The heterogeneity of headache patients who self-medicate: a cluster analysis approach.

TLDR: This study aimed to identify subgroups of individuals with headache who self-medicate, as this could be helpful to tailor intervention strategies for prevention of medication-overuse headache.

Abstract: Patients with headache often self-treat their condition with over-the-counter analgesics. However, overuse of analgesics can cause medication-overuse headache. The present study aimed to identify subgroups of individuals with headache who self-medicate, as this could be helpful to tailor intervention strategies for prevention of medication-overuse headache. Patients (n = 1021) were recruited from 202 community pharmacies and completed a self-administered questionnaire. A hierarchical cluster analysis was used to group patients as a function of sociodemographics, pain, disability, and medication use for pain. Three patient clusters were identified. Cluster 1 (n = 498, 48.8%) consisted of relatively young individuals, and most of them suffered from migraine. They reported the least number of other pain complaints and the lowest prevalence of medication overuse (MO; 16%). Cluster 2 (n = 301, 29.5%) included older persons with mainly non-migraine headache, a low disability, and on average pain in 2 other locations. Prevalence of MO was 40%. Cluster 3 (n = 222, 21.7%) mostly consisted of patients with migraine who also report pain in many other locations. These patients reported a high disability and a severe limitation of activities. They also showed the highest rates of MO (73%).

Figures

Table 1: Characteristics of the three subgroups of patients identified by cluster analysis.

Figure 1: Flow scheme of participant recruitment and data analysis.

Table 2: Attitudes about pain medication across the three patient clusters.

Figure 2: Graphical results of the factor analysis for mixed data (FAMD). X and Y axis represent the two first dimensions of the FAMD (these two dimensions explain 25.9% of the point variability). Individuals are subdivided into three separated clusters (green, orange and red dots).

Figure 3: Summary of main features of the patient clusters.

Full text

biblio.ugent.be
The UGent Institutional Repository is the electronic archiving and dissemination platform for all
UGent research publications. Ghent University has implemented a mandate stipulating that all
academic publications of UGent researchers should be deposited and archived in this repository.
Except for items where current copyright restrictions apply, these papers are available in Open
Access.
This item is the archived peer-reviewed author-version of:
Title: The Heterogeneity of Headache Patients Who Self-medicate: a Cluster Analysis Approach
Authors: Mehuys, Els, Koen Paemeleire, Geert Crombez, Els Adriaens, Thierry Van Hees, Sofie
Demarche, Thierry Christiaens, Lucas Van Bortel, Inge Van Tongelen, Jean Paul Remon
In: Pain 157 (7): 1464–1471
To refer to or to cite this work, please use the citation to the published version:
Mehuys, Els, Koen Paemeleire, Geert Crombez, Els Adriaens, Thierry Van Hees, Sofie Demarche,
Thierry Christiaens, Lucas Van Bortel, Inge Van Tongelen, Jean Paul Remon. 2016. “The
Heterogeneity of Headache Patients Who Self-medicate: a Cluster Analysis Approach.” Pain 157
(7): 1464–1471. DOI: 10.1097/j.pain.0000000000000541
1

1
THE HETEROGENEITY OF HEADACHE PATIENTS WHO SELF-MEDICATE: A
CLUSTER ANALYSIS APPROACH
Els Mehuys
1
, Koen Paemeleire
2
, Geert Crombez
3
, Els Adriaens
1
, Thierry Van Hees
4
, Sophie Demarche
4
,
Thierry Christiaens
5
, Luc Van Bortel
6
, Inge Van Tongelen
1
, Jean-Paul Remon
1
and Koen Boussery
1
1
Pharmaceutical Care Unit, Ghent University, Belgium
2
Department of Neurology, Ghent University Hospital, Belgium
3
Department of Experimental-Clinical and Health Psychology, Ghent University, Belgium
4
CIRM, Clinical Pharmacy Unit, University of Liege, Belgium
5
Department of Family Medicine and Primary Health Care & Heymans Institute of Pharmacology,
Ghent University, Belgium
6
Heymans Institute of Pharmacology, Ghent University, Belgium
N° of text pages: 13
N° of figures: 3
N° of tables: 2
Keywords: headache, medication-overuse headache, self-medication, community pharmacy, cluster
analysis, risk factor, chronic pain
Corresponding author:
E. Mehuys, Pharmaceutical Care Unit, Ghent University, Harelbekestraat 72, B-9000 Ghent, Belgium.
Tel: +32 92648043. Fax: +32 92228236. E-mail: els.mehuys@ugent.be

2
INTRODUCTION
Headache is one of the most common health complaints. According to the Global Burden of Disease
Study 2013, tension-type headache and migraine are the second and the seventh most prevalent
chronic disorders in the world (with, respectively, 1.6 billion and 850 million incident cases in 2013)
[35]. Headache can be associated with substantial morbidity and impaired quality of life [18]. For
example, migraine is the 6
th
leading cause of disability worldwide [35].
There are medications available that effectively alleviate headaches [32], but these may be used
incorrectly [10; 28]. In particular, both frequent and long-term use of acute headache medication
may have a paradoxical effect, worsening headaches rather than relieving them. This condition,
called medication-overuse headache (MOH), is a growing problem, with an estimated prevalence of
at least 1% in the general population and up to 30-50% in tertiary headache centers [10; 28].
Detection of MOH is not easy: many patients rely solely on self-medication with over-the counter
(OTC) analgesics, and have limited contact with physicians [19]. A study in community pharmacies
found that about 25% of individuals who purchased OTC analgesics specifically for regular headaches
overused that acute headache medication, and were thus at risk for developing MOH [27].
Community pharmacists may therefore be in a strategic position to identify and advise individuals
who self-treat, particularly with respect to prevention and early detection of medication overuse
(MO) in order to avoid development of MOH. To accurately assess the risk of MO, it is important to
consider a broad profile of patients, as chronic headache is often associated with other pain
complaints, such as musculoskeletal pain [15; 37]. In fact, there is evidence that the use of analgesics
for other pain complaints, may also cause MOH in individuals with a primary headache syndrome,
especially migraine [2; 36]. As yet, there is no clear picture of the types of individuals who self-
medicate their headaches, and the prevalence of MO in these subtypes. This information is necessary
to develop community pharmacist intervention programs for prevention of MOH.
This study aimed to identify subgroups of individuals with headache who self-medicate. A
hierarchical cluster analysis was used to group patients as a function of sociodemographics, pain,

3
disability and medication use for pain. In addition, we compared attitudes about pain medication
between the identified subgroups.

4
METHODS
Study design
This observational study was carried out from December 2012 till May 2013 in 202 community
pharmacies in Belgium. Of note, in Belgium, the sale of OTC medicines is limited to pharmacies.
Therefore, pharmacies are an ideal setting to recruit a representative sample of persons with
intentions to self-medicate. Approval for the study was granted by the Ethics Committees of Ghent
University Hospital (for Flanders) and CHU Liege (for Wallonia), and all patients gave written
informed consent.
Participants
Pharmacy customers purchasing an OTC systemic analgesic were approached consecutively by the
pharmacist and invited to participate in the study (OTC systemic analgesics available in Belgium:
paracetamol, acetylsalicylic acid, ibuprofen 200-400 mg, naproxen 200 mg, and fixed dose
combinations of simple analgesics with caffeine). They were eligible when meeting the following
inclusion criteria: purchasing the analgesic for themselves, being aged ≥18 years, and suffering from
pain during at least one full day in the last month. From each of the pharmacies, ten patients were
planned to be recruited.
For the present analysis, only data of patients with headache were included. A headache patient was
defined as a patient marking the head on a full body manikin (see section ‘Data collection – pain
characteristics’).
Data collection
Pharmacy customers who agreed to participate completed a self-administered questionnaire,
developed by the multidisciplinary research team (a neurologist, a pain psychologist, a general
practitioner, a clinical pharmacologist and pharmacists) on the basis of literature and a previous
study on this topic [27]. The questionnaire was piloted, before use, by one community pharmacist. It

Frequently asked questions

Q1. What is the role of pharmacists in preventing and treating headache?

In addition, pharmacists should be vigilant for risk factors for pain chronification and could proactively identify and address concerns about pain medication, which may help to enhance adherence and detect (fear of) addiction. 

Q2. What is the common chronic disorder in the world?

According to the Global Burden of Disease Study 2013, tension-type headache and migraine are the second and the seventh most prevalent chronic disorders in the world (with, respectively, 1.6 billion and 850 million incident cases in 2013) [35]. 

Q3. What is the grade of pain intensity and activity interference?

Total scores classify respondents into one of 4 levels of pain intensity and activity interference (grade 1: low disability and low pain intensity, grade 2: low disability and high pain intensity, grade 3: high disability and moderate limitation of activities, grade 4: high disability and severe limitation of activities). 

Q4. What is the role of community pharmacists in the management of headache?

when community pharmacists have the necessary communication and counselling skills, the pharmacy may also prove to be an effective and efficient setting to deliver adequate information about headache and its management. 

Q5. What is the v-test for categorical variables?

For categorical variables, the v-test evaluates whether a cluster is characterized by the categories of the categorical variable. 

Q6. What was the definition of a headache patient?

A headache patient was defined as a patient marking the head on a full body manikin (see section ‘Data collection – pain characteristics’). 

Q7. What is the relationship between socio-economic status and pain problems?

It may well be that the relationship between socio-economic status and pain problems is reciprocal, potentially fuelling a vicious circle of more pain and suffering. 

Q8. What is the effect of the FAMD on the clustering of variables?

FAMD balances the influence of both continuous and categorical variables and as a consequence, they are compared on an equal footing [29]. 

Q9. What is the prevalence of the older patients with migraine?

It seems that this cluster consists of the oldest patientswith mainly non-migraine headache, with on average two other pain locations and an overall low disability.