Medical Informatics

Introduction
Medical informatics is as much about computers as cardiology is about stethoscopes. For those who have studied the application of information technologies in medicine, the last decade has delivered one unassailable lesson. Any attempt to use information technology will fail dramatically when the motivation is the application of technology for its own sake rather than the solution of clinical problems (Coiera !""#$ %yatt !""#$van der &ei$ !""#'. (he role of the information sciences in medicine continues to grow, and the last few years have seen informatics begin to move into the mainstream of clinical practice. (he scope of this field is however enormous ) informatics finds application in the design of decision support systems for practitioners (e.g. Miller, !""#', in the development of computer tools for research (e.g, *unter, !""+', and in the study of the very essence of medicine) its corpus of knowledge (e.g. ,eravnou, !""-'. It is likely that the study of informatics in the ne.t century will become as fundamental to the practice of medicine as the study of anatomy has been to the present. It is with two seemingly contradictory themes in mind / apparently unbridled technological 0romise against less than satisfying practical achievement ) that recent advances will be [Link]. %ork will be assessed against three criteria ) its possibility, its practicability, and its desirability. 0ossibility reflects the science of information ) what in theory can be achieved1 0racticability addresses the potential for successfully engineering a system) what can actually be built given the constraints of the real world1 2esirability looks at the fundamental motivation for using a given technology. (hese criteria are suggested because we need to evolve a framework to 3udge the claims made for these new technolowes, and those who seek to profit from them. 4ust as there is a long)standing symbiosis between the pharmaceutical industry and medicine, there is a newer and conse5uently less [Link] relationship between medicine and the computing and telecommunication industries. Clinicians should try to 3udge the claims of these newcomers in the same cautious way that they would [Link] claims about a new drug (%yatt, !"67'. 0erhaps more so, given that clinicians are far more knowledgeable about pharmacology than they are about informatics and telecommunications. (he first part of the article reviews a variety of activities that collect under the telemedicine banner. 8ince this is a new area, ma3or research themes are only appearing now. 9e.t, protocol based decision support systems are discussed. (hese may be the first substantive clinical information systems to appear in routine clinical practice. Finally, the current state of clinical coding is [Link]. (he terminology and coding enterprise represents the first ma3or attempt to uniformly describe the structure, content and nature of medical knowledge.

Telemedicine 2efinitions of telemedicine abound. (he essence of telemedicine is the [Link] of information at a distance, whether that information is voice, an image, elements of a medical record, or commands to a surgical robot. It seems reasonable to think of telemedicine as the communication of information to facilitate clinical care. And it is not a new enterprise)
:

:inthoven [Link] with telephone transmissions using his new invention, the electrocardiograph at the beginning of the century (9ymo, !""#'. At its inception, telemedicine was essentially about providing communication links between medical [Link] and remote locations. It is now clear that the healthcare system suffers enormous inefficiencies because of its poor communication infrastructure and telemedicine is seen as a critical way of reducing that cost. ;ne estimate suggests that the <8 health system could save =+> billion per annum with improved telecommunications (&ittle, !""-'. Conse5uently, telemedicinehas now become a significantarea for researchand development. As one might [Link], the renewed interest in telemedicinealso has muchto do with the [Link] of new technologies. At present the press is

floodedwith articles about the information superhighway, the Internet,and the rapid growth ofmobile telephony. (elemedicine is often presented in the. guise of sophisticated new communications technologyfor specialist activities like teleradiologyand telepathology. (hese are championedby telecommunication companiesbecausethey have the potential to become highly profitable businessesfor them (?owles, !""#'.0erhaps influencedby these forces, much ofthe research in telemedicine is driven by the possibilitiesoftechnologyrather than the needs ofcliniciansand pallients. @et the communications infrastructure used by health care will not need to be special. (he telecommunications market is competitive and the evolving options are numerous. *ealthcareproviderswill be able to utilise the services of cable television, mobile cellular carriers, and [Link],communications technologydoes not need to be sophisticatedto deliver benefit. Appropriateuse oftodayAs telephone can make significantimprovements to e deliveryofcare. For [Link], patient follow up can often be done on the telephone (Bao, ! "#'. Bapid communication ofhospital discharge information using [Link] electronicdata ansfer mechanismsis beneficialfor general practitioners (?ranger, !""-'.(he combin ion ofmobile telephonyand paging systems can reduce the C!> minutes out ofevery hour any clinicians spend answerpagers (FitDpatrick, !""+'. 0erhaps more interestingly, in .pensivevoice messagingsystems can deliver simple but powerful services over [Link] telephone networks. Eoicemailfor [Link],has significant potential for improvingthe 04AF ess of care (Constable, !""#'. &eirer et. al. (!""!' used a voicemail system to automatically phone medication reminders to elderly people at home, and showed that it reduced both tardiness and complete forgetting. As more patients get access to electronic mail, this will offer further avenues for innovative health services. Already in some populations, access to electronic mail is high. Fridsma et. at. (!""#' in California found that #GH oftheir patients at clinic already used email, 6"H ofwhich was through their place of work.
All these data points suggest ~at the potential for the clinical application of communication technologies is indeed great, but e5ually that there is much still to learn. In particular, the relationship between telemedicine and informatics needs to be [Link] in greater detail. Informatics focuses on the use of information and telemedicine on its communication. Although seemingly disparate endeavours they are intimately linked, since the goals of communicating information and deciding on its content cannot be separated (McCarthy and Monk, !""#'. Further, there is little clinical value in information systems built simply to gather data for administrators, forgetting that the essence ofdelivering health care is the communication of information between members ofthe clinical team. Coupled together, the technologies of information and communication can enhance access to information, whether it is stored electronically or is in the possession of a colleague. 8everal key research 5uestions are apparent. Firstly, clinical practice already revolves around communication, often by telephone, and important information [Link] in this way is often lost because it is not documented (8toupa, !"">'. Capturing the informal information currently lost in healthcareAs communication channels may soon become an important issue for those developing the formal electronic patient record. *ow one decides what information is important and how that information is made available are non)trivial 5uestions involving issues of confidentiality, security, as well as the technology of storage and retrieval of voice recordings. 8econdly, our understanding ofthe effects oftechnology on communication is still in its infancy. Besearchers in the field ofhuman)computer interaction feel that before these technologies can be successfully introduced, the way in which individuals communicate needs to be understood(McCarthy, !""#'. In one recent study, the presence ofa computer during doctor)patient consultations had detectable negative effects on the way doctors communicated (Ireatbatch, !""+'. %hile they were at the computer, doctors confimed themselves to short responses to patient 5uestions, delayed responding, glanced at the screen in preference to the patient, or structured the interview around the computer rather than the patient. ;n the positive side, recent [Link] in 9orway have identified benefits to remote telemedical consultation. 8ervices that provided isolated general practitioners with access to specialist [Link] had an [Link] side)effect. (he skill level of practitioners was raised through repeated interactions with specialists and the management of cases that were previously referred (Akelsen& &illehaug, !""+'.(his may arise through the dynamics of the relationship between

remote practitioner and specialist. <nlike most educational settings, both are motivated to form a coach and apprentice relationship for the immediate management of a patient. 0robably the most important issue for research will be to understand the effect of introducing technologies thatiallow asynchronous [Link] present, devices like telephones and pagers interrupt individuals when communication is desired ) these are known as synchronous methods. Th~messages sent across asynchronous systems like electronic mail and voicemail do not n d to be answered immediately and so have the potential to significantly reduce the num . r of interruptions [Link] by clinicians. 8uch messages may nevertheless carry important information. It will becritical to understand how such systems can be designed to ensure that healthcare workers do not miss critical information, and e5ually are not inundatedlwitha flood of irrelevant messages. Finally, along with new communication possibilities, there come new medico)legal implications. In the <nited 8tates the courts have decided that radiologists are negligent if they fail to personally inform clinicians ofa diagnosis. JCommunication of an unusual finding in an K)ray, so that it may be beneficially utilised, is as important as the finding itselfJ. Further, leaving a message with an intermediary is not enough ) Jcertain medical emergencies may re5uire the most direct and immediate response involving personal consultation and [Link] (,line, !""-'.(he fact that such communication re5uirements are beginning to be mandated reflects the communityAs changing perceptions of best medical practice. (he rapid arrival ofAtelemedicine suggests that the healthcare community is beginning to identify the benefits of good clinical communications practice, and realising the costs of poor communication. (he ne.t few years should see the research in telemedicine mature. (he main focus will become the application of communication technologies rather than their development. (his represents the same shift in focus that was re5uired of medical informatics, which initially spent much effort in developing technologies specifically for medicine.

Protocol-based Decision Support Many see the development of protocol)based medicine as the essential cultural change in clinical practice that will pe it the design of useful clinical information systems (2urinck et aI., !""#'. It was rightly seen inappropriate when early computer system designers sought to regularise clinical practice suit the nature of their systems. (he move to evidence based medicine now begins to make it acceptable for clinicians to follow standard assessment and
treatment protocols (Mulrow, !""#'.In this case it is 5uite appropriate for clinicians to use information systems to assist them. (he ultimate goal of a protocol)based decision support system is to provide a set of tools that allow a clinician to access up to date guidelines,and then apply these in the management of their patients. It seems likely that simple protocol systems will appear in clinical practice by the end of the decade (Benaud)8alis, !""#'.In some sense, first generation systems have already appeared, since one can now begin to access treatment guidelines and clinical trial data on the Internet (Ioodlee, !""#'(see te.t bo.'. *owever,evidence suggests that even when guidelines are available, clinicians forget to follow them or deviate from them without clear cause (Benaud)8alis, !""#'. Forgetting preplanned managementtasks seems to be especially likely in high stress clinical decision making situations (Coiera et aI., !""#'.*owever, it probably will be unacceptableto uniformly enforce adherence to guidelines, given the [Link] of individual patient cases. It should be possible however, to make it as easy as possible for clinicians to access them during routine care, making it less likely that steps will be inadvertentlyforgotten or altered. (his will re5uire the design ofmore sophisticatedsystems that will be integrated into the electronic patient record. (hese will not only be repositories for protocols, but will allow them to be manipulatedby clinicians. For [Link], best practice recommendations may need to be customised for local conditions or for individual patients. Further, guidelinesmay

be incorporateddirectly into patient records. As elements ofthe guideline are completed, they could be automatically noted. (he records ofcare generated in this manner might ultimately be used for population)based outcomes analysis. 8ome researchers advocate the use ofcomputerisedprotocols in even more sophisticated settings. ;ne group use a set of ventilation protocols to ad3ust tidal volume and ventilator rate settings for patients with Adult Bespiratory 2istress 8yndrome(AB28'((homsen, !""+'.(hey report using the system for over C>,>>> hours on !C>AB28 patients (Morris et aI., !""#a'. In one trial with !- patients,"#H of #,C+! protocol)generated recommendations were followed by staff. (he survival rate ofAB28patients supported with computerised protocols was four times the [Link] rate from historical controls (Morris et aI, !""#b'. (wo key problems will be faced as such systems become more commonplace.(he first is the arduous but essential collation ofbest)practice guidelineswhich needs to be carried out by bodies like the Cochrane Collaboration(Ioodlee, !""#'. In the absence of such collections, the value ofprotocol systems will be minimal. (he second will be an issue which is at the heart of informatics ) the problem ofdefming, managing and updating medical terminology,

Terminological Systems
Medicalcoding systems like IC2)")CM, lC2)!>,89;M:2 and Beadare becoming increasingly familiar to clinicians. (heir rationaleis as follows. ;nce capturedelectronically clinical data should be availablefor subse5uent aggregation and analysis. *owever, the words usedto describeconditions vary so muchthat simpleanalysis is often not possible. Further,the meaningsattachefl to terms mayvary. Ifthere was an agreed set of terms to describethe process of care en data analysiswould be simplified(Ackerman et aI., !""#'. (he goal of research intom ical terminologies is to arrive at a consensus on the most appropriate set of terms and e way they shouldbe structured. (he fundamental advancein rminological researchover the last year or so is the realisation that the goal of constructing a completeand universal thesaurusof medicalterms is ill)posed. (erminology evolves in a [Link] use, and attemptingto define conte.t independent terminologies is ultimatelyimplausible. Coupledwith this view comesthe pragmatic understanding that a more robust scientificapproach needsto be broughtto the enterpriseof terminology construction. :ach of these issuesdeservesto be [Link] some detail. Universal Terminological systems (he ideal terminological system would be a complete, formal and universal language that allowedall medicalconceptsto be describedand reasoned about. 8omeresearchers have [Link] buildingsuch a singularand JcorrectJ medical language is their goal (Cimino, !""#$ :vans et al., !""#'. (his task emphasises two clear re5uirements the ability for the terminological language to cover all the conceptsthat need to be reasonedabout, and the independence of the terminology from any particularreasoningtask.A further goal occasionally articulatedis that wherethere are alternative terminologies, they must be logically related such that one can be translated into the other. 2espitethe enormoushealth care investment currentlydevoted to achievingthese goals, currentevidence indicates that they are not possible. (here is no pure set of codes or terms that can be universallyapplied in medicine. (hereare two fundamental and related obstacles
i

to devisinga universal termin~logical system.(he first is the modelconstruction problemterminologies are simplyaw, of modellingthe world,and the world is always richer and more [Link] any modelhumanscan devise. (he second is the symbolgrounding problem ) the wordswe use to label ob3ectsdo not necessarily reflect the way we think about the ob3ects, nor do they necessarily reflect defmedob3ects in the real world (9orman, !""+'. (he cumulativeevidence 11 recent thinking in cognitivescience,computerscienceand artificial intelligence provide a formidable set of supporting arguments.

Cognitivestudies of the way ople form categories have shiftedfrom the view that [Link] ob3ectively, to the notionthat conceptsare relative and structuredaround
probabilistic prototypes (Bosch, !"66'. (he 5ualities of prototypical categories are only generally true ofthe [Link] they classify. For [Link], most people would happily say that flight was a property ofbirds, and cope with the fact that some birds are flightless. (he category AbirdA has no pure definition. (he way in which people use family resemblances to create such categories from e~ples remains an area ofresearch (e.g. Aboand Medin,

!""-'.Manyartificial intelli~ce (AI' researchers also contend that there is no ob3ective model ofmedical knowledgel Much ofthis is based on their [Link] in constructing and maintaining knowledge)based systems (Clancy, !""+'. Further, people choose categories at a level ofdescription that is appropriate for thinking about an ob3ect in most situations (Bosch, !"66'. Categories are formed entirely based upon their utility. MedicineAs terminologies have evolved over many years and are also sub3ect to the same process of cognitive evolution. Conse5uently disease entities [Link] for as long as they are useful mental constructs, and are replaced as better concepts emerge ) there is no static body ofmedical knowledge. 9ot only are new concepts added, often the very structure ofmedical knowledge changes as concepts are internally re)organised (Clancy, !""+b' (&aporte, !""C'. IC2)" and !> are substantially different systems, partly because ofthe changes in medicine over the !C year period in which IC2)I> was built (lMI, !""C'. Any attempt at modelling medical knowledge by the imposition of a structure on its terms will thus decay in accuracy over time (*ogarth, !"6G$ (uttle and 9elson !""#'. Conse5uently it does not make sense to think ofterminological systems developing independently ofa conte.t ofuse. :ven those who seek to build a canonical medical terminology are forced to select a clinical application to set a conte.t before they can meaningfully proceed (Friedman et. aI, !""C'. :5ually, there is no reason to [Link] that thete is any uniform mapping between terminological systems developed in different [Link] of use (Ilowinski, !""#' ((uttle, !""#'. :ven when the systems are of similar construction, problems are encountered when one tries to translate knowledge [Link] in one form into another. (he authors ofone study concluded that the sharing knowledge between terminological systems Jdoes not seem to be easily achievableJ (*einsohn, !""#'. Building Maintainabl. and Testable Terminologies %hile coding systems can never be truly canonical, they still provide a practical basis for managing the language ofm icine ) as long as it is understood that they defme a limited and consensual language that will have to be continually modified. (his modification is a predictable conse5uence of sub3ectivity ofknowledge. %henever a knowledge base is applied to a task outside of it intended use, it will re5uire change (Clancy, !""+'.
89;M:2, for [Link] was initially developed to classify pathological items. It has now been [Link] to produce a gleneral purpose system for all ofmedicine. *owever a study of 89;M:2As utility in coding nursing reports found it coded only about G"H ofterms (*enry et. al. , !""#' ) with the implication that the missing terms would need to be added. 8uch additions are re5uired every time a terminology is applied to a new area, making the task of updating problematic (Cimind and Clayton, !""#'. :ventually, as a terminology is continually [Link] into new areas, its fundamental organisational structure will altered to reflect the different structure ofthese new areas (Clancy, !""+b'. (he process ofterminology growth and alteration introduces huge problems of maintenance, and the very eal possibility that the system will start to incorporate errors, duplications and contradictions. Ifwe simply think ofterminological thesauri as computer programs then we already know that continued modification is a poor development strategy. 8oftware engineering tells us that the best time to modify a program is early in the

development cycle. Introducing changes into a mature system becomes increasingly [Link] over time (&ittlewood, !"67'. Conse5uently, we have probably reached the stage where uncontrolled addition ofterms to [Link] thesauri is no longer acceptable. (hose who pay for their maintenance will be faced with ever increasing costs. (o manage these costs, one would need to measure the performance of a thesaurus on a particular task, and then determine whether proposed additions or alterations will improve that performance, and at what cost. Compositional Terminologies In the longer term, new approaches are needed. Most [Link] coding systems are enumerative, listing out all the possible terms that could be used in advance. A compositional approach in which terms are created from a more basic setofcomponents may be more practical to build, maintain and update (Ilowinski, !""#'. For [Link], a practitioner may ask, does Jsevere discomfort in the fifth left metacarpophalangeal 3ointJ in a patient record correspond to Jsmall 3oint syraptomsJ in a clinical protocol1 An enumerative system would have to have a pre)[Link] code for the clinical fmdings, but a compositional system would generate the fmdings from a set ofcomponents (Figure !'. Indeed it should be able to generate many such specific con3unctions, as long as they are medically sensible (Ilowinski et ai, !""!'. (hus rather than developing static terminologies, the combinatorial approach tries to construct dynamic terminology servers to produce answers to a variety of terminological 5uestions (9owlan et. aI., !""#'.

(here are two hypothesesbehindthe compositional proposal. (he first is an engineering hypothesis) that compositional systems are easier and cheaperto maintain and update than enumerative ones. As we have seen, current enumerative systems continuallyre5uire [Link] will over time introduce inconsistencies to the system.(he compositional approach starts from scratch, defming a core of components that constitutea AdeepA model of medical knowledge. (he [Link] is that terms can be generatedfrom that model.?y definition,since they are generatedfrom the same core of knowledge and the method of generationis known,terms can be mapped onto one another logically. Further, as medicine changes these changescan be madeto the core and be immediatelyreflected in any new term generated. Compositional systems should also allow the use of sophisticated internal checks on the correctnessof their contents (Iobel, !""#'. Compositional systems should also be more efficient to use. (he power of a compositional system is its compactnessand maintainability, while the cost ofusing it is that each answer has to be derivedfrom Afirst principlesAand this takes computertime. *oweverthe more [Link] completean enumerativesystem is, the slower it is to use (*einsohn, !""#'. ;ne ofthe engineeringtrade)offs to be [Link] in the future will be to decide whether a compositional system is 5uicker to interrogatethan a larger enumerativeone. (he evidence from other disciplines is that e compositional approachwill eventuallybe fastest, as enumerativesystemsgrow to e too large. For [Link] in computerengineering,so called reduced instructionset comp er chips (BI8C' have a small set ofbasic operations which can be combinedto do more com le. operations. (hese chips are much faster than traditional ones that have a large enume tion ofoperationsto cover many eventualities. (he secondcompositional hypothesis is a scientific one and is morecontroversial ) that there is such a thing as a deep or core set of medicalknowledge from whichterms can be generated (Friedman, !""C'. Compositional systems, like their enumerative counterparts are only modelsof the world. Theysufferthe same issuesof modelfidelityand sub3ectivity. *encethere is no greaterAdepthAto the knowledge they encode) it is either3ust moredetailed or moregeneral(Coiera, !""L'.

(he way forward In the shortterm administration agencies keento obtainaggregate clinicaldata are drivento

adopt [Link] systems, even if they are imperfect. (his has leadto muchdebateamongst those supporting particularsystems of their meritsover competing ones (e.g. (uttleand 9elson, !""#'. (he<, is nowbeing asked!> adoptthe Eersion + Beadcodes,both for use in personal clinical systems as well as for audit, research, outcomes and guidelines (CaIman, !""#'. 8uch a decision can now be seen to be necessarily interim. %hat is really neededto help rationale choicesin the longerterm is impartial empirical research, comparing the cost and efficacyof differentsystemsin supportof well defined tasks and contexts. For [Link], in a recent studycomparing the utility of differentcodingschemes in classifying problem lists from medical records, none of the ma3or systemswere found to be comprehensive. <M&8and 89;M:2 were foundto be superiorto Beadand IC2)")CM (Campbell and 0ayne, !""#'. In contrastto the <,approach, the ?oardof 2irectors of the American Medical Informatics Association have suggested that it is not necessary or desirable to have all codescoming from a singlemastersystem. (heysuggestthat one should [Link] tested approaches, despitetheir imperfections, to progress 5uickly. A first phase system could be createdby borrowing from the different [Link] systems, each created for and therefore better suitedto, differentsub3ectdomains (Ackerman et al., !""#'. (he longer term need will be to introduce more maintainable and [Link] systems, as the cost of supporting [Link] systems becomes insupportable. A solution based in part on multiple compositional systems would seem to be the most desirable one. 8ince any general medical terminology will on~ cover a small part of the specific vocabulary of any medical speciality, separate systems may need to developed for use between specialities and within specialities ) Jvocabularies need to be constructed in a mannerthat preserves the conte.t of each discipline and ensures translation between disciplines (?rennan, !""#'J. Indeed over a century ago when Farr constructed the classification system ultimately resulting in lC2, he noted that Jseveral classifications may, therefore, be used with advantage$ and the physician, the pathologist, or the 3urist, each from his own point of view, may legitimately classify the diseases and the causes of death in the way that he thinks best adapted to facilitate his en5uiriesJ (IC2)", !"7C'.
Compositional systems will thus need to be constructed that agree on a restricted subset of terms necessary for the passage of information between specialities ) an :speranto as it were, between different cultures. %ork on such communication standards is at present still in itAs infancy (e.g. Ma, !""C' and more substantive work should be [Link] in the future. 0resently terms are created without [Link] tasks in mind, in the hope that all unseen eventualities will be served. Inter)speciality systems would probably need to be tightly task based to ensure [Link] utility. It is at this point that the importance of protocol)based medicine becomes very clear. 0rotocols are constructed with an [Link] task and conte.t in mind. (hey are written by a [Link] within a speciality, who arrive at a consensus on the management of a specific condition. In the process of doing so they have to define their terms. (he communication of information to another speciality can also be defmed in the same manner ) in the conte.t of a patient on a protocol, what information is needed by an allied specialist1 %hile it is clearly the case that good terminologies will be needed to construct computer)based protocol systems (Ilowinski, !""#', the discipline of writing protocols will constrain the terminology problem sufficiently such that a well defmed and relevant set ofterms can be agreed upon.

Conclusion
(his paper has reviewed threeapparently 5uite separate areas ) telemedicine, protocol based decision support systems, ad terminologies. (hey can now be seen to be [Link] entwined, since the goals of communicating information and deciding on the content of

information cannot be separated. *uman communication involves information [Link] in a conte.t (McCarthy, !""#'. %hat is said depends on the intended message, the method used to convey the message, who is Ml4Caking, and who is being spoken to. %ith the development of protocol based systems and their supporting terminological systems, we have the perfect [Link] ofthat symbiosis. ([Link]. ) The Internet At present, the Internet is an open and unregulated community of individuals communicating freely across an international electronic computer network. (he number ofmedical sites 3oining the Internet increases monthly, as does the number of information resources available on it. Indeed, it would re5uire a lengthy article to enumerate what is currently available, and that listing would be out of e well before it appeared in print. (here is now good evidence at such services are valuable and in constant use. (he ;nco&ink information resour e for [Link], provides oncologists with up to date trial and treatment information, as weI as acting as an educational resource for cancer patients and

their families. ;nco&ink was reportedly accessed +G,>>>times in March !""#(?uhle, !""#'. (he figure for 4anuary !""C was -6#,#!- accesses. (he %orld %ide%ebis perhapsthe most important innovation on the Internet in the last few years,and is a software layerthat provides Internetuserswith a simpleway of accessing information. (he%eballowsusers to createand [Link] te.t, imageand video documents. (he 5ualityof these documents are nowso high that the %ebis used by some medical educational institutions. (he <niversity of <tah, for [Link], has an [Link] anatomical pathology images called %eb0ath for its students. At the <niversity ofIowa a set of teachingmaterials is assembled in their Eirtual*ospital. (he9ational&ibraryof MedicineAs AEisible*umanA pro3ectaims to createa complete, anatomically detailed, threedimensional representations of the male and female humanbodyand to makethis available on the Internet. (he pro3ect is collecting transverse CA(,MBIand cryosection imagesof male and female cadaversat one millimetre intervals. Medical research is also taking advantage of the %ebas 3ournals beginto appearon the %eb in preference to, or in advance ofprint. For [Link] the Journal ofMedical Imaging will publishpaperson the net. Its field is movingso 5uicklythat printed mediaare now seen only as the archival form of knowledge. (he form best suitedto rapid dissemination is electronic, withthe additional advantage that one can create paperswhichcontainte.t, graphics, sound, and movingimages. 8imilarly, the moveto evidence basedmedicinewill be able to use the %ebfor rapid distribution of important clinicalmanagement data. (he Cochrane Collaboration, which seeksto collect, reviewand disseminate high 5ualityoverviews of the effectsof healthcare, has already set up a publiclyaccessible resourceon the %[Link] intent
is to make the updated syste~atic reviews it collects accessible via the Internet (Ioodlee, !""#'.