A Knowledge-based Clinical Decision Support System for Headache

Disorders Management

Maria Carmela Groccia

1 a

, Rosita Guido

1 b

, Domenico Conforti

1 c

and Rosario Iannacchero

de-Health Lab-Laboratory of Decision Engineering for Health Care Delivery, Department of Mechanical,

Energy and Management Engineering, University of Calabria, Ponte Pietro Bucci 41C, 87036 Rende (Cosenza), Italy

Headache Clinic, Neurology Department "Pugliese-Ciaccio" General Hospital,

Viale Papa Pio X 83, 88100 Catanzaro, Italy

Keywords:

Clinical Decision Support System, Headache Diagnosis, Headache Disease, Computer-aided Diagnosis.

Abstract:

Headache is one of the most common neurological problems faced by General Practitioners (GPs) and neu-

rologists. Most of GPs ﬁnd the diagnosis of headache rather difﬁcult: paper-based guidelines are long and

the diagnostic criteria are complex. Thus, many headache patients do not have an early accurate diagnosis of

headaches’ type and an appropriate treatment. In order to overcome this burden, we present a knowledge-based

Clinical Decision Support System (CDSS) speciﬁcally devoted to support GPs in the headache diagnosis and

in the appropriate selection of the diagnostic-therapeutic path. The proposed CDSS has been designed and

developed based on internationally validated guidelines and clinical protocols. The knowledge base contains

the medical-clinical knowledge appropriately formalized in several set of rules. Communication interfaces

compliant with HL7 DSS (Health Level Seven Decision Support Service) international standard were devel-

oped enabling interoperation with other healthcare applications. The CDSS has been tested and assessed in

the GPs’ daily practice of the Calabria Cephalalgic Network. During the evaluation period, a reduced number

of requests for neurological visits and unnecessary and expensive instrumental examinations was registered.

The results obtained from the evaluation period demonstrate that the CDSS turns out to be effective in the

management of headache patients.

1 INTRODUCTION

Headache disorders represent a worldwide problem,

affecting people of all ages, races and geographical

areas (World-Health-Organization, 2016). In Italy,

according to a survey by National Institute of Statis-

tics, the recurrent headache is one of the chronic dis-

eases with higher incidence. It is fourth after arthri-

tis, arterial hypertension and allergic diseases (Istat,

2015). Notwithstanding, many patients with chronic

headaches do not receive effective treatment, mainly

due to the underestimation of the impact of the disease

and to the complexity of diagnostic criteria (World-

Health-Organization, 2016). As a negative conse-

quence, headache attacks are very often self-managed

by patients that progressively increase the assump-

tion of analgesics until they reach a condition of

medication-overuse (Elkind, 1991).

https://orcid.org/0000-0001-7570-8458

https://orcid.org/0000-0003-1744-2166

https://orcid.org/0000-0002-4816-4333

The General Practitioner (GP) has a central role

in the headache diagnosis process as the GP has the

ﬁrst contact with the patient making a ﬁrst diagno-

sis. The GP can effectively treat the majority of pa-

tients with headache disorders; only a minor group

requires higher medical specializations as a neurolo-

gist or advanced instrumental examinations (Steiner

et al., 2007). For this reason, it is strategically impor-

tant that the GP be able to make an accurate diagnosis

and therefore has a deep knowledge of the diagnostic

criteria. In this way, patients can immediately get an

effective treatment, reduce pain and discomfort.

A Clinical Decision Support System (CDSS)

could get a great advantage over paper-based proto-

cols and guidelines. Paper-based guidelines to sup-

port headache diagnosis are typically long and difﬁ-

cult to memorize or refer during patient interviews. A

CDSS instead, has many advantages because it can:

(1) assists the GP for an accurate symptoms’ detec-

tion; (2) provide an accurate support in processing a

large number of data, and (3) offer a level of speci-

Groccia, M., Guido, R., Conforti, D. and Iannacchero, R.

A Knowledge-based Clinical Decision Support System for Headache Disorders Management.

DOI: 10.5220/0010238604010408

In Proceedings of the 14th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2021) - Volume 5: HEALTHINF, pages 401-408

ISBN: 978-989-758-490-9

401

ﬁcity for individual patient that is nearly hard to get

by using a paper-based guideline.

In this paper, we design, develop, and evaluate

a knowledge-based CDSS able to support the GP

both in the headache diagnosis and the choice of

best diagnostic-therapeutic path according to patient’s

symptoms. The CDSS uses a rule-based approach.

It has been designed and developed based on inter-

nationally validated guidelines and clinical protocols.

Communication interfaces compliant with HL7 DSS

(Health Level Seven Decision Support Service) inter-

national standard were developed in order to enable

the interoperation with other healthcare applications.

Results from the evaluation of the CDSS in the man-

agement of the Calabria Cephalalgic Network clinical

workﬂow are presented.

The paper is organized as follows. Section 2 in-

troduces the background on headache disorders and

the main works on CDSSs. Section 3 presents

the Calabria Cephalalgic Network healthcare context.

Section 4 describes the proposed CDSS. Section 5

presents results and a brief discussion. Finally, Sec-

tion 6 concludes the paper.

2 BACKGROUND

As reported in the International Classiﬁcation of

Headache Disorders (ICHD) published by the In-

ternational Headache Society (IHS) (IHS, 2018),

there are three main categories of headache disor-

ders: primary headaches; secondary headaches and

painful cranial neuropathies; other facial pains and

other headaches. Each category is divided into

sub-categories. The three major primary headache

sub-categories are migraine (with and without aura),

tension-type headache, and cluster headache. Sec-

ondary headaches are due to other disorders (the

so-called comorbidities) such as infections or trau-

mas cranium. In addition, a secondary headache

can be attributed to a substance or its withdrawal,

e.g. medication-overuse headache. In severe cases,

a headache is often an initial warning symptom of se-

rious diseases such as an ischaemic or haemorrhagic

stroke. In these cases, headaches appear along with

other warning symptoms.

In the literature, several tools have already been

proposed and developed to support health profes-

sionals in the diagnosis of headache disorders (Potter

et al., 2018; Aljaaf et al., 2018). Some tools have

been developed to classify only one type of headache.

For instance, the CDSS presented in (Kopec et al.,

2004) is a rule-based expert system that supports clin-

icians to migraine diagnosis. The rules are rep-

resented in the C Language Integrated Production

System. Other tools support clinicians in the diagno-

sis of multiple headache types. The CDSS developed

in (Yin et al., 2015) is specialized to diagnose

migraine and tension-type headaches. It is based

on case-based reasoning. Previous solved cases are

stored in a built-in library, and a k-nearest neighbour

algorithm is used to measure similarity among cases.

More complex is the hybrid system based on both

rules and case-based reasoning in (Yin et al., 2014).

This system is useful to diagnose the main primary

headaches, i.e., migraine, tension-type headache,

and cluster headache: First, patients’ symptoms

are evaluated in a rule-based module and only if a

diagnosis cannot be suggest, the system searches

for the most similar previous case in a case-based

module. Migraine and other secondary headache

diseases can be diagnosed by the CDSS in (Zafar

et al., 2013). An image processing module is also

included to detect brain tumour from MRI scans. A

knowledge-based system that uses yes/no questions

derived from neurology experts is the CDSS in

(Hasan et al., 2012). More speciﬁcally, neurologist’s

knowledge has been converted in a tree diagram.

Guideline-based systems are presented in (Simi

et al., 2008; Eslami et al., 2013; Dong et al., 2014).

The second version of ICHD criteria were codiﬁed in

the CDSS of (Eslami et al., 2013) to classify the main

types of primary headaches (i.e., migraine, tension-

type headache and cluster headache). This CDSS pro-

poses a questionnaire to patients and implements a

simple human-like algorithmic reasoning. The same

version of ICHD criteria were also codiﬁed in (Simi

et al., 2008); the difference is that the developed

CDSS uses a fuzzy logic as implementation technol-

ogy. The third version of ICHD criteria were imple-

mented in the CDSS in (Dong et al., 2014): clinicians

are assisted to diagnose primary headache disorders

and medication-overuse headaches.

The CDSSs developed in (Krawczyk et al., 2013;

Çelik et al., 2015; Keight et al., 2017) are on machine

learning techniques based. In (Krawczyk et al., 2013),

the authors deﬁned the diagnosis of primary headache

as a classiﬁcation task. They compared three fea-

ture selection techniques and several classiﬁers on the

same set of patients. In (Keight et al., 2017) nine

machine learning based classiﬁers are compared on a

dataset consisting of primary headaches, and the best

predictive performance was achieved by a stacking

classiﬁer. In (Çelik et al., 2015) Artiﬁcial Immune

Systems were used as classiﬁcation algorithms to di-

agnose migraine, tension-type headache, and cluster

headaches. A further extension of this work evalu-

HEALTHINF 2021 - 14th International Conference on Health Informatics

402

ates an ant-colony optimization algorithm on the same

dataset (Celik and Yurtay, 2017).

The CDSS in (Vandewiele et al., 2018) is one of

the most recent systems for the diagnosis and classi-

ﬁcation of the headache disorders. It implements an

automated diagnosis support module that generates an

interpretable decision tree based on data semantically

annotated with expert knowledge.

To summarize, the analysed systems use different

sources of knowledge, architecture and implementa-

tion technology to diagnose headaches. Nevertheless,

these tools support clinicians in headache diagnosis

but not in patient management after a diagnosis. As

a result, in many cases GPs refer patients to headache

specialists for unnecessary examinations.

To the best of our knowledge, there are cur-

rently no systems allowing complete screening of

headaches, that is, primary headache, secondary

headache, and warning symptoms in primary care.

3 CALABRIA CEPHALALGIC

NETWORK: INNOVATION IN

THE CLINICAL WORKFLOW

MANAGEMENT

The Calabria Cephalalgic Network represents a novel

healthcare delivery context. It has been designed in

Calabria (Italy) and it is based on a patient-centred

integrated approach. Its main objective concerns the

management of headache patients through personal-

ized care pathways, ensuring continuity of care at dif-

ferent levels of diagnostic-therapeutic paths.

The architectural organization of the network is

reported in Figure 1. The network is organized in

three levels: GPs, Spokes and Reference Centres. All

the levels interact with Emergency Department.

Figure 1: Organization of Calabria Cephalalgic Network.

GPs are in the ﬁrst level. They represent the core

node of the network since they have the ﬁrst contact

with patients. GPs acquire patient’s signs and symp-

toms through clinical interviews and refer the patient

to the most appropriate diagnostic-therapeutic path.

In case of diagnostic uncertainty and/or ineffective-

ness of the prescribed therapy, the GPs interact with

the second level; in more complex situations, such

as for instance, headache with aura, serious comor-

bidities, cluster headaches, and medication-overuse

headaches, the GPs interact with the third level.

District Spoke and Hospital Spoke are in the

second level. The District Spoke manages patients

needing specialized clinical assessments, whereas the

Hospital Spoke manages patients needing deeper as-

sessments based on instrumental investigations.

The Reference Centre is the third level. It mainly

manages patients with comorbidities, headaches with

aura focusing on rehabilitation of patients with

medication-overuse headaches. Moreover, it man-

ages patients with cluster headaches and coming from

Spoke with diagnostic uncertainty.

It is important to remark that, at each level, health

professionals follow evidence based speciﬁc guide-

lines for patient care, facilitating the coordination

with other health professionals.

Figure 2 shows the GPs’ clinical workﬂow in the

Calabria Cephalalgic Network.

Figure 2: Clinical workﬂow for the GPs in the Calabria

Cephalalgic Network.

A Knowledge-based Clinical Decision Support System for Headache Disorders Management

403

The GPs’ clinical workﬂow can be summarised

as follows. At the baseline visit, a GP has to im-

mediately assess possible emergency conditions. In

case of warning symptoms, the GP refers the patient

to the Emergency Department. Otherwise, the clini-

cal investigation continues and the GP assesses aura

symptoms. If each individual aura symptom lasts less

than 60 minutes, the patient should be referred to the

Reference Centre; otherwise, to the Emergency De-

partment. If there are not aura symptoms, the GP

assesses comorbidities and medication-overuse: se-

rious comorbidities or medication-overuse headaches

prompt the GP to refer the patient to the Reference

Centre. The GP will be able to identify primary

headache subtypes only if all of the health condi-

tions described so far have been ruled out: patients

with cluster headaches should be referred to the Ref-

erence Centre; patients in case of diagnostic uncer-

tainty should be referred to the Spoke; with diagnosis

of migraine or tension-type headache are treated by

the GP.

4 CDSS DESIGN AND

DEVELOPMENT

The CDSS was designed to support GPs in the diag-

nosis and choices of therapeutic paths in a primary

care context. It is fully compliant with clinical guide-

lines and diagnostic-therapeutic paths applied within

the Calabria Cephalalgic Network.

The development of the CDSS followed an itera-

tive approach involving engineers, GPs and neurolog-

ical specialists who collaborated to identify functional

requirements. First, the guideline recommendations

were combined with expert knowledge to determine

system responses based on patient data. Then, the

communication interfaces of the system were devel-

oped.

The general architecture of the proposed CDSS is

reported in Figure 3. The CDSS consists of three main

components: knowledge base, inference engine and

CDSS interface. Communication interface and GUI

(Graphical User Interface) complete the architecture

of the system. The CDSS was designed to work in

a collaborative environment. According to this vi-

sion, a user is any external entity that interacts with

the CDSS such as a GUI, another CDSS, an electronic

medical record or any other component and service of

the health environment.

The three main components of the CDSS, the

communication interface and GUI are described in the

next sections. The CDSS interface receives and vali-

Figure 3: CDSS Architecture.

dates the patient data contained in each request.

If the data is correct, it activates the inference engine.

4.1 Knowledge Base

The knowledge base contains the medical-clinical

knowledge appropriately formalized by the relevant

sets of rules. The National Institute for Health

and Care Excellence (NICE) Clinical Guideline for

headache disorders (Carville et al., 2012), the work-

ﬂow of the GPs and health professionals’ experiences

in the Calabria Cephalalgic Network, were appropri-

ately coded as sources of knowledge.

The clinicians’ expertise and their suggestions

have been obtained through various interviews with

both GPs and neurologists.

All rules were coded as IF-THEN statements in

the Java programming language.

Figure 4: Knowledge Acquisition Process.

Figure 4 summarizes the adopted process of

knowledge acquisition for the headache disorders di-

agnosis. First, knowledge from different sources

have been acquired, analysed and expressed in sev-

eral ﬂowcharts. Each ﬂowchart has been validated

by clinicians. Then, headache clinical attributes have

been deﬁned and summarized. The acquired knowl-

edge has been translated in a set of conditional rules.

According to the diagnostic and therapeutic paths

of the Calabria Cephalalgic Network, ﬁve sets of

rules have been codiﬁed for the diagnosis of headache

HEALTHINF 2021 - 14th International Conference on Health Informatics

404

disorders such as primary headaches (migraine, ten-

sion type and cluster headache), warning symptoms,

aura’s symptoms, comorbidities, medication-overuse

Table 1: Headache features used to create the diagnostic

rules.

Diagnosis Features

Warning

events

Headache that begins after age 50

Headaches increasing in

frequency, duration and severity

Headache subsequent to head

trauma in the previous 3 months

Unusual headache accompanied

by fever and rigor nucalis

Headache with physical exertion

(coughing, sneezing)

Paresthesia of lower limb

Anticoagulant therapy

Limb strength deﬁcit

Sensory deﬁcits

Disorders of consciousness

Speech disorders

Visual disturbances

Disorders of vigilance

Aura,

Comorbidity,

Medication-

overuse

Flashes of light

Lines

Scotomas or spots

Aura duration

Paresthesia

Language disorders

Regular overuse for > 3 months

of one or more drugs

Endocrine and metabolic disorders

Cardiovascular disease

Psychiatric disorders

Severe hypertension

Primary

headache

Pain Location

Pain Quality

Pain Intensity

Effect on activities

Duration of headache

Frequency of headache

Red eye

Watery eye

Nasal congestion

Runny nose

Swollen eyelid

Forehead and facial sweating

Constricted pupil

Drooping eyelid

Vomiting

Nausea

Photophobia

Phonophobia

headaches. Rules for the management of the patient’s

therapeutic path have also been coded. Table 1 reports

the headache features used to create the diagnostic

rules. Table 2 reports one rule for cluster headache

diagnosis.

Each headache clinical attribute is designed as cat-

egorical and identiﬁes the allowed values. Table 3

lists the primary headache categorical attributes.

Table 2: One rule for cluster headache diagnosis.

Rule

Name

Rule Condition

Rule 1 IF (Criterion 1 AND Criterion 2

AND Criterion 3 AND Criterion 4

AND Criterion 5 AND Criterion 6

AND Criterion 7)

THEN Cluster Headache

Criterion 1 IF (Pain Location=Unilateral)

THEN true

Criterion 2 IF (Pain Quality=Variable)

THEN true

Criterion 3 IF (Pain Intensity=Severe

OR Pain Intensity=Very severe)

THEN true

Criterion 4 IF (Effect on activities=Restlessness

OR Effect on activities=Agitation)

THEN true

Criterion 5 IF (Other symptoms=Red eye

OR Other symptoms=Watery eye

OR Other symptoms=Nasal congestion

OR Other symptoms=Runny nose

OR Other symptoms=Swollen eyelid

OR Other symptoms=Forehead and

facial sweating

OR Other symptoms=Constricted pupil

OR Other symptoms=Drooping eyelid)

THEN true

Criterion 6 IF (Duration of headache=

15-180 minutes) THEN true

Criterion 7 IF (Frequency of headache=

1 every other day to 8 per day)

THEN true

4.2 Inference Engine

The inference engine receives and manages all the re-

quests coming via the CDSS interface. Based on the

input patient data, it identiﬁes and activates speciﬁc

sets of rules for obtaining a decision support, as illus-

trated in Figure 5. It ﬁrstly searches for all the possi-

ble rules that match with the patient input data in the

knowledge base. Then, the execution of the selected

A Knowledge-based Clinical Decision Support System for Headache Disorders Management

405

Figure 5: Workﬂow of the Inference Engine.

rules starts. It returns the patient-speciﬁc diag-

nosis and therapeutic path on the basis of satisﬁed

rules.

4.3 Communication Interface

The CDSS communication interface was developed

as SOAP web service compliant with the HL7 DSS

international standard (HL7DSS, 2017). It deﬁnes a

standard interface for decision support applications,

regardless of data format, inference mechanisms and

implementation technology. The HL7 DSS stan-

dard guarantees the interoperability of the CDSS with

other applications in the healthcare environment.

Table 3: Primary headache categorical values.

Variable Allowed values

Pain Location Bilateral, Unilateral,

Mono/Bilateral

Pain Quality Pulsating, Variable,

Pressing/tightening

(non-pulsating)

Pain Intensity Mild, Moderate,

Severe, Very severe

Effect on

activities

Not aggravated

by routine activities,

Aggravated

by routine activities,

Restlessness, Agitation

Other

symptoms

Photophobia, Phonophobia,

Nausea, Vomiting,

Red eye, Watery eye,

Nasal congestion, Runny nose,

Swollen eyelid,

Forehead and facial sweating,

Constricted pupil,

Drooping eyelid

Duration of

headache

30 minutes-continuous,

4-72 hours, 15-180 minutes

Frequency of

headache

< 15 days per month,

> 15 days per month,

1 every other day to 8 per day

4.4 Graphical User Interface

The GUI has three main parts including warning

symptoms questions page, aura, comorbidity and

medication-overuse symptoms question page, pri-

mary headache symptoms question page. A decision

page where patient-speciﬁc diagnosis and therapeutic

path are explained, have been implemented. The ﬁrst

diagnostic web page is the warning symptoms ques-

tion page. The GP inserts patient data in appropriate

ﬁelds. The data are thus evaluated by the CDSS. If no

alert is detected, the GP can continue the diagnostic

process and the second question page is shown; other-

wise, the decision page is presented with the patient-

speciﬁc diagnostic path. The second diagnostic page

concerns aura, comorbidity and medication-overuse

symptoms question page. The GP ﬁlls in the ﬁelds

with patients’ clinical characteristics, which are eval-

uated by the CDSS. If no clinical condition is de-

tected, the GP can continue the diagnostic process.

The third page is the primary headache symptoms

question page. After the evaluation of the CDSS, the

ﬁnal decision page is presented to the GP. Figure 6

shows the three diagnostic steps. The web pages are

Figure 6: Diagnostic steps.

both Italian and English language. Figure 7 shows the

primary headache symptoms question page.

4.5 Functional Speciﬁcations

When a patient assessment is required, the GP sends a

request message to the CDSS, via GUI, to activate the

clinical decision support. The requested data are for-

warded to the inference engine through the CDSS’s

interface. The inference engine, based on the submit-

ted data, identiﬁes the more appropriate rules set in

the knowledge base and performs it. The CDSS re-

turns the diagnostic and the patient care path sugges-

tions to the GUI.

The proposed CDSS is a plug and play system

based on the Service Oriented Architecture paradigm

(Erl, 2005); it has been modelled following the Model

Driven Architecture methodology (MDA, 2012).

HEALTHINF 2021 - 14th International Conference on Health Informatics

406

Figure 7: GUI for the diagnosis of primary headache type.

As regards privacy and data protection, the CDSS

does not use patient identiﬁcation data and bidirec-

tional communication data is encrypted.

5 RESULTS AND DISCUSSION

To investigate the performance of the proposed

CDSS, we ﬁrstly performed an assessment of the sys-

tem on available datasets. In order to measure the

effectiveness of the proposed CDSS in the primary

headache diagnosis, we used a publicly available

dataset hosted on http://www.migbase.com/migbase_

dataset.xls. The dataset contains answers to question-

naires of 850 patients with headache problems. A

neurologist labelled each sample in migraine, tension-

type headache, and cluster headache; only one sam-

ple has been labelled as no headache. The pro-

posed CDSS correctly recognized 609/609 (100%)

patients with migraine, 155/184 (84.23%) patients

with tension-type headache, and 55/56 (98.21%) pa-

tients with cluster headache. Overall, the CDSS cor-

rectly recognized 816 (96%) patients; suggested di-

agnostic uncertainty and further clinical investigation

for 25 (2.94%) patients, whereas only 9 (1.06%) pa-

tients were misclassiﬁed. Because no real data was

available, we produced realistic test data for different

types of secondary headaches and warning symptoms.

Although the use of real data was clearly preferable,

the use of this data allowed us to test all rule sets im-

plemented under the supervision of a clinical expert.

The dataset consists of 10 cases. The CDSS classiﬁed

correctly all the cases.

We further explored the use of the CDSS in a real

healthcare setting. The CDSS has been deployed in

the Calabria Cephalalgic Network and experimentally

evaluated in daily practice by the GPs during a period

of 5 months. The real case study involved 10 GPs, 1

Reference Centre, 3 Spokes, and 80 patients. Before

the practical evaluation, the CDSS was presented at

GPs during several training events. The GPs were in-

terviewed during the evaluation period and at the end,

in order to judge the performance of the CDSS and its

impact in a daily practice. Interviews were also con-

ducted with the Spoke and the Reference Centre to

evaluate how neurologists and GPs managed patients.

GPs considered the CDSS useful and were satisﬁed

by its performance: the CDSS is easy to use, patients

data is collected in a short time, the given support

in diagnosis is fundamental especially for complex

cases. Compared to the previous 5 months, the neu-

rologists of the Reference Centre recorded a reduction

of accesses for episodic migraine diagnosis during the

evaluation period, i.e., 12.5% vs 5%. This result sug-

gests an improvement in management of patients with

headache for the GP and therefore, a reduced demand

for unnecessary and expensive instrumental examina-

tions. The proposed CDSS really supports GPs to

manage patients with headaches. As a consequence, a

more appropriate use of healthcare resources there is.

6 CONCLUSIONS

This paper has presented a knowledge-based CDSS to

support GPs for the integrated clinical management

A Knowledge-based Clinical Decision Support System for Headache Disorders Management

407

of headache patients. The knowledge base codes the

NICE Clinical Guideline for headache disorders, the

GP clinical workﬂow and the clinical best practices

from the Calabria Cephalalgic Network. Communi-

cation interfaces are compliant with HL7 DSS inter-

national standard in order to guarantee interoperabil-

ity with other healthcare applications.

The CDSS has been assessed in the GPs’ daily

practice of the Calabria Cephalalgic Network. The

preliminary results are promising. They conﬁrm an

improvement in the management of patients with

headache within primary care facilities. The CDSS

effectively supports GPs in dealing with a patients

headache diagnosis by reducing diagnosis time on the

one hand, even during the anamnesis phase, inappro-

priate accesses to the Spokes and Reference Centre,

and patient’s expenses on the other hand for headache

treatments.

REFERENCES

Aljaaf, A. J., Mallucci, C., Al-Jumeily, D., Hussain, A.,

Alloghani, M., and Mustaﬁna, J. (2018). A study

of data classiﬁcation and selection techniques to di-

agnose headache patients. Applications of Big Data

Analytics, page 121.

Carville, S., Padhi, S., Reason, T., Underwood, M., Group,

G. D., et al. (2012). Diagnosis and management of

headaches in young people and adults: summary of

NICE guidance. BMJ, 345:e5765.

Celik, U. and Yurtay, N. (2017). An ant colony optimization

algorithm-based classiﬁcation for the diagnosis of pri-

mary headaches using a website questionnaire expert

system. Turkish Journal of Electrical Engineering &

Computer Sciences, 25(5):4200–4210.

Çelik, U., Yurtay, N., Koç, E. R., Tepe, N., Güllüo

glu,

H., and Erta¸s, M. (2015). Diagnostic accuracy com-

parison of artiﬁcial immune algorithms for primary

headaches. Computational and mathematical meth-

ods in medicine, 2015.

Dong, Z., Yin, Z., He, M., Chen, X., Lv, X., and Yu, S.

(2014). Validation of a guideline-based decision sup-

port system for the diagnosis of primary headache

disorders based on ICHD-3 beta. The journal of

headache and pain, 15(1):40.

Elkind, A. H. (1991). Drug abuse and headache. The Med-

ical clinics of North America, 75(3):717–732.

Erl, T. (2005). Service-oriented architecture: concepts,

technology, and design. Prentice Hall.

Eslami, V., Rouhani-Esfahani, S., Hafezi-Nejad, N., Re-

faeian, F., Abdi, S., and Togha, M. (2013). A comput-

erized expert system for diagnosing primary headache

based on international classiﬁcation of headache dis-

order (ICHD-II). SpringerPlus, 2(1):199.

Hasan, M. R., Hasan, M. S., and Siraj, F. (2012). An ex-

pert system based headache solution. In Computer Ap-

plications and Industrial Electronics (ISCAIE), 2012

IEEE Symposium on, pages 287–292. IEEE.

HL7DSS (2017). http://www.hl7.org/implement/standards/

product_brief.cfm?product_id=12.

IHS (2018). http://www.ihs-headache.org/ichd-guidelines.

Istat (2015). http://www.istat.it/it/ﬁles/2015/09/

Dimensioni-salute.pdf?title=Le+dimensioni+della+

salute+in+Italia+-+16%2Fset%2F2015+-+Volume.

pdf.

Keight, R., Aljaaf, A. J., Al-Jumeily, D., Hussain, A. J.,

Özge, A., and Mallucci, C. (2017). An intelligent

systems approach to primary headache diagnosis. In

International Conference on Intelligent Computing,

pages 61–72. Springer.

Kopec, D., Shagas, G., Selman, J., Reinharth, D., and

Tamang, S. (2004). Development of an expert sys-

tem for aiding migraine diagnosis. J Inform Technol

Healthcare, 200:355.

Krawczyk, B., Simi

c, D., Simi

c, S., and Wo´zniak, M.

(2013). Automatic diagnosis of primary headaches by

machine learning methods. Open Medicine, 8(2):157–

165.

MDA (2012). http://www.omg.org/mda/.

Potter, R., Probyn, K., Bernstein, C., Pincus, T., Un-

derwood, M., and Matharu, M. (2018). Diagnos-

tic and classiﬁcation tools for chronic headache dis-

orders: a systematic review. Cephalalgia, page

0333102418806864.

Simi

c, S., Simi

c, D., Slankamenac, P., and Simi

c-Ivkov, M.

(2008). Rule-based fuzzy logic system for diagnosing

migraine. In Hellenic Conference on Artiﬁcial Intelli-

gence, pages 383–388. Springer.

Steiner, T. J., Jensen, R., Katsarava, Z., Linde, M., MacGre-

gor, E. A., Osipova, V., Paemeleire, K., Olesen, J., Pe-

ters, M., and Martelletti, P. (2007). Aids for manage-

ment of common headache disorders in primary care.

The online Journal of Headache and Pain, 8.

Vandewiele, G., De Backere, F., Lannoye, K., Berghe,

M. V., Janssens, O., Van Hoecke, S., Keereman, V.,

Paemeleire, K., Ongenae, F., and De Turck, F. (2018).

A decision support system to follow up and diag-

nose primary headache patients using semantically en-

riched data. BMC medical informatics and decision

making, 18(1):98.

World-Health-Organization (2016). http://www.who.int/

mediacentre/factsheets/fs277/en/.

Yin, Z., Dong, Z., Lu, X., Yu, S., Chen, X., and Duan,

H. (2015). A clinical decision support system for the

diagnosis of probable migraine and probable tension-

type headache based on case-based reasoning. The

journal of headache and pain, 16(1):29.

Yin, Z., Min, L., Lu, X., and Duan, H. (2014). A clinical de-

cision support system for primary headache disorder

based on hybrid intelligent reasoning. In Biomedical

Engineering and Informatics (BMEI), 2014 7th Inter-

national Conference on, pages 683–687. IEEE.

Zafar, K., Anwar, Z., Anwar, M., Ahmed, W., Khan, A.,

Rehman, S., and Halim, Z. (2013). Clinical deci-

sion support system for the diagnosis of migraine and

headache. Journal of Basic and Applied Scientiﬁc Re-

search, 3(10).

HEALTHINF 2021 - 14th International Conference on Health Informatics

408