Architectural Key Dimensions for a Successful Electronic Health Record

Implementation

Eduardo Pinto and Ant

onio Carvalho Brito

INESC TEC, Faculdade de Engenharia, Universidade do Porto

Rua Dr. Roberto Frias, 4200-465 Porto, Portugal

Keywords:

Electronic Health Record, e-Health, Systems Architecture, Information Systems Architecture.

Abstract:

The availability of patient clinical data can be vital to a more effective diagnosis and treatment, by an healthcare

professional. This information should be accessible regardless of context, place, time or where it was collected.

In order to share this type of data, many countries have initiated projects aiming to implement Electronic

Health Record (EHR) systems. Throughout the years, some were more successful than others but all of them

were complex and difﬁcult to materialise. The research involves the study of four international projects – in

Canada, Denmark, England and France – launched with the goal of fostering the clinical data sharing in the

respective countries, namely by implementing EHR-like systems. Those case studies served as data to identify

the critical issues in this area. To address the challenge of sharing clinical information, the authors believe

to be necessary to act in three different dimensions of the problem: (1) the engagement of the stakeholders

and the alignment of the system development with the business goals (2) the building of complex systems of

systems with the capability to evolve and easily admit new peers (3) the interoperability between different

systems which use different conventions and standards.

1 INTRODUCTION

Healthcare is one of the areas where new Infor-

mation Technologies (IT) caused huge impact.

Throughout the years, the society’s dynamic changed

appearing an increasing necessity for sharing clinical

information among different healthcare institu-

tions (Haux, 2006). Several projects were launched

in many countries, such as in England or Canada

for instance, with the goal of creating the well-know

Electronic Health Record (EHR) systems. An EHR

is a “longitudinal collection of electronic health

information about individual patients and popula-

tions” (Gunter and Terry, 2005). The main objective

is to provide clinical information about a patient

where it needs to be consulted, independently of its

origin or location, helping to avoid clinical errors

or duplication of efforts and resources. Usually,

these initiatives are not conﬁned to the EHR concept.

In this sense, the last years brought an emergent

paradigm in which the patients have the control

over the information about themselves as long as

their involvement in the process also grows (Ball

et al., 2007). The Personal Health Records (PHR) –

which are systems whereby individuals can access,

manage and share their health information which

can be accessed by others in a private, secure, and

conﬁdential environment (Connecting for Health,

2003) – are the best example of it. On the other

hand, the excess of information might be a trouble in

different situations. For that reason, some countries

adopted concepts like Patient Summary (PS) (epSOS,

2012) which is a set of information that allows an

healthcare professional to have a quick and easy

overview over a clinical history.

This research intends to understand the extent

of the challenge that constitutes the implementation

of a system like an EHR as a facilitator of clinical

data sharing among several institutions, from an

architectural point-of-view. In order to do so, four

international cases were selected: Canada, England,

Denmark and France. Basically, these projects were

selected for different reasons: Canada has a lot of

information publicly available and that stands as

ofﬁcial information; the English project was always a

very famous, controversial and apparently not so well

succeeded project; Denmark is known as one of the

most successful countries in this area and France has

been investing in this ﬁeld for the last decade. These

case studies were analysed and dissected based on

136

Pinto E. and Carvalho Brito A..

Architectural Key Dimensions for a Successful Electronic Health Record Implementation.

DOI: 10.5220/0004867301360143

In Proceedings of the 16th International Conference on Enterprise Information Systems (ICEIS-2014), pages 136-143

ISBN: 978-989-758-027-7

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

information from ofﬁcial on-line documentation and

scientiﬁc articles. Along with that, the authors expose

the current methodologies, models and technologies

that better address those challenges.

The paper starts with a description of the interna-

tional case studies and their experience of these im-

plementations at Section 2. From that analysis, the

authors extract the main challenges in these projects

and describe the state-of-the-art that allows to respond

to those challenges, at Section 3. Finally, the paper’s

conclusions are presented at Section 4.

2 RELATED WORK

Several initiatives to implement EHR projects have

been created over the years in several countries. Some

countries have achieved more than others, but the

common experience says that it stands not only as a

technological challenge, but way beyond that (Ander-

son et al., 2006; Jha et al., 2008; Protti et al., 2009).

2.1 Canada Case Study

In Canada, its First Ministers created an non-for-proﬁt

organization called Canada Health Infoway back in

2001 and the effort would exceed 1.5 billion dol-

lars (Canada Health Infoway, 2009). In order to

guide the development of the systems in each dif-

ferent province, Infoway provided a set of princi-

ples, guides and components, called EHRS Blueprint.

EHRS Blueprint advocated the creation of a shared

reference information source that is populated by sev-

eral health-care organizations around Canada and is

maintained externally from every health-care organi-

zation (or Points of Service, as designated in EHRS

Blueprint). The Points of Service (PoS) are able to

reference or pull data from the shared repository. In

fact, there is not one single EHR, there are several

(EHRi) replicated through the country, with the same

structure, that are responsible for interacting with the

local entities (PoS). The EHRi information is stored

as copy of the original one and the EHR Blueprint

was built following a Services-Oriented Architecture.

The critiques started to raise up when one of the

Infoway’s goals of having 50 per cent of Canadians

with an electronic health record available by 2010 was

not met. McGrail et al. (McGrail et al., 2010) ar-

gued that “after billions of dollars and nearly a decade

of work by Infoway and the provinces, Canada is

well back of the electronic records pack” and that

Canada’s “laggard position is certainly not a question

of software design or lack of access to user-friendly

technologies”. On the other hand, the same study

pointed out some issues of lack of clinicians engage-

ment, e-health policy and a focus on national perspec-

tive rather than the regional one. Along with that,

six provincial auditors joined and produced a report

raising the possibility that the programme needs more

funds (Legislative Audit Ofﬁces, 2010). More re-

cently, Webster (Webster, 2011) advocated a change

of strategy following British recommendations for not

keep going with a top-down approach that is crafted

with insufﬁcient engagement of clinical users.

2.2 Denmark Case Study

At the end of 1994 the project organisation MedCom

was created and mandated to establish a nationwide

health care data network based on the Electronic Doc-

ument Interchange (EDI) concept. The project was to

be completed within a period of two years (The Dan-

ish Ministry of Health, 1996) and was the beginning

of the Danish Health Data Network. In Denmark,

the doctors known as general practitioners serve as

the “gatekeepers” of patients to other specialists and

health professionals (Kuo et al., 2010). This context

have enhanced the message-like solution that took

place later. More than a national-wide solution, the

programme was concerned about achieving value in

the local regions by allowing fast and simple com-

munication between the different stakeholders. The

communication ﬂow was based on six interactions,

always with the general practitioner as a nodal point:

request laboratory results from hospital; prescription

with pharmacies; reimbursement by assurance pub-

lic health care; messages to emergency community

care; radiology exams; referral and discharge infor-

mation. This approach led to a “tremendous rise in

messaging from 3 million per month in 2005 to 5

million per month in 2009 was much higher than ex-

pected” (Protti and Johansen, 2010). Denmark has a

National Patient Registry (DNPR) which has served

as a data set of hospital contacts since 1977. Despite

of being an administrative tool, the DNPR allowed the

creation of a number of shared services, including the

National Patient Index (NPI) that gathers all relevant

information about individual patients.

Some authors argue that there were two key de-

velopments that led to the success of the Danish pro-

gramme: “(1) creation of a nation-wide electronically

accessible patient summary record, and (2) creation

of a secure national health data network” (Kuo et al.,

2010). Moreover, in 2006, Edwards (Edwards, 2006)

states some critical factors that led to success, such

as: monetary incentives to the adoption of MedCom

standards; precise standards worked out with the clin-

icians; gradual approach and realistic time frames;

ArchitecturalKeyDimensionsforaSuccessfulElectronicHealthRecordImplementation

137

incentives to vendors; culture of consensus and a

project-based approach. In the late years, Denmark

has shine with the national health network (MedCom)

being “used by over 3/4 of the healthcare sector, alto-

gether more than 5000 different organizations” (Protti

et al., 2009).

2.3 England Case Study

The National Health Service National Programme for

IT (NPfIT) was initiated in 2005. It was born as the

world’s largest civil information technology project,

committing 12.4 billion pounds over 10 years (Bren-

nan, 2005). The NPfIT was made of eight speciﬁc-

purpose systems: a big and national healthcare data

repository, a national healthcare network, an elec-

tronic appointment booking system and ﬁve local ser-

vice providers covering England’s territory. In order

to implement the ﬁve local clusters, ﬁve providers

were contracted and made responsible for deliver-

ing the local services. The program deﬁned a set of

standards, frameworks and implementation to guide

and favour the interoperability between local systems

and across them, called NHS Interoperability Toolkit

(ITK). The step-by-step maturity model was meant to

allow the organization to incrementally progress from

sharing binary data to sharing fully-coded CDA docu-

ments. The NHS Care Records contemplated two dif-

ferent types of records: 1) Summary Care Records —

records held nationally and is an implementation of

the concept known as Patient Summary; 2) Detailed

Care Records — records held locally and a more com-

prehensive record which might store data from past

exams and details, avoiding the necessity for repeat-

ing them, for example.

Since its beginning the project was the target of

many critiques. Back in 2004, John Powell advo-

cated the importance of involving the clinicians in

the changes as well as showing the value to patient

care (Powell, 2004). A report by the King’s Fund in

2007 also criticised the government’s “apparent reluc-

tance to audit and evaluate the programme”, question-

ing their failure to develop a capable strategy (Wan-

less et al., 2007). Several authors wrote about the

problems’ causes and things that could and should be

learned (Coiera, 2007; Clegg and Shepherd, 2007),

from lack of clinicians’ engagement till wrong met-

rics, weak political leadership and one-size-ﬁts-all so-

lutions.

2.4 France Case Study

France is a country with one of the best healthcare

systems in the world, having been classiﬁed inclu-

sively as the “best health system in the world” by

the World Health Organization (WHO) in June 2000.

Following a set of initiatives in the late-nineties, in

2004 a law is approved to establish a Personal Medi-

cal Record (Dossier M

edical Personnel - DMP) which

aims to enable: (1) coordination of care (2) improve

quality of care (3) continuity of care. However, that

law was not only about the DMP, it had other direc-

tives to reform all the hospital sector that were “im-

portant enablers of healthcare delivery modernisation

in France” (Artmann and Giest, 2010). In 2009, the

“Rapport Fieschi” outlines semantic interoperability

as the key goal and challenge of health information

systems.

The General Practitioner (GP) assumed a criti-

cal role in the DMP, being able to access it via the

GP software or via the Internet but also accredited to

transfer important documents into the DMP and hide

document data upon request by the patient. Moreover,

a number of patient-centred services are also envis-

aged (Artmann and Giest, 2010): consumer portal al-

lowing the patient to access their healthcare record,

to see the list of professionals who accessed their

DMP and be informed of data updates; patient be-

ing able to manage the access rights of health profes-

sionals and update their personal information space as

well as masking data. In terms of the storage of the

EHR, France is “the best example of a country that

went with a host-based electronic health record sys-

tem” (Stroetmann et al., 2011). The French patients

are allowed to choose whichever data-host they want

for their health record. Obviously, as prescribed by

the French Decrees on Data Hosts, hosts have to be

certiﬁed in order to be allowed to store clinical infor-

mation. Another interesting fact is that, in France, an

electronic health record can only be created after the

consent of the patient (“opt-in” strategy). In terms

of standards, the French Dossier Medical Personnel

adopted the IHE standards that were combined with

the use of HL7 CDA for the Clinical Document Ar-

chitecture.

The Table 1 shows a brief summary of the four

countries presented, from an architectural perspec-

tive. There are four classiﬁcation parameters: (1)

strategy - classify the process of implementation in

terms of management approach, (2) architecture - un-

derstand where it stands between the two extremes

(completely distributed or completed centralized), (3)

communication - state the approach used to allow the

share of information and (4) standards - some of the

used standards in that country either nomenclatures

or encoding ones. It is easy to understand the va-

riety of solutions that is possible to adopt only by

these four examples. In terms of strategy we have

ICEIS2014-16thInternationalConferenceonEnterpriseInformationSystems

138

Table 1: Brief summary of the four countries from an Architectural Perspective.

Canada Denmark England France

Strategy

Top-down Bottom-up Top-down Big-bang

Architecture

Distributed:

fetched in real-

time

Distributed

Hybrid: PS held

national; the rest

held locally

Distributed: host-

based, fetched in

real-time

Communication

SOA Message-oriented SOA

Document ex-

change

Standards

DICOM;

SNOMED CT;

HL7 v3; ICD10-

EDI (internal);

CEN; SNOMED

CT; DICOM

HL7 CDA;

SNOMED CT

HL7 CDA; IHE

proﬁles; SNOMED

two projects (Canada and England) that were con-

ceived from high-level to low-level contexts. By the

opposite, in Denmark the strategy sprouted from the

developments that were being made at a local level.

In terms of architecture, at least with regard to these

cases, the tendency is clear about the distribution of

the data. That is, there is not a completely centralized

repository with all the clinical data. Usually, most of

the data is held locally, where it was produced, and

some set of it is selected to be accessible in a more

centralized and easy way. In terms of communica-

tion, the solutions vary a bit, since it is possible to

ﬁnd a strategy based in messages like in Denmark,

two SOA-based solutions and France relying on the

exchange of structured clinical documents. Finally, in

terms of standards we can identify SNOMED CT as

the common one, but then it is difﬁcult to ﬁnd signiﬁ-

cant similarities.

3 KNOWLEDGE DOMAIN

MODEL

The presented case studies show that the prob-

lems/challenges vary from high-level strategy to cho-

sen standards. Despite of some of the issues discussed

are not properly architectural problems, the fact is

that all these problems are coupled, depend or affect

some-how the architectural decisions. Therefore, the

authors summarized the fundamental problems into

three research questions:

• How to maintain the closeness between the sys-

tem and the business goals guaranteeing the sys-

tem’s acceptance/adoption?

• How to build such complex and component-

crowded systems in a way they would be able to

change and evolve?

• How to make such different systems to communi-

cate and understand each other bringing forth pa-

tient care value?

3.1 Aligning the Business Goals with the

System Development

Twenty years ago, the systems complexity was grow-

ing with an exponential velocity. The software devel-

opment was facing two problems at a time: in one

hand, the systems were becoming huge and hugely

complex; on the other hand, the systems were de-

veloped with few concerns about business orienta-

tion (Sessions, 2007). At that time, the concept of

Enterprise Architecture appeared to address this prob-

lem. A lot of enterprise architecture models ap-

peared and disappeared over the years. Three of the

most known Enterprise Architecture frameworks are:

The Zachman Framework, The Open Group Archi-

tecture Framework (TOGAF) and The Federal Enter-

prise Architecture Framework (FEAF). The Zachman

Framework (Zachman, 1987) aims to guarantee that

all stakeholders’ perspectives are being taken into ac-

count when developing a complex software system.

In general terms, it is important to understand if all

the artefacts are sufﬁciently focused and if the ex-

isting artefacts clarify all the players, from the busi-

ness owner till the database designer, keeping all the

visions aligned. First developed in 1995, TOGAF

was based on the US Department of Defense Tech-

nical Architecture Framework for Information Man-

agement (TAFIM) (Josey, 2011). TOGAF might be

seen as a process for building an Enterprise Archi-

tecture. This framework states this building process

as a continuous process of building multiple architec-

tures from highly generic to highly speciﬁc ones, un-

til reaching the organizational architecture level (Ses-

sions, 2007). Finally, the Federal Enterprise Architec-

ture Framework appeared with the objective of serv-

ing as a platform for sharing processes, information

and documentation among the U.S. Federal Agen-

cies and other government agencies. This framework

ArchitecturalKeyDimensionsforaSuccessfulElectronicHealthRecordImplementation

139

gathers two main characteristics of the two previous:

in one hand, it deﬁnes a taxonomy – similar to the

Zachman Framework – for artefacts classiﬁcation; on

the other hand, it suggests a process for building and

implementing the architecture like TOGAF does.

3.2 Architecting Complex-Crowed

Systems

The concept of Systems-of-Systems (SoS) is closely

related to this idea of achieving value through the con-

necting and integration between several peer systems.

A system is classiﬁed as an SoS when there is an

“assemblage of several components which individu-

ally may be regarded as systems” and with two ad-

ditional properties (Maier, 1998): (1) operational in-

dependence of the components – each component is

able to fully continue operating even when disassem-

bled from the system (2) managerial independence of

the components – the component systems do operate

independently in fact. Maier suggests some architec-

tural principles to deal with this kind of system like

stable intermediate forms, policy triage, leverage at

the interfaces and ensuring cooperation. In addition

to the obvious dependence of several systems, the fact

of being a complex and large kind of project also de-

serves special attention and methodologies. In this

ﬁeld, the Ultra-Large-Scale Systems (Northrop et al.,

2006) and the Large-scale complex IT systems (Som-

merville et al., 2012) are two terms that refer to the

same type of systems. Independently of the desig-

nation used, these are systems identiﬁed by extreme

size in every imaginable dimension: lines of code,

stakeholders, number of systems, etc. Northrop, et

al. state some characteristics as decentralized con-

trol,“unknowable” and conﬂicting requirements, con-

tinuous evolution and others (Northrop et al., 2006).

If we lower the abstraction level, it makes sense

to refer some architecture styles like Service-Oriented

Architectures (SOA) and Resource-Oriented Archi-

tectures (ROA). SOA is an “architectural style that

emphasizes implementation of components as mod-

ular services that can be discovered and used by

clients” and that “emphasis on loose coupling be-

tween interacting services” (Srinivasan and Tread-

well, 2005). On the other hand, the ROA deﬁnes an

easy access to the entities as well as the way that ac-

cess should be done (Overdick, 2007).

3.3 Transforming Data into Information

Several registries are created and updated all over the

healthcare institutions. The process of transforming

that data into information is an essential step. Plus,

the share of that information creating the knowledge

necessary to offer better healthcare services is the ul-

timate challenge before creating real value to the pa-

tient. The pursuit of interoperability is not possi-

ble without a clear deﬁnition of common languages

and communication channels. In this sense, multiple

standards can be found and compared (Atalag et al.,

2010).

One of the most known organizations in this area

is Health Level Seven International (HL7), which

produced HL7 Messaging Standard. The Integrating

Healthcare Enterprise (IHE) Proﬁles are other exam-

ple, in this case these proﬁles deﬁne the systems in-

volved, the speciﬁc standards used, and the details

needed to implement the solution. Also, the Digi-

tal Imaging and Communications in Medicine (DI-

COM) is an worldwide used standard for medical

image communication, providing data structures and

services and allowing the exchange of medical im-

ages and related information. Another standard is

openEHR which develops speciﬁcations for imple-

menting full EHR systems, pronouncing more in per-

sistence as opposed to messaging, with the goal of

achieve lifelong, patient centred, secure and shareable

EHR. The standards referred above are more about

how to transfer the information. Another underlying

question is the codiﬁcation of that clinical informa-

tion. In that case, the research is about terminologies

and ontologies: SNOMED CT, ICD, LOINC.

3.4 The Peer-Generated Value

Despite the healthcare arena is an old-fashion one, it

is interesting to understand how the new models of

development could ﬁt it. As known, the businesses

are evolving from a product-oriented perspective to

a relation-oriented one. This change of paradigm

brought the client to the middle of the business, help-

ing to create value. The Metropolis Model (Kaz-

man and Chen, 2009) appears as an attempt of de-

scribing really huge complex systems built from

two basilar concepts: Open-source Software (OSS)

and Community-Based Service Systems (CBSS). The

Metropolis Model presents a new uniﬁed vision be-

tween the CBSS and the OSS, focusing deliberately

in the crowd value generation through the deﬁnition

of two levels: the kernel services and the periphery

services. In the healthcare, there is also a vast com-

munity ready to produce value and to be involved in

the improvements that need to be done. The question

is if the systems will be able to support and incentive

that contribute.

ICEIS2014-16thInternationalConferenceonEnterpriseInformationSystems

140

Figure 1: The Knowledge Domain Model proposed as an integrated view of the clinical data sharing domain.

3.5 Discussion

From the literature review it was possible to ﬁnd some

methodologies that address a subset of the three ques-

tions. For instance, the BITAM (Chen et al., 2005)

which stands as a methodology for aligning busi-

ness and IT strategy with special focus on “detect-

ing, gauging and ameliorating the effects of misalign-

ment”. In fact, it address in broad terms the ﬁrst two

concerns elicited above but does not goes into details

of interoperability questions. However, we believe

that the three dimensions are critical. The Figure 1

presents an integrated view of how all this knowl-

edge can be gathered. With it, we pretend to show

that the process of transforming the existing clinical

data into information is, in fact, a very complex pro-

cess. It starts with the production of the data at a local

level, usually in an hospital or a primary care unit. At

this level, the main concern is about aligning the sys-

tems with the healthcare processes. That alignment is

what improves the quality and quantity of data avail-

able to be shared. The next step is about fetching the

information at the local repositories and transfer or

make it available to other professionals. Obviously,

this step requires a signiﬁcant effort since it can re-

quire operations like locating the origin repository or

transform the information into a codiﬁcation the des-

tiny will be able to interpret. Not to mention that this

central structure might need to balance a high volume

of requests. Thus, the construction of a platform like

this requires a wide knowledge from the concepts of

Systems-of-Systems to the Architectural Patterns that

work at a much lower level, as presented in the ﬁg-

ure. All these processes aim to transform the data

into information, which basically means to present

the data in a way and within a proper context to be

consumed by a healthcare professional. The interpre-

tation and use of that information by the professional

is implicitly transforming it into knowledge and ulti-

mately into value to the patient when the healthcare

services are improved because of that. Once again,

when providing that information to the profession-

als one big concern must be the alignment between

those systems and the processes, otherwise all the ef-

fort may be in vain.

From the analysis of the English project, it is pos-

sible to say that the ﬁrst dimension might have been

overlooked. The strategy followed (top-down) im-

plies from the beginning a certain distance between

the implementation team and the local stakehold-

ers. Despite the division of the project in 5 “local

providers”, it would be necessary a much greater ef-

fort to bring clinicians and other healthcare profes-

sionals to the centre of the implementation process.

Apparently, the lack of this engagement had as conse-

quence the raising of multiple resistances either sup-

ported in the system’s existing fails or the natural in-

ertia to change and adapt. The other two dimensions

– the architecture of the hole system and the inter-

operability – appear to have been better studied. Al-

though, some articles report the appearance of criti-

cal failures which might result from architectural is-

sues. The step-by-step maturity model to the adop-

tion of fully-coded CDA documents is an example of

the importance given to interoperability, indeed evi-

ArchitecturalKeyDimensionsforaSuccessfulElectronicHealthRecordImplementation

141

denced by the existence of the NHS Interoperability

Toolkit. Most of the critiques raised against the Cana-

dian project are related to the ﬁrst dimension also, ac-

cusing the responsible of insufﬁcient engagement of

clinical users. Plus, some authors criticized the focus

on national perspective rather than the regional one

which is highly inﬂuenced by the chosen architecture.

In terms of the third dimension, the chosen standards

and the architecture based on PoS was not subject to

major criticism and seemed a good concept. In Den-

mark, the lower-proﬁle project was built from the lo-

cal perspective to a national one, which seems to have

facilitated the engagement of the stakeholders (ﬁrst

dimension). Moreover, the standards and conventions

were negotiated with the clinicians and its use was

encouraged with monetary incentives. In fact, the

strategy was able to distribute the efforts between the

three dimensions making this project one of the most

successful case studies in this area. The French case

is another example of balance between the three di-

mensions. Despite the launching in a national reform,

the responsible were able to: 1) engage the different

players of the market by fostering a solution of local

hosts 2) ﬁnd an interesting solution in architectural

terms to balance the data centralization/distribution

issue 3) select and deﬁne the standards of communi-

cation (IHE proﬁles and HL7 CDA).

The Knowledge Domain Model presented above

refers an extra dimension which we believe can im-

prove the implementation of these projects in terms

of adoption velocity and stakeholder participation.

From the case studies referred the only one where

this dimension might have been a bit explored was

in France. In fact, the establishment of standards

and the allowance of private certiﬁed hosts for clin-

ical data storage stands as a good small example of

what a more open strategy can bring to these complex

projects.

4 CONCLUSIONS

The implementation of an EHR or a project that aims

to foster clinical data sharing in a country is a chal-

lenge that congregates wide issues, from strategy and

business processes to technology and interoperabil-

ity problems. Despite the necessary technology ex-

ists, most country healthcare systems might not seem

ready to fully exploit all its beneﬁts.

The authors believe that these projects have three

critical dimensions: (1) keep the stakeholders en-

gaged as well as guaranteeing the system to be aligned

with the business goals, using principles of Enterprise

Architectures (2) adopt an agile architecture that is

able to gathers new peers but also to allow the sys-

tem to evolve and get better, following research in

System-of-Systems and Software Architecture (3) al-

low the institutions (peers) to effectively share clin-

ical data creating value to the patient care services,

taking advantage of the several standards and conven-

tions existing in this area. It is not about technology

that needs to be invented, it is about how one is able

to bring all this knowledge together and deliver not

the “biggest computer program in the world” but the

most valuable one. From the literature review, most

of the projects usually addressed two of the three con-

cerns referred above. The more high-level ones usu-

ally describe the components of the system and may

or may not describe the strategies to engage stake-

holders. The ones more focused on local perspectives

are usually more concerned about deﬁning the ways

of interaction and interoperability but fail to deﬁne a

transversal strategy.

4.1 Future Work

This work is the basis for the research project in which

the authors want to specify an Architecture Frame-

work (AF) for this area of clinical data sharing at a

large scale. Thus, the next steps involve a progressive

breakdown of this model towards the creation of that

AF as a guide to the implementation of these kind of

systems. The AF is expected to: 1) describe the model

in terms of a set of building blocks 2) show how the

building blocks ﬁt together 3) provide a common vo-

cabulary 4) include a list of recommended standards

5) include a list of compliant products that can be

used to implement the building blocks. Meanwhile, a

Quality Attribute Workshop (QAW) (Lattanze et al.,

2003) with Information Systems specialists from Por-

tuguese healthcare institutions will be organized in or-

der to validate the main strategy and the key aspects

of the model. In addition, the validation process of

the framework may include the development of a pro-

totype.

ACKNOWLEDGEMENTS

The authors would like to thank to the Portuguese

Ministry of Health, specially to SPMS which has been

always very supportive. Also, thanks to FCT for fund-

ing this research (SFRH/BD/90838/2012).

ICEIS2014-16thInternationalConferenceonEnterpriseInformationSystems

142

REFERENCES

Anderson, G. F., Frogner, B. K., Johns, R. a., and Reinhardt,

U. E. (2006). Health care spending and use of infor-

mation technology in OECD countries. Health affairs

(Project Hope), 25(3):819–31.

Artmann, J. and Giest, S. (2010). Country Brief: France.

(October).

Atalag, K., Kingsford, D., Paton, C., and Warren, J. (2010).

Putting Health Record Interoperability Standards to

Work. electronic Journal of Health Informatics,

5(1):1–17.

Ball, M. J., Smith, C., and Bakalar, R. S. (2007). Per-

sonal health records: empowering consumers. Jour-

nal of healthcare information management : JHIM,

21(1):76–86.

Brennan, S. (2005). The NHS IT project: the biggest com-

puter programme in the world - ever! Radcliffe Pub-

lishing Ltd; 1 edition (April 2005).

Canada Health Infoway (2009). Building a Healthy Legacy

Together - Annual Report 2008/2009. Technical re-

port.

Chen, H.-M., Kazman, R., and Garg, A. (2005). BITAM:

An engineering-principled method for managing mis-

alignments between business and IT architectures.

Science of Computer Programming, 57(1):5–26.

Clegg, C. and Shepherd, C. (2007). The biggest computer

programme in the world. . . ever!: time for a change

in mindset? Journal of Information Technology,

22(3):212–221.

Coiera, E. W. (2007). Lessons from the NHS National Pro-

gramme for IT. The Medical journal of Australia,

186(1):3–4.

Connecting for Health (2003). The Personal Health Work-

ing Group Final. Technical report, Markle Fundation.

Edwards, J. (2006). Case Study : Denmark’s Achievements

With Healthcare Information Exchange. Gartner In-

dustry Research Publication, (May).

epSOS (2012). epSOS: Patient Summary.

http://www.epsos.eu/epsos-services/patient-

summary.html. [Accessed: 11/11/2013].

Gunter, T. and Terry, N. (2005). The emergence of national

electronic health record architectures in the United

States and Australia: models, costs, and questions.

Journal of Medical Internet Research, 7(1).

Haux, R. (2006). Health information systems - past,

present, future. International journal of medical in-

formatics, 75(3-4):268–81.

Jha, A. K., Doolan, D., Grandt, D., Scott, T., and Bates,

D. W. (2008). The use of health information technol-

ogy in seven nations. International journal of medical

informatics, 77(12):848–54.

Josey, A. T. O. G. (2011). TOGAF Version 9.1 Enterprise

Edition. Group, pages 1–13.

Kazman, R. and Chen, H.-M. (2009). The metropolis model

a new logic for development of crowdsourced sys-

tems. Communications of the ACM, 52(7):76.

Kuo, M.-h., Kushniruk, A., and Borycki, E. (2010). Ad-

vances in Electronic Health Records in Denmark :

From National Strategy to Effective Healthcare Sys-

tem Implementation. In EFMI Special Topic Confer-

ence in Reykjavik, pages 96–99.

Lattanze, A. J., Stafford, J. A., and Weinstock, C. B.

(2003). Quality Attribute Workshops, Third Edition

(CMU/SEI-2003-TR-016). Technical Report August.

Legislative Audit Ofﬁces (2010). Electronic Health Records

in Canada - An overview of federal and provincial au-

dit reports. Technical report.

Maier, M. W. (1998). Architecting principles for systems-

of-systems. Systems Engineering, 1(4):267–284.

McGrail, K., Law, M., and H

ebert, P. C. (2010). No

more dithering on e-health: let’s keep patients safe in-

stead. CMAJ : Canadian Medical Association journal,

182(6):535.

Northrop, L., Feiler, P., Gabriel, R., Goodenough, J.,

Linger, R., Longstaff, T., Kazman, R., Klein, M.,

Schmidt, D., Sullivan, K., and Wallnau, K. (2006).

Ultra-Large-Scale Systems: The Software Challenge

of the Future. Software Engineering Institute,

Carnegie Mellon University, Pittsburgh, PA.

Overdick, H. (2007). The Resource-Oriented Architecture.

2007 IEEE Congress on Services (Services 2007),

pages 340–347.

Powell, J. (2004). NHS national programme for informa-

tion technology: changes must involve clinicians and

show the value to patient care. BMJ: British Medical

Journal, 328(7449):1200.

Protti, D. and Johansen, I. (2010). Widespread adoption

of information technology in primary care physician

ofﬁces in Denmark: a case study. Issue brief (Com-

monwealth Fund), 80(March).

Protti, D., Johansen, I., and Perez-Torres, F. (2009). Com-

paring the application of Health Information Tech-

nology in primary care in Denmark and Andaluc

ıa,

Spain. International journal of medical informatics,

78(4):270–83.

Sessions, R. (2007). A Comparison of the Top Four

Enterprise-Architecture Methodologies.

Sommerville, I., Cliff, D., and Calinescu, R. (2012). Large-

scale complex IT systems. Communications of the

ACM, 55(7):71–77.

Srinivasan, L. and Treadwell, J. (2005). An overview of

service-oriented architecture, web services and grid

computing. HP Software Global Business Unit, 2.

Stroetmann, K., Artmann, J., and Stroetmann, V. N. (2011).

European countries on their journey towards national

eHealth infrastructures. Technical Report January.

The Danish Ministry of Health (1996). A Danish Health

Care Data Network In Two Years. Technical report.

Wanless, D., Appleby, J., and Harrison, A. (2007). Our

future health secured. A review of NHS funding.

Webster, P. C. (2011). Centralized, nationwide electronic

health records schemes under assault. CMAJ : Cana-

dian Medical Association journal, 183(15):E1105–6.

Zachman, J. a. (1987). A framework for information sys-

tems architecture. IBM Systems Journal, 26(3):276–

292.

ArchitecturalKeyDimensionsforaSuccessfulElectronicHealthRecordImplementation

143