A Practical Medical Experience of Successfully Mixing Model-Driven

Paradigm and Business Process Management Principles

V. Cid-de-la-Paz

, L. Morales

and J. M. Ramos

IWT2 Group, University of Seville, Spain

SOLTEL S.A., Spain

Keywords: Clinical Decision Support, Model Driven Engineering, Clinical Practices Guidelines.

Abstract: The Model-Driven Paradigm has been successfully used in several different software contexts and there are

a lot of literature offering approaches, techniques and tools to guarantee its application in different areas,

such as software design, software testing, and so on. But, this paradigm can be also used in other contexts

offering very good results. In this paper, we illustrate the power of using models and transformations to

make an effective and efficient management of clinical guides in medical environments. The paper shows

how using business process management to represent clinical guidelines, principles of Model-Driven

paradigm can be successfully used. The paper presents the experiences in the IDE4ICDS, which is framed

into the medical context to provide a solution to manage the life cycle of clinical guidelines. This project

presents a methodology that allows the management of clinical guidelines to be automated, as well as a

software platform to support it. This platform has been validated with health professionals from the Hospital

Virgen del Rocio (Seville), obtaining promising results. Nowadays, this platform is been validated by

healthcare professionals of Primary Care with patients suffering from Diabetes Mellitus Type 2.

1 INTRODUCTION, CONTEXT

AND NEED

BPM (Business Processes Management) could be

considered as a management strategy that includes

methods, techniques and tools to support the

business processes lifecycle, which includes design,

enactment, management and analysis of operational

BPs (Van-der-Aalst.2002). BPM aims to reduce

costs and improve processes (through a cycle of

continuous improvement) in many organizations

(PMI.2008). In fact, some studies, such as

(ISO/IEC.ISO 9001.2008), conclude that the reasons

for adopting BPM can be grouped into three main

needs: (i) understand and assimilate the intrinsic

knowledge of processes; (ii) know the employees’

performance during the execution processes; and

(iii) monitor and measure processes. Controlling

these needs improves ROI (Return on Investment)

parameter through reducing production costs

(Trkman P. 2010) in any kind of organization.

There are also many institutions that promote, by

means of their standards and guidelines, the

application of BPM as a process-oriented

mechanism to improve productivity,

competitiveness, quality and efficiency at

organizations (OMG.2011, Martinez-Ruiz T et al.

2008, ISO/IEC.ISO/IEC TR24744.2007, Ponce J et

al.2013). Such parameters have been followed by a

large number of companies in all areas of business.

Healthcare organizations are not an exception and,

in fact, the Healthcare Process (HP) management is

essential to ensure adequate patient care, as well as

facilitate the work of healthcare professionals in an

area where it is essential to take of decisions based on

the best available biomedical knowledge. This

practice is known as Evidence-Based Medicine

(EBM) (Tonelli M et al.2018) and it is applied within

a framework of quality of care, patient safety and

efficiency. The management of HP is also closely

related to the term of clinical guidelines, which are

usually are textual and systematic statements of

information on HP, clinical records, recommendations

and clinical decision rules.

The Clinical Guidelines (CG) themselves aim to

improve the quality and safety of patient care,

reduce variability in clinical practice and reduce

healthcare costs. In this context, in recent years,

different research groups in the field of medical

410

Cid-de-La-Paz, V., Morales, L. and Ramos, J.

A Practical Medical Experience of Successfully Mixing Model-Driven Paradigm and Business Process Management Principles.

DOI: 10.5220/0007484704100416

In Proceedings of the 7th International Conference on Model-Driven Engineering and Software Development (MODELSWARD 2019), pages 410-416

ISBN: 978-989-758-358-2

informatics have developed conceptual models to

represent executable CG. However, the

implementation and maintenance of these conceptual

models in software systems are hard, complex and

expensive tasks because these systems are usually

designed and developed ad hoc what implies

hindering interoperability among organizations, the

lack of standardization, and increasing the inter-

center or even inter-professional variability.

Moreover, at present, there are not

methodological frameworks that define how

managing comprehensively all phases of the CG

lifecycle: from its modeling (including its HP,

clinical records, clinical rules, etc.) to its execution,

monitoring, evolution, and integration with

Healthcare Information Systems (HIS) and Clinical

Devices of Patients (CDP, such as glucometer, blood

pressure monitor, clinical sensors, etc.). For this

reason, designing and developing software systems

to CG management are complex task and, in many

cases, these systems are developed ad hoc by each

health organization that decides to implement a

specific CG what implies that the same CG can be

developed (and evolved) in different ways in

different health organizations.

In this context, it is necessary to research,

propose and define mechanisms to ensure the correct

and successful execution and management of CG in

order to reduce developing costs of software systems

and reduce the variability of application of CG in

similar medical situations in different patients.

However, it is important to mention that although

these systems are defined correctly, it is necessary to

give solution to another important aspect related to

the maintenance of CG because this one evolves

frequently. Consequently, proper management of

change, maintenance of traceability between the

definition and implementation of CG, as well as the

achievement of effective continuous improvement,

become fundamental and key tasks within health

organizations.

This paper describes a practical experience in a

real R&D project which aims to propose a

technological solution to solve the previous need in

health organizations. This project (as well as its

support software platform) is named IDE

ICDS

(Integrated Developing Environment for Improving

Clinical Decision Support based on Clinical

Guidelines) and it is subsidized by the Ministry of

Economy and Competitiveness and co-financed with

FEDER funds, in the call Challenges-Collaboration

of the State Program of Research, Development and

Innovation Oriented to the Challenges of Society,

within the framework of the State Plan for Scientific

Research and Technique and Innovation 2013-2016.

The Consortium of public and private entities

that is carrying out the project are: the IWT2 Group

of the University of Seville, Soltel IT Software SLU,

Serviguide Consultoría S.L. and the GIT Group of

the FISEVI Foundation.

IDE

ICDS is based on the Model-Driven

Engineering (MDE) (Schmidt D.C.2006) paradigm

and defines a methodological framework to

comprehensively and collaboratively manage CG, as

well as systematic mechanisms for deploying and

maintaining these CG on a web platform. Moreover,

our experience in transferring knowledge to

companies confirms that naive MDE-based solutions

are more likely to succeed because they enable

designing and implementing transformation rules

(Escalona, M. J et al. 2007, García-García J.A., et

al.2015, Gutierrez, J.J, et al.2015, Dominguez, FJ et

al.2010, Dominguez, FJ et al.2014, García-García

J.A et al.2014, García-García J.A., et al.2017).

The features of IDE

ICDS are: (i) centralized,

i.e., a single nucleus of information is stored and

traceability maintained between all the components

of a clinical guide (processes, simple elements and

decision rules); (ii) integral, i.e., a single platform

(IDE

ICDS) provides modules to define, execute,

monitor and interoperate the CG with HIS and CDP;

and (iii) collaborate, i.e., if a health professional

considers it necessary to improve or evolve a CG

(based on his or her experience), he or she can do it

intuitively and friendly, and that modification could

be used by another health organization. In addition,

it is important to mention that this project will be

tested and validated in a real scenario of patients

with Type 2 Diabetes Mellitus

After this introduction, this paper is structured as

follows. Section 2 describes our proposal to clinical

guidelines management. Section 3 describes our

technological solution. Finally, Section 4 states

conclusions and introduces future lines of research.

2 THEORETICAL

FOUNDATIONS OF IDE4ICDS

BASED ON AN IMPROVEMENT

CONTINUOUS LIFECYCLE

From a general point of view, process management

could be considered a management strategy with a

Andalusian Regional Ministry of Health. Integrated

Welfare Process Diabetes Mellitus. Last access 2018.

A Practical Medical Experience of Successfully Mixing Model-Driven Paradigm and Business Process Management Principles

411

clear multidisciplinary nature, as it can be applied to

different contexts (e.g., healthcare domain) and can

be used by different user profiles (Hill J.B, et

al.2017). This situation has conditioned the

appearance of different views, definitions and

perspectives of management lifecycles as well as

their continuous improvement (Van-der-Aalst.2004),

which define a management model for continuous

business implementations and incremental problem

solving. Although clinical guidelines have more

feature than simple process, the Clinical Guideline

Management (CGM) is similar because both ones

need to be modelled, executed, orchestrated,

measured, improved, etc.

However, CGM has several particularities which

have to be properly supported by decision-support

software systems. As mentioned in previous section,

these particularities are related to improve the

maintenance and evolution of CG, streamline the

day of health professionals and reduce the variability

in the clinical practice and sanitary cost.

For this purpose, the project defines a model-driven

theoretical framework to support a continuous

improvement lifecycle of CG based on four phases

which are:

1. Modeling Phase. In this phase, healthcare

professionals can model his/her CG in a

structured manner (i.e., identifying roles,

activities of the healthcare process, clinical

rules, clinical recommendations, for instance).

For this purpose, we have carried out different

studies of international clinical guidelines to

extract and analyse its textual structure. Once

analysed health documents, we have defined a

simple, flexible and highly-semantic

metamodel (which takes the form of a MOF-

compliant metamodel) to model any aspect of

CG. MOF (Meta-Object Facility) is a set of

standard interfaces that can be used to define

and manipulate a group of interoperable meta-

models and the corresponding models.

We offer a flexible language to model CG with

two main goals: (i) facilitate the application of

MDE-based mechanisms and extensible of our

metamodel in future; and (ii) reduce users'

cognitive overload when they are utilizing our

modeling language. From a MDE perspective,

this simplicity helps us to successfully apply

our solution to health service and open new

research lines related to testing and simulating

health processes. This simplicity is not seen as

a drawback since our proposal has extension

mechanisms.

Due to the complexity of CG modelling

metamodel, below we are going to briefly

describe show the most important metaclasses.

However, Figure 1 shows the general

metamodel of CG. This metamodel is related to

2 secondary metamodels: «ClinicalProcess»

and «ClinicalRule». These metamodel are not

explained in detail but can be consulted in

(García-García et al. 2015; García-García et al.

2018).

Before going further, it is worth clarifying that

the syntax used is not enough to semantically

define our metamodel. In fact, we use OCL

(Object Constraint Language) (28) to add

formal semantic constraints that validate

process models.

The metaclass «ClinicalPracticeGuideline» is

the epicenter of the metamodel and is the

element around which the rest of the elements

of the metamodel are orbiting. With this

metaclase it is possible to represent any CG.

To describe the necessary actions since a

person, with a certain pathology, requests

assistance until it ends, each CG is composed

of a clinical process «ClinicalProcess». To

model these actions, we have activities (human,

automatic and complex) and gateways. In

human activities, the health professional

registers patient data. For it, we associate the

metaclass «Variable».

In addition, given that one of the objectives of

this project is to monitorize the CG, with the

metaclass «Indicator» we can associate

indicators to activities and processes. These

indicators, when the GC is running, will record

the values and will be displayed in the

monitoring module.

To help in making clinical decisions in

activities or to choice with path of actions the

patient should follow, we associated the

metaclass «ClinicalRule» to activity and

gateway. These rules are formed by clauses,

constituted in turn by the variable to be

evaluated, a mathematical operator and the

value with which to compare. Logical

operators are used to relate one clause to

another.

2. Execution and Orchestration Phase.

Nowadays, this phase could be considered as

critical and essential task because health

organizations are being driven by the need to

extensively and continually automate, evolve

and maintenance their CG. For this purpose,

MODELSWARD 2019 - 7th International Conference on Model-Driven Engineering and Software Development

412

CG have to include execution information (i.e.,

such as execution parameters for the communi-

cation and integration with external systems,

for instance). This information are essentials to

deploy and execute CG models into execution

engines, such as process engines or BPMS

(Business Process Management Suite (Meidan

A, et al.2017)), rule decision engines, etc.

IDE4ICDS provides MDE mechanism to solve

this situation. This mechanism allows systema-

tically generating the executable version of CG

from its model (previous phase). It is based on

model-to-text (M2T) transformation rules using

MOFM2T (OMG.MOF.2017). This transfor-

mation protocol will not be explained here,

since they are out of the scope of this paper and

it would become too extensive, but it is

possible to find further information of

application in other context into (García-García

JA, et al.2017). Anyway, we have been able to

generate executable code (based on BPMN-

XML and Java code) from the definition model

of the GC. On the one hand, BPMN-XML code

is generated because most BPMSs support this

standard format (Meidan A, et al.2017) and it

is used but the process engine selected in our

project (see Section 3). On the other hand, we

generate Java code to execute each clinical

decision rule (which is modelled in previous

phase) in the decision rule engine selected in

our project (see Section 3).

Although these kinds of code are related to our

design and technological solution, it is

important to mention that our MDE-based

framework is independent of the platform.

3. Monitoring Phase. Once CG and its

healthcare processes are modelled and

executed, it is time to evaluate its effectiveness.

This evaluation provides a granular view of the

overall productivity of each CG and it is based

on the definition of key performance

indicators.

Figure 1: Metamodel of Clinical Guideline.

«metaclass»

ClinicalPracticeGuideline

background: String

introduction: String

lastupdate: Date

scopeAndGoal: String

subtitle: String

title: String

version: Float

«metaclass»

Auth orshi p

category: CategoryKind

description: String

speciality: String

work ing Group: S tring

«metaclass»

Formu latedQuestion

description: String

«metaclass»

ScientistEvidence

assessmentCriteria: String

description: String

organization: String

«metaclass»

ClinicalR ule

«metaclass»

Disease

classification: String

definition: String

diagnosticCriteria: String

«metaclass»

Str ategy

description: String

type: StrategyKind

«metaclass»

Indicator

description: String

formula: String

«metaclass»

Reference

description: String

id: String

«metaclass»

Appendix

description: String

id: String

subtitle: String

title: String

«metaclass»

FutureWork

description: String

«metaclass»

Variable

value: float

variableDescription: String

variableName: String

variableTy pe: VariableKind

variableUnit: VariableUnit

«metaclass»

ClinicalPr ocess

1..*

isBasedOn

1..*

isAssociatedTo

1..*isProducedBy

1..*

isBasedOn

0..*

includes

0..*

isComposedBy

1..*

isAssociatedTo

0..*

isAssociatedTo

0..*

follows

0..*

formulates

1..*

0..*

isComposedBy

0..*

isComposedBy

A Practical Medical Experience of Successfully Mixing Model-Driven Paradigm and Business Process Management Principles

413

In this case, we have included two mechanisms

in order to support this phase in the IDE

ICDS

platform. Firstly, our CG modeling metamodel

includes concepts (such as metric and

indicator) that help the healthcare managers to

measure each CG. Indicators are defined and

configured during the modeling phase.

Secondly, and once modelled CG and

indicators, these models are systematically

transformed to in executable code using M2T

transformation rules. These executable code is

composed of: (i) SQL (Structured Query

Language) scripts, which update the

measurement database of IDE

ICDS; and (ii)

code scripts, which connect each model of the

CG (i.e., definition model, execution and

orchestration model, indicator model, etc.)

between itself, and calculate each defined

indicator.

4. Continuous Improvement Phase. This phase

aims to achieve higher quality, efficiency,

effectiveness and performance levels during

CG execution what imply to improve the

patient assistant. This improvement of CG

could emerge from two situations. The first one

could success after evaluating CG performance

(through assessment indicators and metrics),

i.e., a healthcare organization could start an

internal improvement process to improve its

services to patients and increase or optimize its

resources. The second situation could append

when healthcare professionals identify any

improvement in the definition model of the

CG. Anyway, after appending any these

situations, organization could iterate over our

CG lifecycle as many times as necessary in

order to achieve its goals.

3 THEORETICAL

FOUNDATIONS OF IDE4ICDS

BASED ON AN IMPROVEMENT

CONTINUOUS LIFECYCLE

Previous section has briefly presented our theoretical

framework to make easier the CG management.

However, it is required to offer a tool-based

mechanism to support this framework in order to

reduce costs and improve its applicability in real

healthcare environments during the definition,

design, implementation and validation of clinical

guidelines. The IDE4ICDS platform has been

designed and developed to achieve the previous goal

under a service-oriented architecture with a user-

centered design. For this purpose, our platform

provides five functional modules (Figure. 2) as

follows.

The first one is the Definition & Traceability

Module (M1), which provides three graphic editors

to model each aspect of a GC (i.e., general

information, healthcare processes, clinical

information records and clinical decision rules).

These editors has been implemented as plugins on

Enterprise Architect (EA)

. These plugins have

different functionalities, such as, UML Profiles to

friendly instance our GC metamodel (Figure. 1),

plugin to guarantee well-defined models verifying

each OCL constraint of metamodel, plugin to

automatically execute each transformation rule, etc.

In addition, it is important to emphasize that these

models are stored in a central repository of models

(supported on a MySQL database).

The second module is the Execution Module

(M2), where the executable version of the CG is

systematically deployed and, later, executed by

healthcare professionals. Once modelled a GC, the

deployment method of this one has to be carried out

in two steps. Firstly, process engineer applies

transformation rules from M1 to automatically

obtain BPMN-XML and Java code which are used

Figure 2: Conceptual model of the IDE4ICDS platform.

SparxSystems. Enterprise Architect. Website. Last

access 2018.

MODELSWARD 2019 - 7th International Conference on Model-Driven Engineering and Software Development

414

by BonitaOS

(process engine) and Drools

(decision

rule engine), respectively. These tools were chosen

according to the requirements of the project and

taking into account the characterization scheme

proposed in (Meidan A, et al.2017), which was very

useful. However, it is important to mention that our

MDE-based framework is independent of the

platform. Therefore, other engines could be chosen

within great efforts. Secondly, process engineer has

to configure and compile these codes manually. This

configuration includes to set execution parameters

with HIS and CDP. The configuration of this

information cannot be automated because it depends

on the specific system and chosen device.

Moreover, the fourth module is the Monitoring

Module (M3). This module has been specially

developed for the project using web technology

(cakePHP, HTML5, CSS3, jQuery, etc.). Once

modeled a GC using the M1 module, the process

engineer can automatically execute the set of M2T

transformations to generate cakePHP code from

definition models of the GC. This cakePHP code

allows calculating each indicator associated with the

healthcare process (which is defined in the clinical

guideline). For this purpose, the M3 module has to

be communicated to the M2 module in order to

calculate indicators during the execution of the GC

(such as number of patients attended, percentage of

emergencies, average time of patient care, maximum

time of patient care, etc.). Once calculated each

indicator, this module provides scorecards, timelines

with the evolution of each indicator, and

alarms/notifications, among other functions.

As mentioned above, M3 and M2 have to be

communicated. For this purpose, we have designed

the Integration Module (M4), which is developed

as an enterprise services bus based on OpenESB

M4 allows internally connecting each module with

another one, as well as externally communicating

the M2 module with HIS and CDP.

Finally, the last module is the Traceability

Module (M5). An important aspect when MDE is

used is to ensure traceability among generated

models. This is essential in the context of IDE

ICDS

because it allows identifying each GC in a unique

way in the platform and its modules. In addition,

traceability allows enhancement points such as

version management of a GC.

BonitaSoft. Website. Last access 2018.

Drools. Website. Last access 2018.

OpenESB. Website. Last access 2018.

4 CONCLUSIONS AND FUTURE

WORKS

Today’s world economic situation is ruled by issues

such as reducing cost, improving quality,

maximizing profit and improving and optimizing

processes at any kind of organization. In this

context, BPM have been confirmed as an essential

and successful strategy. However, over last years,

research community have started to combine the

process management and MDE within controlled

contexts as software testing processes, requirement

processes, etc.

However, the application of MDE in health

contexts is a work and research line that has aroused

the interest of research teams and organizations

interested in transferring research results to real

environments. In addition, the application of MDE

in health environments is also a new research line,

little treated over last years and innovative in terms

of potential results that allows its application. In this

context, this paper has presented the IDE

ICDS

project which proposes a MDE-based solution to

solve two main goals: (i) improving the application

and management of CG in real healthcare

environments; and (ii) reducing variability in clinical

practice during the application of a specific CG, as

well as reducing healthcare costs. At present, this

project will be tested and validated in a real scenario

of patients with Diabetes Mellitus, but we plan to

extend and apply our solution to other medical

pathologies as future works.

ACKNOWLEDGEMENTS

This research has been partially supported by

POLOLAS project (TIN2016-76956-C3-2-R) and

IDE

ICDS (RTC-2016-5824-1) of the Spanish

Ministry of Economy and Competitiveness, and by

the VI PPIT-US of the University of Seville (Spain).

REFERENCES

Domínguez-Mayo, F. J., Escalona, M. J., & Mejías, M.

(2010, July). QuEF (quality evaluation framework) for

model-driven web methodologies. In International

Conference on Web Engineering (pp. 571-575).

Springer, Berlin, Heidelberg.

Domínguez-Mayo, F. J., Escalona, M. J., Mejías, M.,

Ross, M., & Staples, G. (2014). Towards a

homogeneous characterization of the model-driven

A Practical Medical Experience of Successfully Mixing Model-Driven Paradigm and Business Process Management Principles

415

web development methodologies. Journal of web

engineering, 13(1-2), 129-159.

Escalona, M. J., Gutierrez, J. J., Villadiego, D., León, A.,

& Torres, J. (2007). Practical experiences in web

engineering. In Advances in Information Systems

Development (pp. 421-433). Springer, Boston, MA.

Garcia-Garcia, J. A., Meidan, A., Carreño, A. V., &

Risoto, M. M. (2017, October). A Model-Driven

Proposal to Execute and Orchestrate Processes:

PLM4BS. In International Conference on Software

Process Improvement and Capability Determination

(pp. 211-225). Springer, Cham.

García García, J. A., Escalona, M. J., Martínez-García, A.,

Parra, C., & Wojdyński, T. (2015). Clinical Process

Management: A model-driven & tool-based proposal.

Information Systems Development: Transforming

Healthcare through Information Systems. ISBN: 978-

962-442-393-8. 2015.

Garcia-Garcia, J. A., Enríquez, J. G., Garcia-Borgoñon, L.,

Arevalo, C., & Morillo, E. (2017, July). A MDE-based

framework to improve the process management: the

EMPOWER project. In 2017 IEEE 15th International

Conference on Industrial Informatics (INDIN) (pp.

553-558). IEEE.

García-García J. A., García-Borgoñón L., Escalona MJ,

Mejías M. A model-based solution for processes

modeling in practice environments: PLM4BS. Journal

of Software-Evolution and Process. pp.1-23.

https://doi.org/10.1002/smr.1982 2018.

García-García, J. A., Escalona, M. J., Domínguez-Mayo,

F. J., & Salido, A. (2014). NDT-Suite: A

methodological tool solution in the Model-Driven

Engineering Paradigm. Journal of Software

Engineering and Applications, 7(04), 206.

Gutiérrez, J. J., Escalona, M. J., & Mejías, M. A Model-

Driven approach for functional test case generation.

Journal of Systems and Software, 109, 214-228. 2015.

Hill J. B., et al. Cool Vendors in Business Process

Management, Gartner Research, 2007.

ISO/IEC.ISO 9001:2008. Quality management systems,

Requirements. International Organization for

Standardization, 2008.

ISO/IEC.ISO/IEC TR24744:2007 Software and systems

engineering lifecycle management guidelines for

process description. International Organization for

Standardization, 2007.

Martínez-Ruiz T., et al. Towards a SPEM 2.0 extension

define process lines variability mechanisms. Software

engineering research, management & applications,

115–130, 2008.

Meidan, A., García-García, J. A., Escalona, M. J., &

Ramos, I. (2017). A survey on business processes

management suites. Computer Standards & Interfaces,

51, 71-86.

OMG. BPMN, Business Process Modelling Notation,

Version 2.0. 2011.

OMG.MOF Model to Text Transformation Language

(MOFM2T). Website. 2017.

PMI. A Guide to Project Management Body of

Knowledge. ISBN: 978-1933890517. 2008.

Ponce, J., Borgoñon, L. G., García, J. G., Escalona, M. J.,

Domínguez-Mayo, F. J., Alba, M., & Aragon, G.

(2013). A model-driven approach for business process

management. Covenant Journal of Informatics and

Communication Technology, 2(1).

Schmidt D.C. Model-Driven Engineering. IEEE

Computer, 39, 2, 25-31, 2006.

Tonelli M., et al. The role of experience in an evidence-

based practice. The Medical Roundtable General

Medicine Edition, 2018.

Trkman, P. The critical success factors of business process

management. International Journal of Information

Management, 30(2),125-134. 2010.

Van-der-Aalst WMP. Business process management: a

personal view. Business Process Management

Journal, vol.10 (2), 2004.

Van-der-Aalst WMP. Making Work Flow: On the

Application of Petri Nets to Business Process

Management. LNCS, vol. 2360, Application and

Theory Petri Nets, pp 1-22. 2002.

MODELSWARD 2019 - 7th International Conference on Model-Driven Engineering and Software Development

416