Background
Episodic Migraine (EM) and Chronic Migraine (CM) have a considerable impact on patients’ daily lives in terms of personal suffering, reduced quality of life (QoL) and disability [
1‐
7], with female and CM sufferers reporting higher disability [
8]. In particular, CM frequently presents with medication overuse headache (MOH): as shown in some literature findings, an average of 62.6% (ranging from 50.5% to 68%) of patients with CM present MOH [
9‐
12] and whether MOH is a consequence or a cause of CM has not been clarified [
13]. Migraine disorders determine a considerable burden on societies, which is usually addressed in terms of reduced work productivity and cost [
14‐
18]. The most recent studies on the cost of headache disorders, and of EM and CM in particular, showed that most of the cost of such conditions is due to indirect cost, i.e. to reduced work productivity [
14,
15,
19]. When addressing the issue of indirect cost, two elements have to be acknowledged: the lost workdays (absenteeism) and the workdays in which people with migraine worked with reduced productivity (presenteeism). Presenteeism is the main driver of migraine cost and burden: in fact, for each lost workday, patients with EM and CM work three to four days with reduced productivity [
20,
21], and the cost associated to presenteeism is higher than that associated to absenteeism [
14,
15,
22]. Therefore, addressing presenteeism in terms of both frequency of days and impaired productivity is of importance to measure the burden of EM and CM.
While absenteeism can be addressed with a simple and direct question, presenteeism may involve difficulties with the interpretation of content. In fact, the degree to which migraine headaches impact over work-related tasks can be highly variable and is underlined by three elements: a) headache severity, b) the kind of activity or the multiplicity of activities, that constitute one’s own job profile, and c) the context in which one’s own job is carried out. The last two elements may allow the identification of the different tasks and activities as well of contextual elements of the job in terms of interpersonal relationships, and of physical elements that might act as triggers of migraine headaches. A literature review was specifically devoted to understanding the degree to which work-related difficulties are recognized and considered in headache research [
23]. In brief, this review was grounded on a previous work in which the International Classification of Functioning, Disability and Health (ICF) [
24] was used as a term of reference to describe a set of difficulties that are relevant to migraineurs. Fourteen topics, that could be referred to difficulties with work-related activities, were transformed into MESH terms and were used to search for relevant publications in which these difficulties were experienced by patients with EM, CM, chronic daily headache or MOH. A total of 23 publications were selected and the results showed that there was poor recognition of the topic of work-related difficulties, which was limited to a restricted set of activities such as problem solving, speaking, driving and on “remunerative employment”. The latter topic was generally expressed in terms of reduced ability to “perform job activities” or of reduced ability to “work as usual”, and the meaning of these definitions was less than clear in available literature. The presence of contextual elements was a completely neglected issue.
The main reason for the paucity of information on this topic is, in our opinion, the lack of patient-reported outcome measures (PROMs) specifically aimed to capture the presence, the severity and the type of work-related difficulties in patients with EM and CM. We therefore launched an initiative to develop a new questionnaire, the HEADWORK Questionnaire. Given the paucity of literature data, we ran a qualitative study with the aim of exploring which were the most relevant difficulties experienced by patients with their work activities and which were the factors that contributed most to these difficulties, getting indications directly from employed patients with EM and CM. In this qualitative study we ran three focus groups with 14 patients, that were asked to discuss the main issues that constitute difficulties with work-related activities and factors that contributed to these activities [
25]. The results of this qualitative study showed that 27 were the most relevant themes reported by patients, and that they referred to: activities (e.g. reading, writing, speaking), personal factors (e.g. attention, stress), correlated symptoms (e.g. pain, being numb), and contextual elements (e.g. office, colleagues, noise, light). The joint results of the literature review, and of this qualitative analysis enabled us to define a set of relevant themes which referred to 13 activities and 12 factors impacting on these difficulties that were used to develop the preliminary version of the HEADWORK Questionnaire. The aim of this paper is to validate this new questionnaire and report its measurement properties.
Discussion
With this paper we present the validation of the HEADWORK questionnaire, a 17-item PROM specifically designed to assess the impact of EM and CM on work-related tasks and the factors that may contribute to such difficulties. Our results showed that the different dimensions regarding the negative influence of migraine on work activities, i.e. the amount and severity of difficulties in work-related tasks and the factors that impact over them, can be measured by two distinct scales. The first scale, named “Work-related difficulties”, is composed of eleven item dealing with the degree to which migraine headaches determine a difficulty in general skills, such as solving organizational problems or starting a new work task, or in specific tasks, e.g. using the computer or talking and interacting with other people. The second scale, named “Factors contributing to work difficulties” is composed by six item, and addresses the degree to which some factors, such as noise of brightness of workplace, or the attitudes of colleagues, negatively impact on difficulties with work-related tasks. Thus, with the validation process, we reduced the amount of items from the initial number of 25 to the final number of 17, and both HEADWORK subscales showed good measurement properties, with higher scores being associated to higher impact levels.
The assessment of migraine-related impact on work-place activities is a relevant research and healthcare topic because migraine is recognized as one of the most burdensome diseases [
37‐
42]. Of notice, the studies published on the Global Burden of Disease (GBD) confirmed that migraine is more prevalent among females and in both sexes in the most productive age, and acknowledged migraine as the seventh position in the rank of top causes of Years Lived with a Disability (YLDs) in 2010, and then to the sixth in 2013, and eventually to the second in 2016 [
37‐
39]. Such an increased in GBD ranking is likely due to the fact that MOH was kept distinct from migraine in the first GBD reports, while in the newly published GBD study, the burden of MOH was partly assigned to migraine and partly to tension-type headache, with the result that migraine ascended to the second rank in the causes of YLD, being responsible for 5.6% of all YLDs [
41]. The reasons for such a change are shareable, as MOH is a complication of a pre-existing headache disorder and it does not occur otherwise [
42,
43]. Thus we think that it is correct to assume that there is continuity in terms of disease burden between EM and CM with MOH, the impact over work-related tasks being the main domain for negative impact [
44,
45]. Our results are in line with such a hypothesis: HEADWORK scores were higher in CM patients for both scales, as compared to their episodic counterpart (with medium to large ES): these findings can be explained by different aspects characterizing CM patients, i.e. higher headache frequency, more severe pain intensity, but mainly the presence of MOH, which in fact was present in most patients of this group. MOH is present in more than 60% of CM patients, as shown by previous literature findings [
9‐
12] and is presumed to be a concause of CM development [
13], but is a distinct feature as not all CM patients present with MOH. Mixing primary and secondary headaches may be problematic but, in our opinion, the problem is mostly taxonomic, as HEADWORK is intended to capture work-related difficulties due to the presence of EM and CM, irrespectively from the presence of MOH. It has also to be noted that the single item of the preliminary version of our questionnaire addressing MOH (i.e. Need to take an excessive amount of symptomatic drugs) was not retained in the final version: so, we believe that HEADWORK can be used by both the two subgroups of CM patients, with and without MOH.
HEADWORK is intended to fill in the existing gap on the measurement and better understanding of reduced productivity, a task that presents relevant challenges. As far as we know, three are the main available instrument for this task: the Migraine Work and Productivity Loss Questionnaire (MWPLQ) [
46,
47], the MIDAS, and the Work Productivity and Activity Impairment (WPAI) [
48]. The MWPLQ is the only migraine-specific tool to assess difficulties in the workplace [
46,
47]. It is aimed at measuring the impact of migraine headache on work, in terms of hours of work lost or of hours worked with migraine symptoms, which also includes a set of questions assessing the different activities and influencing factors. These items included difficulty in getting to work, working in proximity to environmental triggers of migraine symptoms, difficulty in handling physical aspects of jobs, visual tasks, mental aspects, and interpersonal issues at work. A grading of limitations in each investigated activity is required, on a 6-point scale, from “no difficulty” to “so much difficulty couldn’t do at all”. This questionnaire was developed to assess the positive impact of acute medications: all questions are focused on in the most recent headache attack, and some of them specifically ask the number of hours missed before and after the medication was taken. The MIDAS includes two questions on the number of days with total or partial impairment in work activities experienced in the previous three months. The question on days with 50% or more impairment in work activities does not allow to capture the whole range of possible productivity reduction which may be lower than 50% [
14,
15,
49]. Addressing the full range of limitation is more relevant than missed workdays in migraine patients, as it is the main driver of migraine cost and burden [
14,
15,
20‐
22,
45]. Furthermore, the value of MIDAS seems problematic in those patients with high frequency migraine and CM, because patients are likely to approximate responses to MIDAS questions by multipliers of 5 or 10 [
21]. Finally, the WPAI is a generic instrument addressing the negative impact of different diseases on work productivity [
48]. Questions of the WPAI investigate the number of lost working hours and of hours worked with partial productivity limitations (as assessed on a 10 point scale) in the past seven days, and includes two questions inquiring how much did the underlining health condition affected productivity while working – as well as it affected other regular daily activities – on a 10-point scale (from “no effect” to “completely prevented from working”). A migraine-specific version of this tool can be found on the developer’s website [
50] and its use has been suggested by recent guidelines for randomized trials in CM [
51].
In synthesis, the different available PROMS that can be used in migraine patients are not comparable to HEADWORK, as none of them systematically enable to address a set of activities that are relevant to carrying out work-related tasks. The MWPLQ includes some “qualitative” information on the different types of activities and on influencing factors in the work-place, and both the MWPLQ and the WPAI include questions on the degree of impairment while continuing to perform work activities with migraine. However, the MWPLQ has the specific aim to assess difficulties in relation to the use of an acute medication: therefore, the main focus is the amount of time with difficulties in productivity before and after the intake of medication during a single migraine attack. Despite the WPAI was recently used to address the role of nausea and vomiting in determining the economic burden of migraines [
52] and in a RCT on the anti-CGRP antibody fremanezumab in CM (data reported at the 2017 International Headache Congress [
53]), it has never been formally validated for migraine patients to date. Finally, the time-frame of reference of these questionnaires may be too long (such as the three-month period for MIDAS which may determine reporting bias, particularly in CM patients [
21]) or too short (such as the most recent headache episode for MWPLQ, and the previous seven days for WPAI) in order to assess clinically meaningful data for epidemiological and outcome research. On the contrary, the HEADWORK questionnaire is likely to give an appropriate insight on the different dimensions of work-place difficulties in subjects with migraine in a clinical relevant period of time (one month). It addresses not only the degree of work-related limitations, but also the impact on specific work tasks, and the evaluation of whole range of possible degree of impairment (by a scale from “no difficulty” to “I cannot do it”), thus offering an evaluation of the reduced work productivity while experiencing a migraine episode, which is the most relevant driver of the total costs of migraine [
14,
15,
19]. In reason of these features, we recommend it is used as a measure of migraine impact over work activities, to produce work-related disability weights in studies evaluating the burden of EM and CM, and as a secondary outcome measure in clinical research.
Some of the results we found were expectable and represent a confirmation of the content of HEADWORK items. Among these, our study confirmed that women and CM patients showed a higher difficulties in work activities and reported more factors contributing to these difficulties than men and episodic migraine patients. In addition to this, the fact that HEADWORK questionnaire showed higher correlation indexes with the WHODAS-12 than with the MSQ and with the MIDAS, was expected in consideration of the similarity in the formulation of item and questionnaire construct.
Other results shed light on the value and novelty of HEADWORK as a measure of impact on work-related activities under different aspects. First, the fact that headaches frequency showed higher correlation with the scale “Factors contributing to work difficulties” than with the scale “Work-related difficulties”. Second, the fact that headache intensity showed higher correlations than headache frequency with HEADWORK scales. This is somehow novel, as one could expect that the presence of an higher number of headaches is a factor associated to the presence of more difficulties, while pain intensity is generally considered as a secondary outcome. Third, the fact that both HEADWORK scales showed little correlation with the number of lost workdays and with the number of days worked with reduced productivity. This aspect constitutes a step forward in the understanding of migraine impact, because previously used parameters, such as absenteeism and presenteeism, may provide only an indirect information on the extent of work-related difficulties: what cannot be inferred with such indirect procedures is the extent of reported limitations with reference to the specificity of the task constituting one’s own work duty. HEADWORK fills in this gap, and the little correlation with commonly used indicators, such as the number of lost workdays and the number of days worked with reduced productivity, is a proof of the fact that the content of HEADWORK is not transposable with them. The strong correlation between the HEADWORK scales and the self-reported productivity in the days worked with reduced ability is a further confirmation of the unique information produced by HEADWORK questionnaire. In our opinion, all of these aspects show the ability of HEADWORK to disentangle the problems due to migraine as a disease – which is accompanied by an ensemble of socio-cultural representation, such as the need to use drugs to function, and stigma (which is particularly affected by the ability to work [
54]) – and the presence of single headaches, which may have a “more or less” severe impact depending on several factors. These factors can be connected to the subjective response to therapies, but also to the features of the context in which the person works, in terms, for example, of environmental triggers (like noise or light) or of possibility to quit working or attitudes of colleagues.
Some limitations have to be acknowledged. Sample size was wide enough, as showed by KMO and BTS, but was entirely derived from specialty headache centers: the primary effect of this was the high presence of patients with CM (around 25%) compared to what could be expectable based on the epidemiological presentation of this condition. Second we did not test the stability of the questionnaire, i.e. whether few days after the first administration patients would report similar responses. Similarly, sensitivity to change, i.e. the degree to which changes in patients’ responses are consistent with changes in the disease profile, was not addressed as a longitudinal design would be needed. Such an aspect is of particular relevance, and might constitute important information for clinicians and patients in the process of decision making on the best therapeutic options. In fact, it will be very interesting to understand what may be the main drivers of HEADWORK scales change, considering the potential role of different variables, such as frequency (which is generally considered as the major outcome measure in headache research – but showed a modest correlation with HEADWORK scores), or severity of headaches, but also presence of treatment-related side effects, particularly such those that may have an important role on work-place activities and productivity, such as somnolence, sedation dizziness or fatigue, and which are relatively common with preventive anti-migraine medications. Third, the questionnaire is not designed to distinguish the impact of migraine on work-related aspects in ictal and interictal phases as patients are required to fill in HEADWORK with reference to the previous 30 days, thus taking into account good and bad days. Fourth, headache diaries were used when available, but we do not have track of how many patients had. Diagnosis was clinical and based on ICHD-3Beta criteria for EM with and without aura and CM with and without MOH: however, we cannot exclude mixed diagnoses, i.e. presence of tension-type headaches, for some cases. Finally, among the next steps to further on implement HEADWORK, the definition of cut-off scores is surely the most relevant one. Further clinical and labor-related aspects would however be needed to perform such a task. Frequency of access to emergency departments, recurrence of relapses into MOH and presence of comorbidities, that have been showed to negatively impact on disability and QoL [
55], may be relevant clinical indicators, and presence of disability benefits and – prospectively – risk of unemployment may be relevant labour-related indicators for grading of HEADWORK questionnaire.