UC 577,1; 61 ISSN 0354-3447

Jugoslov Med Biohem 25: 39–46, 2006 Stru~ni rad Professional paper

Jugoslov Med Biohem 2006; 25 (1) 39 DOI: 10.2298/JMB0601039M

Introduction

Client/Server/Internet is at present one of thegreatest sources of information, knowledge, referen-ce materials, and educational and training programsfor the general population, and especially for the pro-fessionals from clinical laboratories.

In the clinical laboratory, we have developed astrategy for optimizing the analysis of laboratory data-bases. The database analysis processes and compa-res all the laboratory data characterizing a web brow-ser interface using standard Internet technology. Thepatient identification and serial comparison of the re-sults of each patient are accurate and easy to per-form. The database could be helpful in a specializedlaboratory assay and could be connected with thehospital network.

Ironically, networking computing now predomi-nantly supports human rather than computer commu-nications. For the clinical laboratory, the internet pro-

vides high-speed communications through e-mail andallows the retrieval of important information held inrepositories. All this capability comes at a price, includ-ing the need to manage a complex technology and therisk of intrusions on patient privacy.

One of the major trends in computing for the1990s is the move towards distributed systems basedon Client/Server/Internet architecture. Although a re-cent survey has suggested the need of planning toadopt this new technology, there is little evidence atpresent of similar progress in the field of clinical la-boratory computing (1). Standardization of interfaceshas been proposed to secure a common frameworkcompatible with different types of information. Net-work databases allow users to retrieve data in multi-ple ways and resolve the problems associated withhaving multiple copies of the same data by setting upmultiple access pointers to a single data element.With the aim of achieving best practice in LaboratoryAutomation System (LAS) or Laboratory InformationSystem (LIS), they present the results of a question-naire survey of LAS or LIS (2, 3).

LISs have evolved into complex applications tomeet the specialized needs of laboratories. As inte-grated delivery systems (IDSs) continue to emerge as

CLIENT/SERVER/INTERNET COMPUTING AND STANDARDIZATION: THIS IS THE FUTURE DIRECTION FOR THE CLINICAL LABORATORY

Maria Markova

Department of Clinical Laboratory and Immunology, Medical University, Sofia, Bulgaria

Summary: This paper describes how, since 1990s, we can use new technologies, object-oriented soft-ware tools and standardization for development in clinical laboratory. The paper will also explain how a labora-tory could easily set up an automatic facility for general practioners to use Internet and browser tools to accessresults directly from the Laboratory Systems report files, and from there automatically capture the requiredpatient information for integration. The files of clinical laboratory database (tests) are facing an increase in thenumber of test items, as well as a corresponding diversification due to the demands of medical institutions andthe improvement in analytical techniques. To respond to this situation, medical institutions have been promo-ting systematization of their procedures. Information exchange among the institutions has likewise expandedwith the use of media such as on-line systems, World Wide Web (WWW) and the Internet. This experience sug-gests that this approach will be the way forward for the high performance but user-friendly laboratory systemsof the future.

Key words: server, internet, object-oriented software, World Wide Web, network, e-mail, LAN, WAN,HTML, URL, hypertext, Java, fuzzy logic

Address for correspondence:

Maria MarkovaDepartment of Clinical Laboratory and ImmunologyMedical University, Sofia, Bulgaria

40 Markova: Client/Server/Internet computing and standardization

the dominant model for healthcare delivery, clinicallaboratories serving them face two imperatives thataffect IDS decisions. First, laboratories in IDSs mustconsolidate to reduce costs and duplication, yet deli-ver services across a region in both in-patient andout-patient settings. Second, laboratories that suc-cessfully perform reference laboratory testing willincrease revenue, generate referrals, and leverageexcess capacity.

Computer networks: What is World WideWeb? How will we use it in LaboratoryInformation Systems?

A computer network, or network, is a conceptseparate from the network database model. A net-work is any interconnection between computer sys-tems or between a computer terminal and a compu-ter laboratory system that supports the interchangeof data. If the network is physically located within a li-mited locale, such as 1 to 2 miles, it is called a localarea network, or LAN. If the network spans a largerarea, it is called a wide-area network, or WAN. High-speed LAN technology is inexpensive and often user-installed. Most computers include Ethernet as anintegral part of the hardware, and such systemsrequire networks to communicate with other com-puters, such as the network for Adapter Data Transfer(ADT) and billing functions. Interchanging data bet-ween different computers requires software to sup-port a common protocol, or an agreed-upon way tointerchange data. Unfortunately, incompatibilitiesbetween software vendors, hardware vendors, andapplications software often make this interchangehighly complex (4). The Web is not identical to theInternet. It is one of many Internet-based communi-cation systems. World Wide Web (WWW) and theInternet function are using the analogy with the glo-bal road system. The basic idea of WWW is to mergethe techniques of computer networking and hypertextinto a powerful and easy to use global informationsystem. Hypertext is text with links to further informa-tion, based on the model of references in scientificpapers (5). Hypertext Markup Language (HTML) orhypertext is a way to organize documents so that yourcomputer can help you find items of interest.

As networks advance in speed and capability,computer systems implementing a Client/Serverarchitecture have emerged. This term describes aspecific interaction between two networked comput-ers. The server computer holds the database fromwhich clients retrieve and process data for their use.If a clients updates the data, that will automaticallyupdate the data on the server. Uniform ResourceLanguage (URL) address is the name of the address.It includes the information on where the file is locatedand what a user can do with the browser. Every file inthe Internet has a unique URL. The client/serverapproach has become popular with LIS systems in

two areas. With smaller systems, some vendors haveimplemented LISs with inexpensive personal comput-ers and Client/Server software, replacing more expen-sive minicomputer-based systems. Larger systemshave implemented a server architecture to allow userswith personal computers to download data so thatthey can process it independently (1).

The other group of tasks which are used isEmbedded Systems are computers or microchip-ba-sed control systems that are dedicated to performingspecific tasks. Embedded systems have been appliedto a variety of medicine tools and laboratory tools.Embedded systems are cheaper and more reliablethan today's software used in PCs. With the advancein the Internet technology, a lot of electronic goods –laser printers, computers, security monitors, medicaltools will be connected by a global network called theEmbedded Internet. The Embedded Internet will per-form many tasks, as the automated control/monitor-ing on production lines. This connection can be fur-ther extended to include LAN, WAN, CAN (ControlArea Network) and CEBus (Commercial ElectronicBus). CAN is a European protocol that provides astandard method for real time. When Java applets are

Figure 1 Hypertext as a way to organize documents

Figure 2 Client/Server architecture

Jugoslov Med Biohem 2006; 25 (1) 41

loaded into the MCUs (microprocessor), CPUs (Cen-tral Processor Unit), and other microchips embeddedin the target systems connected to the Internet viaLAN, WAN, CEBus, or CAN topologies, they makethe Embedded Internet.

Microwave, satellite, and other frequency sys-tems are supporting mobile information-gatheringactivities. For example, some vendors are providinghand-held portable units for phlebotomists to in-crease the accuracy of collection times, and for bet-ter patient identification. Advanced paging deviceslinked to the LIS allow physicians to have immediateaccess to patient data as soon as it has been verified.These systems may also be used to interconnectcomputer systems where cables cannot be used orare inconvenient. Most of these mobile technologiesare substantially more expensive than cable-basedsystems. The effect of all these network technologieshas been to allow the integration of information fromphysically diverse locations. In many institutions, net-works are critical in supporting satellite laboratoryfacilities, allowing the satellites to have access to pa-tient data, and in many cases allowing the creation ofstructures that support a decentralized physical andorganizational operation. Such satellite operationsare becoming increasingly important as laboratoriesgrow to meet expanding health care alliances (6).

Development of Embedded Internet using Javahas emerged as the most widely adopted program-ming language for the Internet applications. Thereare four advantages over other methods for usingJava applets for distributed control over the Internet:1) web browsers are readily available and offer a hig-hly programmable and configurable user interface;2) Java code size is up to 50% less than machineinstructions, making it highly attractive to MCU-basedcontrols where RAM memory size is limited; 3) mostof the hardware and software is readily available, andno extra investment or development is needed; and4) there is no need for developers to reinvent softwaresince most MCUs' kernels or PCs operating systemswill support Java and TCPI/IP (network protocol)which helps the computers in the Internet talk to eachother and HTTP.

The electronic mail (e-mail) was popular with thecomputers networked, and scientists used it to facili-tate their communications. Different computers oftenused different communication protocols, and this pre-vented exchange of information. TCP/IP (transmissioncontrol protocol-internetwork protocol) was developedto overcome this, and it gradually became the domi-nant protocol. The name Internet originates from IP,the abbreviation for »internetwork protocol« (6).

A simplified Embedded Internet consists of webbrowsers at one or more sites, target embedded sys-tems (such as of laboratory tools) at same or differentgeographic locations, and the MCUs embedded inthese target systems. The MCUs are the devices thatstore Java applets to execute the controlling or moni-toring functions on the controlled targets, using eitherclassical method (feedback) or fuzzy logic method (7)(see Figure 4).

In such a networked system, transferring HTMLpages and executing Java applets achieve distributedcontrol. A user at one central site can use a webbrowser to access HTML pages from one or more tar-get embedded systems, along with Java objects thathave been inserted into them, and execute Javaobjects to perform the control operations program-med. After executing one or more such Java objectsfrom a web browser, the user can then send tasks tothe targets, which are represented in HTML pagesplus same Java objects, to perform remote or distri-buted control of the targets (LAB tools). To createnew tasks or to update old tasks at a target, the userat the central site can send via FTP (F Transfer Pro-tocol) software objects (Java applets) to the targetsite where they are to be stored inside the MCU andexecuted when needed (7).

A database in LIS with a network data relationship

This data model very efficiently utilizes comput-er resources and can be very effective in many situa-tions. Unfortunately, the relationships between vari-ous data elements can become very complex and dif-ficult to understand. Such complexities could beco-me particularly problematic should data relationshipschange; that is, systems built using network modeldatabase managers are easily created but difficult tomaintain. Most common database managers are ba-sed on the network model (8).

The relational database management systemhas become popular because of its easy use. Thedata appear as tables, and database operations be-come operations on tables that users can performwith ease (9). This flexibility has earned relationaldatabases a large role in situations where data aremanipulated frequently, as in a research setting. Aside effect of this flexibility is that relational databases

Figure 3 The connection of Embedded systems with LAB

42 Markova: Client/Server/Internet computing and standardization

utilize computer resources less efficiently than sim-pler models do, and this may present problems withvery large datasets. Some examples of relationaldatabases are IBM's DB2, Oracle, and Sybase. Manyother databases often claim to be relational, but arebased on the network model. Few LISs use relationaldatabases for high-volume data transactions becauseof performance and cost issues, though some sys-tems transfer data from a no-relational productiondatabase manager to a relational database managerfor ad hoc queries. These hybrid approaches maycompletely move to a relational database manager ashardware resources become cheaper (10).

Data structures are critical to the design of pro-grams. First, stored patient data in an LIS consumemost of the disk and memory storage. In an LIS, theamount of laboratory data easily exceeds the pro-grams needed to run a system by several orders ofmagnitude. The design of the data structure mayimpose limitations on stored information. One sys-tem might be limited to storing data on 2,000patients, whereas a second system might be limitedto 64,000 patients. The vendor of the smaller systemmay have made this restriction so that their LIS runsfaster and more efficiently and requires less disk sto-rage, thus producing a system less expensive thanthe larger one. But this efficiency is irrelevant if theuser's needs exceed 2,000 active patient files (10).

Integrating and presenting clinical outcome data

Clinical information system integrates varioussources of clinical data and facilities outcome assess-ment for patients evaluated in a lumbar spine service.During an encounter with a patient, the physician for-mulates a hypothesis regarding appropriate forms oftreatment, and he or she may then use this system toexplore previous treatment outcomes for similarcases. The availability of a clinical tool that presentsinformation in an outcome-oriented format may behighly relevant for the delivery of cost-efficient, high-quality health care and also create a formal mecha-nism for detecting practice variability.

Automated laboratory protocols

A protocol is a program which controls, moni-tors and modifies the requests for laboratory servicesduring the diagnostic work-up and/or monitoring of apatient. A protocol language and an OS/2 based sys-tem for the compilation, interpretation and executionof laboratory protocols written in this language arepresented. The system is easily interfaced with anypatient database that supports the structured querylanguage (SQL). A compiled protocol may be assig-ned to a patient and executed as specified in the pro-tocol itself (regularly and/or when certain events suchas test requests or arrival of results occur). In the lab-oratory protocol language, a patient's data are viewedas a set of test procedure groups, each comprisingdata (request time, result, etc.) describing the statusof one or more simultaneously made laboratory testrequests. A pattern specification is a statement sayingthat a sequence of test procedure groups of specifiedtypes and ages is present in the data. Pattern specifi-cations are linked to Boolean variables. If a patternmatching a pattern specification is found in the pa-tient's database, the corresponding Boolean variableis set equal to TRUE. The Boolean variables are uti-lized in the decision logic of the protocol (2).

Standardization

Some embodiments include: (1) Interface Stan-dards on Clinical Laboratory Information. For informa-tion exchange, the format and reporting commentsused in the media systems were standardized under thesponsorship of the Medical Information System De-velopment Center, with a publication issued in 1993;(2) Standardization of Laboratory Test Code Standardsdevelopment and data interchange.

To address the high cost of software develop-ment and the lack of coordination between the vari-ous information systems, the health care computerindustry, LIS and Hospital Information System (HIS)users, and the federal government have created vo-luntary groups to set standards, particularly for theinterconnection between information systems and

Figure 5 A database with a network data relationship. The data are organized so that they can be retrieved

either by the patient or by abnormal results.

Figure 4 A database with a network data relationship. The data are organized so that they can be retrieved

either by the patient or by abnormal results.

between an information system and computerizedequipment. For the clinical laboratory, standardsaddress the interchange of data (1) with bedsidemonitoring systems, (2) with administrative systems,including those associated with payers and govern-ment agencies, and (3) between laboratory systems.Of particular importance are those functions coveringthe interchange of laboratory test requests and re-sults. Three voluntary groups that have created stan-dards in the laboratory arena include the AmericanSociety for Testing and Materials (ASTM), already

mentioned earlier with regard to automated instru-mentation, the Institute of Electrical and ElectronicsEngineers, and the HL7 committee (3, 11). The HL7committee was specifically formed to develop stan-dards for supporting the interchange of informationbetween computers in the hospital environment.Areas covered by the HL7 committee have includedADT, billing, patient demographics, orders, and re-sults. The ASTM El 238 protocol has become thebasis for systems interchanging laboratory data bet-ween different systems. Many reference laboratory

Jugoslov Med Biohem 2006; 25 (1) 43

Figure 6 Laboratory integrated system

systems exchange data with hospital laboratory sys-tems using this standard. Interfaces between an HISand an LIS often use the El 238 standard or a similarsubset as defined by HL7. These hospital standardsfor information interchange carry profound implica-tions for LISs, for they offer large system-wide bene-fits at low additional LIS cost. For example, significantsavings result from physicians directly performingorder entry and retrieving results electronically, a taskthat requires integration of the HIS and the LIS. Theultimate goal of integrating interfaces is the migrationtoward a reduction in paper use and toward the elec-tronic patient chart. This will be difficult until medicalterminology becomes more consistent (5).

Software standards simplify communicationand cooperation between applications provided bydifferent software vendors. However, adherence to arigid standard can stifle innovation. Open standards,which are not proprietary to any particular vendor, areparticularly useful. Many standard-setting organiza-tions have provided useful tools for laboratory infor-

matics, including ASTM, Institute of Electrical andElectronics Engineers, National Committee for Clini-cal Laboratory Standards, and the College of Ameri-can Pathologists. Data communication standards,such as ASTM 1238, HL7, and Digital Image Com-munication of Medicine, have reduced the cost andimproved the timeliness of interfaces. Nomencla-tures, particularly the Systematized Nomenclature ofMedicine (SNOMED), have facilitated the structuringof medical information, and decision rules can becommunicated with ASTM E1460. Laboratory Obser-vation Identifier Names and Codes has eased theproblem of identifying tests during interchange oflaboratory data. Implementation standards are usefulreferences for those deploying information systems.Standards have provided multiple benefits to labora-tories and to the health care organizations of whichthey are components.

Recent advances in tools for scientific data acqui-sition, visualization, and analysis have lead to growinginformation management problems for medical re-

44 Markova: Client/Server/Internet computing and standardization

Figure 7 Design of Laboratory Information System

search laboratories. An exponential increase in the vo-lume of data, combined with a proliferation of hetero-geneous formats and autonomous systems, has driventhe need for flexible and powerful Experiment Ma-nagement Systems (EMS). This paper provides a de-tailed analysis of the informatics requirements of anEMS, and proposes a new type of middleware called anEMS-Building Environment (EMSBE), which enablesthe rapid development of web-based systems for ma-naging laboratory data and workflow. We describe theWeb-Interfacing Repository Manager (WIRM), an open-source application server for building customizableexperiment management systems. WIRM is being usedto manage several ongoing experiments, including anatural language (12–16).

Conclusion

In the new technologies, information will be avalue as well as a commodity. In the case of globalcommerce, this web information may be about cor-porate development. The aim is to develop a distrib-uted hypermedia system over high-speed network. Thehypermedia system supports: high-level nodes; dyna-mic links; sessions; multiple concurrent users; intelli-gent information customization; search and browsingover large hyperbases; and automatic hypertext engi-neering. The targeted applications for the systeminclude computer-aided learning and distributed infor-mation system.

Possibly the greatest potential for changing thehuman–computer interaction comes from new com-

puter technology being introduced today. These tech-nologies have the advantage of being potently easier touse than the keyboard, but reliability of these devicesmust increase before they are acceptable for use inmost areas of the clinical laboratory.

The use of computers in the clinical laboratoryis still in the development stage. The potential of LISsystems to help manage the automation flowrequired for good patient care and to help effectivelymanage the resources required (costs) has barelybeen used. We can imagine a future where informa-tics will be as important as the analytical result.

Major problems associated with the Internetinclude exposure of computer systems to maliciousattacks, possible loss of privacy in computerizedpatient information, complexity in supporting net-works and networked computers, and the inaccuracyof some information being distributed. Any computeron the Internet is potentially vulnerable to datadestruction. Break-ins by malicious users are proba-bly the biggest problem. Data destruction by virusesand worms acquired through the network is a lowerrisk event.

As the use of the Internet spreads and usersbecome more heterogeneous, its distinction as an eliteresource will likely disappear. Moreover, the Internetcan provide so much information that users are over-loaded. Despite these negative, characteristics, theInternet remains a valuable resource for communica-tion and for information, and permits the laboratoryuser to work in a manner more consistent with today’sneeds.

Jugoslov Med Biohem 2006; 25 (1) 45

KLIJENT/SERVER/INTERNET KOMPJUTERIZACIJA I STANDARDIZACIJA:TO JE BUDU]A SMERNICA ZA KLINI^KU LABORATORIJU

Maria Markova

Odeljenje klini~ke laboratorije i imunologije, Medicinski univerzitet, Sofija, Bugarska

Kratak sadr`aj: Ovaj rad opisuje kako od 1990-tih mo`emo koristiti nove tehnologije, objektno-orijenti-sanu (neproceduralnu) softversku opremu i standardizaciju, za razvoj u klini~koj laboratoriji. U ovom radutako|e je obja{njeno kako bi laboratorija mogla jednostavno postaviti automatsku olak{icu za op{tu upotreburadi kori{}enja pristupa internetu i sredstvima za pregledavanje, za pristup rezultatima direktno iz fajlova labo-ratorijskog sistema, i da odatle, radi integracije, automatski pokupi sve potrebne informacije o pacijentu. Fajlovibaza podataka (testovi) klini~kih laboratorija se suo~avaju sa porastom u broju detalja za testove, kao i sa odgo-varaju}im raznolikostima usled zahteva medicinskih ustanova i unapre|ivanja analiti~kih tehnika. Kao odgovorna ovu situaciju, medicinske ustanove unapre|uju sistematizaciju svojih procedura. Razmena informacijaizme|u ustanova se tako|e odvija uz pomo} medija kao sto su on-line sistemi, World Wide Web (WWW) i inter-net. Ovo iskustvo preporu~uje da }e ovakav pristup biti put ka unapre|enju visoko performatnih ali i pre-dusetljivih laboratorijskih sistema budu}nosti.

Klju~ne re~i: server, internet, objektno-orijentisani softver, World Wide Web, mre`a, e-mail, LAN, WAN,HTML, URL, hipertekst, Java, nejasna logika

References

1. Wells IG, Farnan LP. Client/Server computing. Clin-Chim Acta 1996; 15: 245 (1): 31–8.

2. ASTM designation El 381-91, specification for low-levelprotocol to transfer messages between clinical labora-tory instruments and computer systems. 1994 Annualbook of ASTM standards, vol. 14.01, Philadelphia,1994, American Society for Testing and Materials, pp311–317.

3. ASTM designation E1394-91, standard specificationfor transferring information between clinical instru-ments and computer systems, 7994 Annual book ofASTM standards, vol. 14.01, Philadelphia, 1994, Ame-rican Society for Testing and Materials pp 335–349.

4. Grimson W, Brender J, Grimson J. Specifying an openclinical laboratory information system. Comput-Met-hods-Programs-Biomed 1996; 50 (2): 95–109.

5. Bakker AR. Security in medical information systems. InBemmel H, McCray AT, editors: 7993 Yearbook of me-dical informatics, sharing knowledge and information.Stuttgart, Germany, 1993, Schattauer, pp 52–60.

6. Winkel P, Ravn HN. Automated laboratory protocols.Comput-Methods-Programs-Biomed 1996; 49 (1): 69–83.

7. www.aptronix.com/fuzzy/applote/java.htm.

8. Groth T, Grimson W, at al. Open Labs advanced instru-ment workstation services. Comput-Methods-Pro-grams-Biomed 1996; 50 (2): 143–59.

9. De-Moor-G, Fiers T, Wieme R. The research in seman-

tics behind the OpenLabs coding system. Comput-Methods-Programs-Biomed 1996; 50 (2): 169–85.

10. Born G, Given P, O'Moore R. Patient result validationservices. Comput-Methods-Programs-Biomed 1996;50 (2): 161–8.

11. ASTM designation E1466-92, standard specification foruse of bar codes on specimen tubes in the clinical la-boratory, 1994 Annual book of ASTM standards, vol14.01, Philadelphia, 1994, American Society for Tes-ting and Materials, pp 407–409.

12. ASTM designation E1238-94, standard specification fortransferring clinical laboratory data messages betweenindependent computer systems, 1994 Annual book ofASTM standards, vol. 14.01, Philadelphia, 1994, Ame-rican Society for Testing and Materials, pp 132–210.

13. McDonald CJ, Hammond WE. Standards formats forelectronic transfer of data. Ann Intern Med 1989; 110:333 –5.

14. McDonald CJ. Standards for the electronic transfer ofclinical data: progress, promises, and the conductor'swand. Proceedings of the 14th Annual Symposium onComputer Applications in Medical Care, Los Alamitos,Calif, 1990, IEEE Computer Society Press, pp 9–14.

15. Barnett GO, Jenders RA, Cheuh HC. The computer-based clinical record – where do we stand? Ann InternMed 1993; 119: 1046– 48.

16. Butch SH. Computer software quality assurance. LabMed 1991; 22: 18–22.

46 Markova: Client/Server/Internet computing and standardization

Received: June 18, 2005

Accepted: November 18, 2005