Design and the implementation of a webenabled. haematological system

Design and the implementation of a webenabled haematological system Lefteris Gortzis, MSc Medical Physics Laboratory Department of Medical School Uni...
Author: Ambrose Hodges
2 downloads 0 Views 237KB Size
Design and the implementation of a webenabled haematological system

Lefteris Gortzis, MSc Medical Physics Laboratory Department of Medical School University of Patras, Patra 26500 Hellas Mail address: [email protected] Tel number: +32610997882

Stauros Koubias, Assistant Professor Applied Electronics Laboratory Department of Electrical Engineering University of Patras, Patra 26500 Hellas Mail address: [email protected]

George Nikiforidis, Professor of Medical Physics Medical Physics Laboratory Department of Medical School University of Patras, Patra 26500 Hellas Mail address [email protected]

Abstract This paper describes and facilitates the methodology and the implementation of a fully integrated electronic Haematological System, named (e-HS). The proposed system runs on a set of distributed network nodes providing useful haematological services via the web. These services include patient-oriented management, Digitized Histopathological Slides (DHS) acquisition, teleconsulting facilities and etc. The objective of e-HS is to supply web-enabled services according to haematological requirements, implement a distributed storage scheme for DHS, and provide a common database that consists of all haematological laboratory results by using eXtensible Markup Language (XML) and grid services. Our implementation can be available to every authorized physician at the distributed nodes without any hard- or additional software. The only software required for the user is the regular Internet Explorer (Version 3.02 or higher). Besides, by using intuitive and self-explaining user interfaces and HTML-techniques such as hyperlinks the necessary amount of training by the physicians-side is reduced to a minimum. A first implementation of the e-HS, has been established at the Medical Physics Laboratory of the University of Patras (master node of the system), and has been tested with great success from the medical staff of the Hospital Departments of the University of Patras and Thessalonica that served as distributed nodes of the system.

Author Keywords: Haematological services; common database; ASPX; Geographically distributed medical data;


1. Introduction Detailed literature search shows that years ago when medical data distribution began, many point to point systems were established where needed. But when the number of connected physical locations was growing, due to an increasing electronic support handling of many individual links, became intricate [1, 2, 3 and 4]. During the design phase of our system, we have analysed the clinical practice in typical environments of Hellenic haematological departments in order to stimulate basic research as well as improve the handling and the management of medical data. This analysis indicated that even though, departmental systems have sometimes been interfaced in order to transfer e.g patient demographic data between a patient registration system and laboratory system, this was usually based on ad-hoc activities without any global communication concept. Considering the above, we believe that healthcare needs integrated network systems to which each node is connected. Due to the decentralization of the staff working places an efficient and effective network system could provide the necessary link between data collection, storage, and analysis. The objective of the present project is to implement a fully integrated electronic Haematological System, named (e-HS) that runs on a set of distributed network nodes providing the useful haematological services via the web. In accordance to this, we focused on the management of the distributed data that is used by haematologists and we propose a basic methodology in order to enable this data to be linked for the purpose of medical diagnosis, analysis and comparison. In contrast with the existing systems, the proposed methodology has increasingly acknowledged the role of human factors and `usability engineering' in services design and evaluation [5, 6].

Furthermore, our system enables medical research on plethora of data, efficiently and costeffectively providing predefined procedures and knowledge [7].


Analytically, e-HS supports the major haematological requirements, implements a distributed storage scheme for Digital Histopathological Slides (DHS), and provides a document-based common database of total of haematological laboratory results. The proposed common database is a crucial informational link in the development of haematological information networks, in the establishment of multi-node systems, and in the vertical integration of nodes and physicians. As a result, each system’s node seems to provide all the e-HS’s functionalities even if some of them are not locally available, as illustrated in Figure 1. A first implementation of the e-HS, has been established at the Medical Physics Laboratory of the University of Patras (master node of e-HS), and has been tested with great success by Haematologists of the Hospital Departments of the University of Patras and Thessalonica that served as distributed nodes.

2. Background Several efforts of our research team have addressed the issue of remote handling the data contained in distributed collaborative systems [8, 9, 10]. These efforts targeted to propose seamless connectivity communication-skins between deferent networking structure providing an acceptable Quality of Services (QoS) level. In the field of healthcare, accordingly additional research has proved that in order to achieve a useful wide-area integrated system, there are a number of complementary issues that must be fully addressed in whatever way federation is to be accomplished. The architectural structure is one of these issues. The architecture of the existing systems has been focused so far on two main approaches. One uses object-oriented methodologies for building the architecture model. This approach was first used in the GEHR (Good European Health Record) project [11] in the early nineties and has subsequently been used by a number of projects and groups such as Synapses [12], , SynEx [13], and CorbaMed [14].

The second and more recent approach has used

document-oriented methodologies as seen in the US HL7 Kona [15] and the UK Oswestry project (a documentation project designed to help implement the electronic patient record) [16].


There are undoubtedly benefits to be gained from the combination of to above approaches. Considering this observation, our project has been focused on a hybrid methodology of the two above approaches. It is based on the ideas of distribution and openness. The objectives of this methodology are i) to achieve communication integration within distributed nodes ii) and to develop a network system that provides useful haematological services to its endusers within these nodes. However, it overcomes on one hand, the disadvantages of CORBA-based architectures for web communication [18] and the problems regarding the platform/object dependence and the communication, rising through the Internet on the other (e.g. firewall security restrictions) [19,20]. The key-feature of our methodology is the melding of the recent trend towards remote nodes, common database and web emerging standards and technologies within a web enabled system [21, 22]. This fully integrated electronic haematological system, named e-HS, supplies a powerful set of integrated haematological services, it provides a document-based common database that consists of all node’s haematological laboratory results by using XML adhering to published W3C standard [23], and it facilitates the dissemination of DHS maintained locally in the distributed nodes. e-HS system consists of the following five web subsystems: ‰

a subsystem, named MCGrid for data managment,


a subsystem, named DHSV, for DHS viewing


a subsystem for searching data that is contained at common database


a subsystem for DHS replicating at master node database and finally


a teleconsulting subsystem.

The present work should consist of parts with contents according to Instructions to Authors: Design consideration; System description; Status report; Lessons learned; Future plans.


3. Design consideration Physicians, who are the target group of our efforts, rather than becoming "power users" of a narrowly defined software package, they access a wide variety of systems and resources. Therefore, routine use of computers by them can be most easily achieved if a system offers them a critical mass of services that are smoothly integrated, via collaborative medical nodes, and at the same time useful for essentially every patient encounter [24]. Before building a system for the integration of the distributed data, it is obviously imperative to have a clear and detailed functional requirement’s specification for the physicians. Equally, we believe that it is imperative to have a clear and detailed functional requirement’s specification for using, retrieving, updating and sharing medical data, independently of the technology used to implement a web-enabled system. Especially in the field of haematology, a detailed study of the end-user’s requirements can be defined by outlining the intended role of a Clinician Haematologist. Some of the basic roles of Haematologists are summarized as follows: 1. In-patient management of patients with various blood disorders, including anaemias, abnormalities in white cells and platelets, marrow failure, leukaemias and lymphomas,






splenomegalies, bleeding disorders and thromboembolic disorders 2. Chemotherapy and supportive management of patients with leukaemias and lymphomas 3. Competence in interpretation of morphological haematology (e.g. DHS), routine and specialized haematology tests 4. Consultation by other specialties (e.g. physicists) on general haematology, bleeding and blood transfusion problems 5. Ambulatory care of patients with blood disorders 6. Anticoagulant clinic for acquired and inherited thrombotic diseases 7.






Anaemia)Clinic; Bone Marrow Transplant Clinic, etc.




8. Working knowledge in the following specialized areas: (i) Bone marrow and stem cell transplantation (ii) Plasmapheresis and apheresis procedures (iii) Blood component collection, processing and blood banking (iv) Routine and specialized haematological laboratory procedures. Our implementation was designed to provide distributed services on the medical stuff according to all the above tasks by focusing on the quality of haematological-care and patientcare data. It is important to note that the collection of highly specialized haematological services is not the only problem when building a medical system. The major difficulty lies the integration of those services into a coherent environment. Our system provides this integration into a coherent environment and embeds the point-and-click technologies as a result we can see that our system promise to be more attractive to clinicians The architectural structure of our implementation has a hybrid format locating on one hand all the haematological laboratory textual results in a common database (centralized approach) and the DHS in distributed network nodes on the other (decentralized approach) as shown in the figure . This solution provides two key benefits (i) functional integration of the medical nodes via web and (ii) increased efficiency of provided services by reducing the time spent by health professionals collecting the existing data. The common database is the core-element of our system and is a possible solution to the current difficulties that exist in the field of data management and data delivery in large-scale environments. This document–based file meets three major requirements: (i) it contains in an efficient way all information on the patient’s current and past conditions, procedures that have been performed and responses to therapy (ii) it bridges the gap between unstructured and structured content sources. A synoptic view of the data-tags that is embedded at the common database are: Patient-ID, Exam-ID, Date-exam, Middle cell, Basophils, Eosinophils, Monocytes, Lymphocytes, Neutrophilic, White, MCHC, MCH, MCV, Blood type, Indirect Coombs, Coombs, Ret, Hb, Htc. EPO, Inclusion with incubation, Inclusion without incubation, Folic acid, Other Hb, HbS, HbF, HbA2, HbA, Rest blood exams, Rest blood Windfalls, Cell structure, Transferrin, TIBC, Fe, Β12, Ferritin, ESR, Blood-platelets, Immature RBCs, and Destination of DHS.


However, Haematologists of regional departments of Patras and Thessalonica in Greece who were involved to the evaluated phase of our system, observed the clinical benefits that the combination of the e-HS’s characteristics provide. A few of their observations regarding the system are following: i) Time saving through reducing searching and clerical activities, consultation between departments, handwritten information transfer ii) Greater accuracy and speed of information transfer iii) Additional services for better diagnosis (e.g. DHS processing via web) iv) More complete patient record and data and global access v) Clear presentation of data vi) Rapid switch to different service 4. System description e-HS involves bringing together people, data, process, and technology to provide services that would not occur if the integration were not present. Since the focus is upon the integration, instead of upon the items being integrated, one cannot build a system as a singe task. Instead, one must create an environment that redirects and coordinates a variety of individuals so that they come together to form the e-HS. 4.1 The Interface Manager As we mentioned earlier, the whole system is based on the idea of distribution and openness. The e-HS achieves communication integration and allows the exchange of data and services between nodes linked by a network, but hides the distribution aspects to its users. The exchange of highly specialized haematological services is not the only problem when building a medical system. The major difficulty lies the integration of those services into a coherent environment. e-HS allows access to services, through a single browser and a homogeneous user interface, irrespectively of the storage place and the format of it. The interface Manager (IM) is responsible for this presentation integration. It is therefore the unique user-interface for handling all the services that compose the e-HS system. The IM hierarchy, thank to the inheritance principle, ensures a common look and feel among all the users and facilitates the interaction of a Haematologist with the e-HS system. The basic environment of this IM starts with a graphical web environment that embeds the MCGrid component. All the other services can be also launched from this module by clicking the corresponding buttons.


4.2 e-HS Services The e-HS is intended to gather the following services that meet the major haematological requirements: i) Tele-monitoring and management service for laboratory textual results: Integrating computer used with the workflow of busy clinicians is inherently challenging, but when such use requires data entry by physicians, few systems present the “willingness” to be successfully adopted. Considering this observation, we implemented a component, namely: MCGrid that displays and manages the content of the common database in an easy, user-friendly and secure fashion. Specifically, this component enables a graphical grid scheme that displays the main predefined haematological variables and other related aspects (e.g. physical location of corresponding DHS or diagnosis etc.) as shown in figure ----------. More specially MCGrid fixes one column per blood parameter’s value and each bloodexamination per row and chooses to organize the information within each episode in reverse chorological fashion. In this case, the most recent episode of care or note is positioned at the top of each section. If neither of these options is used, analysts checking the health record for completeness simply include all data in given section of record. Although this speeds up the assembly process, it impedes quick data retrieval. As a result, Haematologists can easily work on the patient’s data, simply by clicking on the appropriate contextual links. This interaction via simple clicks on the MCGrid interface, provides operations for collection, maintenance and deletion, operations for data creation, reload, data replication and movement, access control and data deletion as shown in appendix in figure 2a, 2b, 2b. ii) Monitoring service for remote DHS: Medical images (e.g. DHS, PET, SPECT, MRI, CT, X-ray, etc.) are frequently shared by networks and used for diagnosis [25, 26]. In this framework, we have developed a viewer that enables the analysis of remote DHS within a browser-based interface, named VDHS. It is the module in charge of the remote handling of DHS. By using this component the physicians can compare simultaneously DHS (previous and current exams) without cumbersome tasks and


determine in a quantitative way the abnormalities of a specimen in order to increase the diagnostic or prognostic value of biopsy figure 3. iii) DHS’s replicate services: At this point, we should mention that another significant service of e-HS is the replicate of distributed DHS at the Medical Physics Laboratory of the University of Patras for image processing by the experts. In other words, e-HS allows the remote DHS to be replicated and to be stored in underlying physical location at the master-node’s database, for image processing. The clinicians at regional nodes, can easily generate DHS’s folders at master database, denote their name (usually–Patient’s ID) and move a DHS- transcripts into them. This procedure may be dynamically accompanied by a metadata generation, which can be used by the system in order to form the list of patient’s folders. as shown in figure 4 iv) Consultation by remote physicists on DHS During the above service, e-HS is able to establish a teleconsulting

facility by a

complementary service, among the remote haematologists and the physicist at master node, in order to support the DHS’s processing-procedure. The physician selects meaningful regions in each DHS and indicates histopathological features for quantitative analysis. The image processing of DHS’s transcript takes place at the master node (Medical Physics Laboratory of University of Patras) by the experts of our lab who perform image segmentation by isolating the structure requested [27, 28 and 29]. Finally, the remote haematologist grabs with improved accuracy the suspected histopathological problem by using the embedded DHSV . The sequence of this procedure is illustrated in appendix -------- figures 4a, 4b, 4c, 4d, 4e, 4f]. v) Data Discovery Service Discovery in large-scale databases has been defined as a non-trivial process [30]. This service allows the searching to common database and the merging of the results into a sub-catalogue.


The principal aim of this service is to provide a modern discovery-mechanism for the haematological laboratory results for a specifics patient. The module includes a subject gateway that directs search in easy way to the common database. This service is critical for the system’s performance, as the physicians have a large variety of searchable regional collections with clinical data. The physicians only need to enter the patient’s ID or patient’s date exam; the service takes care of the rest. This approach reduces the complexity of the system. Typical scenario of this for patient with ID equal to 10023 is presented in figures 5. 4.3 Development issues We have implemented the whole system based on a server-based approach. Master-node’s machine was constructed upon a MS WINDOWS 2000 Advanced Server environment that is supported by MS Internet Security & Acceleration Server (ISA). Moreover, each node’s machine was constructed upon a MS WINDOWS 2000 environment, which is supported by an Internet Information Server (IIS v5.0). IIS is used to cluster the machine with the others remote system’s parts; as a result the whole system appears as a virtual unified node with a total capacity of about 0.63 terabyte high performance data cache. The use of the above server-based solution allows the web-based system to exploit the full power of a high-level programming approach and provide a set of services through the HTTP protocol. The system is accessed compliant to the HTTP protocol via an active server pages (ASPX) and programming script in order to overcome previous deficiencies and functional insufficiencies [31]. ISA provides two tightly integrated modes - a multilayer firewall and a high performance web cash server. Specifically, it provides unified management, including logging, alerting, monitoring, and reporting, providing a comprehensive and accurate picture of the system’s usage and the connectivity needs. This means that our implementation of an e-HS can be available to every authorized physician at the distributed nodes without any hard- or additional software. The only software required for the user is the regular Internet Explorer (Version 3.02 or higher). Besides by using intuitive and self-explaining user interfaces and HTML-techniques such as hyperlinks the necessary amount of training of physicians-side is reduced to a minimum.


e-HS is constituted from several independent and cooperative pieces of software, called modules, devoted to the management of particular modalities of medical information both when services are being developed and being run. Another aspect of source’s design of the system is that it allows pre-compiling all business logic so that it does not buy any performance gains because modules compiles code-behind pages the first time users access the module.

e-HS is made up of the following modules: a)Modules for Data storage ‰

Regional DHS’s databases


Common database at master node

b) Modules for Data management ‰



DHS Viewer (DHSV)


Resource Discovery (RD)


Teleconference module


Slides acquisition module

These individual modules are analyzed along with its relationships and interfaces in the following sections. 4.4 Databases of the e-HS The data storage module consists of two types of components: i) Regional DHS’s databases: The system runs on a collection of distributed databases that contained DHS. These databases are made available via local servers. DHS have been set up, according to agreed interoperability guidelines, including a decided format (e.g. JPEG, TIFF, etc). As a result, it is only necessary to access the individual participating node’s servers and link to the corresponding path in order to grab DHS. ii) Common database: The ordinary Database Management Systems (DBMS) have focused only on structure data, laid out in tabular form .


In contrast to this, our approach supplies a possible solution, for the connection of the structure and non-structure data by using the common database to the back-end of master node [32]. The proposed common database is a non-traditional approach and can be used from future haematological systems, applications or systems providing integrated ways to web enabled services. This database is a document based file that identifies all patients who have been admitted or treated by a health care department. All patients who registered to receive care as inpatients, outpatients, emergency care patents are listed individually in the common database. In addition, the common database is the key to locating DHS. Its schema maps the values of blood parameters and provides information about the physical location of the distributed DHS at regional network-nodes. By using this approach, system is becoming more common, saving space and more importantly, facilitating the retrieval of information. With located throughout the haematological facility, the system makes the identification information available to all nodes. This hybrid approach also reduces errors in filling and increases the accuracy of data from a distributed medical node. As a result, a unified, document-based file is created that may be indexed for quick, efficient retrieval and universal access to all medical information sources of the system. There are three different types of data to the current scheme of XML based database: ‰

Haematological laboratory blood results,


physical slide’s location


diagnosis annotation and commentary

A typical dataset scheme for a specific patient is presented below 34567 17/10/2002 46.4 15.4 - - AB 92 32.4 31.3


8400 65% 24% 8% 2% 260000

4.2 Data management Interactions with the common database and the services are performed through the MCGrid interface. We have developed a first version of this MCGrid module that is fully implemented in MS Visual Basic.Net and has already received positive evaluation feedback. This module must be as transparent as possible and will maybe a possible integration for the previous approaches for the collaborative systems [33]. Data management subsystem consists of the following components: i) MCGrid module: As mentioned before, we have developed an entity with a web oriented interface (ASPX page), which facilitates the data-generation and the datamanagement at the common database in a secure, effective and easy fashion. The size and quantity of datasets at common database will grow in the coming years so this structure is a possible solution that can handle millions of datasets. The Hyper Text Transfer Protocol (HTTP) mainly aims at the web-exchange of information [34] while the MCGrid is concerned with the exchange of computer power, data-storage, and accessing the distributed databases, without forcing haematologists to search for these resources. Several techniques and programming languages (C# and MS Visual Basic.NET, and scripts) is used at source-side of MCGrid for the

sorting, editing,

deleting, and inserting haematological laboratory results into the common database. ii) DHS Viewer (DHSV): The DHSV module grabs DHS from a specific locations (paths) at regional server-based database and display them in easy fashion. The DHS may be accessible by standard protocols as HTTP using a pop up viewer. This embedded pop-up window uses the mechanism of web browser and Active-X controls for the displaying remote DHS. Each row-record at MCGrid is accompanied by DHSlocation’s data and the corresponding link. As a result, the DHS maintained locally in


the remote network-nodes are accessible through this module in a trouble-free way, and according to logical cataloguing scheme independently of the physical location. This first version of DHSV grab DHS with JPEG2000 format; but it is not difficult to extent the module to display and other DHS-formats. The recently adopted JPEG2000 standard is based on wavelet transform technology, an approach that is deemed superior to the Fourier-transform based JPEG approach. With JPEG2000, an image can be compressed about three times more than with JPEG, and still achieve the same image quality. We believe that additional studies must be performed in order to determine the minimal DHS requirements for accurate remote diagnosis and may be able to take advantage of adaptive colour reduction-algorithms to reduce DHS’s file size (without under grading slide quality). iii) Resource Discovery module (RD): As we mentioned before, the Resource Discovery module allows Haematologists to search at common database with two types of variables. In our approach for example, a haematological query, according to the requirements, is able to be “Find the patient with the specific ID and in specific date exam”. Haematologist fills the text-boxes of RD-web environment and press the searchbutton. The results are presented easily and effectively in a new sub-grid environment. This module is an ASPX page that it’s source is fully implemented in C#. Besides the search, mechanism is based on a XML filter, named XmlDataTableFilter. Current module can easily extend in order to submit various other types of queries or to allotted nodes for discovery binary object. iv) Teleconsulting module: With a piece of code and by using the embedded Netmeeting v3 Haematologist, through an automate process, is able to consulting the physicist at master node (with standardized IP address) in order to locate together the suspect area of DHS. iv) Slides acquisition module: Digitizing an entire microscope slide at diagnostic resolution is a formidable technical challenge. DHS should be captured with a scanning system that provides high quality images at the resolution required to illustrate the features of interest. These are


compressed in JPEG 2000 loss-less format. The compress-format is important for the medical images, especially if the these could be used for diagnosis. This module uses the file-browse mechanism of Windows to identify the local slides and the upload method (MSVB.Net) in order to transmit them to master node’s database. The patient folders are listed automatically and the physicist can easily get the DHS for processing. Files from MS WINDOWS 2000 and XP have been successfully transmitted using this slides acquisition module. 4.3 Security issues According to existing healthcare standards, security plays an important role to our methodology and as a result, e-H.S offers strong security available on the Internet. The combination of e-HS, ISA Server and its underlying MS Windows 2000 Advanced Server brink security and acceleration of data flow to regional nodes. All attacks and unauthorized accesses are rejected. MS ISA works at various communication levels to protect the corporate network of e-HS. When ISA Server processes a request from any external client, it checks IP packets filters, publishing rules, and routing rules to determinate if the request is allowed and which internal or external server should service the request. The system, by using ISA, also supplies the Domain Name Server (DNS) application filter that analyzes all the incoming traffic for specific intrusion against the corresponding servers. Moreover allows smart service filter, integrated intrusion detection, secure service publishing and broad service support. Specially, we have adopted the client-certificate authentication method (for incoming requests) and the server-certificates (for the published node-server) It should be emphasized that the real patient’s identification data (except ID) and other practitioner’s special useful data are packed into a 128-bit hash code at a non-shared directory at each node for security issues. 5. Status report The first version of this server based system, has been established at the Medical Physics laboratory of the University of Patras ( which served as master-node, and adapted for pilot use in two distributed nodes (University Haematological Departments of Patras and Thessalonica) and has already receive positive evaluation feedback.


The common database-the database file that identifies all patients who have been admitted treated by the health care facility- is the key to locating DHS and is crucial data storage source for the facility. The system by using this type of database is becoming more common, saving space and more importantly, facilitating the retrieval of information. With located throughout the haematological facility, the system makes the identification information available to all nodes. This hybrid approach also reduces errors in filling and increases the accuracy of data from a distributed medical node. During the experimental use of this system: ‰

78 patients (150 record-sets) were included in the medical catalogue. 48 patients were extracted to the common database from the existing MS Access database of University haematological hospital department of Patras as shows the figure ++++ and 30 patients were inserted at MCGrid from distributed points using remote access.


200 searching events that the physicians successfully performed to common database (using the embedded Discovery services) and 100 DHS grabbing events


16 DHS were captured at hospital haematological department of Patras and replicated at the master node database.


20 teleconsulting sessions among haematologists and physicist

Concerning usage and usability, the pilot version of e-HS system was generally well accepted,

by the haematologists, of the two distributed departments who accessed the

proposed system. They also noted that a point especially important is the DHS-replicating procedure in order to image processing by remote physicists, which provides a feeling of security at the diagnosis on the Haematologists side, within minimal time and without the need to post by the classic way -or sent them by email- to the laboratories [35]. Specially, they took all the useful data correctly, from the improved DHS (JPG 2000 lossless format) and easily interpreted them. Just 5% Haematologists reported difficulties in handling the e-HS, 95% found it easy to use. These difficulties were encountered when they had to move the DHS. In addition, the unified solution was not seen to provide sufficient protection of


rights in their data by a few Haematologists. Besides, a few navigation-difficulties were encountered with the viewer, but these were simplified in the `next version'. Just 5% Haematologists reported difficulties in handling the e-HS, 95% found it easy to use. These difficulties were encountered when they had to move the DHS. In addition, the unified solution was not seen to provide sufficient protection of rights in their data by a few Haematologists. Besides, a few navigation-difficulties were encountered with the viewer, but these were simplified in the `next version'. 6. Lessons learned The e-HS is a system that makes the distributed patient health record available at all times to authorize users from every node and provides a set of useful services. The system uses a variety of distributed data with the end result being an efficient and effective method of collecting and retrieving haematological information. Whatever the system merge, preserve, and provide the haematological folder rendered by various professionals during an episode of care. The variety of services and the unified methodology being mentioned previously, points out that the end users would benefit from this quality-filtered, comprehensive web-based system. A very important aspect of e-HS is the given ability of remote diagnosis performed intra or inter nodes. Unified approaches will move the healthcare information systems away from the current server-centric Web model, with relatively few servers and vastly greater numbers of clients (browsers), and towards distributed models in which every connected node can share its resources with every other. 7. Future plans We should make it clear that e-HS by itself is not enough to enable a worldwide haematological system infrastructure yet. However, it definitely is a first step towards this. A universal approach needs predefined haematological a data-model and facilities. Besides there are significant obstacles inherent in common database

and in the other system’s

components as well. Even if this methodology will eventually become accepted as a standard approach for delivering haematological services, it will take years for the proposed technologies to mature and be widely adopted. Finally, there are significant technical and


regulatory issues surrounding the privacy, security, and confidentiality of individually identifiable health information, which need to be settled. 8. References [1] E. Coiera, Guide to Medical Informatics, the Internet and Telemedicine, in Artificial Intelligence in Medicine, Chap.19 ( Elsevier 1997). [2] K. Clarke, R. O'Moore, R. Smeets, J. L.Talmon, J. Brender, P. Nair, J.Grimson, B. Barber, A methodology for evaluation of knowledge-based systems in medicine, Artificial Intelligence in Medicine 2 (1994) 107-121. [3] B.Kaplan, The influence of medical values and practices on medical computer applications, in Use and Impact of Computers in Clinical Medicine, pp. 39-50 (SpringerVerlag, New York, 1987). [4] F. Gremy and P. Degoulet, Assessment of health information technology: Which questions for which systems? Proposal for a taxonomy, Medical Informatics 18 (3) (1993) 185¯193. [5] S. Jones, Graphical Interfaces for Knowledge Engineering: An overview of relevant literature, The Knowledge Engineering Review 3 (3) (1988), 221¯247. [6] J. Gosbee, The discovery phase of medical device design: a blend of intuition, creativity and science, Medical Device and Diagnostic Industry 19 (1997), 79-82. [7] L.G.Kun, Telehealth and the common health network in the 21st century, Computer Methods and Programs in Biomedicine 64, (2001) 155 – 167. [8 ] E.Topalis,, G. Orphanos, S. Koubias, G. Papadopoulos, A Generic Network Management Architecture Targeted to Support Home Automation Networks and Home Internet Connectivity, IEEE Trancactions on Consumer Electronics 46, (2000), 44-51 [9] V. Kapsalis, S.Koubias, G. Papadopoulos, OPC-SMS: A Wireless Gateway to OPC Based Data Sources, Computer Standards and Interfaces, Elsevier 24, (2002), 437-451 [10] E.Topalis, L. Mandalos, S. Koubias, G. Papadopoulos, I. Nikiforakis, A Novel Architecture for Remote Home Automation e-Services on an OSGi Platform via High-Speed Internet Connection Ensuring QoS Support by Using RSVP Technology, IEEE Transactions on Consumer Electronics 48, (2002) 825-833. [11] [12]


[13 ] [14 ] [15 ] [16] J.Leeming,D.Garnett,A.Roberts,The 'Oswestry' EPR Project (1994-1999) : A Final Report,(NHS Executive,United Kingdom,1999) [17 ] [18] C.H.Kim, H.M.Choi, S.M. Park, Medical image conference system using CORBA for Web client design, Multimedia Conference in Korea, (1998) 339¯343. [19]






Architecture for networked electronic patient record systems, International Journal of Medical Informatics 60 (2000) 161-167. [20] M Dugas, K Kuhn, N Kaiser, K Uberla, XML-based visualization of design and completeness in medical databases, Medical Informatics and the Internet in Medicine 26 (2001) 237-250. [21] A. B. Chaudhri, Objects, XML and databases, Information and Software Technology 44 (2002) 195-196. [22] S.Alagi, Institutions: integrating objects, XML and databases, Information and Software Technology 44 (2002) 207-216. [23 ] [24] W.D Bidgood, Y al Safadi, M. Tucker, F.Prior, G.Hagan, J.E. Mattison, The role of digital imaging and communications in medicine in an evolving healthcare computing environment: the model is the message, Journal of Digital Imaging 11 (1998) 1-9. [25] M.H Doolittle, K.W Doolittle, Z Winkelman, D.S Weinberg, Color images in telepathology: how many colors do we need?, Human Pathology 28 (1997) 36-41. [26] Heung-Kook Choi, Se-Myung Park, Jae-Hyo Kang, Sang-Kyoon Kim, Hang-Mook Choi, Tele-medical imaging conference system based on the Web, Computer Methods and Programs in Biomedicine 68 (2002) 223-231. [27] P. Spyridonos, P. Ravazoula, D Cavouras, K. Berberidis, G. Nikiforidis, Coputer-based grading of haematoloxylin-eosin stained tissue sections of the urinary bladder carcinomas, Medical Informatic & Internet in Medicine 26(3), (2001) 179-190


[28] P. Spyridonos, P. Ravazoula, D. Cavouras, G. Nikiforidis, Neutral

networks based

segmentation and classification system for the automatic of histological section of urinary bladder carcinoma, Analytical and Quantitative Cytology and Histology 24(6) (2002) 317-324 [29] P. Spyridonos, P. Ravazoula, D. Cavouras, G.Nikiforidis, A computer-based diagnostic and prognostic systems for accessing urinary bladder tumour grade and predicting cancer recurrence, Medical Informatic & Internet in Medicine (27) (2002) 111-122. [30] [31] J.M. Ivler, K. Husain, CGI Developer's Resource, NJ (1997) 33¯90. [32] Y.Kambayashi, M.Kitsuregawa, A.Makinouchi, S.Uemura, K.Tanaka, and Yoshifumi Masunaga, XML Databases Non-traditional Database Systems, The Information Processing Society of Japan, Chap.11 (JAPAN, 2002). [33] P. Mattioli, P.J. Klutke, F. Baruffaldi, M.Viceconti, A.Toni, Karl-H, A study of the application sharing capabilities in telemedicine, Computer Methods and Programs in Biomedicine 58 (1999) 89 – 97. [34] F.Gagliardi, A biomedical user point of view, 5th EU DataGrid Project Conference, CERN PILISCSABA, HUNGARY (2002). [35] D.H Okada, S.W Binder, C.L Felten, J.S Strauss,A.M Marchevsky, Virtual microscopy" and the internet as telepathology consultation tools: diagnostic accuracy in evaluating melanocytic skin lesions, American Journal of Dermatopathology 21 (1999) 525-531.



Suggest Documents