Hospital information systems–Past,Present, Future

International Journal of Medical Informatics (2006) 75, 282—299

Hospital information systems—–Past,present, future�

Peter L. Reichertz

Institut fur Medizinische Informatik, Medizinische Hochschule Hannover,Carl-Neuberg-Str. 1, D 30625 Hannover, Germany

KEYWORDSHospital informationsystems;

Summary This article about the state of the art of hospital information systemsand about future directions was written in 1984 by the late Peter Reichertz. It nowserves as a companion paper to a paper with the same title, written by Reinhold

System architecture Haux 20 years later (this issue). In that paper Reinhold Haux starts where PeterReichertz ended. The original text of Peter Reichertz has been retyped and thefigures redrawn. Possible errors may have resulted from this process.© 2005 Elsevier Ireland Ltd. All rights reserved.

1. Introduction

Hospital information systems have been in exis-tence for almost two decades now. Therefore, itis possible to speak about their past. This contribu-tion is intended to examine some of the featureswhich we could observe during this evolution andintends to project some of the present observationsinto the future. However, the analysis will not be aformal one. Especially, it will not add new classifi-cations and typologies to already existing ones like,e.g. the typology of Ball [1] according to the dif-ferent stages of achievement and communication.Instead, some general aspects will be examined andsome projections will be made out of the presentstage as hospital information systems present them-selves today.

DOI of original article:10.1016/j.ijmedinf.2005.08.002.� Key-note address during “Medical Informatics Europe 84”,

Fifth Congress of the European Federation for Medical Informat-ics, Brussels, September 10—13, 1984.

It has to be noted that some of the features dis-cussed for the future already have been seen in thepast and have vanished in the present. They maycome back in the future in a ‘recycling’ develop-ment as technology advances. Earlier ideas, whichcould not be implemented successfully due to lackof software and hardware techniques, may again beavailable tomorrow. Also, some features, which aredescribed as features of tomorrow, can already beseen here and there in present systems of prototypecharacter without being generally available.

Since earlier some contributions to the sametopic have been made, it may be permitted touse some of the ideas expressed earlier resp. fig-ures used elsewhere, (e.g [9—14]). As it happens,some observations and projections made in the pastare still of value today. Therefore, the respectiveagreements and graphics may be allowed in thisreview for a more comprehensive picture.

In general, an analysis has to orient itself accord-ing to the axes:

- software and software technology;
- hardware and development of equipment;
1386-5056/$ — see front matter © 2005 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.ijmedinf.2005.10.001

Hospital information systems—–Past, present, future 283

- system concept, e.g. the general approach andthe architecture developed;

- environment and its changing aspects and- system ecology as the new ecology evolving in

an environment into which an information systemis placed and which begins to react with peopleand organization which themselves have a stronginfluence on the development of the system assuch resulting in a new ecological balance whichdiffers from the old environment and, in mostcases, also from the original system design.

However, contrary to previous analyses [11—14]the following reflections will not be groupedaccording to these axes though they will be dealtwith under various aspects. Also it is not intendedto make some programmatic statements with a fer-vent belief in all the good things technology canbring to medicine. But hopefully an analysis froma more remote standpoint may reveal one or theother global aspect of value.

2. General aspects

Imsost

model given in Fig. 1 has the components of database and base of methods for:

- acquisition of information;- processing information and- presentation of information [10,13].

Information systems do not have the right ofexistence out of themselves. They have to serve apurpose. The functions of such a system (see also[8]) can be described as:

1. reporting and regular collection and collating ofdata;

2. analysis, e.g. comparison and description of theevolution of data;

3. trend recognition as the result of the evaluationof continuous analyses;

4. prognoses in terms of projection of trends intothe future;

5. problem identification and classification and6. definition and selection of alternatives for

action.

Hospital information systems very doubtlesslyhave reached the second level in this list of func-tsrqio

ation

nformation systems have to map the real environ-ent into a formal representation, e.g. data base

tructures, through the bottleneck and restrictionf data acquisition [10,13]. The information thustored has to be processed and presented back tohose who want to use the system. Therefore, the

Fig. 1 Model of Inform

ions, though naturally in one or the other case,ome functions of the higher levels may be alreadyecognized. This means, however, even thoughuite successful solutions have been created, theres still a long way to go to surpass the mere logisticsf data, as important as this function may be.

Systems (see [10,13]).

284 P.L. Reichertz

Fig. 2 Global tasks of information systems.

Fig. 2 attempts a more global approach indescribing tasks of information systems:

(A) The most primitive, though necessary functionis that of retrieval of information as it has beenstored. This is the function of system descrip-tion.

(B) If information is managed properly, it can bemade available, whenever it is needed, whereit is needed and in what presentation it isneeded: the function of information logisticsserves the system management.

(C) Future aspects go beyond system managementto problem solving. This function of system con-trol requires the further development of tech-niques just being experimented with. Expertsystems and methods of artificial intelligence,hopefully combined with advanced statisticalmethods, have the potential to contribute tothe solution of actual problems in the systemenvironment. Here lies the future of informa-tion systems and there are great possibilitiesand needs within the hospital environment.

of resources in laboratories and ancillary depart-ments.

So far, the main thrust of hospital informationsystems has been lying in the area of optimiza-tion of hospital operations as such. Much remainsto be done in optimizing care for the patient. Ulti-mately, this might also contribute to the efficiencyof the hospital operations as such. But, this mayremain wishful thinking. One has to face the prob-lem that care optimization may generate additionalcost. Hopefully information systems within the hos-pital may help to reconcile the conflicting goals.

In the 70’s, systems concentrated more on thefiscal operations of a hospital and the adminis-trative aspects. Gradually, interest is focussing onpatient care, resp. the medical and nursing aspects(for the Federal Republic of Germany see [5,16]).

One aspect of providing better care for thepatient is that of integrating all the informationpertaining to his status and originating from variousplaces and sources within the hospital. Interestinglyenough, these integrating aspects could already befound in very early attempts, which failed due toinsufficient hardware and software technology ofthe late sixties and early seventies (e.g. [4]).

3. What has determined thedevelopment of systems in the hospital?

One has to bear in mind that hospital operationsfollow conflicting goals in terms of optimal use ofresources and functions [9].

The system hospital has to be optimized asan ‘industrial enterprise’, minimizing cost resp.increasing efficiency.

On the other side, the system patient requiresoptimal attention and expects the use of all pos-sible resources and highest technology in order toachieve the best results in re-establishing healthor minimizing disease. Clearly, this does not neces-sarily go together with minimizing costs of hospitaloperations. Also time aspects of serving the ‘sys-tem patient’ do not necessarily coincide with plan-ning aspects of hospital personnel and allocation

Because if one examines how informatics tech-nology ‘penetrated’ into the hospital, it can be seenthat aspects of ‘integrated files’ [4,6] and logisticsof patient care information was one of the majordesign criteria of early developments. What pre-vailed in the seventies however was providing sup-port for administration on one side, or a specificlaboratory operation on the other (see Fig. 3).

In 1978 [11] (also see [12,14]), I described theaspects of an actual information system within a

Fig. 3 The penetration of informatics technology[11,14].


specific health care environment to have resultedfrom three major hierarchies:

• The design objectives as the hierarchy deter-mined by objective criteria and rational argu-ments for the construction of such systems beingobjectives of:- health care planning according to the local and

regional requirements;- specific medical tasks, clinical trials and inves-

tigations;- patient care management and its improvement

and- last, but very often not least, hospital manage-

ment and optimization of hospital operations.• The design ‘drives’ being the ‘hidden’ forces

behind the objective criteria and very often pro-viding the motivational aspects and emotionalengagement for the pursuit of a certain way resp.systems aspect. These include:- the systems hierarchy resp. the attempt to

achieve a certain structure;- the technology hierarchy resp. the influence

of certain technology aspects and the belief insuch;

• Finally the design environment naturally has hadstrong influence on the actual development, beit through:- the aspects of finance and funding;- the structure of, e.g. the actual health delivery

system;- the political environment with its priorities and

preferences and- last, but not least, industry with its various

influence zones, reaching from financial oversystems to political aspects.

4. The present state

Accordingly, hospital information systems dis-play different aspects according to the localand regional health care environment. Nonethe-less, typical features have evolved, being domi-nated by the dualism of hospital goals namely toserve:

• the hospital management in order to optimizeoperations and

•

fi

tem

- the influence of the sociological hierarchy, e.g.the relationship between physicians and thenursing staff and

- the utility hierarchy, e.g. the value placed uponthe utilities as they are to be provided by thesystem (Fig. 4).

Fig. 4 The major hierarchies of sys

the patient resp. to provide better possibilitiesfor patient care and its management.

Fig. 5 tries to extract a conceptional schemarom the various types of applications and theirnter-relation (compare also [19]).

development (taken from [11—14]).

286 P.L. Reichertz

Fig. 5 Conceptual model of hospital information systems.

4.1. The core of hospital informationsystems

The center of information systems and the levels, asthey developed both in pioneer installations and inthe gradually evolving industrial software, is a cen-tral data structure and a means for communication.The patient ‘enters’ the system through the admis-sion, transfer and discharge functions of the coreand leaves the system, at least partially, throughthis main port.

Some solutions have opted for a more distributedconstruction of data bases; nonetheless centralordering principles have to be kept to achieve thenecessary integration of information and the distri-bution to the various points where it is needed, beit in the area of hospital management or in the fieldof care provision (Fig. 6).

This central data base is serving the central oper-ational purposes of the hospital in the context ofits dual goals. The data base structures may bestraightforward, or more complex in a data bank,which certainly is to be preferred because of thevarious requirements and access procedures resp.combinations. Also a chronological hierarchy may

Preferably, hospitals have only one communica-tion system. In history however, it has been nec-essary to keep old communication systems and tocombine them with newer ones. This applies moreto installations that have had to ‘marry’ olderapproaches with newer ones. Industrial systems, asbeing offered, certainly will be based on only onedata base manager and communication system.

4.2. The horizontal service layers

The horizontal service layers (see also [19]) spandepartments both in patient care and hospital man-agement and provide the means for daily opera-tions. They have a broad interface with the core,both in regard to communication and to the database structures.

Consequently, the data exchange between thesesystems and the core is high in volume and in fre-quency.

According to the dualistic structure of hospi-tal information systems, the service layers can befound both in hospital management and patientmanagement. Fig. 7 gives a prototype example forthe possible modules in hospital management. Thelte

have to be developed in regard to mass storageconcerning archiving of data, abstracting and main-taining central summaries.

ist of modules and functions is neither exhaus-ive nor absolute; in reality more functions or oth-rs may be present according to the actual health


Fig. 6 Core of hospital information systems.

Fig. 7 Horizontal service layer of hospital management.

care environment. One application may span differ-ent functions and vice versa the modules may bebroken down according to more subfunctions. Typ-ically, however, these systems require additionalservice data structures, according to their own con-straints to function properly and efficiently. Veryoften, patients carry more than one identificationnumber, e.g. a medical record unit number, or, as inmany instances in the Federal Republic of Germany,a personal data related ID-number and additionalaccount, claim or policy numbers for a stay withinthe hospital, e.g. to associate the data with a cer-tain billing procedure. However, also here the broadcontact with the central communication and database system is necessary and exercised.

Unfortunately, according to the specific tasksof administration, there is a strong tendencyto separate these service layers from the com-mon approach within hospitals. This is not onlydeplorable regarding the necessary additionalresources and redundant storage, but an integratedview is only possible if administrative data arederived from medical actions. Vice versa medicalactions can be reviewed and correlated to admin-istrative functions and consequences. Despite thedhmcct

ualistic aspect of the operations in hospitals, theospital as a whole has a common goal to giveaximum service to the patient with optimal pro-

edures, while maintaining high quality of medi-al care. This common goal can only be reachedhrough a common approach.

288 P.L. Reichertz

Fulfilling a basic prerequisite for modernaccounting procedures, systems addressing them-selves predominantly to these functions weredeveloped by industry in the sixties and found theirway into the hospital in the industrialized coun-tries. For the Federal Republic, e.g. a recent surveyrevealed that approximately 85% of all hospitalshave contact with edp-systems in one way or theother [5,16]. Of those hospitals having more than250 beds this figure can be specified with 97%.Mostly these applications lie in the area of busi-ness and management. Thus, it is naturally, thatmore comprehensive systems start from here pro-viding data entry for business administration, plan-ning and other chores of management.

Fig. 8 is the analogy to Fig. 6 for the ‘patient careworld’ of the horizontal service layer. The same asfor Fig. 7 has to be said here in regard to compre-hensiveness and grouping of the systems described.Typical for these systems is, like for hospital man-agement, that they span services and departmentsand provide a general tool for the scheduling and

processing of patient care. Typical is also, that mostof the procedures planned or supported are able toprovide basic information for cost accounting andbilling, when they are properly constructed and thenecessary information is captured and transferredto the service layer of hospital management. Thismakes the integrative aspects obvious.

More sophisticated approaches may, e.g. com-bine reporting systems (for instance like in radi-ology) with abstracting and coding systems forthe further classification of the care procedure orthe diagnostic-therapeutic classification or process.The same holds true for surgical pathology and his-tology.

As indicated in Fig. 5, the horizontal layerscommunicate broadly with the central data base.Nonetheless, own storage structures are developedfor performance, safety and optimal storage. Inprinciple, the functions described may be imple-mented in separate hardware structures, providedthat resp. software guarantees the necessary inte-gration and communication.

Fig. 8 The horizontal layer of pat
ient management objectives.


Fig. 9 Vertical layer of departmental systems in patient care.

4.3. The vertical departmental layers

The vertical departmental layers (see also [19]) inFig. 5 support the functions of a department or asubsystem with the hospital. In hospital manage-ment, these systems process accumulated patientdata, communicate with the outside in regard tosupply and finance etc.

In patient care, this vertical layer serves themedical departments and services as described inFig. 9.

It is typical, that here the necessity for owndata structures increases resp. the exchange withthe central data base of the core decreases, resp.develops towards defined interfaces and parame-ters.

4.4. The level of independent systems

Like in the universe, far-away from the core, inde-pendent structures develop like own galaxies. Fig. 5describes these as patient proximity systems forpatient management and as office systems for the‘world’ of hospital management. These applica-tions have increased dramatically with the adventof PCs. In principle, a communication with the coreis not necessary, though communication may beadvantageous.

Such ‘satellites’ or ‘new galaxies’ are describedin Fig. 10. The development in the future will liein the area of expert and pictorial systems. Butstill great research and development is necessarybefore they will become tools for routine use (seealso [15]).

5. Principles and problems ofcommunication

This development of structural layers serves thenecessity to provide adequate utility to the vari-ous structures and objectives. However, the morecpTiwnipd

shca

omplex the system grows, the smaller will be theossibility for changes and dynamic development.his is especially true for the necessary integrat-

ng aspects. Communication becomes a problem,hen information has to be presented where it iseeded and when it is needed. Synchronization andntegrity of the various data pools is necessary forroper functioning and in order to avoid repetitiveata capture and update.

Fig. 11 gives a general model of the relation-hip between complexity and flexibility. Flexibility,owever, is also required in regard to the necessaryhanges resulting from the advances in medicinend the social changes reflected in new proce-
Fig. 10 Systems of the patient proximity level.

290 P.L. Reichertz

Fig. 11 The relationship between complexity and flexi-bility.

dures for accounting, costing, restrictions and otherevents of influence upon the health care deliverysystem.

Flexibility is also necessary in adapting hospitalinformation systems to local environments. It can-not be expected, and this is one of the lessons ofthe past, that one system can be transferred toanother location without a pre-analytic phase inorder to define the requirements of the new siteand an adaptive phase where these requirementsare translated into the mechanisms of the system.The otherwise necessary mirror-like copy of organ-isational and internal structures is only possible ina very few cases. Consequently, systems have to beready for changes and this preferably without com-promising their principles of architecture.

Communication is necessary between the vari-ous layers within the systems. When only financialsystems are created and in operation, e.g. it is notnecessary to leave the system in order to go to itsvarious branches of operation. The same is true fororder entry systems without communication to lab-oratory systems or administrative systems or thosein the patients’ proximity.

The communication within a system in its various

Fig. 12 Principle of communication: ants type.

low the human mind in addressing various issues,like those for order entry and switching over toreceiving and interpreting results and finally, e.g.to address the question of patient billing.

The thoughts and issues therefore have to jumpfrom one system aspect to a completely new struc-ture like Tarzan swings from one tree to anotheruntil he reaches that particular system where hewants to dwell or follow a particular purpose.Fig. 13 illustrates this principle of inter-structuralcommunication.

But Tarzan, if we want to stay with this allegory,may jump from one tree to a very different otherone, he still remains within a common environmentof trees of a tropical nature. The human mind,on the other side, is not only able to jump fromone particular aspect of an application to anotherwithin the same overall environment, but is ableto leave one system completely to enter anotherone with completely different rules of weightingand reasoning.

Future systems have therefore to provide thisintersystem type of communication, like travellingfrom one galaxy to another in a sort of ‘hyper-jump’as described in science fiction (Fig. 14).

branches goes up and down those branches like antscrawling up and down in a tree and looking aftertheir different chores while exchanging experienceon their way. This type of communication, the antstype, is the intra-structural communication withinsystems for a single purpose (see Fig. 12).

The past has shown that such an intra-structuralcommunication does not suffice for hospital infor-mation systems. The system aspects have to fol-


Fig. 13 Principle of communication: Tarzan type.

Expert systems in medicine indicate such achange between fields of reasoning. In one context,symptoms are weighted in a particular way, butwith changing fields of interest, these weights haveto change completely and systems have to adaptthemselves to these different views. Human mindscan do this very easily. Computer systems haveonly started to learn this in very crude approaches.

Mostly their process of reasoning crawls up anddown a tree following the principles of ants. Theresulting slowness and awkwardness of the processexplains why most expert systems, e.g. have notyet gained routine usage in practical medicine.

Surely, this intragalactic type of communicationis the domain of research and future application.But very certainly hospital information systems

Fig. 14 Principle of communic
ation. Intra-galactic type.

292 P.L. Reichertz

even today need the Tarzan type of communicationand this is possible through a common architec-ture and modern technology. Carrier systems whichinstall different applications within the same struc-ture have the technological requirements for suchan approach. Another approach is the developmentof networks in hospitals with connections betweenthe different layers as shown in Fig. 5.

But technical means alone do not suffice. Infor-mation in the context of a hospital information sys-tem has to serve very different purposes, has to behanded from tree to tree in Tarzan’s type of commu-nication and even has to be sent from one galaxy toanother. One principle to deal with these require-ments is that of twin packaging, preferably withabstraction.

Fig. 15 tries to illustrate what is meant by thisprinciple. Mostly certain patient data in a networkhave to flow to the address of a certain depart-ment or service. Here the data in detail are neededpertaining to a certain patient. However, both forthis department and for the communication withother services, an abstraction of the data is needed.In radiology, e.g. the full description of the X-rayimage is necessary for the process of reading and

Fig. 15 Principle of twin-packaging with abstraction.

cate the situation between systems, e.g. those ofinternal medicine to those of surgery (Fig. 15).

It really does not matter, whether the communi-cation process and the integration are based on onecentral system or on the combination of central sys-tems and peripheral storages, as indicated in Fig. 5.The principle of twin-packaging with abstractingallows for a central version as well as a peripheralone.

Fig. 16 shows a central version of such integra-tion. The abstracts are sent to a central synthesissystem, where the information is available both tothe peripheral system as such as well as to othersystems for their particular purpose. Fig. 17 showsa peripheral version of information synthesis. Herethe abstracts or communication parts of the resp.information elements are stored within the periph-eral systems and are made available to a centralsynthesis system for further reference.

It is necessary to define these processes of infor-mation exchange when networks are intended tomove from a mere combination of informationsources with information sinks to an integratinginstrument in patient care and hospital manage-ment. The solution lies neither in the develop-m

is wi

finding the diagnosis. In the patient summary, how-ever, only the radiological diagnosis is needed resp.the final evaluation. Such is the case also for anexpert system, which has to develop a final andcomprehensive diagnosis. For an administrative sys-tem it is only necessary to know that a certain typeof examination has been performed.

Abstracting and providing information for a com-prehensive view is mainly necessary in the medicalrealm of hospital information systems or the upperside of Fig. 5. But also abstracting is necessaryfor financial evaluation and overall planning resp.cost accounting. In medical systems, however it willbecome of particular interest if observations of alongitudinal matter are to be made available forthe reasoning process. This means the condensationof observations of one episode to make it availableas information for subsequent ones or to communi-

Fig. 16 Central version of synthes

ent of monopolistic or monolithic great multi-

th the principle of twin-packaging.


Fig. 17 Peripheral version of integration according to the principle of twin-packaging.

purpose central systems providing the integrationdue to comprehensive databases nor in the desin-tegration into single purpose dedicated systems,but in the development of multi-purpose networkssystems.

Fig. 18 shows the resulting principle as describedearlier [11—14]. Resulting from these require-ments, the core of hospital information systems hasto be broader than described in Fig. 6.

Fig. 19 consequently shows the updated aspectsof the advanced core in hospital information sys-tems. Besides the basic components shown in Fig. 6here the communication system is expanded by net-work aspects and the functions of a central inquirysystem is added, be it in the peripheral or centralform of realization.

The importance of architecture has already beenstressed several times. Architecture in this con-text means the ‘art of construction’, the philosophybehind an approach including also aspects of userpsychology and system ecology, which is meant asthe eventual new biotope developing out of theinteraction between information systems, organi-zational structures and the human beings who liveand work in this new environment. Systems are

Fig. 18 Multi-purpose and single-purpose elements of asystem for a multi-purpose network system.

changed by this system’s ecology, human behaviouradapts to systems aspects and develops new abili-ties to interact with the system.

The resulting consequence is the necessity toprovide means for adaptations of functions of thesystem to the new requirements of the developingsystems ecology. ‘Hard-programming’ is no longersufficient to cope with these requirements. ‘Car-rier systems’ have been described as to be able toprovide the tools for change to meet these require-ments (see [11—14]).

Fig. 20 describes the principle of such carrier sys-tems. The system logic is no longer within the pro-gram itself, but is separated from the program andis defined in a formalized structure kept separatefrom the application logic as such. Modern hospitalinformation systems follow this philosophy of car-rier systems: they provide for a separate definitionof the logic of screen events, printing and processusing high level approaches for defining the ele-ments and steps of screen definition, informationdisplay, storage and regrouping. These approachesattempt to make this technique independent fromcertain machines or, in other words, to make thesystem runnable also on smaller hardware, so thattstsd

6

IIwobidt

he logic may be developed or used on differentites and on different machines (an own attempto transfer the software developed within the ownystem onto System/1 and smaller types had to beiscontinued because of lack of support).

. Architecture and user psychology

nformation systems are not placed into the void.t has been stressed several times that they reactith their environment and form a new user ecol-gy. This development, however, has to be plannedy architectural approaches. Such an approachncludes the recognition and implementation of theevelopment of user psychology and behaviour intohe constructional process. Organizational and psy-

294 P.L. Reichertz

Fig. 19 Advanced integrating core of hospital information systems.

Fig. 20 The principle of carrier systems (taken from [20]).


Fig. 21 Phases of user reaction during planning, imple-mentation and routine use (see [8,17,18]).

chological obstacles are considered to be the mostimportant ones on the way towards the successfulimplementation of information systems in the hos-pital (see [3,8,17]).

A time dependent development of attitude canbe seen during almost every introduction of a sys-tem into an environment, which as such has notdirectly been involved into the acquisition. Even ifsuch an involvement has been the case, changesbetween insecurity, hypersensitivity and realisticapproach to pragmatism characterizes the devel-opment from emotional to rational attitude. Fig. 21shows such a time-attitude matrix, broken down inphases and components.

This stresses the necessity of coping with userpsychology in the implementation of systems in hos-pitals. Not only the system has to be taught assuch but one also has to prepare for the reactionwith the system. This stresses also the necessityof a pre-introductory phase which is not only nec-essary to prepare the system for the introductioninto the new environment according to its par-ticular aspects, but also to prepare the environ-ment for the system, both in organizational andpsychological-sociological aspects.

qthmittibAptptae

7. How will new technologies changehospital information systems?

The major changes due to new technologies will becaused by the rapid deployment of more and morepowerful small computers. The same computerpower can be bought today at a price of around1/150 of the price 12—14 years ago. This certainlyenhances the possibility of creating independentsystems in the satellite layer (see Fig. 5), bothon the side of hospital administration and patientcare. This enhances also the production of problem-oriented software. The more people have access tothese tools, the more hidden talents will be wak-ened and tapped for the production of software tosolve immediate problems. Some of this softwarewill be ready for generalization and general distri-bution. Only a very small part will meet the require-ments for routine usage and exportation to othersites. But since more attempts will be made, theresulting number of available software will increasedramatically. This leads to the phenomenon of an‘exploding universe’, with the various new galac-tic systems moving more and more away from acentral hierarchical system, but also from centraliithaai

ceiotuacwo

ttt

diutavnr

For the immediate future, this has the conse-uence that appropriate installation means haveo be available when a major change is made in aospital which, e.g. is the change from a completeanual and traditional system to a fully computer-

zed type of operation. Whether such a technologyransfer team is formed by the hospital itself underhe leadership of medical informatics, is broughtn from an independent consulting firm or providedy the vendor, is basically of no major concern.ll varieties and combinations are possible. And inreparing the transition, one has to be remindedhat major changes sometimes are easier to accom-lish than subtle differences and adaptations inhe daily way of life. And these types of changesre the gist of a transition into a new systemscology.

ntegrating forces. The tasks of medical informat-cs therefore will move away from the construc-ion of individual systems towards more compre-ensive system analysis, construction of networksnd devising schemes for twin-packaging and hier-rchies of information to provide exchange andntegration.

With new developing technologies such as opti-al disks, higher storage capacity, fiber glass optics,tc. there will be many incremental improvementsn storage, transmission and speed. However, veryften a technology is hailed as a major break-hrough, but still does not find its way into practicalse. An example for this is, e.g. the optical disklready mentioned. But then, after some time ofrawling around, the new technology will attackith a big jump. Where and when it happens is veryften difficult, if not impossible to predict.

Fig. 22 gives a few highlights of new trends andechnologies as they emerge at the moment and ashey get ready to conquer the field of new applica-ions in the hospital environment.

Pictorial data processing systems are underevelopment, combining multiple sources of (dig-tized) data into single display and manipulationnits. They start emerging in radiology for digi-al X-ray pictures, computer tomograms, NMR-scansnd sonograms. The technology of today is alreadyery advanced with high resolution and good tech-ical quality. The work stations being developedight now try to mimic the possibilities of viewboxes

296 P.L. Reichertz

Fig. 22 Highlights of new technology.

with overlays, comparisons, zoom-effects and def-initions of areas of interest. But also new fea-tures appear: the processing of information intoa 3D-synthesis. It does not seem to be only thestereoscopic views that cue a 3D-impression: morepowerful seem to be the depth perception due tomovements of the head (and the position of theviewer) and the resulting changes in perspectives.Laboratory devices try to simulate such changes inperspectives: sensing devices at the head of theviewer indicate the movement of the head and leadto resulting changes of the picture, vibrating mir-rors create depth perception and rotating diodesgive an almost eary picture of the image to bepresented. But also good stereoscopic and perspec-tive presentation of skulls and skeletal images ofCT-scans give almost unbelievable clear pictures ofimpression fractions, distortions, etc. which so far,were indicated only by additive linear silhouettesin X-ray projections.

Initially these pictorial data systems will be onlyavailable within specialized departments. Later on,however, a linkage over a hospital network maymake it possible to combine these technologies withdigital information systems giving the physician on

disk may serve as convenient storage medium in theevolution to replace conventional archives. Ideally,also these networks should be linked together withdigital networks, both for information processingand display as well as, in the future, transmissionof pictorial data.

These developments necessitate also the emer-gence of new types of work stations, which makethe access possible to the different informationsystems of digital, pictorial or video-text nature.Overlays have to be possible, windows into vari-ous systems, paging backwards and forwards andsideways with 3D-manipulation and other means ofpresenting data in new ways and flexible forms.These work stations have to provide access to localcomputer power for dedicated tasks like word pro-cessing and personal computing, but also have togive access to more advanced statistical software,e.g. graphical systems and other services offeredby central computing facilities.

The last item given in Fig. 22 is conceptually ofgreat importance, though the realisation to a usabletool will be very difficult and requires scientificanalysis and development efforts. Nonetheless, thelinkage of a data to counselling and expert systemsh‘[

oamnttvrI

the wards the possibility to look at the presenta-tions during his work on the patient, in his officeor whenever this information is needed. In someway, these technologies will eliminate the needfor specialists that have learned to interpret two-dimensional projective X-rays during a long time ofresidency and subsequent experience.

Videotex applications may be soon around thecorner. The experiences gained now in commercialsystems may make it possible to develop in-housenetworks for storage of static data, such as oldpatient records and other data, both of adminis-trative and patient care origin. Here the optical

as already been attacked in some instances. TheHELP’ system may be considered as a prototype18].

In expert systems in the future, a synthesis willccur between the old probabilistic and algorithmicpproaches and the new ways of using inferenceachines, rule based approaches and other tech-

iques formalizing the experience and approach ofhe human expert. Research is necessary in ordero investigate the determining and discriminatingariables, the interrelation between causes and theesulting changes and further factors of influence.t cannot be assumed that all this can be verbal-


Fig. 23 The binding of resources by increasing mainte-nance.

ized by experts out of their experience. But thecombination of analytical research and formaliz-ing empirical experience will lead to better resultsthan using only one of these great reservoirs of cog-nition. Advantages can be expected, when routinedata bases can be linked with expert systems bothfor the exploration of facts, generation of hypothe-ses and the application of decision rules onto dataobtained by routine examinations and specific diag-nostic procedures. Here lies a wide field for creativeresearch in the field of Medical Informatics and Clin-ical Medicine.

8. Who carries the development?

In the past, with ‘straightforward systems’, manydevelopmental efforts were found in larger univer-sity and care providing hospitals. Industry madeseveral approaches which were unsuccessful, aslong as not adaptable carrier systems were avail-able.

Fig. 23 describes how here the maintenanceproblem by accumulating residual maintenancesgradually devoured all developmental resources.

otpitpi

aet

Fig. 24 The different forms of development.

ments and certainly some of the prototypes havebeen taken by existing industry in an attemptto market hospital information systems. Some ofthese attempts have been successful, others havefailed. Naturally out of old industries developednew industries, either continuing to exist or join-ing with existing industries to new ventures. Butalso out of hospitals and universities cooperativeshave resulted, still functioning as cooperatives andsometimes developing into new industries. Themain thrust now goes from existing industry andcooperatives to hospitals and the routine applica-tion in these hospitals takes strong influence onthe systems being developed and maintained forthe market. Existing industries still have recourseto universities for new ideas and developments.Also a shift of interest has occurred from largeruniversity hospitals to common hospitals and thiswill increase with the resulting demand for moreroutine-oriented and ‘of-the-shelf’ systems which,nonetheless, will need adaptation and introduc-tion.

9. Concluding remarks

Tihbodah

msi

One way to cope with the problem was the devel-pment of carrier systems, centralizing the main-enance tasks for systems aspects and making itossible to spread out development over severalnstallations. The industrial distribution of a sys-em with the resulting spreading of error discovery,ressure for adaptation and maintenance is anothermportant factor.

Fig. 24 shows what has happened in the pastnd is happening at present. Whereas most of thearlier developments started in universities, indus-ry has taken an early influence on these develop-

his analysis has looked at various points of interestn describing the past, present and the future ofospital development. It does not consider itself toe comprehensive. On the contrary, it is selectiven a purpose. Nonetheless, a few traces could berawn from the projection of past developmentsnd the present state into the future, attempting toighlight some requirements and new possibilities.

When looking again at the global tasks of infor-ation systems as presented in Fig. 2 and if con-

equently one tries to reflect the role of hospitalnformation systems in fulfilling the tasks of a hos-

298 P.L. Reichertz

Fig. 25 Pyramid of hospital objectives.

pital, one is reminded that the goals of hospitalsare dualistic in nature.

The system ‘patient’, be it in terms of the indi-vidual patient or a population, is the primary targetof the delivery of health care. Efforts here maynot be coincidental with the goal of running thehospital efficiently in terms of hospital economyand employees satisfaction. Professional recogni-tion somehow centers between both goals.

So far, the development of hospital informationsystems has started at the zero point of Fig. 25 andhas moved along the axes of hospital management.It then has turned and is nowadays gradually mov-ing towards patient management in order to maxi-mize both objectives. Necessary in the future is themovement up to the top of the pyramid towardsquality of care.

Certainly quality of care does not automaticallyoptimize hospital management in terms of econ-omy. Hospitals have to have a broader goal in mindthan only their own businesslike operations.

Future hospital information systems will haveto concentrate more and more on the quality ofmedicine. Systems will increase in the area ofpatient management, but also of direct problem

Fig. 26 Historical development of hospital systems.

of architectural hierarchies and the developmenttowards the objectives of solving problems for thecomplex and sometimes antagonistic goal structureof hospitals.

The development of hospital information sys-tems therefore should be considered to be ‘socialengineering’ as well as the pursuit of scientificobjectives in medicine for the increase of qual-ity of medical care. This requires comprehensiveresearch and the cooperation of clinical medicine,medical informatics, industry and administration.The development certainly is still quite at thebeginning, not yet approaching its end of possi-bilities. However, because it is social engineer-ing in changing systems ecology, it is a fieldwhich deserves careful attention and concern of allinvolved in order to create a useful future.

References

[1] Ball, Marion J Medical Information systems in the USA,1980, in: F. Gremy, P. Degoulet, B. Barber, R. Salamon(Eds.), Medical Informatics Europe 81. Third Congress of

management with the help of pictorial data andexpert systems.

Fig. 26 tries to summarize the development sofar from a systems point of view. We presentlyare in the transition between integration phase Iand the proliferation. Integration phase II must beattempted next and some tentacles of the presentdevelopment point into this direction. The prolifer-ation, however, can lead to a galactic isolation. Thishas to be avoided by controlled structures in terms


the European Federation for Medical Informatcis, Proceed-ings, Toulouse/France, March 9—13, 1981. Series: LectureNotes in Medical Informatics, vol. 11, Springer, Berlin, Hei-delberg, New York, 1981, pp. 22—32.

[3] P. Koeppe, Die Bedeutung der Manware beim Aufbau vonKrankenhausinformations-systemen, in: P.L. Reichertz, G.Holthoff (Eds.), Methoden der Informatik in der Medizin,Springer, Berlin, Heidelberg, New York, 1975, pp. 25—32.

[4] D.A.B. Lindberg, in: Charles C. Thomas (Ed.), The computerand medical care, Springfield, 1968.

[5] W. Lordieck, P.L. Reichertz, Die EDV in Krankenhausern derBundesrepublik Deutschland. Das Ergebnis einer Umfrage.Reihe: Medizinische Informatik und Statistik. S. Koller, P.L.Reichertz, K. Uberla (Eds.), Bd. 45, Springer-Verlag, Berlin,Heidelberg, New York, Tokio, 1983.

[6] P.L. Reichertz, K. Sauter, J. Mohr, B. Kreslak, W. Zowe,Konzeptioneller Aufbau eines integrierten Patientenfiles,in: G. Fuchs, G. Wagner (Eds.), Krankenhausinformation-ssystemen, Schattauer, Stuttgart, 1972, pp. 73—81.

[8] P.L. Reichertz, Informationssysteme in der Medizin, IBM-Verlag, Bonn-Bad Godesberg, 1975.

[9] P.L. Reichertz, Hospitals and health care, IFAC 75,in: Proceedings of the Sixth Triennial World Congressof the International Federation of Automatic Control,Boston/Mass/USA, Aug. 20—30, 1975, Instrument Society ofAmerica, Pittsburgh, 1975, pp. 81—23.

[10] P.L. Reichertz, Sinn und Kriterien einer zeitgerechtenDatenverarbeitung im Dialog, in: G. Wagner, C.O. Kohler(Eds.), Interaktive Datenverarbeitung in der Medizin. Verh.

[

[

[13] P.L. Reichertz, Structure and content of Information Sys-tems in the Hospital environment, in: R.H. Shannon (Ed.),Hospital Information systems. IFIP-working Conference,Cape Town: April 2—6 1979, North-Holland, Amsterdam,1979, pp. 83—98.

[14] P.L. Reichertz, Forecast of hospital information systems,Maydaon 4 (1983) 3—11.

[15] P.L. Reichertz, A Physician’s View of Medical Informatics,in: IEEE Computer Society, Technical Committee on Com-putational Medicine: MEDCOMP’82, IEEE Catalog No. 82CH1797-0, Los Angeles, 1982, pp. 118—127.

[16] P.L. Reichertz, W. Lordieck, EDV-Einsatz in Krankenhausern:Ergebnis einer Umfrage. Das Krankenhaus 76 (1984/9) 390-395, (1984/10) 443—449.

[17] O. Rienhoff, P.L. Reichertz, User Psychology: or Who isUsing Whom in Medical Informatics, in: D.B. Shires, H. Wolf(Eds.), MEDINFO’77, North-Holland, Amsterdam, 1977, pp.893—987.

[18] H.B. Warner, T.A. Pryor, S. Clark, J. Morgan, Integration ofComputer Support for Institutional Practice: The Help Sys-tem, in: C. Weller (Ed.), Computer Applications in HealthCare Delivery, New York, 1976, pp. 121—133.

[19] B.G. Winfree, The Logistics of Patient Information, in: S.Stipa, A. Cavallara, C. Medini (Eds.), Informatics and Bio-engineering in Medicine, Serono Symposia Review No. 5,Ares Serono Symposia, Rome, 1984, pp. 351—354.

[20] E. Wolters, P.L. Reichertz, Problem directed interactivetransaction management in medical systems, Meth. Inform.Med. 15 (1976) 135—140.

20, Jahrestagung der Deutschen Gesellschaft fur Med.Dokumentation. Informatik u. Statistik, Schattauer Verlag,Stuttgart 1976, pp. 21—39.

11] P.L. Reichertz, Forecast of Hospital Information Systems,in: H. Abe, H. Inada, et al. (Eds.), Proceedings of MEDIS’79,International symposium on Medical Information Systems,Osaka/Japan, 2—6, 1978, The Medical Information SystemDevelopment Center and the Kansai Institut of InformationSystems, Tokyo/Osaka, 1979, pp. 157—168.

12] P.L. Reichertz, Medical Informatics—–fiction or reality?

Further reading

[2] C.Th. Ehlers, R. Klar, Future Trends in Hospital InformationSystems, in: O. Fokkens (Ed.), MEDINFO’83 Seminars, North-Holland, Amsterdam, New York, Oxford, 1973, pp. 112—116.

[7] P.L. Reichertz, Medical School of Hannover Hosptial Com-puter System (Hannover), in: M.F. Collen (Ed.), Hospital Com-puter system, Wiley, New York, 1974, pp. 598—661.

Documents

Hospital information systems–Past,Present, Future