten von der Industrie, und des Gesundheitswesen sowie an Curricula der HTA Bern und von ausländischen Bildungsinstitutionen. Das NDS Medizin-Informatik-Management (MIM) umfasst 4 Semester, wobei das vierte Semester der Diplomarbeit zu widmen ist. Die wichtigsten für Medizin-InformatikCurricula typischen Fächer sind vertreten: Grundlagen der Biologie und der Medizin, Medizinische Dokumentation, Medizinische Methodologie, Wissensverarbeitung in der Medizin, sowie, im Hinblick auf Biometrie Mathematik und Statistik. Branchen-Kenntnisse sind sehr wichtig und werden im Block „Gesundheitswesen Schweiz“ vermittelt. Der aktuelle Stand der Informatik im Gesundheitswesen – Krankenhaus- und Praxis-Informationssysteme, Bilderfassungsgeräte und digitale Bildverarbeitung sowie Archivierung, Kommunikationsstandards und Telemedizin, etc. – wird ebenso vermittelt wie die eher traditionellen Informatik-Themen Software-Engineering und Datenbanken.

SGMI SSIM SSMI

den. Bei genügender Anmeldungen wird der Start im Jahr 2004 erwartet. Die aktuelle Version des Studienführers NDS MIM kann von der untenstehenden Adresse herunter geladen werden. Erstaunt hat uns bei der Anmeldungen die Heterogenität der Zielgruppe der Studierenden. Das Interesse am NDS MIM von Seite medizinisch ausgebildeten Personen war grösser als erwartet. Diese Personengruppe gehört nicht zur traditionellen Zielgruppe einer technischen Fachhochschulen, welche Ingenieure anspricht. In diesem Stammmarkt der Fachhochschule hat eine Markterhebung das Interesse an Medizin-Informatik nachgewiesen. Da jedoch schon einzelne Mediziner die SWS absolviert haben und sich schliesslich etwa so viele Personen mit medizinischen Ausbildungen für NDS MIM interessierten wie Leute mit technischem Hintergrund, wird das Curriculum vom NDS MIM modifiziert und besser nach den Marktbedürfnissen ausgerichtet. Das verbesserte Konzept ist auf Herbst/ Winter 2003/04 geplant. Die unterschiedliche Vorbildung der NDS MIM Studierenden wird besser berücksichtigt indem der erste Studienblock voraussichtlich customisierten Unterricht erlauben wird. Weitere Informationen finden sich unter: http://www.hta-be.bfh.ch/~wwwmed/

Starkes Gewicht wird auf Projektmanagement, auf Gesundheitsökonomie und das Verständnis für komplexe Organisationen gelegt. Selbstverständlich bleiben rechtliche Fragen wie der Datenschutz und die für Medizin-Informatik relevanten Normen und Vorschriften nicht ausgeklammert. Für diesen Kurs resp. für die entsprechenden Lerninhalte der ersten Version des Curriculum konnten sich 10 Studierende begeistern und meldeten sich an. Da die NDS der HTA Bern jedoch selbsttragend sein müssen, werden pro Kurs-Durchführung mindestens 15 Anmeldungen benötigt. Der Beginn des ersten Kurs konnte deshalb nicht wie vorgesehen bereits im 2003 erfolgen und musste verschoben wer-

SMI 2003; No 51

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

Computers for Health (C4H): a Swiss Virtual Campus project for pre-graduate medical informatics learning modules.

Bengt Kayser Antoine Geissbuhler Faculty of medicine, University of Geneva.

In the Swiss medical curricula, learning opportunities in medical informatics are too scarce. To improve knowledge and skills in the field of medical informatics the universities of Basel, Bern, Geneva, Lausanne and Zurich, in collaboration with the Lausanne polytechnic school (EPFL) and the SGMI/SSMI develop a modular e-learning course in medical informatics within the framework of the Swiss Virtual Campus. This course will be available for medical and other students from the academic year 2003-2004 onwards at http://www.c4h. ch/. The general learning objectives are to make the student aware of the ubiquitous nature of information technology in today’s and tomorrow’s health care, teach the basic concepts, and show the benefits of medical

informatics. Specific learning objectives and learning content are provided in a modular and thematic way: Introduction to medical informatics (an introduction to the discipline), Legal aspects (of the use of computer technology for health), Information literacy (how to search, find and critique medical information), Basic imaging (basic concepts of medical digital imaging), Advanced imaging, Clinical decision support (computers as an aid to the clinician), and Hospital information systems (the use of computer technology and informatics in a modern hospital). Further modules will be added in the future. For more information confer with: http://www.virtualcampus.ch

Die Mitgliedschaft bei der SGMI beinhaltet folgende Dienstleistungen: • das Abonnement der Zeitschrift «Swiss Medical Informatics» (Publikationsorgan der SGMI) • reduzierte Gebühren an der Jahresversammlung der SGMI • das «Yearbook of Medical Informatics» der IMIA

Beitritt zur SGMI

Anmeldung als Kollektivmitglied Jahresbeitrag sFr. 400.– Name

als Einzelmitglied Jahresbeitrag sFr. 120.–

in Ausbildung Jahresbeitrag sFr. 60.–

Vorname

Anrede Institution Adresse PLZ/Ort Postfach

Land

Telefon

Fax

e-mail Einsenden an: Sekretariat SGMI-SSIM, c/o VSAO, Dählhölzliweg 3, Postfach 229, 3000 Bern 6

SMI 2003; No 51

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

SGMI SSIM SSMI

Introducing Peer-to-Peer (P2P) in Healthcare

Claus Eikemeier

Abstract

Faculty of Mathematics and Computer Science, University of Bremen

Advances in technology are enabling building blocks for future IT solutions. One of them is the Peer-to-Peer (P2P) paradigm. The notion of P2P describes a class of distributed, dynamic, re-organizing systems, means to empower the user on multiple levels of work and to implement systems that use idle resources on the “edge of the net”. This facilitates services and enables a complete new way of collaboration and resource sharing. Major software companies are doing research on the field of P2P as it potentially solves a couple of problems of today’s health information systems. Even if media industry is challenged massively by the P2P paradigm, the healthcare sector is yet slightly touched by the P2P paradigm: only few systems are available serving as a prove of concept. The paper outlines P2P technology and its possible impact in the healthcare domain. If the character of P2P is understood, healthcare (biomedical research, electronic health records, communications, and other fields) will benefit from using it. But it doesn’t work in all cases. The underlying problems are solved „quite good“, when the paradigm fits – independent of the domain. The P2P paradigm will change today’s healthcare systems in various ways. This paper will describe some cornerstones of the P2P paradigm and explain their usage in healthcare systems.

Author’s address: Dipl.-Ing. Claus Eikemeier University of Bremen Bibliotheksstr. 1 D-28359 Bremen Germany Phone: +49 421 218 8260 Fax: +49 421 218 8661 E-mail: cei@informatik.uni-bremen.de

Lessons Learned from The Napster Case: The Immense Power Of P2P Music media industry faced a big challenge during the last years. The combination of the MP3 music data format in combination with advanced networking abilities of modern programming languages and an increase in available bandwidth to the

end user made it possible to share pieces of music just as easy as writing an email or downloading a program from a website. For quite a long time the music industry didn’t realize the threat for its traditional business model. When the tremendous impact of the P2P paradigm was realized, it was much too late: nearly every modern song was immediately available for free download on the P2P music exchanges like Napster [1] or Gnutella [2]. Each member of this virtual community offered very little to the common goal but these small contributions resulted in a very efficient way to move files from one computer to another, with or without using central system component. It circumvented the traditional business structures in a way that threatens the complete music business. Other well known P2P applications[3] use the idle CPU time of (client) computers to perform complicated computations (signal processing [4], weather, mathematical problems, pattern matching). In general P2P applications deal with communications, collaborations and sharing resources [5, p. 23], [6]. The systems are distributed in a way that a central server is not needed any more – every “peer” has the same capabilities, being it acting as a client or at the same time as a server (“Servent”- concept). This architecture resembles to the somehow „chaotic“ structures of real world communities [7] and offers each of the members (the peers) an increase in possibilities and performance. The media P2P case shows that using the P2P paradigm in a correct way offers enhancements to established systems. From the users point of view, using P2P systems results in a superior performance of the overall system. It is interesting that the service provider can offer better service with less resources: much of the work is done by the peers themselves. The case of Napster [1](P2P system with a central component) and Gnutella [2] (P2P system without central components) has

SMI 2003; No 51

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

been analyzed well [e.g. 8]. The “lessons learned” can be used to critically compare the changes in the music industry to other domains. In general, the P2P paradigm challenges existing systems (infrastructure, organization, business model, …). Only if the lessons learned from the known cases are assessed correctly, the P2P paradigm can be controlled to offer superior effects in other fields, e.g. like in management sciences or in this case in healthcare. Defining Peer-to-Peer When talking about Peer-to-Peer systems, it strongly depends on the personal context, what is understood by that notion since there is no strict definition of Peer-to-Peer [6, p.71] and no sharp borderline to nonP2P systems. In general most definitions [9],[10] rely on the technological features of the discussed systems. But this doesn’t result in a consistent and proper definition. In most cases sharing of CPU power (e.g. SETI@home [4], Intel cure project [11]) are typical Client/Server applications from a technological point of view and hence not of P2P type, but are commonly considered to be. This arises from the fact that the overall task of one client is processed by a lot of servers. This is one reason why technology based definitions are generally not sufficient. If we have a notion of “sociological kind of P2P applications”, meaning that the peers belong to a community that globally works on a task, we can extend those pure technical explanations. Especially powerful concepts like informally organized self-help groups belong to the success factors of P2P systems. A third domain that has to be mentioned to form a holistic comprehension of P2P is the economical view[12]. It includes the economical impact of the change of existing systems (markets, value chains). This view particularly mirrors community effects that are important driving forces in the different socio-technical networks. Summing up, we have to mention technological, sociological and economical effects of the P2P paradigm to understand the phenomenon. Depending on the domain, the focus varies. And even still other aspects

SMI 2003; No 51

SGMI SSIM SSMI

(like a juridical perspective) can get into the focus, too. Is P2P an issue In The Healthcare Sector? Healthcare systems are facing a big challenge by the increasing digitalization of the domain. More and more patients are used to use the possibilities of the internet [13], [14]. They demand for efficiently working infrastructures for a better support of the healthcare processes. It is no question that P2P has an impact on the healthcare sector [5, Chapter 15] as first P2P systems are already running. Even if some conditions are different from the media cases mentioned above, we find similar structures like Communities of Practice (physician, nurse, pharmacy), the need to exchange files (prescriptions, healthcare data records, 2D and 3D images, billing data), to communicate securely and collaborate (second opinion) and so on. Client/Server solutions are commonly viewed as only partly resulting in sustainable solutions. Systems in the field of healthcare are based on similar pillars like the P2P systems: technology, sociology and economics and so we can translate findings from existing P2P cases to new healthcare systems. Even if mentioning other concepts like trust or reputation, we have to consider a use in the healthcare domain [15]. Of particular importance in understanding the application of P2P systems is the role of the central provider of a service. Traditionally it was the center of the system and had to announce, start, run and maintain the service. Within the P2P paradigm, there is no need for a central maintainer of the service any more. Each peer is doing the house-keeping itself. Comparable to real life, some nodes are better cared of, some less, some offer more to the system than they pull from the network, others more or less only exploit it and do not contribute. This last case commonly is called “free-riding”[16]. Free-riding is a good example to look at when porting the P2P cases to other domains: one can learn from a problem in one domain to avoid it in the other context: Offering the patient a superior treatment based on a comprehensive, fast to access

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

and easily to integrate access to his patient record and access to up-to-date global knowledge will be accepted as a reason to offer new services [17]. Using P2P systems is not reduced to the patient but there will be reasonable systems supporting physicians, pharmacists, researcher and so on. In the next section some examples of using P2P systems in the medical domain are sketched. They are not standard examples like known from biomedical research (protein folding, gene sequencing) but are more related to the patient-physician relation and electronic healthcare records. Examples of P2P Systems in Healthcare This section explains a few systems that adopt P2P aspects in the healthcare domain. The first case considers „patients” that form a group of peers. Health portals with a forum / chat corner resemble the way of how P2P software can be used. Think of building reputation of the primary care physician in the patient community [5, chapter 16]: by using P2P software: opinions will be gathered on an application where the patients can discuss illnesses and treatments, can exchange opinions about the physician, about its office etc. This kind of community support will increase knowledge on the patient side since there is an open information pool that can be used by the patients (e.g. when choosing a physician) and - of course - by the physician (he will use it as a means to get feedback on his work). This second functionality is often neglected. The availability of such tool will result in better transparency on the market of health. The patient is not only dependent on the personal opinion of a few acquaintances but has access to an opinion that is based on a much broader audience. This is only one single aspect of enabling the patient with a „P2P means“. The relation of community informatics systems to the P2P paradigm is currently under research as there might be strong links between them. If considering the eHealth issues like „reliability“, “completeness”, „patient empowering“ or the „informed patient“, especially when considered with the question of costs, the physicians could take advantage of the

SGMI SSIM SSMI

P2P paradigm. Think of a scenario where a patient asks the physician to comment on an article from a newspaper. Probably this physician is not the first to answer the question, but in general he is not able to get this answer. So, when using P2P knowledge management systems, one could easily put the answer into the forum, where a general professional opinion could be discussed and published. This would be a possibility for the community of physicians to tackle the increasing information flood and hence offering a superior service in today’s world of information overload. When talking about P2P, it must be analyzed if the media case of Napster / Gnutella can be directly translated into the domain of interest [18]. In healthcare, this means for example that different hospitals and physicians offer data of a patient and the actually treating physician can combine the pieces of information from these different sources into one integrated view on the patient. One can consider such a system to be a distributed electronic patient record ( [19]). Seminal work is related to Freenet [20]. That is a Gnutella-like pure P2P system (no central entity is used) that stores data encrypted and anonymously and – just to mention another feature- splits it up into different pieces that generally are not held on the same peer. So the retrieval of a document needs to combine different parts of it that on their own are not usable in any way. Each piece of demanded data is pulled from the peer where it is actually stored on the demanding peer. This is equivalent to a physical moving of document copies from one physician to another (e.g. when the patient moves). The access to the data is much faster, when the “new” physician accesses the data a second time. One (still unpublished) P2P project of this type is currently under development in Northern Germany: it will use a central index to store meta data from the health record and will allow other registered health professionals to access data stored on the computers of the peers. The access is logged likewise. Physically there will be a special computer (always online) in each physicians office that serves as a gateway into the secured P2P network. The main target of this network is enabling an electronic data exchange in daily routine work (e.g. as a

SMI 2003; No 51

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

way to implement the digital prescription) and in the case of emergency. It is a safe system (according to today’s standards) and even if not every information is accessible in it, the system will improve the normal healthcare processes considerably. This project belongs to the field of Grid applications that are probably are the most promising applications of P2P type in healthcare at this moment. Even if in this field exist many issues, too (security, merging of different data chunks, patient participation, etc.), there are a couple of tasks (e.g. image and biosignal processing) that can be solved on a grid more efficiently [21,22]. Commonly work-arounds help to overcome the existing problems. These grid applications are more oriented towards technology than the sociology oriented P2P systems mentioned before that have to deal with other, often irrational effects. As mentioned above, there are medical P2P systems designed, implemented and running [5]. Common to all scenarios is the need for an in-depth understanding of the characteristics of P2P paradigm, the new way of using it and the astonishing performance of the overall systems.

Conclusion P2P is a new paradigm that offers superior new possibilities in many domains, including healthcare. The cases of Napster and Gnutella offer valuable help in analyzing and designing new software and hardware systems whose technical, sociological and economical architecture are based on the P2P paradigm. SETI@home enables for intense usage of remote resources (c.f. Grid applications, [23]). The P2P characteristics offer a couple of new solutions but -of course- do not solve all problems in the healthcare domain. There often exist P2P solutions to those general problems but they have to be evaluated if they also fit to the needs of healthcare. As there exist already healthcare related cases, healthcare is and will be a domain for P2P solutions. Depending on how the different communities are using the P2P systems, it could be a threat or benefit for the stakeholder – being it the patient, physician or another member of the system. Besides those systems for (bio-)medical research and the more technical grid applications, the biggest benefit will probably arise from using it in the field of electronic health care records, as this problem fits well to the P2P paradigm. Nevertheless not all open issues will be solvable by applying the P2P paradigm. Acknowledgement I thank Henning Müller, University Hospital of Geneva, for some valuable links to medical grid applications.

Fig. 1: Simplified example of a healthcare network using a modified Napster technology: Coordination is based on an overlay network with a central repository that links to the different providers. Data is exchanged on a peer-to-peer base, hence storage of data and the transmission load is distributed among all participants.

SMI 2003; No 51

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

SGMI SSIM SSMI

References

Shirky, C. (2001) Listening to Napster, in Andy Oram: Peer-to-Peer – Harnessing the Power of Disruptive Technologies“, O`Reilly, Beijing

Kan, G. (2001) Gnutella, in Andy Oram: Peer-to-Peer – Harnessing the Power of Disruptive Technologies“, O`Reilly, Beijing

N.N., Homepage www.rechenkraft.de (in German), accessed 17.11.2002

SETI@home (search for extraterrestrial intelligence), downloaded at http://setiathome.ssl. berkeley.edu/ on 05.07.2002

Hassan M. Fattah, „P2P – How Peer-to-Peer Technology Is Revolutionizing the Way We Do Business“, Dearborn Trade Publishing, Chicago, Il, 2002

Eikemeier, C. and Lechner, U. (2002), Peer-to-Peer – Eine verteilte Anwendung auf der Suche nach einem zentralen Verständnis. Proceedings of the GeNeMe 2002, Eul Publ., Köln (in German)

Bonifacio, M., Bouquet, P. and Traverso, P. (2002) Enabling Distributed Knowledge Management: Managerial and Technological Implications – Swiss Informatics, Special issue on Knowledge Management, 1/02 :p. 23-29

Lechner, U. and Schmid, B. Communities – Business Models and System Architectures. In Proceedings of the Hawaiian Int. Conf. On System Sciences (HICSS 2001). IEEE Press

Schollmeier, R. (2001), A Definition of Peer-to-Peer Networking for the Classification of P2P Architectures and Applications. Extended Abstract in Proceedings of the IEEE 2001 Int. Conf on Peer-to-Peer Computing (PIP2001), Linköping, Sweden.

Shirky, C. (2000) Clay Shirky on P2P, http://davenet.userland.com/2000/11/15 (accessed 03.08.2002)

Harrison, L. SETI founder speaks about Intel P2P cancer project, downloaded at http: //www.theregister.co.uk/content/6/18129.html on 04.07.2002

Cohen, G. (2001) Building a successful P2P Business Model, Talk, The O’Reilly P2P and Web Services Conference, Washington, D.C., http://conferences.oreillynet.com/

Blackmore, M. Consumers Health Forum of Australia: Empowering Consumers and Communities for Better Health, http://www.chf.org.au/issues/Speeches/ (22.10.2001)

Geiger, M., Eikemeier, C. and Grütter, R. Gesundheitsportale im Internet (health portals on the internet), Working Report (accessible at http://www.ta-swiss.ch)

Eggs, H. Sackmann, S., Eymann, T., Müller, G. (2002) Vertrauen und Reputation in P2PNetzwerken, in Schoder, Fischbach, Teichmann (Eds.), Peer-to-Peer, Springer (in German)

Adar, E. and Huberman, B. (2000) Free Riding on Gnutella, FirstMonday.dk, http:// citeseer.nj.nec.com/article/adar00free.html (accessed 10.08.2002)

Moody, D. and Shanks, G. (2002) Using On-Line Medical Knowledge To Support Evidence-Based Practice: A Case Study of a Successful Knowledge Management Project. In Grütter, R. (Ed.) Knowledge Media in Healthcare, Idea Group Publishing

Hess, T., Anding, M., Schreiber, M. (2002), Napster in der Videobranche? Erste Überlegungen zu P2P-Anwendungen für Videoinhalte, in Schoder, Fischbach, Teichmann (Eds.), Peer-to-Peer, Springer

Eckhardt, A. (2000) Computerbasierte Patientendossiers. TA-Study, Center for Technology Assessment (ZTA), 36/2000, Bern.

Clarke, I. (1999). A Distributed Decentralized Information Storage and Retrieval System. Report, Division of Informatics, University of Edingburgh.

Müller, H, Garcia, A., Vallee, J.-P. And Geissbuhler, A. (2003) Grid Computing at the University Hospitals of Geneva, Proceedings of the First HealthGrid Conference, Lyon, 2003.

Homepage of Healthgrid Cluster, www.healthgrid.org

David Needle, „Latest P2P Effort Focused on Cancer Research“, downloaded at http: //siliconvalley.internet.com/news/article.php/3531_732661 on 05.07.2002

SMI 2003; No 51

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

Integration der elektronischen Kommunikation zwischen Spital und Arztpraxen ins elektronische Patientendossier Stefan Hunziker 1, Cécil Cheah 1, Peter Steinmann 2, Guido Schüpfer 3 1 Abteilung Medizininformatik 2 Informatikabteilung 3 Institut für Anästhesie Kantonsspital Luzern, 6000 Luzern, Schweiz

Abstract Die Kommunikation zwischen verschiedenen Teilnehmern im Gesundheitswesen wird immer wichtiger. Dabei müssen verschiedene komplexe Umgebungen zusammengeführt werden. Am Kantonspital Luzern wurde zwischen dem elektronischen Patientendossier (Clinicware ®) und der Kommunikationsplattform HAKOM eine HL7 Schnittstelle implementiert. Das System steht seit eineinhalb Jahren im Einsatz und findet eine zunehmende Verbreitung unter den Hausärzten und den Spitälern der Region. Ein grosser Vorteil liegt in der Möglichkeit der asynchronen Kommunikation zwischen den Teilnehmern. Mit dem HAKOM System wird eine strukturierte externe Datenablage ermöglicht, als Basis für ein späteres externes Patientendossier. Einleitung Eine Umfrage unter Hausärzten bezüglich der Leistungen fokussiert auf Chirurgen eines Zentrumsspital zeigte sehr gute Resultate für die Kompetenz der erbrachten ärztlichen Leistung. Hingegen gibt es deutliche Abstriche, was die Erreichbarkeit der Chirurgen und den Informationsfluss betrifft [1]. Diese Ergebnisse decken sich mit eigenen Erfahrungen. Wir haben gefragt, ob sich die Kommunikation mit Arztpraxen mittels neuer Informationstechnologien verbessern lässt und gleichzeitige die Integration in ein bestehendes spitalinternes Klinikinformationssystem möglich ist.

Korrespondenzadresse: Dr. med. Stefan Hunziker Kantonsspital Luzern Spitalstrasse CH-6000 Luzern Switzerland Tel.: +41 (0)41 205 2524 E-mail: stefan.hunziker@ksl.ch

SMI 2003; No 51

Methode Die Kommunikationsplattform (HAKOM) ist auf einer Lotus Notes Lösung aufgesetzt, wie sie bereits von Tschudi et all [2] und Freiermuth et all [3] beschrieben wurde. Das Konzept beruht auf einem Client/

Server System mit verteilten Datenbanken basierend auf Lotus Notes (Figure 1). Der Arzt im Spital kann aus dem Patienteninformationssystem (Clinicware®) den Bericht per “Knopfdruck” an den Hausarzt senden. Nach der Freischaltung wird der Bericht in eine PDF-Datei umgewandelt. Die notwendigen administrativen Daten werden in einer HL7 Headerdatei mitgegeben. Beide Dateien werden über die Datendrehscheibe an Lotus Notes übermittelt und dort entsprechend den Angaben an die Empfängern weiter geleitet. Der Hausarzt erhält ein Benachrichtigungsmail an seine normale Mailadresse mit einem Link, der auf den eingegangenen Bericht verweist. Dazu benötigt er lediglich einen Browser mit Internetanschluss um den Bericht weiter zu bearbeiten. Den Teilnehmern stehen drei Datenbanken zur Verfügung (Berichte, Anmeldungen/Hotline und Forum). Dieses IT-Layout bietet grösste Datensicherheit und genügt hohen Ansprüchen an den Datenschutz (Patientengeheimnis). Resultate Zur Zeit sind 171 Ärzte angeschlossen. Dabei handelt es sich um 77 Hausärzte und 94 Ärzte in sechs angeschlossenen Spitälern. Im Kantonsspital Luzern haben 449 Benutzer Zugang zum HAKOM, davon sind 294 Ärzte. Am meisten ist die Berichtsdatenbank im Gebrauch. Bis anhin wurden 3581 Berichte übermittelt, mit einer starken Zunahme ab Mitte 2002 (Figure 2). So werden im dritten Quartal 2002 pro Tag im Schnitt 12,9 Berichte versandt, gegenüber 4,7 Berichte pro Tag vor Jahresfrist. Nach Freischaltung des Berichtes durch den Arzt geht es durchschnittlich 24 Minuten (minimal 16/maximal 36 Minuten) bis der Bericht für den Hausarzt zur Verfügung steht. Der Hausarzt kann frei über die Berichte verfügen (inkl. löschen).

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

Für Anmeldungen und medizinische oder organisatorische Anfragen steht eine weitere Datenbank zur Verfügung. Hier sind 21 Abteilungen des Spitals direkt elektronisch erreichbar. Von dieser Möglichkeit wird von den Zuweisern noch nicht den Erwartungen gemäss Gebrauch gemacht. So sind in den ersten zehn Monaten 2002 erst 153 elektronische Anmeldungen eingegangen. Die Forumsdatenbank steht allen Teilnehmern für den freien Meinungsaustausch offen. Schlussfolgerung Die Schnittstelle mittels HL7-Header und PDF Attachment ist einfach zu realisieren. Das System lässt sich aufgrund der modularen Schnittstellentechnik leicht in ein bestehendes Klinikinformationssystem integrieren. Durch die Verwendung von zwei Informationssystemen (spitalinternes Patientendossier und Kommunikationsplattform) wird die Abhängigkeit von einem Produkt/Lieferant reduziert. Bei gleichzeitiger Anwendung von Standards (Schnittstelle, Technologie) kann die Komplexität reduziert werden.

Figure 1. Kommunikationskonzept.

Die Übermittlung erfolgt übers Internet, dabei werden alle Daten verschlüsselt. Die Sicherheitsvorkehrungen entsprechen dem aktuellen Stand der Technik. Berichte sind für die Hausärzte deutlich schneller verfügbar als über den üblichen Postweg. Das System ermöglicht zudem eine zeitlich asynchrone Kommunikation zwischen den Teilnehmern. Trotzdem können mit Benachrichtigungsmails oder wenn gewünscht mit SMS die Reaktionszeit verkürzt werden. Die strukturierte, spitalexterne Datenablage ist einem einfachen Mailsystem vorzuziehen [3]. Sie dient als Basis für ein späteres externes Patientendossier, welches mehreren Beteiligten offen steht. References 1

Heberer M. Strategic positioning of a University Hospital. M.B.A These GSBA Zürich/New York 1997.

Tschudi P. Internetkommunikation zwischen Hausärzten und Universitätsspital. Praxis; 2002;257-60.

Freiermuth O. Die elektronische Kommunikation zwischen Arztpraxen und Spitälern wird zum Standard. Schweiz Ärztezeitung 2001;82;2237-48.

Figure 2. Anzahl der übermittelten Berichte.

SMI 2003; No 51

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

SGMI SSIM SSMI

Carte de santé: un premier pas vers le futur Réseau d‘informatique sanitaire dans une logique de changement culturel Marzio Della Santa Ufficio del Medico Cantonale, Bellinzona

E-mail: marzio.dellasanta@ti.ch

La santé est une des préoccupations principales de la population en Suisse [1]. Bien que le fonctionnement du système de santé satisfasse largement la majorité de nos citoyens et de la classe politique, l’évolution des coûts de la santé menace l’intégrité de ce même système. Afin de pouvoir maintenir ses prérogatives, il est donc nécessaire de trouver un équilibre entre efficacité, efficience, et équité. Un système de santé est un dispositif complexe, multifonctionnel, et se compose d’une multitude d’acteurs aux compétences variées et aux intérêts spécifiques, structurellement complémentaires, mais parfois conflictuels. En pratique, ça se traduit par une fragmentation du système. Les implications de ce morcellement ne sont pas seulement financières. Les difficultés dans l’échange entre structures sanitaires des informations qui concernent le patient sont souvent cause des souffrances psychologiques et physiques. Pour un malade chronique, une simple interaction entre médicament induite par une connaissance insuffisante du patient, peut signifier l’isolement, la dépression et des nouvelles souffrances qui vont aggraver une qualité de vie déjà compromise.

Correspondance: Dr Marzio Della Santa Chef de projet Rete Sanitaria Ufficio del Medico Cantonale Via Dogana 16 6500 Bellinzona

SMI 2003; No 51

Pour cette raison le CdE du Canton du Tessin déjà en 1998 décida d’élaborer un projet apte à «créer les conditions idéales pour favoriser l’utilisation correcte des structures et des services de santé» [2]. Par le projet on a fournit aux principaux acteurs de santé du Canton l’inspiration nécessaire pour des choix stratégiques à moyen et long terme, qui soient en mesure de promouvoir la coordination et la coopération entre professionnels de santé travaillants dans des structures différentes. Ce projet veut – en substance – servir de détonateur et catalyseur d’un processus de changement culturel [3] à réaliser avec les

principaux acteurs de santé du Canton [4]. Le recours aux nouvelles technologies de gestion électronique de l’information, par nature neutre en ce qui concerne les équilibres de pouvoir du monde de la santé, ont été identifiées dès le début comme constituant la meilleure stratégie pour atteindre les objectifs posés. Etant donné la complexité et la fragmentation du système de santé, il apparaît essentiel de rendre plus efficace l’échange d’informations entre ses acteurs, en permettant ainsi une amélioration de l’efficience mais aussi une plus grande sûreté (privacy) et meilleure qualité des soins. Plusieurs instruments de santé électronique (e-health) ayant un possible impact sur le dispositif organisationnel ont été retenus: de la télé-médecine au dossier patient informatisé, de la carte santé au réseau de santé informatisé. Chaque instrument a ensuite été évalué en termes de faisabilité et non seulement d’opportunité. Après un processus de concertation, l’Etat et les partenaires du projet décidèrent de procéder de façon pragmatique, en concentrant leurs énergies sur la carte de santé électronique. Instrument banal à première vue, et donc acceptable du profil des émotions, celui-ci possède en revanche de grandes potentialités pour modifier radicalement les habitudes des patients et professionnels de la santé. Soit en tant que moyen de transport des données (contenues dans un microprocesseur), soit comme future clef d’accès électronique au réseau d’informatique sanitaire, la carte représente un moyen de pilotage et d’harmonisation du réseau informatique naissant et plus encore, un symbole de la centralité du patient dans le système de santé. C’est justement au niveau de cette dernière propriété, c’est-à-dire le fait d’être un instrument technologique apte à soutenir le

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

processus d’autodétermination du patient (empowerment), que ce trouve un potentiel pour révolutionner, à moyen long terme, le rôle actuel que ce dernier joue dans le système, lui permettant ainsi d’assumer entièrement ses responsabilités.

carte offre implicitement la garantie d’une communication sûre des informations d’un acteur à l’autre. En d’autres termes, le réseau de santé informatisé se destine à un développement harmonieux autour de l’élément fédératif que représente la carte.

Ces considérations sont confirmées par l’expérience acquise dans la majeure partie des pays européens (de façon particulière en Allemagne, en France, et en Italie), mais aussi au Canada, aux Etats-Unis, et au Japon, où la carte est introduite non seulement dans le but de faciliter les procédures administratives (Carte d’assuré), mais aussi d’améliorer la qualité des soins administrés (Carte sanitaire).

L’opportunité de garantir la complémentarité de la Carte de santé avec le Réseau informatisé se trouve renforcée par l’Accord de collaboration paraphé en août 2002 entre le Département de la santé et de la socialité du canton du Tessin et le Département de la santé et de l’action sociale du canton du Canton de Genève. Ce dernier est effectivement en train de définir les conditions pour la création d’un Réseau communautaire d’informatique médicale (RCIM), accessible par l’intermédiaire d’une clef (la Carte de santé), dans le respect des droits du patient.

Sur le plan nationale, une proposition pareille est actuellement à l’étude des deux Chambres dans le cadre de la deuxième révision LAMal. Toutefois, la politique de santé étant de compétence cantonale, c’est au niveau cantonal – et non fédéral – qu’un tel instrument peut être légitimement développé au delà des limites d’une simple carte de comptabilité et d’assurance. Une carte de santé est en effet un véritable instrument de pilotage du système de santé, dont les considérations d’assurance forment uniquement un sous-groupe. Le Conseil Fédéral, sensible à ce problème, a invité le Canton du Tessin – dont le projet est explicitement présenté dans le message relatif à la carte d’assuré (art. 42 a) – à collaborer au développement de la Carte d’assuré, en utilisant les compétences accumulées durant ces dernières années. La future Carte de santé, basée sur les standards internationaux, offrira par exemple au patient la possibilité d’enregistrer certaines données d’urgence, telles que les allergies, les vaccinations, la thérapie pharmacologique, ainsi que la liste des principaux antécédents de santé qui ont caractérisé la vie du patient. La possibilité d’accéder à ces informations permet aux professionnels de la santé de sauver des vies et d’éviter des souffrances inutiles, en contenant sensiblement les coûts de la santé. La possibilité d’utiliser la Carte de santé comme clef d’accès électronique au réseau informatique sanitaire naissant, promet cependant des résultats encore meilleurs. Durant cette phase de constitution du réseau, l’introduction de la

SGMI SSIM SSMI

Pour le projet tessinois la composante technologique occupe, en fin de compte, un rôle mineur. Le vrai défi est de nature culturelle: aujourd’hui rares sont les services et les structures qui recourent à l’informatique dans le déroulement des pratiques médicales. Pour cette raison, ses qualités et ses défauts nous sont encore en grande partie inconnus, ce qui peut provoquer facilement incompréhensions et préjudices. Un exemple: la protection de la sphère privée. L’introduction de la Carte de santé et la mise en réseau des opérateurs de santé permettra au patient de décider systématiquement à qui autoriser l’accès, de manière sûre, aux informations le concernant. L’expérience, par exemple celle du milieu bancaire (carte ec), montre tout de même que l’introduction d’un nouveau système implique toujours un processus d’apprentissage et d’appropriation. Pour cette raison, et en accord avec les différents partenaires du projet, il a été décidé de démarrer le processus avec un projet pilote dans la région urbaine de Lugano, destiné à favoriser le changement culturel et à confirmer les choix opérationnels qui concernent l’emploi de la Carte de santé. Actuellement nous prévoyons la participation d’au maximum 3000 patients et 300–400 opérateurs sanitaires qui travaillent dans des cabinets médicaux, SMI 2003; No 51

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

pharmacies, structures stationnaires publiques et privées, services d’ambulance et d’assistance et soin à domicile. Pendant cette période, plusieurs initiatives permettront de sensibiliser l’opinion publique. La participation volontaire de groupes spécifiques de patients permettra en outre de comprendre les difficultés pratiques et psychologiques existantes dans l’utilisation de ces instruments. A la fin de la phase pilote, grâce à l’obtention d’une meilleure compréhension et acceptation des instruments de santé électronique, il sera possible modifier la base légale existante, de manière à étendre l’utilisation de la carte à le Canton dans son ensemble et de même, à soutenir la création homogène du réseau informatique de santé. Il s’agit d’une stratégie pragmatique, nécessitant du temps, mais qui offre la possibilité à l’Etat de définir les règles permettant de poursuivre des intérêts collectifs et de préserver les droits des patients. Cette stratégie qui nécessite le consentement de tous les différents acteurs est conforme à la culture helvétique de recherche du consensus.

Dans cette optique nous sont parvenues les encouragements de divers observateurs externes. C’est le cas, par exemple, de l’Office fédéral de la technologie et de la formation professionnelle (Département de l’économie) qui, dans le cadre du projet soft[net], a voulu soutenir le projet en raison de sa validité et de sa capacité d’innovation, son rôle pionnier au niveau national, et sa recherche du consentement des différents partenaires. Actuellement on est dans phase préparatoire du pilote, pendant laquelle, avec des partenaires technologiques, on est entrain de préparer les cartes, développer les logiciels e identifier les solutions Hardware adaptes aux différents contextes. Etant l’adhésion volontaire, cette phase est délicate surtout parce qu’on doit recruter les patients et les professionnels de la santé qui veulent participer. Une opportunité unique qui permettra à chacun d’entre eux de contribuer à un processus culturel et structurel auquel tout le monde sera prochainement confronté.

Références

SMI 2003; No 51

Baromètre des préoccupations 2001, GfS, Bern, 2001.

Rapport du Grand Conseil sur la troisième mise à jour des lignes directives et du plan financier 1996–1999. CdS, novembre 1998 (page 20).

Une action pragmatique dictée par les résistances mises en évidence par des groupes de travail auxquels on pris par des professionnels de la santé, des patients, des juristes et des membres de l’Administration cantonale.

ACPT – Association des cliniques privées, APSI – Association des patients de la Suisse italienne, ARODEMS – association romande et tessinoise des directeurs des établissements medico sociaux, santésuisse, EOC – Association des hôpitaux publics, FCTSA – Fédération du Canton du Tessin des services d’ambulance, OFCT – Association cantonale des pharmaciens, OMCS – Association cantonale des médecin, SACD – Conférence cantonale des services d’aide et soin à domicile.

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

Anlässlich der Generalversammlung der SGMI in Luzern traten Ruedi Tschudi, Sekretär, und Hansruedi Straub, Chefredaktor SMI, von Ihren Ämtern zurück. Beiden soll hier nochmals für ihre jahrelange wertvolle Mitarbeit gedankt werden.

SGMI SSIM SSMI

À l‘occasion de l‘assemblée générale de la SSIM à Lucerne, Ruedi Tschudi, le secrétaire, et Hansruedi Straub, le rédacteur de chef du SMI, se sont retirés de leurs fonctions. Nous les remercions encore tous deux vivement pour les années de travail au service de la société.

Die neu gewählten Vorstandsmitglieder / Les nouveaux élus du comité

Benno Sauter

Christian Lovis, MD MPH

Geboren zum Sommerbeginn 1968 in Basel. Matura 1987 und Beginn des Medizinstudiums. Unterbruch und Arbeit im Bau- und Architekturgewerbe sowie Aufbau der Informatik-Kenntisse im Bereich Betriebssysteme, Netzwerk und Hardware. Wiederaufnahme des Medizinstudiums und Abschluss mit Staatsexamen (1997, Basel). Ab 1998 Medizininformatiker im Schweizer Paraplegiker-Zentrum Nottwil, Integration der interdisziplinären, prozessorientierten und ubiquitär verfügbaren elektronischen Krankengeschichte. Berufsbegleitende

Ausbildung zum Informatik Projektleiter (2000). Leiter der Gesundheitsinformatik (2001). Seit Mitte 2003 Aufbau und Leitung des Bereichs Wissensmanagement. Fachgebiete: Elektronische Patientenakte mit interdisziplinärer Prozessabbildung / Funknetzwerke im Spitalumfeld / Auswertungen und Datenanalysen medizinischer Datenbanken / medizininformatische Projekte in der Spitalinformatik / Wissensmanagement im Spital Hobbies: Alles rund um elektronische Geräte und Medien, Musik, Kino.

Christian Lovis is responsible for the Clinical Information System of the Division of Medical Informatics (Prof. Antoine Geissbuhler) at the Geneva University Hospitals. Born in 1962 in Jura, Switzerland, he was trained as internist and staff physician at the Department of Internal Medicine, University of Geneva (Prof. Francis Waldvogel) with special emphasis on emergency medicine. In parallel he studied Medical Informatics at the University of Geneva (Prof. Jean-Raoul Scherrer) with special emphasis on clinical information systems. Between 1998 and 2000, Christian Lovis worked on a two-year post-doctoral fellowship at the Veterans

Affairs Puget Sound, Seattle, WA as Senior Research Fellow. His research focused on electronic order entry. During that time he also gained a Master in Public Health from the University of Washington. His double background in Medicine and Informatics allowed him to develop a new line of research on knowledge representation and information tools for medical specialists, which in 2001 was awarded a full professorship by the Swiss National Science Foundation. He is author of a large number of peer-reviewed papers and member of the Editorial Board of the Journal of the American Medical Informatics Association.

Ich habe an der Universität Zürich Mathematik, Informatik und Operations Research studiert und 1984 mit dem Diplom abgeschlossen sowie 1997-99 das Nachdiplumstudium in Unternehmensführung der Universität St. Gallen absolviert (Exec. MBA HSG). Beruflich habe ich mich als Informatiker im

Generalstab in Bern mit Simulation, Führungsinformationssystemen und Informatikstrategie befasst und bin seit 1999 Informatikleiter (Chief Executive Officer) des UniversitätsSpitals Zürich. Ich wohne mit meiner Frau Christa und meinem Sohn Lukas in Zürich.

Felix Heer

SMI 2003; No 51

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

SGMI SSIM SSMI

Clinical study on the acceptance of a Web-based self-help guide for bulimia Tony Lam 1, Isabelle Carrard 2 , Maria Damoiseau Carola 3, Anne-Christine Volkart 3 1 NetUnion Sarl, Lausanne 2 Liaison Psychiatry Unit, University Hospitals of Geneva 3 Institutions Psychiatriques du Valais Romand, Hôpital de

Phone: +41 (0)21 331 1577 E-mail: lam@netunion.com

Abstract Eating disorders are a common source of psychiatric morbidity among young women. In order to respond to such problems, it is urgent to develop and evaluate more accessible treatment methods. Self-help manuals could bring a solution. Research on self-help manuals for the treatment of bulimia has already been conducted in Anglo-Saxon countries, where efficiency of such methods is recognised. These manuals are based on cognitive and behavioural therapy and provide step-by-step methods to face eating disorders by progressively recovering self-control over one’s food behaviour. The study is part of Salut! (IST-200025026), a 39-month project funded by the European Commission under the 5th Framework Programme and by the Swiss Federal Office for Education and Science (OFES). This project, started in January 2001, aims at developing Web-based and mobile tools for diagnosis, treatment, and prevention of eating disorders. One of the main developments of the project is a Web-based self-help guide for bulimia. Several European countries are participating in clinical trials aimed at evaluating this guide developed by the liaison psychiatry unit of the University Hospitals of Geneva and by NetUnion in Lausanne. Project description

Malévoz, Monthey Corresponding author: Tony Lam NetUnion Sarl 20, av. des Figuiers 1007 Lausanne Switzerland

SMI 2003; No 51

Salut! uses advances in information and telecommunication technology to design, prototype and validate innovative tools and cost effective strategies for the prevention, diagnosis and treatment of eating disorders. The project has two main objectives: (a) to develop and validate online tools and mobile components for supporting the prevention and treatment of bulimia; (b) to

facilitate the exchange of reliable information about eating disorders between health professionals, researchers and the general public. A main project component is the implementation of an online multi-lingual “selfhelp” guide (SHG) for outpatient treatment of bulimia. This guide contains evaluation and treatment modules designed to progressively deliver users techniques to get through their illness by themselves. The main modules of the SHG are expanded to form the basis of a more generic platform for supporting other online and mobile applications based on Cognitive Behavioural Therapy (CBT). Salut! also set up a network of regional Web portals to facilitate the access and exchange of information between health professionals, researchers and the general public. These portals [1], launched in 2002, seek to encourage the dissemination of unbiased information about eating disorders, listing of current events, and local resources for people searching for information about eating disorders. Self-help guide for bulimia The online version of the “Self-Help Guide” [2], developed by the University Hospitals of Geneva (HUG) and NetUnion, is composed of seven sequential steps (cf. figure 1): (1) preparing yourself for change, (2) observing yourself, (3) changing your behaviour, (4) changing the way you think, (5) identify and solve your problems, (6) self-assertion, and (7) conclusion. Steps are composed of lessons, exercises and several examples illustrated by a virtual character called Sarah. One of the most important exercises of the SHG is the food diary and its weekly summaries: users have to record their meals, binges, binge triggers, etc. daily. At the end of each week, they are invited to analyse their eating behaviour

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

person during pre-, post-treatment, and follow-up evaluation sessions. Technical aspects The online version of the SHG is based on a modular and multi-lingual e-health/elearning platform developed by NetUnion in Java (JSP/Servlets) and XML with a PostgreSQL database backend. The current release contains six main modules: (1) user management, (2) editing/publishing, (3) evaluation, (4) program, (5) results analysis, and (6) messaging.

Figure 1. SHG steps

using a series of reports generated from their food diary (cf. figure 2). Clinical trials Medical partners in France, Spain, Sweden, and Switzerland are conducting clinical trials on the SHG using a common research protocol accepted by the psychiatry department’s ethical committees of Swiss partners. Participants in the study are selected according to strict inclusion criteria. The main goals of this first study are the evaluation of acceptance and efficiency of the SHG in a population of patients suffering from bulimia, purging type (and more precisely “vomiting type”). Trials in Switzerland are currently being conducted by HUG and IPVR (Institutions Psychiatriques du Valais Romand) in the French speaking part of Switzerland. German and Italian versions of the SHG are being prepared to anticipate extension of the trials in other Swiss cantons. The clinical trials have a six-month cycle (four months self-treatment, two months follow-up). All trial participants maintain regular contact with their assigned coaches. Coaches can monitor patient progress via the result analysis module of the SHG and by using standardized questionnaires (EDI-2 [3], SCL-90(R) [4]) administered in

The user management module was designed in order to provide flexible, hierarchical and compartmented administration of users. The different user types have restricted rights according to their role(s). The overriding design criterion is to maintain confidentiality between coaches and participants. The editing module is a powerful tool that allows rapid deployment of multi-lingual versioning of the SHG. This module includes interfaces for editing program content (lessons, exercises, examples), user interfaces (menus, buttons, etc.), evaluation questionnaires and the online help. Using this module, researchers can publish new program content directly on a pre-production server, setup new or edit existing research questionnaires (based on an XML template) to support clinical research and data collection. Other questionnaire-specific parameters – dimensions and norms – can also be edited using this interface. Questionnaires can be viewed and printed using XSL style sheets. The result analysis module enables easy monitoring of patient progress by providing up to the minute results of participant response to questionnaires and exercise results. Results from questionnaires and exercises are presented graphically by creating SVG (Scalable Vector Graphics) files dynamically and rendering it on the fly as PNG images. The usage analysis module also provides managerial and administrative information tailored to hospital administrators, program directors, coaches, and researchers. Information from this module is intended to help monitor information such as clinical workload, and potential impact on health care

SMI 2003; No 51

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

delivery. Prototyping of specific mobile components is ongoing. The aim is to provide anytime / anywhere access to specific exercises or components of the SHG. Privacy and security During the evaluation process, access to the SHG is restricted to users who participate in the clinical trial. Otherwise, the lessons, exercises, and other program modules are not made available to the general public. A main design concern is protecting the confidentiality of the coach / patient relationship. Several steps were taken to assure this. For example, only coaches have access to the following information: patient responses to questionnaires and exercises, patient contact information, and messages exchanged between patient and coach. These measures were achieved using technical and procedural solutions. The participants are only identified by a pseudonym. Real name and patient files are kept separately under lock and key. An internal messaging module allows coaches and patients to exchange messages without exposing their

SGMI SSIM SSMI

e-mail address. The messaging module also implements rules restricting who can receive messages from whom. For example, coaches can only write to participants in their own groups. The message is written using the pseudonym. The application respects prevailing European and local legislation on privacy and data protection [5]. However, a balance between security and ease of use had to be found in order to provide a reasonable level of security without asking too much IT knowledge from users. Again, appropriate technical and policy solutions were developed to meet this challenge. Collected information is transmitted over a secured connection (SSL). Safe password rules are implement to request users to choose “more secured” passwords. User passwords are stored as MD5 hash in the database. Intermediate results and conclusion The online version of the SHG was released in September 2002 and is currently available in French, Spanish, Swedish, and English. German and Italian versions, as

Figure 2. Evolution of number of meals and binges along the weeks

SMI 2003; No 51

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18iĂ¨mes journĂŠes annuelles

well as, prototypes of mobile components, will be available first quarter 2003. Clinical trials were launched in October 2002 and will continue throughout 2003 in four countries. Final results are expected by first quarter 2004. The Self Help Guide research is already providing interesting feedback on the use of online tools for collaborative research, and novel cost effective treatment strategies. The multi-lingual scope of the application gives us the opportunity to deploy the SHG in most European countries and to easily expand its use to other languages.

SGMI SSIM SSMI

The opportunity for providing service to a minority population is also perceived to be an unexpected benefit (e.g. providing the Spanish version of the SHG to the Spanish speaking communities in Sweden). Moreover, the CBT platform, by its modularity, allows the creation of other self-help applications based on cognitive behavioural therapy and a similar clinical framework. Insuring and promoting trust was also an issue. The project chose the strategy of using HON Code of Conduct (HONcode) [6] accreditation, an external quality benchmark indicating compliance to prevailing best practice [7] for online dissemi-

nation of health-related information.

SMI 2003; No 51

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18iĂ¨mes journĂŠes annuelles

SGMI SSIM SSMI

Bridging the gap between medical narratives and structured data in the computerized patient record Christian Lovis, Alexander Lamb, Anne-Marie Rassinoux, Antoine Geissbuhler Division of Medical Informatics, University Hospitals of

Phone: +41 (0)22 372 6201 E-mail: christian.lovis@hcuge.ch

Abstract Care documentation is a central activity of care delivery and a mandatory step to the development of predictive and supportive care informatics in a collaborative paradigm. Beside its importance for classical data processing in healthcare such as reimbursement claims, scientific research or teaching, care documentation must also fit within the daily work of healthcare providers without intrusion and remain a precise and life biography of the patient. In this view, human-machine interfaces and philosophy behind data acquisition and restitution interfaces are of major importance. There have always been some antagonisms between narratives and structured data entry, both having advantages and disadvantages, supporters and detractors. In real practice, most documents used in clinical settings are made both of typed or structured data and narratives or free texts. In order to try to have a common source for all these information, we developed a unified representation, acquisition and storage system for medical information. To use this system in our computerized patient record, we use a middleware based on HTTP and XML that permits standardized exchanges between applications and data repositories. This paper is devoted to the description of some part of our system as well as its real implementation in a CPR working in the five Geneva University Hospitals.

Geneva Corresponding author: Prof. Dr. med. Christian Lovis University Hospitals of Geneva 21, rue Micheli-du-Crest CH-1211 Geneva 4 Switzerland

SMI 2003; No 51

Introduction We do present a concept that permits the unification of semi-structured information such as what can be found in free texts and strictly typed data such as fields in question-

naires, allowing a common representation at several levels, from user interfaces to data storage. This concept is currently being implemented and already used at the university Hospitals of Geneva. The ultimate goal of documentation is to provide accurate and timely clinical information for patient care and complete documentation for all stakeholders [1]. The diversity of stakeholders is a characteristic of a large health care organization. Each professional group deals with its personal and often partial view of the global system, with its knowledge, culture and terminologies. Moreover, the same professional can have multiple roles: a nurse can be planning the care of a patient, checking vital signs at the bedside or rounding with physicians. In each role, the information needs and the ergonomics of the system will be different: a computer at the nursesâ&#x20AC;&#x2122; station would be necessary for the care planning while a handheld device would work best at the bedside. However, computer-based systems have historically been built to help each professional group to deal with their information needs, thus respecting or reinforcing the functional boundaries of the groups. These systems have then been interfaced to others in order to form hospital-wide information systems, generally in a way that mimicked existing paper-based transactions. The development of healthcare delivery networks is increasing the diversity of different stakeholders, thus complicating the task of integration. The need for resource management and integrated clinical pathways requires a transversal understanding both of care structures and of data representation, management and acquisition. The main challenge is the ability to represent the knowledge in a way that is usable, maintainable and meaningful to diverse users. However, this implies numerous requirements that are sometimes in contradiction. One of these apparent contradictions concerns the discussion of structured data versus free texts [2, 3]. The underlying problem is far 21

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

from trivial and major consequences may have to be faced in the long term. On one extreme, one could consider to have a CPR by just scanning all paper-based documents. Such a system would answer immediate

and important needs, such as ease of access, ubiquitous record, centralized management and right accesses, etc at reasonable cost. However, it would obviously not answer some important needs such as data processing and analysis, decision support or clinical pathways. At the other extreme, a patient record could be entirely model driven with a “Point&Click” user interface, such as developed within the Pen&Pad project [4]. In this system, a complete patient record can be mapped to a semantic model and the user interface for patient data acquisition is driven by the underlying model. Between these two extremes, many clinical information systems (CIS) do have to face with the coexistence of both documents and structured data management. There is a challenge in trying to bridge the gap between structured data and free text at several levels, from data acquisition to information processing and knowledge representation. •

Structured versus narrative data acquisition (questionnaires versus documents) There is a long tradition of controversy between the advantages and disadvantages of using structured data entry compared to free text. The two major arguments underlying the use of structured data entry are a) the analytical power of structured data and b) the ease

and speed of data entry. Both points are of high importance in medicine, as they allow analysis of data, decision support, knowledge coupling, etc. On the other side, free text entry as a far better power of expression, allows layout formatting such as emphasis and is easier to use in numerous situations [5, 6]. •

Closed versus open user interfaces (Menus versus command-line) historically, one have moved from commend-line driven systems, such as MS-DOS to menus and dialogs driven which are at the core of MS-Windows, Apple MacOS, etc. The explosion of computers proves better as anything that menus and dialogs driven-systems are easier to use. However, this does not prove to be true when very large number of items must be available, such as classification, order entry, etc. Several systems offer now again command-line entry, or their “modern” equivalent such as completers, natural language analyzers, parsers, etc. that allow users to type in a way more or less similar to free text and have the system interpret the entry.

•

Patient oriented versus data-type oriented storage (patient record versus data warehouse) Most computerized patient record are built upon relational databases, which are most often structured regarding data type attributes rather than patient centric. So, for instance, there might be some tables for laboratory data, some other for administrative information, some other devoted to drug prescription, etc. Accesses to the whole record of any patient might require consolidating data coming from numerous tables or databases. On the other hand, some document oriented system can store all data of any single patient at a unique place, therefore facilitating and speeding accesses for patientcentric queries, but making cohort

SMI 2003; No 51

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18iĂ¨mes journĂŠes annuelles

studies almost not possible, or at the cost of high processing power. Merging questionnaires and documents In a move to merge document-based medical narratives and questionnaires data acquisition, we consider that questionnaires and free texts documents have the same structure and representation; they are built

data acquisition and has more a structure of questionnaires. Both interfaces do use the same data representation and same repositories and all transactions are made using a common http/xml based middleware. Questionnaires use mostly basic attributes types, such as Boolean, dates, lists or numeric whereas documents are mostly formed using paragraphs. Documents have therefore been highly structured based on their paragraphs, such as patient history, discussion, conclusions. All documents and questionnaires do share a set of similar characteristics used for document management and workflow, privacy control and versioning. Data model The data model is based on the idea that all documents are built using a set of attributes (the data fields) grouped with a document class (the data context) (see figure 2). The model is separated from the references to attributes. It is based on the idea that all documents are built using a set of shared attributes with a given type and meaning. Therefore, the fields of a document are only references to attributes. The values are represented as lists of attribute-value pairs connected to an instance of a document for a given patient at a given date. Once filled,

upon a description of fields. Two different user interfaces allow the acquisition of data. Both do allow the acquisition of structured and typed data as well as blocks of free texts. One interface has a strong emphasis on layout control and is more devoted to such documents as discharge letters or reports. The other is more devoted to fast

SMI 2003; No 51

the resulting values will be stored in a separate repository, which contains attributesvalues pairs. The elements of figure 3 are explained thereafter: â&#x20AC;˘

Data context Logical model that regroups fields. It is, in fact, the list of all fields needed to create a questionnaire, a document or a specialized record such as the record of anesthesiology consultation.

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

•

Field It is the label of an attribute in a questionnaire or a document. Each field is linked to a specific attribute. For example, in a given questionnaire, it might be a field “Smoker?” linked to the attribute tobacco_use Attribute An attribute is one entry in the common dictionary of classes of facts. For each attribute classes, several properties can be specified, such as data type, description, links to external models, etc. Within the dictionary, each attribute has a unique internal identifier that cannot be removed. Each attribute can have dates of validity, so that attributes are never deleted but only inactivated. Each attribute belongs to one of the seven basic data types, which are enumerated, date, decimal, image, integer, long text and short text. The values that can take an attribute can optionally be limited by a code value, such as “Yes”, “No”. If needed, these possibles values can be aggregated in Groups such as [“Yes”, “No”] for example. Attributes can be linked to external classification, and numerous are already linked, to ICD10 for example. View A view is a way a data context will be displayed. It does not define which fields (and linked attributes) are used, but the format, layout, authorization schemes and components used for display. An example of view is PDF for Adobe Acrobat Portable Document Format®. Another of the Views of a data context could be a Microsoft Rich Text Format template that can allow complex editing and layout. Page Page allows to group fields and will be used to produce automatically acquisition user interfaces. It is useful to avoid this kind of very long document that must be

scrolled. A page can be assimilated to panels in a web-based questionnaires or to pages in a Word document. •

Once a data context has been instantiated, such as a questionnaire or a document for example, attributes will receive data, that is, the attributes’ values. All data of a data context represent its data set. A link is maintained between data and attribute to allow attributevalues to be retrieved, as well as between data set and data context, so that the whole context in which data have been acquired is kept.

Semantic model A semantic model acting as a “semantic” shield over the list of attributes has rapidly proven to be necessary with the increase of the size of the number of attributes and the apparition of synonyms or duplicates for example. One of the main problems encountered has been to have a way to create rapidly new attributes without loosing the added-value of a semantic representation. However, there is a large pressure for having new attributes fast whereas organizing these attributes in a semantic model can take long time and discussions. We therefore completely separate models, the data model and the semantic model. The dictionary of attributes can be considered as a flat list, or almost equivalent. However, the process of building a model on the top of this dictionary is ongoing. Attributes are linked to concepts with only four relation types: •

isA. It is a subsumption relation that can be used in various cases, such as cl_femur isA cl_bone or cl_headache isA cl_pain.

•

partOf. is the usual partitioning link.

•

equiv. It is an equivalence relation used to express synonymy or medical equivalence. It allows the creation of “grapes” attributes that are similar.

•

isNot is a negation that can be used to ease the matching of similar concept, but that would have Fig-

SMI 2003; No 51

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

ure 2. Overall architecture (dashed = not yet implemented) been expressed using negation in the Attributes. The same relations are used to built a network of concepts, avoiding a deep model whenever possible [7]. The pragmatic goal is to reuse data in the CPR and to be able to help decision-support. Storage The storage is made in the two main repositories (in darker in figure 2) of the CPR, the patient-centric database, DoMed (documents médicaux) and the data-centric database, BDOC (Banque de Données Opérationnelle Clinique). All accesses to the data are made through the middleware MUSIC (Middleware Unifié du Système d’Information Clinique) in lighter in figure 2 using HTTP / XML messages. Both databases do share a common set of components for management, accesses, auditing, etc.

cal character recognition (OCR) to DoMed. All data sets that have been validated in the BDOC database will also be saved as consolidated documents, in general using PDF format, in the DoMed database. So is it also for the laboratory results too. This allows being able to display the document of the CPR very fast to the users, without having to do complex queries in order to consolidate data. However, data are available in a data-centric database, therefore permitting analysis and decision-support. Conclusion The patient record is made of very heterogeneous documents originating from numerous sources. Most of these documents are made both of structured data and narratives. Some of them, such as admission notes from general practitioners, will not be available largely in electronic form for a long time and must be scanned. We have developed a way to have a unique repository for all these documents and data in implementing a dual storage architecture that is tightly integrated. This system is part of our n-tiers component-based architecture and can be used with XML formatted messages and the HTTP protocol. Up to now, more than 6’000’000 documents are stored in the DoMed database and more than 100’000 data sets in the BDOC database.

Beside questionnaires and documents, BDOC does hold all structured data on clinical activity about patients, such as order entry and laboratory. All these data have the common characteristics to have very defined acquisition pathways. In the contrary, the DoMed database is much more heterogeneous and can hold pretty much any kind of data being file oriented. So, the scanning The separation of the semantic model and system that will be used in the near future to the pragmatic attributes dictionary layer has scan all documents still on paper will send allowed a fast growth of the number of data all its outputs, scanning and texts from opticontexts, although the building of the semantic model remains an important problem and challenge that will have to be solved. References 1

Tang PC, LaRosa MP, Gorden SM. Use of computer-based records, completeness of documentation, and appropriateness of documented clinical decisions. J Am Med Inform Assoc 1999;6(3):245-51.

Sideli RV, Johnson SB, Clayton PD. Full-text document storage and retrieval in a clinical information system. Comput Methods Programs Biomed 1999;60(3):153-81.

Lovis C, Baud RH, Planche P. Power of expression in the electronic patient record: structured data or narrative text? [In Process Citation]. Int J Med Inf 2000;58-59:101-10.

Kirby J, Rector AL. The PEN&PAD data entry system: from prototype to practical system. Proc AMIA Annu Fall Symp 1996:709-13.

Stein HD, Nadkarni P, Erdos J, Miller PL. Exploring the degree of concordance of coded and textual data in answering clinical queries from a clinical data repository. J Am Med Inform Assoc 2000;7:42-54.

Tange HJ, Hasman A, de Vries Robbe PF, Schouten HC. Medical narratives in electronic

Acknowledgment Part of the work has been funded by the Swiss National Science Foundation 632 -066041

SMI 2003; No 51

SGMI SSIM SSMI

RAFT: un réseau de télémédecine en Afrique francophone Ousmane Ly Antoine Geissbühler

Télémédecine avec l’Afrique francophone: l’expérience malienne

Hôpitaux Universitaires de Genève

Les programmes de télémédecine et de téléenseignement menés depuis deux ans en collaboration entre les Hôpitaux Universitaires de Genève et les centres hospitaliers du Mali, et, plus récemment, de Mauritanie, ont permis de mieux cerner les besoins et possibilités qu’offrent les technologies de l’information et de la communication dans le domaine de la santé.

www.emh.ch

verlag@emh.ch

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

Ces projets ont débuté sur un mode «traditionnel», dans lequel les connaissances du Nord sont diffusées vers le Sud sous la forme de cours de formation ou de consultations à distance, en utilisant le réseau Internet comme outil de distribution, permettant ainsi d’atteindre et de faire participer des correspondant non seulement dans les capitales, mais également dans les régions. Il est progressivement apparu que des adaptations à ce modèle étaient souhaitables pour en augmenter l’utilité, aussi bien au niveau du contenu qu’au niveau de la forme de ces échanges. L’axe Nord-Sud qui prévalait initialement a été rapidement complété par des contributions Sud-Nord au cours desquelles les connaissances du Sud venaient augmenter celles du Nord, donnant à l’échange une dimension de partenariat mutuellement enrichissant. Le concept devrait être développé vers des collaborations Sud-Sud qui sont probablement les plus aptes à répondre aux besoins actuels des systèmes de santé, besoins fortement ancrés dans les réalités culturelles et économiques régionales. Correspondance: Prof. Dr Antoine Geissbühler Division d’Informatique Médicale, HUG Hôpital Cantonal 24, rue Micheli-du-Crest CH-1211 Genève 14 Tél. 022 372 6201 E-mail: antoine.geissbuhler@dim.hcu ge.ch

SMI 2003; No 51

Il est donc nécessaire de développer ces réseaux régionaux afin d’assurer des échanges aussi efficaces que possible, et, par la même occasion, de limiter le risque de voir les technologies de l’information et de la communication creuser encore l’écart qui existe entre les grandes villes et les zones

rurales. Dans ce contexte, la création, la diffusion et la maintenance de connaissances à valeur régionale, notamment sous la forme de contenu en ligne, permet de construire un pont entre les connaissances médicales globales et les réalités locales, facilitant non seulement leur application, mais également un enrichissement réciproque. Le projet malien «Keneya-Blown», ou «vestibule de la santé», débuté en 2000 avec l’aide de l’Etat de Genève, illustre bien ces différents aspects. Ce projet consiste au développement d’un réseau national de télémédecine au Mali, reliant a) les différents centres médicaux de la capitale par un réseau métropolitain sans fil, b) depuis la capitale les différents hôpitaux régionaux, et, c) à terme, les zones rurales, en utilisant Internet. Il propose des outils de communication aux médecins, un portail médical, ainsi que des activités de téléenseignement et de téléconsultation. Ces outils sont basés sur des technologies tenant compte des impératifs d’infrastructure, visant à les rendre accessibles aussi loin que possible sur les réseaux de communication dans les régions. Les activités de télémédecine ont initialement consisté en des séances de formation post-graduée données par des spécialistes européens à la demande des médecins maliens, ainsi que plusieurs téléconsultations médicales durant lesquelles le cas de patients maliens étaient examinés à distance par des experts de Genève. La parcimonie du système permettant de suivre confortablement un cours sur une bande passante de moins de 30 kbits/secondes permet à une audience large d’y assister, de manière interactive. Par exemple, les cours sont suivis non seulement à Bamako, mais également à Ségou et à Tombouctou, faisant déjà profiter à des médecins en périphérie de cours de formation continue. L’échange en ligne de données médicales et d’images radiologiques permet également de réaliser

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

des consultations à distance, notamment dans des domaines de pointe où l’expertise locale est insuffisante, par exemple en neurochirurgie au Mali. La figure 1 présente une jeune patiente malienne, opérée il y a quelques années à Genève, est suivie périodiquement, à distance, en collaboration avec ses médecins maliens

SGMI SSIM SSMI

Mais la complémentarité des compétences régionales est encore plus évidente lorsqu’il s’agit de partager des connaissances réellement applicables sur le terrain. Les stratégies médicales utilisées dans les pays du Nord ne sont pas toujours applicables dans le Sud, pour des raisons aussi bien économiques que culturelles. L’idée est donc de développer un réseau de télémédecine Sud-Sud reliant les centres de compétences des pays d’Afrique de l’Ouest. A ce stade, des centres médicaux de onze pays participent à l’élaboration de ce réseau de compétences (fig. 2). En parallèle à cette approche venant «du haut», c’est-à-dire des capitales, il est nécessaire d’étendre le réseau depuis la périphérie, en utilisant des solutions qui n’impliquent pas d’infrastructures lourdes au sol, dont le déploiement dans certaines régions éloignées va encore prendre plusieurs années. Les technologies satellitaires offrent de ce point de vue des possibilités très intéressantes, et de plus en plus abordables.

Fig. 1

et les neurochirurgiens genevois, afin de s’assurer qu’elle continue de recevoir une prise en charge adéquate sans pour autant devoir quitter son environnement familial. Lorsqu’un des centres hospitaliers de Bamako a été équipé de son propre système d’émission de téléenseignement en 2002, il est apparu qu’en plus de la dimension nationale, à savoir la diffusion d’enseignement depuis la capitale vers les régions, d’autres collaborations régionales et internationales étaient possibles et souhaitables. En effet, les compétences des différents centres hospitaliers sont souvent complémentaires. Il est donc utile de pouvoir échanger et partager les différentes expertises en utilisant les technologies de l’information et de la communication. Par exemple, les médecins en formation de médecine tropicale à Genève peuvent maintenant suivre des cours donnés par des médecins maliens, et un dermatologue genevois, confronté à un cas de lèpre, situation inhabituelle sous nos climats, est conseillé par un spécialiste malien, familier avec cette pathologie et les stratégies de prise en charge.

La motivation du développement de la périphérie est double: d’une part, il s’agit d’amener de l’expertise médicale aux endroits où elle est le plus nécessaire, et, d’autre part, apprendre par l’utilisation sur le terrain à construire des outils et du contenu vraiment utiles au niveau régional. Le contenu local, le savoir local, est un enjeu majeur pour l’identité d’une communauté régionale et pour sa capacité à collaborer efficacement. Il est indispensable pour traduire les connaissances médicales globales pour les rendre applicables aux réalités régionales, en tenant compte des possibilités techniques et financières, des déterminants sociaux et culturels, et des connaissances traditionnelles. La gestion de ce contenu nécessite différents types de formation. Une formation technologique quant à la mise en ligne sur Internet, la gestion des équipements et des utilisateurs. Une formation technique sur les bonnes pratiques de la gestion documentaire, la gestion de projet et l’organisation des modes de collaboration. Enfin, une formation conceptuelle quant à la qualité de l’information médicale sur Internet, et l’accessibilité par les différents

SMI 2003; No 51

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

intervenants, qu’il s’agisse de professionnels de la santé, de patients ou de citoyens. Enfin, cette gestion nécessite des outils nouveaux. On voit déjà apparaître des

Fig. 2

moteurs de recherche régionalisés, tenant compte des aspects géographiques et linguistiques. C’est un début. D’autres outils, basés sur les technologies peer-topeer ou des philosophies de partage telle que l’Open Source, vont certainement jouer un rôle important dans les années à venir. Il faut également s’assurer que l’information produite respecte des critères de qualité et de crédibilité reconnus, ainsi qu’un niveau éthique élevé. Les codes de conduite, tel que celui proposé par la Fondation Healthon-the-Net, le HON code, sont des guides utiles. Le projet RAFT Le Réseau d’Afrique Francophone de Télémédecine RAFT est un projet qui va mettre en relation des institutions sanitaires de formations d’environ douze pays d’Afrique Francophone qui sont: le Benin, le Burkina Faso, le Cameroun, le Mali, le Maroc, la Mauritanie, le Niger, le Sénégal, le Tchad, le Togo et la Tunisie. En pratique, le projet RAFT prévoit d’équiper les institutions et de les aider à se connecter aussi bien vers les réseaux internationaux que vers leurs réseaux nationaux.

SMI 2003; No 51

L’équipement de chaque institution consiste en : •

un système de réception et de projection de téléenseignement sur Internet;

•

un système d’emission de téléenseignement sur Internet, également utilisable pour des téléconsultations et comme serveur Web;

•

une connexion à Internet à au moins 64 kbits/s en émission et en reception;

•

l’accès au serveur de télépathologie développé par le Prof. Martin Oberholzer, également utilisable pour le partage d’autres types d’images (dermatologie, radiologie).

Le projet prévoit également une assistance à l’exploitation du réseau et à son dépannage. Le dépannage matériel de base sera effectué par le correspondant local. En ce qui concerne l’assistance technique avancée, elle sera initialement fournie depuis Genève par la Division d’Informatique Médicale (DIM) des HUG, mais devrait être ensuite décentralisée. De plus, des outils informatiques spécifiques devront être développés afin de compléter l’offre applicative déjà disponible, ainsi qu’un site Web équipé d’outils collaboratifs (forums de discussions, annuaire, bibliothèque des sessions de téléenseignement, etc). Le projet bénéficiera en outre de l’expertise de l’Institut de Pathologie de l’Hôpital Universitaire de Bâle dans le domaine de la télépathologie et l’organisation du travail collaboratif sur Internet. Afin de permettre l’accès à ce réseau depuis des sites distants, inaccessibles par les infrastructures téléphoniques, des points d’accès satellitaires, basés sur une nouvelle technologie particulièrement attractive en raison de son faible coût, seront déployés. Un site pilote, installé dans une commune rurale du pays Dogon (Mali) donne accès à un hôpital rural aux différentes ressources du réseau de télémédecine (accès aux médecins spécialistes de Bamako ou de Genève, accès au portail médical malien ainsi qu’à de nombreuses ressources

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

médicales). Le projet prévoit l’installation d’une douzaine de ces points d’accès dans différents environnements ruraux, ainsi que le développement d’un modèle économique pour une transition vers un financement industriel, en envisageant notamment la possibilité d’utiliser ces points d’accès pour d’autres activités (éducatives, économiques, etc.), ce qui permettrait d’inscrire le projet dans une logique de durabilité et de pérennité. La création et la maintenance de contenu médical adapté aux spécificités locales est un enjeux primordial pour la réussite d’un réseau de télémédecine réellement utilisable. Ces activités nécessitent une formation dans le domaine de la gestion documentaire. Le projet prévoit la formation de documentalistes, afin qu’ils ou elles soient capables de gérer le contenu de bases documentaires et de portails médicaux nationaux, et de collaborer avec les médecins et les techniciens informaticiens. Par contre, le projet ne prévoit pas de financer la création de contenu per se, cette tâche étant confiée aux institutions participant au projet.

SGMI SSIM SSMI

Enfin, le projet prévoit une évaluation des aspects organisationnels, économiques, médicaux et socioculturels, afin de permettre, le cas échéant, une adaptation des stratégies utilisées pour des projets ultérieurs. Dans un premier temps, les stratégies de chaque institution pour déployer, utiliser et pérenniser le système seront comparées, sachant que le projet prévoit que chaque institution devra trouver un financement pour supporter les frais de maintenance du système dès la troisième année d’utilisation. Dans un deuxième temps, une évaluation qualitative des impacts médicaux et socioculturels sera effectuée, avec l’aide d’un doctorant en anthropologie médicale. Parallèlement, des indicateurs quantitatifs seront suivis en continu, notamment grâce aux outils de monitoring centraux : nombre de sessions de téléenseignement et de téléconsultation, nombre de participants (connectés), activité sur le site du projet, activité sur les sites des institutions participantes, etc.

SMI 2003; No 51

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

SGMI SSIM SSMI

Clinical Decision-Support Tool for Acute Appendicitis Stéphane Meystre, MD, MS.

Abstract

Department of Medical Informatics, University of Utah, Salt Lake City, Utah, USA.

The diagnosis of acute appendicitis is still an issue, even with the availability of medical imaging techniques like ultrasonography or computed tomography, and after the publication and evaluation of many clinical scores. The main part of the decision still uses the clinical experience, often limited since physicians taking care of emergency department patients are usually residents or junior attending physicians. Clinical scores rarely comply with all required standards of quality, and are always used separately. In this project, a new approach was tried, developing a decision-support tool with an expert system development application, and using three clinical scores chosen for their qualities. Introduction Acute appendicitis diagnosis issue: Although the treatment of acute appendicitis is simple and straightforward, its diagnosis remains a challenge, and the negative appendectomy rate in large series range from 15% to 33%. Furthermore, in the patients with either atypical history or equivocal physical examination findings, particularly in women of childbearing age, the negative appendectomy rate has been as high as 45%. With an annual rate of 250’000 cases in the US and an incidence of 86 every 100’000 people worldwide [1,2], acute appendicitis is a common acute abdomen condition in the emergency department [3].

Author’s address: Dr. med. Stéphane Meystre Department of Medical Informatics University of Utah School of Medicine 30 North 1900 East – Room AB193 Salt Lake City, UT 84132-2913, USA Phone: 001 801 5871414 E-mail: stephane.meystre@hsc.utah.edu

SMI 2003; No 51

The treatment is surgical: appendectomy (removal of the appendix). It is one of the most frequently performed surgical procedures in the United States and the most common surgical emergency of the abdomen [4]. Acute appendicitis is a common condition in children, accounting for one-third of all

pediatric hospital admissions with acute abdominal pain. This difficult diagnosis has to consider and weigh the morbidity of unnecessary appendectomies against the potential morbidity and mortality of neglected acute appendicitis. The pediatric population, especially children below 5 years, is often subject to higher missed or erroneous diagnosis, due to the lack of specific symptoms and signs. Women are even more at risk, because of the large gynecological differential diagnosis. It was shown that the specificity of the Alvarado score was of 92% with men, 90% with children and 67% only with women [5]. How to improve appendicitis diagnosis? Attempts to increase diagnostic accuracy have included computer-aided diagnosis, imaging by ultrasound scanning and computed tomography, and laparoscopy. Ultrasonography was shown to be helpful in combination of clinical evaluation and score(s) eventually [6,7]. In these studies, GCUS (Graded Compression UltraSonography) was used when the clinical evaluation with the score of Alvarado didn’t have clear results. It was also well mentioned that the accuracy of ultrasonography is dependent on the experience of the ultrasonographer. This type of imaging was also included in a score, reaching excellent sensitivity and specificity [8]. The score included 5 clinical items and the ultrasonography. The sensitivity was 82%, and the specificity 96%. The use of computed tomography has also increased these last years, with variable results, as described in Lee et al. [3]. They showed that computed tomography and ultrasonography didn’t improve the diagnosis, and even caused delays in the treatment, causing avoidable morbidity. Some computer-based systems were developed to try helping clinicians in this difficult task. A group in Norway developed one of these systems to evaluate the physician’s probability of acute appendicitis estimate [9]. The system they developed performed like the

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

experienced surgeons in diagnostic accuracy, but avoided the usual 10% overestimate of initial probability made by surgeons. The area under the ROC (Receiver Operating Characteristic) curve was 0.81, comparable to some clinical scores like Alvarado, Arnbjørnsson and Teicher (area under the ROC 0.81, 0.83 and 0.85 when initially tested). But these scores performed much worse when tested in different groups (0.60-0.66). This system could therefore be used as a decision-support tool [9].

1985[13], based on many items from the history and the examination (duration of symptoms, location of initial pain, migration of the pain, previous attacks of similar symptoms, anorexia or nausea, vomiting, diarrhea, gender, fever, abdominal tenderness, rebound tenderness, rigidity or guarding, abdominal mass, rectal tenderness and rectal mass). The cutoff value also separated patients for operation or observation. Its evaluation showed some good results, but not good enough to recommend it [11].

Various clinical scoring systems have been devised to aid diagnosis (Macklin, Radcliffe et al. 1997). They will be discussed below. Despite the availability of new diagnostic means like ultrasonography, computed tomography or laparoscopy, the frequency of misdiagnosis leading to unnecessary appendectomy has not changed, nor has the frequency of perforation decreased [4]. This study, based on the CHARS (Washington State Comprehensive Hospital Abstract Reporting System) and using ICD-9 codes (International Classification of Diseases), showed that the rate of ruptured appendix and misdiagnosis remained stable for the last ten years. Incidental appendectomies decreased and nonincidental appendectomies remained stable.

The next was published by Alvarado [14] in 1986. Using Bayesian analysis, he assigned a numerical value to each of eight signs, symptoms and laboratory values. He recommended immediate operation for a score higher than 6 over 10, observation for scores of 5 to 6, and discharge with outpatient follow-up for scores lower than 5. A rate of initial negative appendectomy of 7%, and a missed appendicitis rate of 14% were reported then. A prospective study in the UK [15] was performed in 1996, concluding that this score couldn’t be recommended to use in pediatric surgical practice, because of significantly inferior results obtained with the score compared to their current clinical practice. Another evaluation concluded that it was the only score that fulfilled all standardized criteria. These criteria are based on recommendations about acceptable rates for removing normal and perforated appendices [17,11].

Clinical scores: As of yet, at least eleven clinical scores have been published and tested to help acute appendicitis diagnosis. The first was published by Van Way [10], in 1982. The score was applied to patients with appendectomy, and based on only four history items: gender, duration of pain, nausea and anorexia. A cutoff value was fixed, to separate patients to operate from patients to monitor. Evaluations of the score had insufficient results to recommend it for clinical use [11]. In 1983, Teicher [12] published a score also made for appendectomy patients, but with more items: 4 from the history (gender, age, duration of pain and presence of urinary or gynecological symptoms), 2 from the examination (right lower quadrant muscle spasm, rectal mass right side), and one investigation (leucocytes). The cutoff value separated patients for operation or monitoring. It was evaluated and recommended for further large scale testing [11]. Arnbjørnsson published another score in

SGMI SSIM SSMI

Fenyö published a score in 1987 [16], based on a very long list of items from the history and examination, used separately for men and women (19 items for each gender). A cutoff value allowed separating patients for monitoring or operation only. Evaluated by Ohmann et al. [11], the results showed that only two of the standard criteria were fulfilled. In 1988, Lindberg [18] published a score that also fulfilled two of the four standard criteria when evaluated by Ohmann et al. [11]. It originally contained 10 items from the history, examination and investigation, and separated patients in three groups: operation, observation and exclusion. In 2000, a modified version was proposed [19], considering the temperature instead of the aggravation of pain by coughing and modifying the values and score points for each

SMI 2003; No 51

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

item. It was retrospectively tested on 197 pediatric patients (2 to 17 years old) with excellent results: only 2% of initial negative appendicitis rate (unnecessary operations), but 8% would have been discharged wrongly (missed appendicitis rate). They also showed that the sensitivity was much lower for girls (76%) than boys (91%), but the specificity was similar (87-88%). The next score was proposed by Izbicki in 1990 [20], based on 7 items (gender, leucocytes, guarding, rebound pain, migration of pain, duration of pain and type of pain). The cutoff value only separated the patients for operation or observation, and it performed badly when tested by Ohmann et al. [11]: initial negative appendectomy rate of 49%! Then, de Dombal published a score in 1991 [21], to separate acute appendicitis from non-specific abdominal pain, and used 7 items (migration of pain, aggravation with movements or coughing, simultaneously nausea, anorexia and vomiting, skin color, tenderness in right lower quadrant, rebound and guarding, and rectal tenderness). The score allowed separating patients to exclude, observe or operate. Tested by Ohmann et al. [11], it performed very badly with a potential perforation rate of 82%! In 1992, Christian published a simpler score [22], only based on 5 items with 1 or 0 points (abdominal pain, vomiting, tenderness, low grade fever and polymorphonuclear leucocytosis). It separated patients to operate from those to observe, and didn’t perform well when tested [11]. The last score published was developed by Ohmann after his disappointing evaluation of all the previous scores, in 1995 [23]. He based it on 8 items (tenderness in right lower quadrant, rebound pain, no urinary symptoms, continuous pain, leucocytosis, age, migration of pain and guarding), with cutoff values to decide whether to discharge, observe or operate patients. Evaluated in 1999, it performed very well with a missed appendicitis rate of only 0.9% and an area under the ROC curve of 0.901 [24]. Another evaluation showed a missed appendicitis rate of 1.1% (falling to 0% when combined with the clinical evaluation!) [25].

SMI 2003; No 51

Methods Terms definitions: Many terms are used to evaluate the diagnosis and treatment of acute appendicitis. They have been well defined, and should be used with precision to allow comparison of different methods of diagnosis and treatment. Initial negative appendectomy rate:

(means the proportion of patients without acute appendicitis assigned to the operation group). Should be ≤ 15%. Potential perforation rate:

(means proportion of patients with acute appendicitis not assigned to the operation group). Should be ≤ 35%. Initial missed perforation rate:

(means proportion of patients with perforated appendicitis not assigned to the operation group). Should be ≤ 15%. Missed appendicitis rate:

(means proportion of patients with acute appendicitis assigned to the exclusion group). Should be ≤ 5%[11]. The knowledge base: The goal of the diagnosis of acute appendicitis is to avoid unnecessary operations (lowest initial negative appendectomy rate as possible, at least less than 15%), to avoid missed appendicitis (lowest missed appendicitis rate as possible, at least lower than 5%), and after these to have low potential perforation rates and low initial missed perforation rates. Being able to operate only real acute appendicitis lowers the morbidity associated with the surgery and reduces the costs. Avoiding keeping patients without appendicitis in

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

observation and discharging only them also diminishes the costs. To approach these goals, three scores were chosen. The score of Ohmann was selected for its low missed appendicitis rate, to decide which patients to discharge. The modified score of Lindberg was selected for its low initial negative appendectomy rate, to select the patients to operate. All other patients should be observed. The score of Alvarado was also included because of its consideration as a standard to help acute appendicitis diagnosis. In summary, the chosen scores and the rates when they were tested [11, 24, 25, 19]:

Scores: Initial negative appendectomy rate Potential perforation rate Initial missed perforation rate

Lindberg modified 2%

Lindberg modified 14,3-15,6%

Alvarado 5,3-29%

28,7%

37-38,5%

19-76%

not available

40-40,9%

14-65%

Table 1: Chosen clinical scores The expert system application: CORVID™ [26]. This Java™-based application was

Figure 1: Screenshot of the decision-support tool

chosen for its ability to derive applets or Java applications from the system created. Developed by Exsys® Inc., this application allows the development of on-line expert systems for the Web. It has three impor-

tant features: object-oriented structure, Logic Blocks and practical Java delivery. The object-oriented structure is based on Microsoft’s Visual Basic model, making it easy to build systems using methods and properties of Variables, while not requiring the developers to change the way they think and describe their decision-making steps and logic. Logic Blocks can be any combination of rules and trees that have a related function, anything from an entire knowledge base to a single rule. This allows the logic to be organized into blocks that behave as objects. They can be run via forward or backward chaining, and even associated with a spreadsheet file. CORVID applications are delivered by a small applet (about 100 Kbytes) that allows robust interface design options, and is fast for users to download. If special system features are required, the applet can communicate with other applets on the page. For data only available on a server, CORVID can access CGI (Common Gateway Interface), ASP (Active Server Pages) and JSP™ (JavaServer Pages™) pages to perform server-side calculations [26]. In our situation, the knowledge base was created with the three scores described above, introduced into CORVID as Variables and Logic Blocks. Each item tested and its answers were introduced as Variables (Static Lists) and other Variables (Numeric) were created to assign values to the scores. The three total values of the scores were also created as Variables (Numeric). Logic Blocks were then created to ask the user the questions needed and select the right answers with the associated score values. All the Logic Blocks (16) are controlled by a Command Block that runs each Logic Block in a forward chaining manner, asking the user to enter all answers. The scores are then calculated and one of the three possible attitudes with the patient (discharge, observation or operation) is displayed. If the score of Ohmann is lower than 6, the discharge with outpatient follow-up is proposed. (Fig. 1) If the modified score of Lindberg is higher than -3, the operation is proposed. For all other cases, inpatient observation is advised. For information, the results of all three scores are finally displayed, and the details of each score are shown.

SMI 2003; No 51

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

Discussion Because the potential perforation rate and the initial missed perforation rate have not been included in the knowledge base, the possible results of this tool have to be considered cautiously. One could imagine that a score could be developed to have the lowest possible initial negative appendectomy rate, thus having almost no false positives, but a lot of false negatives. It means that a lot of not operated patients would have an acute appendicitis, and the potential perforation rate and initial missed perforation rate would be very high. But, as told before in this paper, the reasonable goals for these two rates are 35% and 15%, respectively. The summary table (table 1) shows that the modified score of Lindberg, chosen to select the patients to operate, has a potential perforation rate of 28.7%, which is below these standards. Even if the initial missed perforation rate is unknown because non-operated patients with perforation are not cited, we can say that the potential issue of observing a lot of patients that should be operated can be minimized. All the scores studied use about the same clinical signs or symptoms and laboratory values, changing the selected ones and their weighs. The resulting scores have different selectivities and specificities; some perform quite well, and some not so well. But maybe some other elements from the clinical evaluation could be considered, like the family history. A prospective study [27] indicated

that heredity (first-degree relatives) is a significant factor in pediatric patients who have appendicitis, showing that children operated on for appendicitis were twice as likely to have a positive family history than those with right lower quadrant pain, and practically 3 times as likely to have a positive family history than those without abdominal pain. Conclusion The stability of the missed appendicitis and perforated ones over the last years [4], even with the increasing availability of CT, US and laparoscopy, shows that some new help is needed to improve the diagnosis and treatment of acute appendicitis. To help clinicians in their difficult task, a lot of different approaches have been considered and evaluated, but none had given results good enough to be considered a real standard and give strong evidence to clinicians. The attempt to combine clinical scores is new, and the results could be promising, but it has not been evaluated yet. Standardized performance criteria should be applied to ensure objective judgment, and whenever possible, a prospective study should be done - for example, by comparing a baseline phase without scoring and a subsequent test phase with scoring [11]. Acknowledgments The author has no relationships with Exsys®, Inc.

References

SMI 2003; No 51

Addiss DG, Shaffer N, Fowler BS, Tauxe RV. The epidemiology of appendicitis and appendectomy in the United States. Am J Epidemiol 1990;132(5):910-25.

Korner H, Sondenaa K, Soreide JA, Andersen E, Nysted A, Lende TH, et al. Incidence of acute nonperforated and perforated appendicitis: age-specific and sex-specific analysis. World J Surg 1997;21(3):313-7.

Lee SL, Walsh AJ, Ho HS. Computed tomography and ultrasonography do not improve and may delay the diagnosis and treatment of acute appendicitis. Arch Surg 2001;136(5):556-62.

Flum DR, Morris A, Koepsell T, Dellinger EP. Has misdiagnosis of appendicitis decreased over time? A population-based analysis. Jama 2001;286(14):1748-53.

Malik AA, Wani NA. Continuing diagnostic challenge of acute appendicitis: evaluation through modified Alvarado score. Aust N Z J Surg 1998;68(7):504-5.

Davidson PM, Douglas CD, Hosking CS. Graded compression ultrasonography in the assessment of the “tough decision” acute abdomen in childhood. Pediatr Surg Int 1999;15(1):32-5.

Douglas CD, Macpherson NE, Davidson PM, Gani JS. Randomised controlled trial of ultrasonography in diagnosis of acute appendicitis, incorporating the Alvarado score. Bmj 2000;321(7266):919-22.

Galindo Gallego M, Fadrique B, Nieto MA, Calleja S, Fernandez-Acenero MJ, Ais G, et al. Evaluation of ultrasonography and clinical diagnostic scoring in suspected appendicitis. Br J Surg 1998;85(1):37-40.

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

SGMI SSIM SSMI

Hallan S, Asberg A, Edna TH. Estimating the probability of acute appendicitis using clinical criteria of a structured record sheet: the physician against the computer. Eur J Surg 1997;163(6):427-32.

Van Way CW, 3rd, Murphy JR, Dunn EL, Elerding SC. A feasibility study of computer aided diagnosis in appendicitis. Surg Gynecol Obstet 1982;155(5):685-8.

Ohmann C, Yang Q, Franke C. Diagnostic scores for acute appendicitis. Abdominal Pain Study Group. Eur J Surg 1995;161(4):273-81.

Teicher I, Landa B, Cohen M, Kabnick LS, Wise L. Scoring system to aid in diagnoses of appendicitis. Ann Surg 1983;198(6):753-9.

Arnbjørnsson E. Scoring system for computer-aided diagnosis of acute appendicitis. The value of prospective versus retrospective studies. Ann Chir Gynaecol 1985;74(4):159-66.

Alvarado A. A practical score for the early diagnosis of acute appendicitis. Ann Emerg Med 1986;15(5):557-64.

Macklin CP, Radcliffe GS, Merei JM, Stringer MD. A prospective evaluation of the modified Alvarado score for acute appendicitis in children. Ann R Coll Surg Engl 1997;79(3):203-5.

Fenyö G. Routine use of a scoring system for decision-making in suspected acute appendicitis in adults. Acta Chir Scand 1987;153(9):545-51.

Hoffmann J, Rasmussen OO. Aids in the diagnosis of acute appendicitis. Br J Surg 1989;76(8):774-9.

Lindberg G, Feny G. Algorithmic diagnosis of appendicitis using Bayes’ theorem and logistic regression. In: Press OU, editor. Bayesian Statistics 3; 1988. p. 665-8.

Dado G, Anania G, Baccarani U, Marcotti E, Donini A, Risaliti A, et al. Application of a clinical score for the diagnosis of acute appendicitis in childhood: a retrospective analysis of 197 patients. J Pediatr Surg 2000;35(9):1320-2.

Izbicki JR, Wilker DK, Mandelkow HK, Muller K, Siebeck M, Geissler K, et al. [Retro- and prospective studies on the value of clinical and laboratory chemical data in acute appendicitis]. Chirurg 1990;61(12):887-93; discussion 893-4.

De Dombal FT. The diagnosis of acute abdominal pain with computer assistance: worldwide perspective. Ann Chir 1991;45(4):273-7.

Christian F, Christian GP. A simple scoring system to reduce the negative appendicectomy rate. Ann R Coll Surg Engl 1992;74(4):281-5.

Ohmann C, Franke C, Yang Q, Margulies M, Chan M, van Elk PJ, et al. [Diagnostic score for acute appendicitis]. Chirurg 1995;66(2):135-41.

Zielke A, Sitter H, Rampp TA, Schafer E, Hasse C, Lorenz W, et al. [Validation of a diagnostic scoring system (Ohmann score) in acute appendicitis]. Chirurg 1999;70(7):777-83; discussion 784.

Zielke A, Sitter H, Rampp T, Bohrer T, Rothmund M. Clinical decision-making, ultrasonography, and scores for evaluation of suspected acute appendicitis. World J Surg 2001;25(5):578-84.

Exsys® Inc. CORVID™, http://www.exsys.com

Gauderer MW, Crane MM, Green JA, DeCou JM, Abrams RS. Acute appendicitis in children: the importance of family history. J Pediatr Surg 2001;36(8):1214-7.

SMI 2003; No 51

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

The global Image Management Strategy within the HUG David Bandon, Christian Lovis, Jean-Paul Vallée, Antoine Geissbuhler, François Terrier University Hospitals of Geneva (HUG)

Abstract We present the overall strategy adopted by the University Hospitals of Geneva (HUG) for the medical image management. The core imaging facilities rely on a commercial PACS system deployed within the radiology department and on an hospital wide distribution through our home-made EMR application (Electronic Medical Record). Additional facilities include a teleradiology experience and image case collection. We describe technically and functionally those facilities and report the difficulties. The on-going challenges reside in the growing image production within radiology, the removal of the traditional film harcopies and the PACS extension to all hospital imaging sources. Introduction Our goal is to deliver a unified hospitalwide image management environment. The implementation strategy focus on the integration and workflow perspectives: IT integration via the use of medical standard such as DICOM (1) or HL7 (2), seamless integration between the different software environments and workflow integration consistent to the global patient care chain.

Corresponding author : Dr. David Bandon Hôpitaux Universitaires de Genève 24, rue Micheli du Crest 1211 Geneve 14 Switzerland Phone: +41 (0)22 372 6262 E-mail: david.bandon@hcuge.ch

SMI 2003; No 51

Our approach is a step-by-step one: we acknowledged the maturity of commercial PACS1 solutions and therefore decided to rely on one. A multidisciplinary group representing radiology users and IT defined requirements and applied a call for tender. For deployment in radiology, we emphasized the process re-engineering, spped performance and IT integration. Once the user acceptance has been successfully reached in radiology, we moved to radiological image distribution to clinicians. We logically promote our home-made EMR2 solution called DPI (Dossier Patient Intégré), as a federated access to medical information. The

next step will be the PACS extension to all imaging sources within the hospital. We are pragmatic in the sense we acknowledge the existence of various mini-PACS needed to be progressively integrated within the institutional PACS infrastructure. In parallel we are implementing added-value services that take advantage of the PACS infrastructure: image case collection is the best illustration. It is fully integrated to PACS and allows creation of teaching files that support the teaching activity. In the future, we intend to offer a wide variety of specific image processing tools. Pacs within the radiology Technical architecture. The PACS is a commercial system acquired via a call for tender launched in 1999. The system has been supplied by a consortium: it includes an image management server supplied by Image Devices GmbH (general contractor) (3), and reading stations delivered by CEDARA (4). Table 1 highlights the main facts. •

Archive. The archive management is a three-tier hierarchical storage: a 500 gigabyte magnetic disk (RAID3) as online, a 5-terabyte near-line archive (drive technology: 9.2 GB MOD4) and a 7 terabyte tape backup (drive technology: 100 gigabyte S-DLT5).

•

Imaging modalities. The PACS is connected to 33 different imaging modalities using DICOM. Their standard compliancy profile includes DICOM worklist to retrieve the patient demographics, DICOM store to push images to PACS and DICOM print for hardcopy. The non-DICOM imaging devices such as our ultrasound units have been attached to a “DICOM box” (video-grabbing and conversion to DICOM).

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

•

Reading station: The reading stations have a dual monitor configuration. Initially, we selected high brightness 3 megapixel monitor but more and more we move to color flat screen which are widely accepted for CT and MRI reading. Two communication modes are handled between the reading stations and the PACS server: either image files are autorouted to station once acquired, either image files are retrieved via network once required and directly loaded. Performance standards are the following: 100 CT images loaded in 35 seconds when remotely accessed versus 15 sec when locally stored.

•

The nuclear medicine and radiooncology sections can not yet fully rely on PACS due to DICOM integration problems and specific imaging visualization expectations. We consider to offer them alternative softcopy solutions specialized for these activities.

•

Problems with the modality replacement. Regularly we are facing some DICOM integration problems impacting the daily routine. For instance some new modalities do not properly transfer information which are useful both for medical practice or automated viewing protocols. This situation remains even if we apply a selective strategy based on high-level DICOM conformance requirements.

The PACS is also integrated to our RIS (Radiology Information System) via a broker (supplier: MITRA) to transmit the patient demographics, study schedule information and signed reports.

Medical practice and user acceptance issues. The PACS is now widely used in a daily routine for more than three years and is fully accepted by radiologists. The major key success factor has been the intuitive and productive tools supplied by the softcopy software (automated viewing protocols, contextual menus). PACS introduction significantly changed the radiological workflow by a direct reading after acquisition. Five clinical rounds are daily performed in a computerized way (internal medicine, neurology, surgery) and CT, MRI are fully softcopy-based. However, some difficulties remain due to functional limitations:

our electronic order entry system (ORMED, home-made implementation).

•

No possibility to build a dedicated worklist for clinical rounds and transfer it from a reading station to the other.

•

Subsistence of hybrid workflow (coexistence of still paper based order with image softcopy) leading to serious organization problems. To address this issue, we started a RIS renovation project to unify all sources of information (order, preliminary and final reports) and worklists. It will be connected to

Daily stored images Part of digital images on global image production Reading stations Image production volume/year

12000 > 90% 25 2.5 Terabytes

Table 1: Facts Hospital-wide distribution Radiological images are distributed within the hospital through our EMR application called DPI (5). Image access is provided with other clinical information and lab results in an integrated way. The clinician logs on the EMR and loads the selected patient folder. The access rights policy is centrally controlled and any access is tracked within our audit trail system. Technically speaking, the image retrieval mechanism is performed in two steps: •

use of informational services to document the radiological record content. The supplied imaging list is displayed to user for study selection within the EMR;

•

launching of a java applet viewer loading imaging studies or series from a central image repository acting as a web server.

SMI 2003; No 51

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

The informational services are delivered through our middleware framework based on an HTTP-XML protocol (10) (queries structured in an XML dataset and transmitted through a http channel). These services provide the EMR with an exhaustive information set concerning the patient’s radiological records (patient, study and series levels). Each series is documented by an icon path to offer a preview facility to doctor. Moreover, the XML service indicates the level of image availability within the PACS storage architecture (online, near line or backup level). By this way, he/she may anticipates the retrieval waiting time. Concerning the image display part, we offer an image viewer applet supplied by Image Devices GmbH. It is launched through the web (call of a servlet method via an url with the patient, studies and series identifications). This applet offers all basic display tools such as zoom, window/level adjustments, convenient browsing modes. We apply a selective viewer mode strategy to address the issue of limited resource availability on some low-end stations. Depending on the amount of free memory, one of the three viewer modes is used: simple HTML viewer based on JPEG images or either a 8-bit, either 16-bit image java viewer applet. That image display capability has been introduced within our EMR application in early 2002 and is progressively expanded. Hundred image displays are daily requested to PACS. Image case collection PACS offers a patient-centered image storage and retrieval solution best suited for reading practice. However it lacks the ability to index studies using advanced criteria such as a medical context for instance. For that specific purpose we implemented Casim@ ge (7) to create some teaching files for teaching and research purposes. This solution is implemented using the 4th Dimension DBMS software from ACI to allow rapid design. With this software, the three main functions are: a data import function fully integrated to PACS, data editing by means of simple and powerful client software and data sharing and distribution with the creation of stand-alone and multiplat-

SMI 2003; No 51

form CD-ROMs and with a Web-server for online access. In the medium term Casim@ge will include content-based image retrieval (CBIR) algorithms to retrieve similar cases for reading activity. Teleradiology •

A specific experience is led with French neighbor hospitals in a neurosurgery application. The goal is to decide whether some trauma patients have to be transferred to Geneva for surgery. For that purpose, surgeons need to consult images acquired in those remote sites.

•

We have implemented a web-based solution called TELIM. Communication is push-based on asynchronous and secured ways: images are initially sent from the imaging modalities to a local relay server where encryption is performed. As a second step, image files are transferred via HTTP to a HUG server installed on a semi-public domain network (behind our firewall). From this point, image files are automatically copied to an intranet server from which they become available to doctors.

•

This preliminary experience will be extended to new sites in the future to scope other needs (second advice or patient transfer).

•

As an alternative we have implemented a secure DICOM communication compliant to DICOM Supplement 15. Implementation (6) is based on OpenSSL library and ensures triple DES encryption and RSA based certificates for user authentication.

Challenges and future projects The increasingly growth of image production within the radiology. The most challenging issue is the huge amount of images generated by the newest generations of functional MRI unit – up to 20’000 images per study - and multislice CT (MSCT), averaging 400

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

to 800 images per study with a 4-channel system. Our two 4-channel MSCT units will be soon upgraded to 16-channel MSCT channel. This heavy production load leads to severe performance problems on the reading stations. In order to accommodate our PACS with the increased image burden (yearly production growing from 2.5 to 3.8 Tera-bytes), we are currently moving to a selective archive strategy based on a thickslice storage (3 to 5 mms). Raw data are only stored temporarily for image processing purposes or interpretation control if required. In a more general perspective, the interpretation of functional MRI or multi

slice CT requires a new image navigation paradigm offering 3D navigation or hierarchical view on image (automated image fusion to have either thin or thick slice navigation). This paradigm requires a major PACS evolution. Key image selection for clinicians. This method promotes the display to clinicians of significant images previously selected by radiologists. Nonetheless clinicians may access all images if wished at any time. This image selection facilitates the clinician consultation by highlighting the significant information. We failed to introduce this feature so far due to a delay in DICOM evolution and the non complete proprietary integration of the PACS suppliers. The new DICOM supplement (key image note) will definitely help us. Hardcopy removal. Second challenge deals with the final removal of the film hardcopies. This removal will be introduced in two steps: removal in hospital once images are fully available within the EMR. For the outside distribution, we are currently investi-

gating a distribution solution via CD-ROM given directly to patient when he/she leaves. This CD-ROM will store both image files embedded with a light DICOM viewer. Seamless integration between RIS and PACS. Our current RIS is based on specialized software modules (scheduling, charge posting, dictation and finally report transcription). These modules share the same patient demographic information. However each module has its own user interface not seamlessly integrated with each other. Moreover they are based on different workflows. Therefore we are in the process of renovating our RIS: •

To offer a transparent access to all information and tools especially in the interpretation activity: transparent access to clinical indication, previous reports and images as well as access to dictation tool and report validation. This capability will rely on RIS / PACS / EMR client synchronizations.

•

To have a central workflow enabling a unique view on any open jobs. We will rely on a generic workflow engine and on IHE6 (11). IHE compliancy will ensure the best use of the DICOM and HL7 standards to integrate HIS, RIS, modalities and PACS.

PACS extension to all hospital imaging sources. We envision to open our PACS system to all medical images produced within the hospital. Indeed, specialties like cardiology, dermatology, ophthalmology, pathology or gastro-enterology heavily rely on image manipulation in their daily routine activity and wish to switch to a digital mode. In that respect, strategy is to provide a unified storage facility within the PACS and a unique distribution point via our EMR application. To reach that goal, we need to take into account the heterogeneity of the installed systems. In that respect, two categories can be listed: •

Specialized image management systems, acting as mini-PACS. Examples are the pathology specialty having a network of microscopes connected to a system offering report tools and image display

SMI 2003; No 51

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

and processing facilities (DIAMIC product supplied by INFOLOGIC) or the cardiology department equipped with a specific mini-PACS adapted to echo-cardiography needs (ENCONCERT solution supplied by PHILIPS). These solutions provide temporary storage. The need is therefore to automatically and regularly export images from these mini-PACS units for their long-term archiving within our institutional PACS. •

Digital devices (endoscopy units, microscopes, stroboscopes or small digital cameras). Such devices may be used to acquire color images –static or video - to be associated to any medical context (diagnosis, therapy preparation, follow-up). Clinicians or any other authorized operators expect to have a software utility to manually associate images to a specific patient and context or document.

•

We are currently seeking the best solution to address these needs and issues. We are considering the following strategy based on image conversion services integrated to our middleware framework: •

Image files and their acquisition context are pushed via HTTP to a server where there are converted to DICOM and finally sent to PACS. The acquisition context is represented in a XML dataset built from a simplified DICOM information model: each XML tag matches a corresponding DICOM data element but we only transfer the most relevant information and all DICOM UIDs (Unique Identifiers) are generated afterwards. Rationale is that a DICOM-XML like format is more easily accepted by non-DICOM suppliers because it is more easier to implement.

•

Static images are converted in a DICOM image format, the modality being either SC (Secondary Capture) or VL (DICOM Secondary Capture). The supported formats are JPEG and TIFF. If the source format is JPEG, we directly store the JPEG dataset within the DICOM image dataset without any decompression step.

•

Video image case is a more complicated issue. Some major compression rates are applied at the acquisition level (roughly 20:1). Since these compression formats are not supported by DICOM we are facing a conversion problem. An uncompressed storage would seriously burden the PACS performance and capacity. In the same time, conversion from MPEG to a DICOM compression format, i.e. JPEG would introduce additional image quality degradation due to successive lossy compression oper-

Moreover we need to address the following issues related to DICOM standard and the large data volume: •

•

SMI 2003; No 51

Unlike the radiology field, the majority of suppliers involved in other medical sectors do not support DICOM. This situation is mainly due to a low penetration of DICOM within those industry segments. But nevertheless DICOM effectively supports the corresponding images, especially through the visible light supplement (1). As an alternative, those suppliers opted for the different image formats used in the video industry. The most frequent formats are MPEG (8) or MJPEG7 for motion image (11) or JPEG, TIFF for static image (9). DICOM lacks the ability to integrate some of these video formats with lossy compression (MPEG or MJPEG compression techniques for instance), even if there are some on-going discussions to support MJPEG codec in the DICOM Visible Light supplement.

Most of imaging sources produce motion image (video) for which the volume can be important. It arises the storage strategy issue: do we need to archive all imaging loops or only selected ones relevant for the patient follow-up?

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

ations. In that respect, we consider to keep the initial format and store it as a complementary file to DICOM studies. We would only extract a representative image from the loops and store it in DICOM in order to keep the DICOM arborescence structure (study and series). Videos would be displayed using freeware viewers. Conclusion Major achievements have been performed in term of infrastructure since the project launch in 1999: (a) a PACS deployed in radiology (1999-2000) and (b) an image distribution within the HUG-EMR since 2002. With the acquisition in 2003 of three new phosphor plate units, the radiology department will be fully digital (expect the mammography). We will therefore move forward to the complete filmless hospital: the last step being the extension to all imaging sources located outside the radiology. This new phase will start in 2003 and will address the issue of dynamic video and integration with heterogeneous specialized mini-PACS solutions. We will follow a customer focus approach by integrating the current working methods of each customer groups into the global patient care process. Our mission is to facilitate the workflow of clinicians using the best of new technology. Thanks The authors are grateful to people who strongly contributed to this project success: Andrew Parrott as system engineer, Michael Dehouck, Cyrille Duret, Marcello Echeverria, Francis Klumb and Marianne Logean and Turid Schou for the implementation part; Pierre-Alain Meche and Irene Ponsolle for support. References 1

DICOM Standard: Diagnostic Imaging and Communication in Medicine. http://medical.nema.org/.

HL7 standard: Health Level 7.

Lovis C, Baud RH, Rassinoux AM, Scherrer JR. Value to add to the patient record when making the EPR. In: 17 Iwg, editor. Eprimp; 1998; Rotterdam; 1998. p. 225-227.

Echeverria M, Bandon D., Parrott A., Vallée J-P. eMEDIBOX: a secure DICOM communication based on internet plug-in. Proc. of the 12th International Symposium in Computer Assisted Radiology, H.U. Lemke, M. Vannier, K. Inamura and A.G.Farman (Eds), 1170, Elsevier Science, Berlin, June 2001

Rosset A., Ratib O., Geissbuhler A., Vallée J-P. Integration of a Multimedia Teaching and Reference Database in a PACS Environment. RadioGraphics. 2002;22:15671577-245

Moving Picture Experts Group. MPEG home page: http://mpeg.

The Graphics File Format Page.

telecomitalialab.com http://www.dcs.ed.ac.uk/home/ mxr/gfx/2d-hi.html. Accessed November 26, 2002. 10

Geissbuhler A, Lovis C, Lamb A, Sphani S: Experience with an XML/HTTP-based federative approach to develop a hospital-wide clinical information system. Medinfo 2001;10(Pt 1):735-9.

Bandon D, Vallée J-P, Geissbuhler A, Initiative IHE: une nouvelle perspective pour les systèmes d’imagerie, Congrès annuel de la SSIM, Bâle, 2001

Footnotes 1 PACS: Picture Archiving and Communication System 2 EMR: Electronic Medical Records 3 RAID: Redundant Array of Inexpensive Disk 4 MOD: Magneto-optical disk 5 S-DLT: Super-Digital Linear Tape 6 IHE: Integrating Healthcare Enterprise. 7 MJPEG (Motion JPEG) video is essentially a sequence of JPEG stills playing. MJPEG is widely used for digitizing analog video.

http://www.hl7.org

Image Devices GmbH (software company). http://www. imagedev.com.

Cedara (company). http://www. cedara.com.

SMI 2003; No 51

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

Tips & Tricks Die Rubrik „Tips & Tricks“ soll dem medizinischen Anwender praktische Hinweise im Umgang mit dem Computer geben. Sie soll mit neutralen Beschreibungen auch SoftwareWerkzeuge vorstellen, die für die tägliche Arbeit von Nutzen sein können. Anregungen für zu bearbeitende Themen sind willkommen. Bitte richten Sie Ihre Vorschläge an den Hauptschriftleiter (siehe Impressum).

La rubrique „Trucs et Astuces“ a pour objectif de donner de petits conseils pratiques pour faciliter l’utilisation de l’ordinateur. Parfois, nous vous proposerons aussi de courtes évaluations ou comparaisons de logiciels qui pourraient vous être utiles. Vos commentaires, suggestions ou encore les petits trucs qui vous aident chaque jour sont les bienvenus ! Adressez les au rédacteur en chef (voir Impressum).

Instant Messaging

Instant Messaging ist eine Art der Kommunikation, die durch folgende Eigenschaften gekennzeichnet ist:

Dipl.-Ing. Claus Eikemeier Universtität Bremen, Deutschland Wenn Mitarbeiter, Bekannte oder Freunde, miteinander kommunizieren möchten, gibt es dafür verschiedene Möglichkeiten. Für uns selbstverständlich sind das normale Gespräch (Face-to-face) oder ein Telefongespräch. Mit der starken Verbreitung des Internets hat sich auch Email zu einem ubiquitären, d.h. allgegenwärtigen, Medium entwickelt: welcher Leser hat nicht schon einmal auf einer Konferenz oder im Urlaub seine Emails über ein Web-Frontend gelesen. Jede Kommunikationsform hat spezielle Eigenschaften, die sie von anderen unterscheidet. Unterscheidungsmerkmale sind Zeit (muss z.B. gleichzeitig kommuniziert werden?) und Raum (müssen sich die Personen an einem bzw. einem bestimmten Ort befinden?), aber auch die Frage, ob der Kommunikationspartner gerade für ein Gespräch bereit ist (Bewusstsein über Zustand des Kommunikationspartner, engl. Kurz: “awareness”), stellen Entscheidungskriterien für die Auswahl des Mediums dar. Für Personen, die ad-hoc Informationen mit Bekannten (Kamerad, engl. “buddies”) austauschen wollen, eignen sich die oben genannten Möglichkeiten evtl. nicht: Bei Nutzung des Telefons sieht man nicht, ob der andere Gesprächspartner genau jetzt gerade nicht gestört werden will, bei Email ebenfalls, man kann noch nicht einmal eine sichere Aussage darüber machen, wann die Nachricht beim Empfänger ankommt (obwohl das meist innert weniger Minuten erfolgt).

SMI 2003; No 51

•

Verwendung von textbasierten Nachrichten (wie Email)

•

Starke Synchronität (Nachricht wird sofort ausgeliefert, Zwischenspeichern der Nachricht nur, wenn der Empfänger temporär nicht verfügbar ist)

•

Kommunikationssoftware läuft “nebenher” (siehe Bild 1), ist dadurch sehr schnell zugreifbar und somit ideal für ad-hoc Fragen nutzbar

•

Unterstützung von Awareness (man sieht den Zustand (nicht erreichbar, vorübergehend nicht erreichbar/ nicht stören, anwesend) seiner Kommunikationspartner in der “Buddylist”)

•

die Art der Kommunikation ist eher informell (kurze Fragen, kurze Antworten)

Instant Messaging wird innerhalb verschiedener Gruppen von am Computer arbeitenden Menschen stark genutzt. So kann es teilweise den Austausch von Informationen innerhalb von verteilt arbeitenden Arbeitsgruppen – entsprechend der Situation “am Kaffeeautomaten” – ermöglichen. Im Bereich der Freizeitkommunikation gibt es eine ähnliche Entwicklung, die Kommunikation per Short Message Service (SMS) von Handy zu Handy. Der Empfänger entscheidet, ob er genau zu dem Zeitpunkt

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

des Sendens die Antwort schreiben will (was häufig gemacht wird, dann entsteht eine “quasi-kontinuierliche” Kommunikation) oder ob er das später machen will. Er kann auch z.B. auch sehr einfach eine Nachricht hinterlassen (wie “bin Essen, komme 13: 30 Uhr wieder”), die dann dem Anrufer angezeigt wird. Wichtig ist die Möglichkeit der schnellen und einfachen Bedienung der Software. Die bekanntesten Instant Messaging Werkzeuge werden kostenlos, z.T. im Verbund mit anderer Software abgegeben. Beispiele sind der AOL Instant Messenger (AIM), ICQ (“I seek you”) oder auch Microsoft Netmeeting. I.d.R. installiert man sich eine Software, die mit den entsprechenden Programmen auf den Rechnern der Buddies kommuniziert. Häufig erhält man daraufhin eine eindeutige Kennzeichnung zugeordnet. So ist die Kommunikation nicht auf einen Rechner beschränkt und man ist weltweit identifizierbar und “ansprechbar”. Für längere Zeit waren die Welten der jeweiligen Tools nicht kompatibel, d.h. eine Person mit einem ICQProgramm konnte nicht mit einem AIM Empfänger kommunizieren. Die Grenzen verschwinden hier jedoch zunehmend. Die Zusatzdienste, die meist über die zugehörige Website angeboten werden (z.B. Suchen nach gleichgesinnten Personen aufgrund von Interessenprofilen), beschränken sich allerdings meist nur auf die Nutzer des gleichen Programmes. Zusammenfassend kann gesagt werden, dass Instant Messaging einen bedeutenden Kommunikationskanal im Rahmen von Projektarbeit darstellen kann. Es ermöglicht die Nutzung einer Chat-Funktionalität, häufig auch Dateiübertragung bzw. sogar das Arbeiten auf dem entfernten Rechner (z.B. bei Computerproblemen). Insbesondere bei stark computer-orientierter Arbeit bietet sich dieses Medium “für die kurze Nachricht zwischendurch” an Freunde, Kollegen etc. an. Initial ist wenig Aufwand notwendig, um an einer IM Community teilzunehmen. Als Tipp sollte man sich in das gleiche Netzwerk einbinden, welches auch von den Freunden genutzt wird, allerdings ist die Verbindung mit mehreren bzw. in andere IM Netzwerke mit entsprechender Software möglich.

Bild: IM-Awareness der Komm.-Partner: Es ist erkennbar, dass 3 Partner im und einer ausserhalb des USZ derzeit gesprächsbereit sind (fingiertes Szenario).

Quellenangaben: 1

ICQ http://web.icq.com/

NetMeeting http://www.microsoft.com/windows/ netmeeting/ Remark: NetMeeting is delivered with the windows operating system

AOL Instant Messenger http://www.aim.com/

Yahoo! instant messenger http://messenger.yahoo.com/

SMI 2003; No 51

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

Trucs et astuces pour Office

Gouttières et tableaux

Prof. Christian Lovis

Une gouttière désigne l’espace réservé à la droite d’une colonne d’un tableau, dans laquelle il est impossible d’écrire.

Faculté de médecine, Université de Genève Comment ajouter un titre à un tableau Rien n’est plus simple que de remplir la première ligne d’un tableau de libellés d’entêtes. Mais lorsque le tableau en question couvre plusieurs pages, faut-il recopier cette ligne manuellement ? Pas du tout. Il suffit de la sélectionner (pour sélectionner rapidement une ligne, cliquez avec la souris sur le côté gauche dudit tableau) puis de se rendre dans le menu Tableau pour activer Titres. La ligne se répétera alors automatiquement sur toutes les pages. A noter qu’il est possible de sélectionner plusieurs lignes pour les utiliser comme titres.

Pour manipuler ce paramètre, sélectionnez les colonnes à modifier, puis faites apparaître le menu. Demandez Tableau/Taille des cellules, en cliquant avec le bouton droit de la souris sur le tableau et en optant pour Taille des cellules ou, plus simplement encore, en doublecliquant sur l’un des taquets de tabulation repérant les colonnes, sur la règle. Rendez-vous ensuite dans l’onglet Colonne et ajustez la taille indiquée dans le champ Espace.

Attention, vous n’aurez la possibilité de voir ces titres que si vous passez en mode Page. De plus, sachez qu’ils ne se répètent qu’après un saut de page automatique et non manuel.

Events in Medical Informatics

Switzerland

Europe

e-Healthcare.ch 2003 Date: 16-17/10/2003 Location: Universität Irchel, ZürichContact: www.ehealthcare.ch

Internet et Pédagogie Médicale (IPM2003) Date: 11-12/12/2003 Location: Marseille, France Contact: www.ipm2003.org

MEDNET 2003 Internet and Medicine Date: 3-6/12/2003 Location: Palexpo, Geneva Contact: www.mednet2003.org OSHCA 2003 Open Source HealthCare Alliance Date: 7-9/12/2003 Location: Palexpo, Geneva World Summit on the Information Society Date : 10-12/12/2003 Location: Geneva, Swizterland Contact: www.itu.int/wsis/ SMI 2003; No 51

World

AMIA 2003 American Medical Informatics Association Fall Symposium Date: 8-12/11/2003 Location: Washington, DC Contact: www.amia.org MEDINFO 2004 Date: 7-11/9/2004 Location: San Francisco, USA Contact: www.medinfo2004.org

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

nächste Ausgabe: Februar 2004 prochaine édition: fevrier 2004

Impressum Publikationsorgan der Schweizerischen Gesellschaft für Medizininformatik Organe de publication de la société suisse d‘informatique médicale Herausgeber / Editeur SGMI, Schweizerische Gesellschaft für Medizininformatik c/o VSAO Dählhölzliweg 3 Postfach 229 CH-3000 Bern 6 Tel. 031 350 44 99 Fax 031 350 44 98 e-mail: admin@sgmi-ssim.ch Internet: http://www.sgmi-ssim.ch/ Vorstand der SGMI / Comité de la SSIM Judith Wagner, Martin Denz, Felix Heer, Benno Sauter, André Assimacopoulos, Ulrich Woermann, Christian Lovis, Antoine Geissbühler Chefredaktor / Rédacteur en chef Rolf Grütter Redaktion / Rédaction Rolf Grütter, Christian Lovis, Ulrich Woermann,

Die nächste Ausgabe des Swiss Medical Informatics erscheint im Februar 2004 und behandelt folgendes Thema: • Nationale Strategie zur Integration der ICT ins Gesundheitswesen

Redaktionsadresse / Adresse de rédaction Rolf Grütter Institut für Medien- und Kommunikationsmanagement Universität St. Gallen Blumenbergplatz 9 9000 St. Gallen e-mail: rolf.gruetter@unisg.ch Layout / Mise en page Jürg Hirsiger Abteilung für Unterrichtsmedien AUM Universität Bern Inselspital 38 3010 Bern Autorenrichtlinien / Directives pour les auteurs http://www.sgmi-ssim.ch/smi/index.htm Verlag / Editions Schwabe & Co. AG Steinentorstrasse 13 4010 Basel Betreuung im Verlag: Dr. Markus Trutmann Tel. 061 467 85 55 Fax 061 467 85 56 e-mail: mtrutmann@emh.ch

La prochaine édition du Swiss Medical Informatics paraîtra en fevrier 2004 et traitera le sujet suivant: • stratégie nationale pour l’intégration des technologies de l‘information dans le système sanitaire

Druck und Versand / Impression et distribution Druckerei Schwabe & Co. AG Farnsburgerstrasse 8 CH-4132 Muttenz Tel. 061 467 85 85 Fax 061 467 85 86 e-mail: druckerei@schwabe.ch Inserate / Régie des annonces Schwabe & Co. AG Chantal Schneeberger Frankfurtstrasse 14, Postfach 340, CH-4008 Basel Tel. 061 333 11 05 Fax 061 333 11 06 e-mail: c.schneeberger@schwabe.ch Abonnemente / Abonnements Sekretariat der SGMI, c/o VSAO, Dählhölzliweg 3, Postfach 229, CH-3000 Bern 6 Tel. 031 350 44 99 Fax 031 350 44 98 e-mail: admin@sgmi-ssim.ch Abonnementspreis / Prix d’abonnement CHF 40.– (zuzüglich Porto / port en plus) Einzelnummer / Exemplaire unique CHF 15.– (zuzüglich Porto / port en plus) ISSN 1660-0436 erscheint 3mal jährlich paraît 3 fois par an

SMI 2003; No 51

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

SMI 2003; No 51

SGMI SSIM SSMI

Swiss Medical Informatics 18. Wissenschaftliche Jahresversammlung / 18ièmes journées annuelles

SGMI SSIM SSMI

SMI 2003; No 51