Preface-2-06.dvi

A roadmap towards healthcare
information systems interoperability
in Greece
Alexander Berler, Anastassios Tagaris, Pantelis Angelidis, and Dimitris Koutsouris Abstract— The advantages of the introduction of information
ment in patient care, have been satisfactorily embraced by and communication technologies (ICT) in the complex health-
the medical community, this has not been the case with care sector are already well known and well stated in the past.
healthcare informatics. Thus, it can be argued that issues It is common knowledge that in order to install any type of
such as data management, data modelling and knowledge information system in healthcare, six main groups of issues
management have a long way to go before reaching the have to be dealt with: organizational and cultural matters
maturity level that other technologies have achieved in the related to healthcare, technological gap between healthcare
medical sector.
professionals and information science experts, legal require-
A variety of reasons could be proposed for this issue, ments on the confidentiality of personal data, of patient re-
lated data and on data privacy, industrial and market posi-
though with a short analysis it becomes rather clear that tion of healthcare informatics and interoperability complexity,
modern ICT present integration problems within the health- lack of vision and leadership of the health care managers and
care sector because of the way the latter is organised.
health authorities and user acceptability and usability of the
Healthcare is a strongly people-centred sector in which ICT proposed information systems. In order to meet these issues
has been considered more as an intruder, as a "spy" to the stated above, a special focus group (Z3) performed an assess-
healthcare professionals' way of doing things and as a com- ment of the situation of healthcare informatics in Greece and
petitor to this people-centred model. Thus, if ICT intend of the main key points that would lead to success. In that sense
to prove its advantages towards establishing an information it is now common knowledge that Greece is lagging informa-
society, or even more a knowledge society, it has to focus on tion and communication technology progress in healthcare be-
providing service-oriented solutions. In other words, it has cause almost none of the above mentioned issues were dealt
to focus on people and this has not been the case in most of with. This assessment is the result of the interaction of more
than 150 decision makers, medical informaticians, healthcare
the circumstances. It is common knowledge that in order practitioners and other individual involved in healthcare. As
to install any type of information system in healthcare, six a conclusion, this focus group resulted in 4 major propositions
main groups of issues have to be dealt with [7, 8]: that will lead to healthcare informatics introduction with bet-
1. The organizational and cultural matters related to ter success chances: focus on terminologies and standards,
healthcare. This issue is rather important, regard- focus on interoperability and information systems sustainabil-
ity, focus on clear goals and system metrics that can create
less of any information system, since organizational a healthcare performance management cockpit, and focus on
models and culture do endorse neither the continuity people and what they have to say, by creating a e-health fo-
of care, nor any type of structured data collection.
rum. These conclusions were taken into consideration by the
Issues such as mistrust between different specialists, Greek government and are incorporated the IASYS project,
between the different healthcare structures or between the national healthcare informatics framework for the next ten
doctors and nurses prevent in many cases the effective sharing of information. Health reforms are currently Keywords— interoperability, HL7, regional healthcare infor-
under way in many countries stressing the will to deal mation system, medical informatics standards, medical codifi-
with this problem.
cations, information and communication technologies develop-
2. The technological gap between healthcare profession- ments in the healthcare.
als and information science experts. Doctors are of-ten reluctant to use information systems which theybelieve are not designed for them.
point of view, healthcare informatics have been in-troduced in healthcare institutions mostly on pilot- The advantages of the introduction of information and com- based projects aiming at addressing specific issues munication technologies (ICT) in the complex healthcare and have proposed solutions addressing a small num- sector have already been depicted and analysed in the ber of healthcare practitioners, resulting in establish- healthcare informatics bibliography [1–6]. It is neverthe- ing a complex map of information niches. This ap- less paradoxical that, although several major technological proach is the consequence of applying information discoveries such as magnetic resonance imaging, nuclear technology to procedures that where not designed for medicine and digital radiology, which facilitate improve- it, thus creating a panspermia of information models Alexander Berler, Anastassios Tagaris, Pantelis Angelidis, and Dimitris Koutsouris which are neither compatible nor interoperable, even countries are in the process of introducing or imple- within a single institution's environment. Efforts menting such business process reengineering projects in creating interoperability standards and protocols, in order to address healthcare delivery in a more in- such as health level seven (HL7), are proposing so- formation flow conformant way. This is a key point lutions to address this issue, thus enabling data ma- in reaching knowledge management, knowledge re- nipulation and knowledge management.
use and sharing, and finally proposing a solution forthe knowledge-based society of tomorrow. This issue 3. The legal requirements on the confidentiality of per- should be dealt with by proposing strategies that fo- sonal and patient related data and on data privacy. It cus on processes and by establishing key performance is clear that if this issue is not addressed at a manage- indicators, balanced scorecards, or other metrics that rial and procedural level by imposing suitable poli- are the upper level of a structured information flow- cies to meet these requirements, there is little chance based model.
that medical data will be kept digitally in a struc-tured manner (thus allowing the transition from dig- 6. User acceptability and usability of the proposed infor- ital islands of clinical data towards a structured elec- mation systems. This issue is the one most strongly tronic healthcare record). The implementation of an related to the problem of dealing with the people- information system, where the electronic healthcare centred approach of the healthcare sector. This is- record is considered to be the core of the system sue deals with information systems' user friendli- (patient-centred model), is the only way to drive data ness and attractiveness, with usability issues such as management towards creating new knowledge. The the time to reach a data entry point, the speed of complexity of the problem can be explained if one information retrieval, the quality of information re- just observes the course of implementation of both trieval, the complex security procedures, etc. In or- the Health Information Privacy and Accountability der to implement information systems and knowledge Act (HIPAA) in the US and Directive 95/46/EC in management systems, education and training must the European Union (EU). The issues seem to have be addressed with high priority since user accept- been dealt with at the strategic level, but still a lot ability is strongly related to them. Service oriented has to be done in the implementation and setup of models and patient-centred information systems have those strategies.
a higher chance of passing the user acceptability test.
A system that is not accepted by the user is often 4. The industrial and market position of healthcare in- a system with poor data quality (or no data at all) formatics. In general the healthcare market is seen by and knowledge management, business intelligence or the industry as large in size but not highly profitable, data warehousing solutions are consequently inoper- mainly due to the lack of standards in implementing able and unsuccessful.
and interoperating healthcare informatics products.
As a consequence, the industry has focused on cre- Taking the above issues in mind, as well as the ongoing ef- ating mostly small scale products (i.e., laboratory in- forts of the Greek Ministry of Health, the Greek e-business formation systems – LIS, radiology information sys- forum1 initiated a new focus group regarding e-health and tems – RIS, clinical information systems) and not on interoperability, which took the codename Z3. This focus evangelising the production of information systems group gathered more than 150 decision makers, medical in- that are dealing with healthcare as a whole. The lack formaticians, healthcare practitioners and other individuals of end-to-end solutions is dealt with by interconnect- involved in healthcare. The focus group started working ing heterogeneous information systems (a rather com- in September 2004 and ended in April 2005 with a one plex task with constant change management issues) day event (workshop) to present publicly its findings and and by introducing solutions from other business sec- recommendations. The following paragraphs are depicting tors (i.e., ERP, SCM, CRM) that have often been re- the result of that effort in Greece.
jected by "key users" as non compliant with theirjob description. Nevertheless, the new web technol- 2. Defining the open issues ogy approaches (web services, XML, etc.) and thenew information technology strategies (i.e., service The focus group prepared an exhaustive questionnaire that oriented architecture) could be the drivers towards was filled by the focus group members. The following list merging information technology and healthcare ser- of issues was depicted from those questionnaires: vices and thus enabling the establishment of serviceoriented products.
1. Political issues are strongly biasing the government's decision making strategy. In that sense, politics tend 5. The lack of vision and leadership of healthcare man- to change continuously, creating a lack of high level agers and health authorities, and the lack of willing- ness to re-engineer healthcare processes for the ben-efits of efficiency and quality of care delivery. Some A roadmap towards healthcare information systems interoperability in Greece 2. There is no national strategy for medical terminology, institutions. Even today with the operation of a gen- information systems security, disaster recovery, data eral secretariat for social security, the Greek gov- interchange protocols, etc.
ernment has not succeeded yet to create the correctenvironment for the citizen, despite the efforts of the 3. Greek medical institutions are understaffed regard- ing their need for the successful adaptation to newinformation and communication technologies.
15. The human factor lacks of expertise and training in ICT, thus making almost impossible to locate the cor- 4. As the public sector is concerned, the Focus Group rect amount of key users or early adopters to pro- noticed that procedures do not comply to the intro- duction of ICT, thus creating a draw-backing inertiaof the national healthcare system.
16. It is extremely difficult to implement business reengi- neering projects in the public sector. Nevertheless, 5. High level leadership mostly focuses on day to day many efforts are in the process of implementation.
management than towards introducing the necessarystructural changes to support ICT.
17. The reaction to change is quite large, since techno phobia has passed from top management to a large 6. There is a strong lack of vision amongst leadership, number of employees, thus creating a hostile envi- starting top down from the high level administration.
ronment for ICT visionaries.
7. The Greek medical ICT market is very small to en- force correct bottom up solutions, thus existing so- 3. Interoperability roadmap prerequisites lutions simply follow the complex and bureaucraticway of doing things in the Greek public medical in- In order to establish an interoperability roadmap, three pre- requisites have to be met: 8. The user requirements and technical specifications proposed to the implementers often lack of severity, – slection of an interoperability architecture; clarity and business scope.
– pilot testing to establish possible open issues and im- 9. There is no follow up of other worldwide best prac- plementation risks; tices, and visionaries are restricted to deploy strate- – defining an information systems sustainability score- gies that never succeeded to overpass the design 10. The proposed time management of the government 3.1. Proposing an interoperability architecture
ICT project is unrealistic and does not take into con- In 2001 a reform of the Greek national healthcare sys- sideration the complexity of the healthcare sector.
tem [9] was introduced in order to enhance the performance 11. Fund management and human resources management and control of healthcare provision in Greece. One of the is not clear and are both mostly spent in unrealistic main changes was the division of the country in 17 au- projects that to not promote ICT as success cases.
tonomous healthcare regions where the regional healthcareauthorities (RHA) are responsible for the regional health- 12. The high level leadership lacks of ICT knowledge and care strategy. In order to support this reform a series of cannot focus correctly upon the benefit of the cor- ICT oriented interventions were introduced. After a period rect introduction of integrated information systems of analysis and design the Greek government started issu- in Greek medical institutions. A large majority of ing a number of extremely detailed (more than 500 paged questionnaire reported a techno phobic approach of each) request for proposals (RFP) for each RHA [10].
the political and administrative leadership.
The integration of existing and forthcoming information 13. The Greek healthcare sector has four decision mak- systems represents one of the most urgent priorities in or- ing groups (Ministry of Health, Ministry of Educa- der to meet the increasing clinical, organizational and man- tion, Ministry of Social Welfare and Ministry of De- agerial needs [11, 12]. In that context, the use of standards fence), thus making the business rules extremely bu- is essential since data processing needs vary widely in the reaucratic, creating a business environment that lacks complex regional healthcare environment. All RHA have of homogeneity in matters of terminology and pro- a major concern in evaluating the existing operational hos- pital information systems and other information system in-frastructure in order make a decision on whether to main- 14. The social security sector is also extremely complex tain or replace them. In Greece, more than ten distinct ven- and not homogenised in procedures, insurance cov- dors have installed healthcare IT related products (hospital erage, and support to citizens. This is due to the information system, laboratory information system, radiol- separate route that each ministry has followed for its ogy information system, etc.) that mostly work indepen-

Alexander Berler, Anastassios Tagaris, Pantelis Angelidis, and Dimitris Koutsouris Fig. 1. Regional healthcare information system basic software components.
dently as IT niches. It is known that the lack of healthcare achieve a regional healthcare information system is to use, information standards is one barrier to the broad application where applicable, an HL7 message-based communication of IT in health care units. The inability to share informa- system implemented by an asynchronous common commu- tion across systems and between care organizations is just nication infrastructure between healthcare sites.
one of the major impediments in the health care business's The proposed information system in the RFP consists of progress toward efficiency and cost-effectiveness, as well a series of subsystems as depicted in Fig. 1, covering infor- as, the absence of a unique national or even regional patient mation management issues in a regional healthcare system.
identifier in Greece. Integration of these existing diverse The system is innovative in the sense that it required the systems with the future information systems to come re- design and implementation of a complete and integrated in- mains problematic with a number of competing approaches, formation system at a regional level that comprises all types none of which alone represents the perfect solution. Cur- of healthcare levels (primary care, secondary care, home rent practice shows that the most promising approach to care, etc.), that includes interoperability issues, that covers A roadmap towards healthcare information systems interoperability in Greece most of the needed components and that could be able towork efficiently in a secure wide area network (i.e., a VPN)to ensure data privacy and confidentiality.
Through the aforementioned RFPs, the need has arisen tomake healthcare information systems in Greece to worktogether as the components of regional healthcare net-work (RHN), where newly introduced information systemsmust communicate with systems already present in varioushealthcare institutions. The proposed solution features theuse of middleware broadcasting systems that are based oninformation exchange via messages utilizing some applica-tion protocol (ISO-OSI level 7).
The proposed architecture fulfils at least the following re-quirements: 1. Existing systems do not need to be altered.
2. No significant extra (hence unanticipated) load on existing systems is introduced.
3. Connecting existing systems is an economical viable The three requirements are met by an asynchronous mes-sage based information exchange infrastructure defining Schematic representation: (a) direct connection; (b) use of middleware broadcasting system.
a uniform interface for any system that must send orreceive information. All systems are connected, througha uniform interface, to an interoperability framework ormore technically to a common communication infrastruc-ture (CCI). In an asynchronous message based CCI, in-formation is exchanged between two systems by break-ing up the information into chunks. These "chunks" arecalled application protocol data units (APDU). An APDUhas an explicit structure that is defined by the APDU (ormessage) syntax. Additional encoding and decoding ruleshelp sending and receiving systems to construct and toanalyze APDUs. Sending systems can insert informationinto APDUs and receiving systems can extract informationfrom the APDUs.
APDUs are not transmitted directly; they are embeddedin so called protocol data units (PDU). APDUs form the"payload" of PDUs. PDUs contain enough information for Fig. 2. Workload produced by connected systems.
Fig. 4. Creating an interoperability framework.
Alexander Berler, Anastassios Tagaris, Pantelis Angelidis, and Dimitris Koutsouris Fig. 5. Regional healthcare based interoperability framework.
the CCI to be able to "route" the information sent to the re- connected easily and even routing of information becomes ceiving application. Additional "meta" data help the receiv- feasible. The latter is very important to connect remote ing side to understand if the PDU has been received intact systems that cannot communicate directly. Clearly the third and contains the APDU anticipated.
advantage is the most important. The fact that two infor- Using (A)PDUs to exchange information between systems mation systems do not need to know each others database brings a number of distinct advantages: schemata or database connection technology, tremendouslysimplifies the task of interfacing these systems. Figure 3 1. All systems can be interfaced in a uniform way with depicts the change that occurs when introducing a middle- ware broadcasting system.
2. There is decoupling between systems which allows Another important feature of the proposed solution is that it information to be routed, stored and forwarded, and creates an interoperability framework that can be replicated processed independently from the actual exchange.
from one healthcare institution to another. In that sense,common interoperability messages can be used to intercon- 3. Information exchanging does not need to reveal their nect heterogeneous information systems within a health- internal structure to each other. This form of "in- care institution or even at a regional healthcare level if formation hiding" significantly improves the con- a centralized information system is in place, as depicted nectability of systems.
As depicted in Fig. 2, the use of a middleware broad- The proposed interoperability framework greatly simplifies casting system is enabling the interconnection of informa- the data exchange issue in a regional healthcare informa- tion systems without creating extra workloads on existing tion system since a lot less interoperability connections are information systems. When a system provides a uniform required and messages used are homogenized between all interface for sending and receiving information it can be involved healthcare institutions.
A roadmap towards healthcare information systems interoperability in Greece Health level seven [13–15] is by far the most widely one software engineer with knowledge of HL7, data- used message based information exchange standard in the base systems and HL7 middleware concepts, one proj- clinical environment. It is in use on all continents of ect coordinator (part time employment), one system ad- the world. Also HL7 is clearly the most mature message ministrator per site/link (part time employment), one based information exchange standard. As a consequence, project manager from the RHA (part time employment) HL7 was set as a mandatory requirement in the selection and one application/database administrator (part time em- process for the implementation of the RHN for each RHA The time consumed for the pilot project is described in Figures 3 and 4 mostly deal with interoperability issues within a healthcare institution, where typically hospitalsare mostly concerned since they produce the wider range of medical data. Figure 5 though is extending and describing Duration of the phases of such project the proposed interoperability framework and clearly depictthe basic interoperability paths required at a regional health- care level. As stated before, a regional healthcare system can be either an aggregation of interconnected distributed Defining the scenario and variable information systems, either a totally central- ized system based upon an application system provider's Defining the events and (ASP model) approach or a combination of the aforemen- tioned architecture. In all cases information flows, patientjourney data, electronic healthcare record data (data col- Selecting the right message types lected from various institutions, in various formats and Extend selected message appointed to each individual based upon a mater patient index – MPI) are creating very important interoperabil- Define the protocol ity issues. It is without saying that data privacy issues Determine implementation are important when transferring or gathering data at a re- gional level and should be dealt with according to EU di-rectives and additional national laws. Data privacy issues Map message fields onto are addressed by the means of creating the proper patient consent mechanism, by creating and imposing strict and Mapping table columns onto firm data manipulation and data storage procedures and by avoiding aggregation of sensible data when not strictly Implement message sending Implement incoming message Figure 5 depicts the interoperability point within a re-gional healthcare information system. The interoperabil- ity framework can be implemented either centrally with Verification and validation one middleware broadcasting system that interconnects all 500 hours
concerned information systems in a regional healthcare net-work (VPN based) setting or with an aggregation of inter-connected and networked middleware broadcasting systems Furthermore, the risk analysis for the implementation of (one for each institution in the regional setting) that all com- such projects is summarized into three categories: data
municate by an agreed numbers of HL7 based messages.
quality, technical and organizational. Table 2 presents
In that sense, the cooperation of such middleware systems an analysis of the risks.
could be expanded nationwide, thus enabling patient mo-bility and data consistency within a nationwide electronichealthcare record.
3.3. Information systems sustainability scorecard
Defining existing information systems sustainability is not 3.2. Pilot testing
an easy task since most of the reasons for disinvestingor reinvesting in information systems is highly subjective, In order to test the proposed framework a small scale pi- mostly based upon criteria such as user friendliness, cost lot project was conceived [16]. The pilot aimed at imple- effectiveness, etc.
menting interoperability among hospital information sys- The proposed scorecard is based upon some initial assump- tems and the management information system (MIS) of a RHA. The implementation of an HL7 link requiresfewer resources when HL7 middleware is deployed, and • There is no issue of sustainability concerning exist- the data are stored in open architecture database man- ing information systems that are to be replaced by agement systems. The required human resources were technologically more advanced platforms.
Alexander Berler, Anastassios Tagaris, Pantelis Angelidis, and Dimitris Koutsouris Risk analysis table Data quality risks
Annotated database There is no annotated database schema at all. Lack of a comprehensive database schema significantly reduces the likelihood of successfully reverseengineering the database schema.
Database schema contents The names used in the schema are unclear or ambiguous; also relations are not interpretable not clearly defined.
Database tables not used as Database table usage has drifted away from the logical or semantic design.
described in the databaseschema Free text fields used for Free text fields are using to store structured data. The database schema structured data in an ad hoc should be redesigned.
Inconsistent use of enumer- In particular in reference tables like COUNTRY, CITY values like "GREECE", "Greece", "greece" all representing the country Greece.
Required data not present Data that according to the database schema are required (NOT NULL) but are null in the tables. The database schema does not reflect the currentstructure of the database.
Semantic analysis of data Data fields does not contain semantically valid data.
Character set encoding There are problems with the character set encoding in the database.
General interface to access Access to information systems' databases through general interfaces that are independent from applications, must be available (and configured) in every(hospital) information system.
Proprietary operating sys- The database system runs on a platform with proprietary operating system (e.g., Not Windows, Linux, Unix, or VMS).
Database not accessible It is impossible to logon to the database.
Exotic communication pro- The platform on which the database runs can only be connected through a non-TCP/IP communication protocol.
LAN not reachable The LAN on which the pilot system is connected to cannot be reached from outside the LAN due to security matters and other reason.
Unstable or failing comput- The computing environment is unstable or failing causing the pilot system to malfunction.
Not enough competent staff IT staff is not qualified or inadequate in quantity.
Lack of individual Individuals in the organization are reluctant to co-operate.
Rules and procedures Rules and procedures are becoming an obstacle or slow down progress especially in public services.
Lack of decision making There are no decision makers that can put the project in progress.
Lack of software vender Software vendors that need to assist do not do this.
Software vendor sabotage Software vendors are actually sabotaging the project out of commercial A roadmap towards healthcare information systems interoperability in Greece • A RHIS is an integration of specific oriented building is due to a lack of codification, terminology and standard blocks that are commonly acknowledged and agreed.
usage for recording, storing and interchanging data. The Those are the enterprise resource planning (ERP), use of medical terminologies allows systemic and proce- the HIS, the LIS, the RIS, the human resources man- dural reuse of information in order to assist medical staff, agement, the document management system, the in- to fill the electronic patient record, to promote prompt and teroperability middleware tool, the portal and the correct diagnosis and to enhance quality of care.
business intelligence tools.
Furthermore coded data are more malleable concerningstatistical analysis and public health monitoring, both at As a consequence the following steps are required: a national and international level. Both administrative man- – defining the building blocks (BB) of RHIS (as agement and medical staff are able to gather any type of data fitting their job descriptions.
The most common codifications are the classifications such – proposing a scorecard for sustainability; as the International Classification of Diseases (ICD)2 pro- – defining the needs for interoperability between build- posed by the World Health Organization (WHO), and the nomenclatures such as Systematized Nomenclature of Hu-man and Veterinary Medicine (SNOMED)3. Other types As a matter of scorecard the following criteria were pro- of codifications are the thesauruses, the taxonomies and the formal terminologies. In Table 3 some of the mostcommonly proposed and used codifications are listed more 1. Technical and logical architectural conformity.
as examples than a complete list. The oldest classifica- 2. Recent technological platform (Windows or Unix tion reported, the "London Bills of Mortality" was con- GUI, Web GUI).
ceived in England for forensic purposes in 1662. WHOstarted ICD in 1901 with Version 1 and today we have 3. Interoperability capability with other BB reached Version 10, finalized in 1992. SNOMED started in 1928 (SNOMED RT) and is been continuously up-dated, now having more than 361 800 medical terms, 4. 80% coverage of ERP standard functionality.
975 000 descriptions and 1 470 000 semantic correlations 5. 60% coverage of HIS standard functionality.
in SNOMED CT (2004).
Organizations as WHO, College of American Pathologists 6. 100% coverage of HRM standard functionality.
(CAP), Health Level Seven and the world organization ofnational colleges, academies and academic associations of 7. 100% coverage of LIS or RIS standard functionality.
family physicians and general practitioners (WONCA) are 8. 80% coverage of the established Greek national not the only bodies that have deployed successful coded healthcare systems procedures.
data sets. It is rather common that national standardizationbodies are either translating most commonly used interna- 9. 100% coverage of required data exchange within BB tional codifications or creating their own national subsets of any type and complexity.
10. Vendor sustainability (ability to deliver and support Codifications by themselves are one of the most important the information systems for at least 3 years).
steps toward public health monitoring, cost containmentand better healthcare services to the citizens. Codifications 11. Fixed budgeting rules.
are also extremely important as input or output of a health-care information system of any range and penetration. Theuse of coded data results in having high quality structureddata that enable better reuse of the knowledge created dur- 4. Medical terminology: a prerequisite ing the day to day process of patient treatment, thus en- for interoperability abling patient history keeping, diagnoses recording and bet-ter healthcare outcomes. Structured data enable statisticians 4.1. The importance of codification
and administrations to better monitor public health, diseaseprevention and strategic policy planning.
Healthcare institutions are creating a huge amount of data of The use of coded data is also the cornerstone of cost anal- any type (administrative, financial, medical, etc.) or format ysis of a well designed healthcare system, making it pos- (reports, medical records, medical images, transcriptions, sible to foresee procurement requirements, institution de- doctor letters, etc.), on a daily basis.
ployment and other important decision regarding healthcare Despite the technological efforts and new proposed tech- management. It is important to state that EU has a strong nologies of our times, a great deal of those data is stillhand written or paper based, thus not enabling the exploita- tion of those rich information sources. Part of this delay Alexander Berler, Anastassios Tagaris, Pantelis Angelidis, and Dimitris Koutsouris Most common existing codifications (copyright: A. Berler) National and European data sets
Financial and administrative data sets
Master patient index Greek GL prerequisites (PD 146/2003) Social security number (SSN) National and EU statistical codification sets NCDP (e-claiming) Patient record: OpenEHR/HL7-RIM Data interfacing protocols
HL7 (Version 2.x/Version 3) Diseases and procedures
Arden syntax (HL7), OWL, GELLO (HL7), HCPCS, CPT, OPCS-4 semantic web, GLIF, XML topic maps focus towards structured medical and clinical data and has • In 1992 at LDS Hospital of Salt Lake City (US) the proposed a series of white papers, green papers, commu- establishment of an adverse drug reaction monitoring nications and directives [17–22].
Codifications are also system recorded 569 cases, saved many extra bed critical for the dissemination of medical knowledge and days and $1 000 000 of the hospital's budget.
information systems interoperability. It is not possible todesign any type of interoperability roadmap without tak- • The drug "Seldane" was approved by the Federal ing into consideration the strategic need at a national or Drug Administration (FDA) in 1985, presented the European level for structured data. Initially medical termi- first adverse reactions with erythromycin (cardiac ar- nologies, clinical classifications, medical procedures and rhythmia) in 1992 and was only withdrawn in 1998 clinical guidelines were proposed as a solution to calculate due to lack of decisive data.
and restrict the number of medical errors or adverse drug The use of codified data into information systems in health- care provides the ability to those systems to interoperate and A large number of studies in the US [23, 24], Australia [25], exchange important medical knowledge in order to estab- Canada, Denmark, Italy, The Netherlands, Sweden and lish a unique electronic healthcare record (EHR) for each New Zealand, all report that a large number of adverse citizen by collecting all important data from each patient drugs events and medical errors have resulted in damages encounter with the healthcare system. EHRs cannot be cre- of the health of patients. In the UK, statistics report that ated with medical codifications since they are the base for about 10% of inpatients have been involved in episodes any type of semantic interoperability. This issue is not new of care where wrong dose or other medication was given since Florence Nightingale stated in 1893: with minor or important consequences in patient's health "In attempting to arrive at the truth, I have applied ev- status. The financial costs of those events are estimated erywhere for information, but in scarcely an instance have at £3 billion only for the extra bed days. As a conse- I been able to obtain hospital records fit for any purposes quence the use of e-prescribing, bar coding and/or computer of comparison. If they could be obtained, they would en- based order entry systems are of critical importance and able us to decide many other questions. . They would show have proven to reduce dramatically the number of medical subscribers how their money was being spent [and] what amount of good was really being done with it. . " In Italy more than 14 000 patients die every year dueto medical errors whilst this number reaches each year 4.2. Medical terminology and codifications:
44 000 up to 98 000 in the US, surpassing death tolls the cornerstones of e-health
that are accredited to traffic accidents, breast cancer,AIDS, etc. [24]. All studies state that those errors could Figure 6 depicts the workflows both external and internal be prevented or at least a large number of them, if medical that have to be met within a healthcare system of any range data collected had the proper quality rate. Medical termi- (from a single institution to a national healthcare system).
nologies and codifications have a lot to offer in that sector: From that figure it is clear that three major structural re- A roadmap towards healthcare information systems interoperability in Greece 2. Urgent involvement of key users in the design process of ICT projects instead of simple top down decisionbased projects.
3. Incorporation of knowledge experts such as the Greek affiliate of HL7 international.
4. Continuous ICT dissemination training programmes.
5. Top down design should be restricted to business planning and strategic objectives clarifications in or-der to make the national business rules crystal clear.
6. Strategic business planning continuity from the Greek Ministry of Health regardless from any po-litical or governmental change. This should be madepossible by employing a number of business expertsfocused towards ICT implementation in healthcarewhich is at least a ten years plan.
7. Strategic cooperation with other decision makers in Fig. 6. Healthcare workflows (copyright: A. Berler).
healthcare such as the Ministry of Social Securityand the Ministry of National Economy.
quirements have to be met to reach a satisfactory level ofcoverage of the healthcare workflows with the use of ICT 8. The establishment of an information authority moni- tored by the Ministry of Health that will be responsi- 1. The deployment of key performance indicators and ble for the implementation of ICT strategies, mainte- public health indicators, which is one of the key rec- nance of medical terminologies and the management ommendations and directives of the European Union of the interoperability framework.
9. The creation of national public health indicators that 2. The design and implementation of an interoperabil- will be in accordance to the EU guidelines and re- ity framework based upon commonly adopted and agreed international standards such as HL7 (alreadyadopted nationally in the US, Canada, New Zealand, 10. Establishment of an e-health forum in order to cre- Australia, The Netherlands, Germany and UK, while ate a constant interaction framework between all key other States such as France, Croatia, Ireland, Italy, players in the healthcare sector (government, medi- etc., are moving towards that direction). EU is also cal institutions, industry, medical informatics imple- in favour of such strategic policies [28–30].
menters, etc.). This forum will be responsible for 3. The implementation and maintenance of national the public concertation of regulations, terminologies, medical terminologies as described in the previous strategies and other policy papers so that the maxi- mum consensus can be reached before the implemen-tation of new strategies and regulations.
5. Focus group suggestions 5.2. Ten recommendations for medical codifications
Taking all the above mentioned issues the focus groupreached a consensus that is described below as a set of The following recommendations were drawn up: recommendations for the Greek medical informatics mar-ket, the establishment of medical codifications and for the 1. Greece has to fully participate in the creation of in- establishment of an interoperability roadmap.
ternational standards and protocols by assigning na-tional delegates to all forums and standard develop- 5.1. Ten recommendations for the Greek medical
ment organizations related to healthcare.
2. The "one size fits all" codification scheme is not ef- The recommendations are as follows: ficient as medical specificities are regarded. Some 1. Urgent involvement of the leadership in favour of clinicians prefer nomenclatures (i.e., pathologists) projects related to the introduction of ICT in health- while others prefer simple classifications (i.e., in- Alexander Berler, Anastassios Tagaris, Pantelis Angelidis, and Dimitris Koutsouris 3. There is an urgent need to select and implement the 5.3. Ten recommendations towards interoperability:
Greek set of codifications since the existing scheme creating the roadmap
of "any code is good" is a major draw back for anynational data quality strategy and national ICT de- There are the following recommendations: ployment for better health and cost containment. This 1. Deployment of an interoperability framework based should be made clear to all decision makers and ad- upon common communication interfaces.
2. Assessment and sustainability of existing information 4. The deployment of medical terminology could be as- systems in medical institutions, based upon a specific signed to the healthcare market itself through scien- tific societies and international standardization bod-ies.
3. The healthcare informatics market should strongly focus towards standards conformance and standards 5. EU directives should be taken into immediate con- Consensus based processes for the sideration regarding medication related errors, the deployment of the basic standards functionality are creation of access-free libraries of codes, the cre- of critical importance (i.e., implementing integration ation of workflow models based upon adopted standards (OpenEHR, HL7-RIM) and abstinencefrom the creation of national standards where in- 4. HL7 is mature enough to solve most of the interop- ternational or European standards are already in erability issues in Greek and many more than simple data interchange.
5. HL7 standards should be refined to meet peculiarities 6. For the successful implementation and use of medical of the Greek healthcare system if such issues exist.
terminologies it is required that medical personnel isimmediately involved in the design and proposition 6. HL7 Hellas can assist the Greek ministry of health in process, constant dissemination and training strate- the required standardization process that is needed to gies are followed, consensus based decision making implement a national interoperability platform (ter- is adopted, job descriptions are refined and incentives minologies, processes, workflows, performance indi- are proposed.
cators, etc.).
7. Each selected codification should be selected for the 7. Specific task forces and standardization teams should specific requirements that need to be covered. All be established immediately, under the umbrella of an codifications should be maintainable and upgradeable information authority or of an independent scientific and have the possibility to interrelate with other ter- society, such as HL7.
8. National interoperability conformance statements 8. Codifications and terminologies should be selected as must be implemented based upon the work done by integrated parts of a wider interoperability platform integrating the healthcare enterprise (IHE) with the so that all type of internal or external workflows can use of HL7 and DICOM conformance statement tem- be completed with the use of ICT.
plates and methodologies.
9. Greece should follow the work done by international 9. A constant dialogue framework must be established task force created by standardization bodies such as regardless of any political matters and governmental ISO, CEN/TC 251, HL7, openEHR, etc. This is es- changes. The proposed e-health forum is an optimum pecially valuable as the creation of a national EHR solution for this clause. A five to ten years consensus is regarded.
is absolutely necessary.
10. Immediate involvement of Greek experts and knowl- 10. Greece has separated the strategic planning of health- edge workers in international standardization pro- care and social security, thus cutting the correlation of healthcare providers from payers. This has cre-ated a duplication of standardization efforts not al-ways pointing to the same direction; Best practices in interoperability and standardization in health haveoften started from the payers rather than from the The result of the focus group was publicly presented providers. Payers, providers and patients should be during a one day workshop with the involvement of all put all together under the same strategic umbrella as key players of ICT in healthcare in Greece. It was not soon as possible to create the needed economies of expected that those recommendations would change the situation in Greece overnight. Nevertheless the situation

A roadmap towards healthcare information systems interoperability in Greece of ICT introduction in Greece is blooming of activity with [11] S. Spyrou, P. Bamidis, I. Chouvarda, G. Gogou, S. M. Tryfon, and more that 15 high level ICT projects in the implementa- N. Maglaveras, "Health care informatics standards: comparison ofthe approaches", Health Inform. J., vol. 8, no. 1, pp. 14–19, 2002.
tion process, with the initiation of consensus based efforts [12] J. Grimson, W. Grimson, and W. Hasselbring, "The SI challenge in in order to reach a national framework regarding uniform health care", Commun. ACM, vol. 43, no. 6, pp. 49–55, 2000.
workflows and processes, medical terminologies and an in- [13] "Health Level Seven (HL7): an application protocol for electronic Concerning the latter, HL7 has data exchange in healthcare environments", Version 2.1, 1990.
been largely adopted by the project implementers and the Chicago: Health Level Seven, 1990.
sustainability scorecards is expected to integrate informa- [14] W. E. Hammond, "Health Level 7: an application standard for elec- tronic medical data exchange", Top. Health Rec. Manag., no. 11, tion niches wherever this is plausible, both technically and pp. 59–66, 1991.
[15] HL7 standards Internet resources, http://www.hl7.org Finally, the Greek government has made tremendous ef- [16] S. Spyrou, A. Berler, and P. Bamidis, "Information system interop- forts into proposing a complete strategy regarding ICT in erability in a regional healthcare system infrastructure: a pilot study Greece, by proposing the IASYS project to be gradually using healthcare information standards", in Proc. MIE 2003, SaintMalo, France, 2003, pp. 364–369.
implemented in the forthcoming years. All projects under [17] OECD, STI Scoreboard, A.6.3, Health-related R&D, 2001.
way at this moment will be integrated into this national [18] CEN/ISSS e-health standardization focus group, 2004, strategy from a technical an procedural point of view. The proposed interoperability roadmap will permit information [19] P. Doupi and P. Ruotsalainen, "eHealth in Finland: present status systems to cooperate efficiently and work together as one.
and future trends", Int. J. Circ. Health, vol. 63, no. 4, pp. 322–327, In order for this important strategic project to succeed an information authority is to be established to manage cen- [20] P. Doupi and P. Ruotsalainen, "Healthcare informatics in Finland: current status and future prospects", British J. Healthc. Comput. trally all ICT projects in Greece and in parallel to create Inform. Manag., vol. 21, no. 10, pp. 20–23, 2004.
the required coded data structure, procedures, workflows [21] "MEDITRAV WP11, Deliverable 1", P. Doupi et al., Eds., in The and quality rules for these information systems. In addi- State of eHealth in Europe, Helsinki, Finland, 2003.
tion, the Ministry of Health is about to announce the oper- [22] M. McKee and J. Healy, "Hospital in a changing Europe", "Euro- ation of an e-health forum covering most of the proposed pean Observatory on Health Care Systems" Series, Buckingham –Philadelphia, Open University Press, 2002.
recommendations of focus group Z3.
[23] HMSO, "An Organisation with a Memory", Report of an expert In that sense most of the focus group recommendations group on learning from adverse events in the NHS chaired by the have been considered by the Greek Ministry of Health. The Chief Medical Officer, England, 2000.
process of the successful introduction of ICT in the Greek [24] L. T. Kohn, J. M. Corrigan, and M. S. Donaldson, To Err Is Human: healthcare system should nevertheless take more than ten Building a Safer Health System, Institute of Medicine. Washington:National Academy Press, 2000.
years to be completed.
[25] QAHCS, "Quality in Australian Health Care Study", 1995.
[26] F. H. Roger, "The minimum basic data set for hospital statistics in the EEC", Document EUR 7162, Luxembourg, Office of Publica- tions of the EEC, 1981.
[27] EU Council, "Decision No 1786/2002/EC of the European Par- [1] L. Eder, Managing Healthcare Information Systems with Web En- liament and of the Council of 23 September 2002 adopting abled Technologies. Hershey: Idea Group Publishing, 2000.
a programme of Community action in the field of public health [2] S. Englebardt and R. Nelson, Healthcare Informatics, an Interdisci- plinary Approach. St Louis: Mosby Edition, 2002.
[28] EU Commission, "e-Europe 2005: an information society for all – [3] A. Harmoni, Effective Healthcare Information System. Hershey: an action plan", COM (2002) 263 final, 28.05.2002.
IRM Press, 2002.
[29] EU Commission, "e-Health – making healthcare better for Euro- [4] T. E. Norris, S. S. Fuller, H. I. Goldberg, and P. Tarczy-Hornoch, pean citizens: an action plan for a european e-health area", COM Informatics in Primary Care, Strategies in Information Management for the Healthcare Provider. New York: Springer, 2002.
[30] EU Commission, "Follow-up to the high level reflection process [5] E. H. Shortliffe, L. E. Perreault, G. Wiederhold, and L. M. Fagan, on patient mobility and healthcare developments in the European Healthcare and Union", COM (2004) 301 final, 20.04.2004.
Biomedicine. 2nd ed. New York: Springer, 2001.
[6] R. Stegwee and T. Spil, Strategies for Healthcare Information Sys- tems. Hershey: Idea Group Publishing, 2001.
[7] I. Iakovidis, "Towards a health telematics infrastructure in the Euro- Alexander Berler was born
pean Union", in Information Technology Strategies from US and the in Lausanne, Switzerland, in European Union: Transferring Research to Practice for Healthcare 1969. He received his degree in Improvement. Amsterdam: IOS Press, 2000.
electrical engineering from the [8] I. Iakovidis, "Towards personal health record: current situation, ob- stacles and trends in implementation of electronic healthcare records Aristotle University of Thes- in Europe", Int. J. Med. Inform., vol. 52, no. 123, pp. 105–117, 1998.
salonica, Greece, in 1995 and [9] N2889/2001 (FEK-A/37/02.03.2001), Greek National Healthcare his M.Sc. degree in biomedical System Reform Act, 2001 (in Greek).
engineering from the National [10] Information Society SA, "Healthcare information system for the Technical University of Athens 2nd regional healthcare authority of Central Macedonia", Requestfor proposal co-funded by the 3rd CSF under the EU decision (NTUA), Athens, Greece, in C(2001)551/14-3-2001, Greece, May 2003 (in Greek).
1997. He is currently pursuing

Alexander Berler, Anastassios Tagaris, Pantelis Angelidis, and Dimitris Koutsouris a Ph.D. degree at NTUA in medical informatics focusing on and newpage data mining as well as business process the design and development of interoperable healthcare in- reengineering, and telemedicine networks. He has pub- formation systems towards the implementation of a citizen lished 6 scientific papers in international journals and virtual medical record. He has been with the Department conferences. His current research interests include health of Electrical Engineering, NTUA, since 1996 as a Research information systems and telemedicine, knowledge databases Postgraduate Student working in the area of biomedical and data mining, application of grids to healthcare systems, engineering in European Union funded projects and he is image content based retrieval from databases, business currently working at Information Society SA as a project process reengineering, and telemedicine networks. Doctor manager in large healthcare informatics government project Tagaris has been active in a number of European and na- co-funded by the European Union under the 3rd Com- tional R&D programs in the field of telematics applications munity Support Framework. He is a student member of in healthcare. He is a member of the Technical Chamber IEEE Computer Society, IEEE EMB society since 1998, of Greece, the IEEE and the Greek Society of Biomedical a member of ACM and European Society on Engineer- Technology as well as an active member of the HL7 Hellas ing in Medicine (ESEM) since 1998, and a BoD member working groups.
of HL7 Hellas, the Greek international affiliate of HL7 Biomedical Engineering Laboratory School of Electrical and Computer Science Biomedical Engineering Laboratory National Technical University of Athens (NTUA) School of Electrical and Computer Science 9 Iroon Polytechniou st National Technical University of Athens (NTUA) 15773 Zografou Campus 9 Iroon Polytechniou st 15773 Zografou CampusAthens, Greece Pantelis A. Angelidis received
his diploma in electrical engi-
neering
Anastassios Tagaris was born
in N. Figaleia, Greece, in 1968.
ing from the School of Elec- He graduated from the National Technical University of Athens (NTUA) Department of Electri- in 1989 and 1993, cal Engineering and Computer respectively. Since May 2001 Science and received his degree he holds the position of the in this area in 1993. He ob- Director of INA – Southeastern tained his Ph.D. degree in med- Europe Telecommunications and Informatics Research ical informatics from the NTUA Institute. He has worked as a technology expert in the in 1999. During 1993–1994 he areas of telecommunications and biomedical engineering was working for the Research Committee of NTUA, for the for the past 12 years. He has served as a project manager development of the financial management system of the in over twenty research projects. He has presented more European projects undertaken by NTUA. From 1999 he is than 40 papers in international scientific conferences and working as a teaching and research assistant in the Biomed- has published 8 original articles in international research ical Engineering Laboratory, Department of Electrical and journals and over 10 in Greece. Doctor Angelidis is Computer Engineering, National Technical University of a member of the permanent consulting committee of the Athens, Athens, Greece. From 2001 to 2005 he used to Technical Chamber of Greece on telecommunications, work in CCS SA in the Medical Informatics Department as a certified consultant on telecommunications of the Tech- head engineer of the Laboratory Information Management nical Chamber of Greece (Central Macedonia chapter), System (Medilab) and the HL7 Communication as well as a certified consultant on telecommunications of the Greek the radiology information system development team. He is Courts (Thessaloniki), member of the CEN TC224/PT06 now working as a medical informatics researcher in the In- technical committee and Vice President of the HL7 Greek stitute of Communication and Computer Systems (ICCS).
He has been active in a number of European and national R&D programs in the field of information and telecommu- INA c/o HL7 Hellas nications technologies in healthcare. His current research Southeastern Europe Telecommunications interests include among others interoperability standards and Informatics Research Institute and applications of informatics and telecommunications in 9th km Thermi Road healthcare and medicine, Internet technology, applications of grid technologies in healthcare, knowledge databases 570 01 Thessaloniki, Greece A roadmap towards healthcare information systems interoperability in Greece Dimitris Koutsouris was born
Professor and Head of the Biomedical Engineering Labora- in Serres, Greece, in 1955.
tory. He has published over 100 research articles and book He received his diploma in chapters and more than 150 conference communications.
electrical engineering in 1978 He is also President of the Greek Society of Biomedical (Greece), DEA in biomechan- Technology. He has been principal investigator in many ics in 1979 (France), Doctorat European and national research programs, especially in the field of telematics in healthcare.
Doctorat d'Etat in biomedical engineering 1984 Biomedical Engineering Laboratory (France). Since 1986 he was School of Electrical and Computer Science research associated on the USC (Los Angeles), Rene National Technical University of Athens (NTUA) Descartes (Paris) and Associate Professor at the School 9 Iroon Polytechniou st of Electrical and Computer Engineering at the National 15773 Zografou Campus Technical University of Athens (NTUA). He is currently

Source: http://users.uowm.gr/paggelidis/publications/publications/J26.pdf