ADAPTIVE CLINICAL PATHWAYS WITH

SEMANTIC WEB RULES

Dimitrios Alexandrou, Fotis Xenikoudakis and Gregoris Mentzas

Information Management Unit, Electrical and Computer Engineering, National Technical University of Athens

9 Iroon Politechniou Street, Athens, Greece

Keywords: Adaptive clinical pathways, semantic web, semantic rules, pathway adaptation, ontology, workflow engine.

Abstract: The increase of treatment quality offered by the healthcare organizations is one of the main challenges of

the modern health informatics. The personalization of treatment presupposes the real-time adaptation of

treatment schemes since the clinical status of the patient and circumstances inside a healthcare organization

constantly change. In this paper we present SEMPATH prototype which aims at providing a solution

concerning the real-time adaptation of healthcare business processes. The prototype consists of a healthcare

process execution engine assisted by a semantic framework for the adaptation. The semantic framework

consists of an ontology enclosing the required knowledge based on which a semantic rule set was created.

During the execution time of the clinical pathways, the system reasons over the rules, the knowledge and

information collected, and provides decisions and recommendations for the next steps of the treatment.

Moreover, the results of the rule-set execution may produce new knowledge objects which are inserted in

the ontology.

1 INTRODUCTION

One of the main challenges of modern healthcare

organizations is to increase the treatment quality. In

order to achieve their goal, they need to utilize

standardized clinical protocols used in many

domains of medicine. Such a protocol contains

detailed medical plans for diagnosis, therapy scheme

and follow-up. Moreover, it encloses the information

required so as to deal with exceptional situations,

which occur during the treatment execution time and

require quick and appropriate modifications of the

treatment of a patient, thus increasing the flexibility

of the treatment processes. One valuable tool to

achieve the above-mentioned objectives is “Clinical

Pathways”.

Clinical pathways can be utilized for the

implementation of medical guidelines in a specific

healthcare environment and decrease undesired

variability of medical practice (Campbell et al.,

1998). In contradiction with the medical guidelines,

clinical pathways enclose multidisciplinary valuable

resources like personnel, education level, medical

equipment availability and other operational and

administrative information. Medical guidelines

require the consensus between medical experts. On

the other hand, clinical pathways require a

consensus between multidisciplinary groups of

hospital personnel taking actions during the

treatment execution. Clinical pathways constitute

treatment process patterns which aim to increase

both the healthcare process quality and the

utilization of resources. Consequently, a clinical

pathway may deviate from a clinical guideline due

to administrative reasons, and a treatment scheme

may deviate from the clinical pathway due to

patient’s symptoms during its execution.

In order to support the execution of treatment

schemes based on clinical pathywas and to relief the

medical personnel, a software system is required

which will handle the healthcare business processes

in an efficient manner (Greiner et al., 2004). Such a

system would be responsible for the observation of

the execution and the current status of the apllied

clinical pathways, offer the characteristic of

automatic recognition of exceptional events and

provide decision support services in order to handle

the exceptions in an efficient and effective way.

Moreover, the software system should be capable to

dynamically adapt the treatment process so as to

control the appropriate modifications.

140

Alexandrou D., Xenikoudakis F. and Mentzas G. (2008).

ADAPTIVE CLINICAL PATHWAYS WITH SEMANTIC WEB RULES.

In Proceedings of the First International Conference on Health Informatics, pages 140-147

 SciTePress

In this paper we propose an approach which

includes a workflow management system combined

with a rule base in order to handle the

abovementioned requirements. The workflow

environment handles the execution of treatment

schemes and the incorporation of user types, data

and peripheral applications. Additionally, the

specific software system needs to support the

dynamic adaptation of clinical pathways in order to

handle the flexibility of treatment schemes (Dadam

et al., 2000), (Miksch et al., 2001). The rule base is

responsible for the handling of the required streams

of knowledge enclosed in the clinical pathways and

is utilized for the detection of exceptional events and

their confrontation.

In this paper we present our software prototype,

SEMPATH, which follows this approach and has all

the required functionality to support adaptive

clinical pathways. SEMPATH performs a rule-based

exception detection with semantic rules (SWRL) and

dynamic clinical pathway adaptation during the

execution time of each pathway.

The rest of the paper is organised as follows.

Section 2 refers to our motivations and related work

performed in the area of our interest. Section 3

overviews the proposed Semantic Approach the

SEMPATH follows, while Section 4 outlines the

SEMPATH conceptual framework and technical

architecture which is being implemented. In section

5 we present our experimental scenario and the

SEMPATH walkthrough. Finally, section 6

concludes the paper combined with our thoughts for

future work.

2 MOTIVATIONS AND RELATED

WORK

At this section we present the motivations, the

related work and our contribution in the area of the

adaptive clinical pathways. The motivations

presented led to the research stream of adaptive

clinical pathways. Moreover, a significant amount of

work has been realized towards the direction of the

optimal handling of the exceptions occurring during

the execution of treatment schemes of a patient.

Finally, our research in the area and the development

of SEMPATH prototype tries to contribute in

specific and focused issues.

2.1 Motivations

The trends in healthcare business processes and their

establishment and utilization in the healthcare

routine are up to now quite mature. Nevertheless,

there are several open issues / challenges that further

motivate our effort (Song et al., 2006):

 Clinical Pathways Adaptability: The

traditional clinical pathways are normally

static and lack of dynamicity. Moreover, they

are standard procedures applicable to a patient

taxonomy not addressing the case of each

patient. Moreover, they do not take under

consideration the most current medical,

operational, and financial knowledge (Colaert,

2007).

 Maintenance: The implementation of Clinical

Pathways is based on medical guidelines and

additional types of knowledge. The

maintenance of the healthcare business

process suffers from the continuous update,

since both the medical guidelines and the

circumstances inside a healthcare organization

change constantly.

 Medical Guidelines Formalization: The

formalization of medical guidelines is being

performed in a specific and per case manner.

Their formalization is required since their

parameters will be able to be processed by an

IT infrastructure that supports their execution.

 Clinical Pathways Modelling: The modelling

of Clinical Pathways lacks a formal structure.

Different approaches exist in the area of

modelling. Their interoperation could be of

major importance since the Clinical Pathway

exchange between healthcare organizations

could facilitate the execution of the treatment

schemes utilized.

 Real-time information capturing:

Information capturing consists one of the

major factors for success of the treatment

scheme executed for each patient. The lack of

real-time information feed to the clinical

pathway creates a major need, since the

information collected could lead to major

reconfigurations of the executed Clinical

Pathway.

 Real-time knowledge recycling: The

knowledge recycling during the execution of a

Clinical Pathway constitutes one of the major

challenges for the area. The knowledge

feedback would be valuable since the

knowledge update is able to redefine the

Clinical Pathway and the model of the

exception rules.

ADAPTIVE CLINICAL PATHWAYS WITH SEMANTIC WEB RULES

141

2.2 Related Work – State of the Art

As (Lenz et al., 2006) states, “healthcare processes

require interdisciplinary cooperation and

coordination”. Towards this direction, he divides the

processes inside a healthcare organization into two

categories: the organizational processes and the

medical treatment processes. The organizational

processes are of equal importance to the medical

treatment ones, since they heavily affect their

execution and effectiveness. Moreover, the medical

treatment processes are influenced by the medical

knowledge and the patient information. So, he

introduces the need for WfMS (Workflow

Management System) inside a healthcare

organization so as to handle the intra-organizational

processes. Moreover, the addition of the appropriate

web-services could lead to the inter-organizational

healthcare processes. The abovementioned concerns

led to the implementation of ADEPT system

(

Reichert et al., 2003) which focuses on the healthcare

processes execution. The ADEPT system enables the

execution, monitoring and management of the

healthcare process running inside a healthcare

organization. Moreover, it offers the functionality of

dynamic changes in the predefined healthcare

processes on execution time. The development of

the specific system lasted for some years and

provided valuable information and experience from

its pilot and productive periods (Lenz et al., 2007),

(Blaser et al., 2007).

Additionally, (Colaert, D., 2007) introduces the

term of adaptive clinical pathways and presents the

research work performed inside Agfa Healthcare. He

stresses out that Adaptive Clinical Workflows are

based on a) Medical, b) Practice, c) Clinical and d)

Operational Knowledge. Agfa constitutes one of the

active members of W3C Semantic Web Health Care

and Life Sciences Interest Group (W3C) which

encloses the “Adaptive Healthcare Protocols and

Pathways Task Force” which aims at the utilization

of semantic web technologies in order to enhance

the adaptable clinical protocols and pathways.

(Abidi and Chen, 2006) introduce another IT

platform that enables the adaptivity of clinical

pathways based on a semantic framework.

CAREPLAN system (Abidi and Chen, 2006) tries to

combine heterogeneous healthcare knowledge

sources with the available patient information. The

system reasons over the knowledge and adapts

standard pathways towards personalized healthcare

plans, utilizing the technology of web-services for

the composition of the integrated pathways.

2.3 Our Contribution

Our approach led to the creation of the SEMPATH

software system which enables the adaptation of

clinical pathways in order to serve the

personalization of the treatment plans for each

patient. Our contribution concerning the state-of-the-

art in the specific domain could be summarized in

the following axes:

 Real-time adaptation of clinical pathways:

SEMPATH approach is based on continuous

reasoning over the current knowledge so as to

adapt each step of the clinical pathway under

execution.

 SWRL Rule Base: SEMPATH encloses a rule-

set created by utilizing SWRL (SWRL)

language in order to integrate the rule-base

with the ontology. The rule-base is able to

create new facts and update the ontology

accordingly, thus creating new knowledge as

each pathway evolves. This feedback

constantly updates the knowledge stored in the

ontology and leads to better results concerning

the adaptation of the pathway.

 Establishment of a meta-model for each

clinical pathway: in our approach, we define

a meta-model for each clinical pathway to be

executed. The meta-model encloses atomic

and complex sub-pathways which are fed to

the process execution engine. The integration

of discrete parts and connections results to the

establishment of the meta-model of the

pathway to be executed.

3 THE SEMPATH APPROACH

The SEMPATH approach followed during the

implementation of the software prototype includes

the design and creation of the Adaptive Clinical

Pathway Ontology. The created ontology is utilized

for the implementation of the semantic web rules.

The implemented rules are inserted into the JESS

System to comprise the rule-base in order to be

executed and extract the appropriate facts that

influence the adaptation of the clinical pathways.

3.1 Adaptive Clinical Pathway

Ontology

The Adaptive Clinical Pathway Ontology constitutes

the main infostructure of the semantic layer of the

implemented architecture. The core of the ontology

is based on the ACPP Ontology (ACPP Ontology)

HEALTHINF 2008 - International Conference on Health Informatics

142

that is implemented by ACPP Task Force (ACPP

Task Force).

The core of the ontology was further extended

and broadened both in subdomains and existing

concepts and instances. The specific ontology is

divided into four (4) main knowledge streams.

The main stream refers to the representation of

the medical knowledge. It contains the semantics

utilized for the description and structure of the

medical part of the clinical pathway. The specific

concepts describe the medical domain knowledge

that semantically enhances each clinical pathway.

Moreover, the second stream of the ontology

comprises the operational knowledge structure. It

contains the concepts utilized for the description of

the operational issues that arise during the execution

of the clinical pathway and may affect its

evolvement. Each healthcare organization encloses

specific procedures and resources which are

combined in order to offer its services.

Consequently, the operational knowledge is one of

the main elements of a clinical pathway since it

affects its execution and success.

Additionally, the third knowledge stream refers

to the concepts and terms that define the financial

issues that affect the execution of a clinical pathway.

The utilization of a clinical pathway aims at the

optimization of the financial resources required for

the treatment path of a patient. Each healthcare

organization aims at both the reduction of its costs

and the increment of the quality of healthcare

services provided. Consequently, the financial part

of the ontology models the financial resources and

rules utilized during the execution of each clinical

pathway.

Finally, the fourth knowledge stream refers to the

modelling of the clinical pathway itself. It contains

the concepts and terms that describe the building

blocks of the clinical pathway. It is utilized by the

software prototype for the design of the treatment

workflow.

SEMPATH Ontology is available online at:

http://www.imu.iccs.gr/index.php?option=com_cont

ent&task=view&id=206&Itemid=90/sempath_onto.

zip.

3.2 SWRL Rules Modelling

The rules implemented for our prototype refer to the

execution of a clinical pathway, and more

specifically to the exception handling procedure.

The exception management procedure constitutes

one of the every-day routine of healthcare

professionals (Kobayashi et al, 2005), (Tucker et al,

2002). As (Han et al., 2006) states, “an exception,

constitutes an abnormal behaviour from the normal

workflow”. The handling of exceptions occurring

during the execution of a clinical pathway encloses

three major issues: a) exception management

representation, b) implementation and execution of

exception management and c) exception analysis.

Concerning SEMPATH prototype, the rules for

exception management are designed in SWRL

format. SWRL enables the integration of the

modelled rules with the Clinical Pathway Ontology.

The interaction between rules and ontology leads to

new knowledge. An indicative set of implemented

rules for the SEMPATH prototype is presented

below:

Rule-1:

TriageAdmission(?t) ∧

DiagnosedNeurologicalDeficit(?s) ∧

hasTask(?a, ?t) ∧

hasPatient(?a, ?p) ∧

hasPatientState(?p, ?s) ∧

EvaluationForThrombolysisEligibility(?t2)

→ hasNextTask(?a, ?t2)

Rule-1 describes the following situation: if the

patient is admitted in the healthcare organization and

there is a diagnosis of Neurological Deficit, then the

patient has to be evaluated for Thrombolysis

eligibility.

Rule-2:

EvaluationForThrombolysisEligibility(?t) ∧

ThrombolysisCandidate(?s) ∧

hasTask(?a, ?t) ∧

hasPatient(?a, ?p) ∧

hasPatientState(?p, ?s) ∧

ConfirmationOfAcuteStroke(?t2)

→ hasNextTask(?a, ?t2)

Rule-2 describes the following situation: if the

Evaluation for Thrombolysis Eligibility is confirmed

and the patient is Thrombolysis Candidate, then the

next examination has to confirm or not the Acute

Stroke episode.

Rule-3:

EvaluationForThrombolysisEligibility(?t) ∧

ADAPTIVE CLINICAL PATHWAYS WITH SEMANTIC WEB RULES

143

AcuteCVA(?s) ∧

hasTask(?a, ?t) ∧

hasPatient(?a, ?p) ∧

hasPatientState(?p, ?s) ∧

EmergentHeadCT(?t2) ∧

CBC-PT-PTT_stat(?t3)

→ hasNextTask(?a, ?t2) ∧

hasNextTask(?a, ?t3)

Rule-3 describes the following situation: if the

Evaluation for Thrombolysis Eligibility is confirmed

and the patient’s state is Acute CVA, then the

patient has to be forwarded to the CT department for

Emergent CT and for CBC-PT-PTT examination.

The above-mentioned and explained SWRL rules

are indicative elements from rule-base which is

utilized by the SEMPATH infostructure. As seen

above, the rules contain classes from the ontology,

both in the antecedent and the consequent parts of

the rules. Any new facts deriving from the rules

execution are being added in the ontology as new

knowledge objects for future utilization.

3.3 Adaptation Methodology

SEMPATH adaptation methodology is based on a

meta-model clinical pathway establishment. Each

clinical pathway to be executed is a meta-model of a

set of atomic and complex sub-processes. The

atomic processes are executable parts of the

healthcare business process forwarded to the

execution engine. The complex processes are sub-

workflows which contain atomic processes and a set

of decisions. The atomic and the complex processes

are interconnected in the meta-model level. Their

connections are based on SWRL rules. Once an

atomic or complex process is executed, the rule-base

is triggered. The knowledge existing inside the

ontology, the current clinical status of the patient

and the rule-set are interoperating in order to select

the next executable part of the clinical pathway.

Thus, the adaptation occurs during each step of the

pathway execution.

During each cycle of execution, the triggering of

the rule-base may result to new knowledge creation

that will be utilized in next steps during the

execution. This fact ensures the constant update of

medical, organizational and operational knowledge

stored inside the ontology and consequently to the

rule-base.

4 SEMPATH PROTOTYPE

The following sections present the conceptual

framework and technical architecture of the software

system prototype that executes the clinical pathways

and performs the required dynamic adaptation.

4.1 Conceptual Framework

As depicted in Figure 1, the conceptual framework

of the SEMPATH system comprises three (3)

distinct architectural layers. The upper layer of the

architecture is called “Semantic Layer” since it

encloses the required semantic infrastructure.

Figure 1: Conceptual Framework.

The core of Semantic Layer is the Adaptive

Clinical Pathway Ontology which encloses the

appropriate knowledge streams required for the

modeling of the clinical pathways, in terms of

structure and content. Moreover, the specific layer

encloses the semantic modeling of the rules that

handle the exceptions inside the clinical pathway

during its execution. These rules are the cornerstone

of the dynamic adaptation performed in the clinical

pathways.

HEALTHINF 2008 - International Conference on Health Informatics

144

The second layer constitutes the “Clinical

Pathway Adaptation Layer”. The specific layer

encloses the components required for the adaptation

of the clinical pathway. The rule engine is

responsible for the execution of the semantic rules

concerning the exception handling of the pathway.

Once the result is produced the “Adaptive Clinical

Pathway” manager is the software module that

updates accordingly the structure of the pathway.

Finally, the last layer of our conceptual

framework is the “Clinical Pathway Layer”. It

contains the workflow – part of the clinical pathway.

The execution of the clinical pathway happens inside

a workflow engine since is constitutes a healthcare

business process (Lenz, 2006). A clinical pathway

repository contains a set of available clinical

pathways so as to select the most appropriate for

each patient.

4.2 Technical Architecture

The technical architecture of the implemented

prototype is presented in Figure 2. The Adaptive

Clinical Pathway Prototype technical architecture

comprises three (3) major components. These three

major components are described in full detail in the

following sections:

a. Semantic Infostructure: The core of Semantic

Infostructure component is the Clinical Pathway

Ontology. As depicted in the following diagram, the

ontology is implemented in OWL (OWL) format.

The ontology encloses the abovementioned streams

of knowledge to be utilized for (1) the creation of

rules, (2) the modelling of Clinical Pathways and (3)

the recycling of knowledge through the dynamic

production of facts by the rule engine.

The Protégé API has been utilized concerning

the implementation and maintenance of the specific

ontology. Moreover, the SWRL designer of the

semantic rules is a Protégé Plugin (SWRLTab)

which enhances the integration between the

Ontology and SWRL rules. Consequently, each rule

created by the specific plugin is consistent in

semantic terms, since the semantics required come

directly from the ontology.

b. Rule Execution Environment: This component

handles the maintenance of the semantic rules as

well as the Rule Engine implemented. The SWRL

rules implemented with the Protégé Plugin are stored

as a SWRL repository. Once the system is triggered,

the appropriate rule set is selected. The SWRL rules

are initially converted into JESS rules in order to be

executed by the rule engine. Once the semantic rules

are executed, the result of the rule engine is

produced in XML format.

The specific XML file is a custom structure

which is utilized by the prototype so as to proceed to

the adaptation of the clinical pathway.

Figure 2: Technical Architecture.

ADAPTIVE CLINICAL PATHWAYS WITH SEMANTIC WEB RULES

145

Moreover, once the rule set is executed, despite

the production of the XML result, a feedback

message is generated which contains new facts and

conclusions that update accordingly the knowledge

stored inside the ontology.

c. Clinical Pathway Execution Environment:

Finally, the last technical component of the

prototype architecture is the Execution Environment.

The core of the specific component is the workflow

execution engine which in our case is ActiveBPEL

workflow environment. The interface between the

workflow engine and the rest of the components is

the Clinical Pathway Manager. Firstly, the specific

component triggers the system once an exception

occurs during the execution of the pathway. The

message produced is forwarded to the Rule Engine

in order to run the complete rule-set for the pathway

and produce the result for the adaptation. Moreover,

the Clinical Pathway Manager interoperates with the

Result Analyser which is responsible for the

processing of the result structure.

The SEMPATH Prototype Software that we have

developed can be found here:

http://www.imu.iccs.gr/index.php?option=com_cont

ent&task=view&id=206&Itemid=90/sempath.zip.

4.3 Experimental Scenario

According to the real-life scenario, a patient

confronts a health problem and decides to be

admitted to a healthcare organization for treatment.

Once the admission is performed, an initial set of

medical examinations is decided to be performed.

The result set of the initial examinations provides

valuable information for the clinical status of the

patient which leads to the decision concerning the

selection of the appropriate clinical pathway to be

executed. The execution of the treatment scheme

produces exceptions which are handled on real-time

basis by the implemented software prototype.

More specifically, once the patient is admitted to

the healthcare organization, its IT infrastructure

should become aware of the data accompanying the

specific patient. So, an initial data entry for the

medical record dataset is performed. This procedure

is performed either manually or automatically if the

patient’s medical record is received from another

healthcare organization.

Once the clinical status of the patient is set,

SEMPATH proposes an initial set of examinations

to be performed. Afterwards, the result set of the test

is inserted into the medical record of the patient. The

system proposes an appropriate clinical pathway

according to the diagnosis. So, the execution of the

treatment scheme begins, under the constant

inspection of SEMPATH prototype. Once an

exception occurs, SEMPATH receives the exception

information, executes the required rule-set and

proposes the next ”step” of the treatment scheme.

The “step” derives from the following two

categories:

 Atomic process: a single step procedure,

executable by the process execution engine.

 Complex process: a multiple step procedure. It

is a set of atomic processes without decision making.

A complex process may contain parallel execution

paths leading to a unified result.

The above-mentioned procedure is repeated during

the execution of the treatment scheme of the patient.

This way, the personalization of treatment for each

patient is highly ensured, increasing the possibilities

for the most suitable treatment.

After the implementation of the SEMPATH

prototype, we are now in close cooperation with a

University Hospital (LAIKO Hospital). We have

modelled five (5) clinical pathways, defined the

corresponding rules and we are in the process of

real-life scenario execution.

5 CONCLUSIONS & FUTURE

WORK

The existing practices and work performed in the

area of clinical pathways has led to significant

results. Nevertheless, clinical pathways are static

procedures based on medical guidelines and

organizational and operational knowledge. The

current trend focuses on the adaptation of these

static structures in case of exception occurrence.

The introduction of semantic technology

provides further opportunities for clinical pathway

adaptation. Semantics enable the representation of

all required types of knowledge by the utilization of

the corresponding ontologies. A significant amount

of work has been done in the area of ontology

creation. The developed ontologies cover several

streams of medical knowledge.

However, this knowledge has to be combined

with rules in order to handle the pathway exceptions.

In this paper we presented SEMPATH prototype, a

software platform based on semantics. SEMPATH

introduces the creation of semantic rules in SWRL

format which provide the basis for the rule-engine

that handles the pathway exceptions. SWRL rules

were created by the utilization of the SEMPATH

ontology. The execution of the semantic rules

HEALTHINF 2008 - International Conference on Health Informatics

146

provides new knowledge objects that can be added

to the existing ontology. Consequently, the

knowledge enclosed in SEMPATH prototype is

constantly updated and maintained by its routine

operation.

Our intensions for further work can be presented

in a three-fold structure:

 Semantics Infrastructure: our main aim

concerning the evolution of SEMPATH semantics

infrastructure is to proceed with the ontology

enhancement. The enhancement will focus on

organizational issues modelling and medical

knowledge representation. Furthermore, our

intention is to integrate existing medical ontologies

so as to enrich our ontology model.

 Pathway Modelling: in the field of pathway

modelling we plan to concentrate on simultaneous

execution activities management and on providing

different views of the pathway for each type of user.

Moreover, our intention is to add semantic

information in activities that will establish a priority

weight model in order to perform more intelligent

resource and activity management.

 System evaluation and usability: since the

SEMPATH prototype is finalized and functional, we

intend to perform real-life stress tests concerning its

performance inside a healthcare organization. A

real-life test will provide valuable results concerning

the usability of the system, the performance and

further enhancement of the implemented ontology

and the further enhancement of the semantic rule-

set.

ACKNOWLEDGEMENTS

This work has been partially funded by the European

Commission with the IST 027065 contract (RIDE: A

Roadmap for Interoperability of e-Health Systems in

Support of COM 356 with Special Emphasis on

Semantic Interoperability).

REFERENCES

Abidi, S., Chen, H., 2006. Adaptable Personalized Care

Planning via a Semantic Web Framework.

ACPP Ontology, http://esw.w3.org/topic/HCLS/

ACPPTaskForce

ACPP Task Force, http://esw.w3.org/topic/HCLS/

ACPPTaskForce

Blaser, R., Schnabel, M., Biber, C., Baumlein, M., Heger,

O., Beyer, M., Opitz, E., Lenz, R., Kuhn, K.A., 2007.

Improving pathway compliance and clinician

performance by using information technology.

International Journal of Medical Informatics, Elsevier.

Campbell, H., Hotchkiss, R., Bradshaw, N., Porteous, M.,

1998. Integrated care pathways. BMJ 316, 133-137.

Colaert, D., 2007. Bringing the pieces together.Towards

semantic interoperability in e-Health Workshop, RIDE

Project, Brussels.

Dadam, P., Reichert, M., Kuhn, K., 2000. Clinical

Workflows – the killer application for process-

oriented information systems?. Proceedings of 4

International Conference on Business Information

Systems, Berlin, Springer, 36-59.

Greiner, Ul., Ramsch, J., Heller, B., Loffler, M., Muller,

R., Rahm, Er., 2004. Adaptive Guideline-based

Treatment Workflows woth AdaptFlow. Proceedings of

the Symposium on Computerized Guidelines and

Protocols, CGP 2004, Prague, IOS Press, 113-117.

Han, M., Thiery, T., Song, X., 2006. Managing Exceptions

in the Medical Workflow Systems. ICSE ‘06, May 20-

28, Shangai, China.

Kobayashi, M., Fussell, S. R., Xiao, Y., Seagull, F. J.,

2005. Work coordination, workflow, and workaround

in a medical context. Conference on Human Factors in

Computing Systems.

Lenz, R., Blaser, R., Beyer, M., Heger, O., Biber, C.,

Baumlein, M., Schnabel, M., 2007. IT support for

clinical pathways – Lessons learned. International

Journal of Medical Informatics, Elsevier.

Lenz, R., Reichert, M., 2006. IT Support for healthcare

processes – premises, challenges, perspectives. Data

and Knowledge Engineering, ELSEVIER.

Miksch, S., Kosara, R., Seyfang, A., 2001. Is Workflow

Management Appropriate for Therapy Planning?.

Proceedings of EWGLP 2000, Amsterdam, IOS Press,

53-69.

OWL, http://www.w3.org/TR/owl-features/

Reichert M., Rinderle S., Dadam P., 2003. ADEPT

workflow management system: flexible support for

enterprise-wide business processes. Proc. Int’l. Conf.

on Business Process Management (BPM’03), pp. 370–

379.

Song, X., Hwong, B., Matos, G., Rudorfer, Ar., Nelson,

C., Han, M., Girenkov, An., 2006. Understanding

Requirements for Computer-Aided Healthcare

Workflows: Experiences and Challenges. ICSE ’06,

May 20-28, Shanghai, China.

SWRL, http://www.w3.org/Submission/SWRL/

SWRLTab, http://protege.cim3.net/cgi/bin/

wiki.pl?SWRLTab

Tucker, A. L., and Edmondson, A., 2002. Managing

Routine Exceptions: A Model of Nurse Problem

Solving Behavior. Advances in Health Care

Management, 3, 87-113.

W3C, Semantic Web Health Care and Life Sciences

Interest Group, http://www.w3.org/2001/sw/hcls/

ADAPTIVE CLINICAL PATHWAYS WITH SEMANTIC WEB RULES

147