Abstract
The Dutch version of the MSQOL was evaluated psychometrically and applied in a convenience sample of 90 migraine patients from the Dutch Society of Headache Patients. Internal consistency and test-retest reliability of the Dutch MSQOL were good (>0.90) and comparable with those of the original English version. The measure also had good validity, with its scores associated as expected with headache intensity, psychological well-being and level of patient functioning. Respondents who employ more passive methods of coping with their headaches, such as ‘worrying’, ‘retreating’ and ‘resting’, had worse quality of life. ‘Worrying’ was particularly associated with diminished quality of life, suggesting that cognitive interventions might be of benefit to migraineurs who use this method of coping.
Introduction
The last decade has shown an increasing interest in the consequences of migraine and tension headache for the life of patients. Previous research using generic instruments demonstrated an impaired health status in migraine patients. Osterhaus et al. (1) showed that migraineurs have comparable physical functioning and perceived health status to patients with arthritis, gastrointestinal disorders and diabetes but worse role functioning, social functioning, pain and mental health. According to Essink-Bot et al. (2), migraine patients have problems with role functioning, household work, social functioning and home-life. In addition, they have lower energy levels and feel less healthy than a matched control population. These differences could not be explained by their higher tendency to be depressed. Similar findings were reported by Lipton et al. (3). In a recent Dutch study on health status in migraineurs of the general population, comparing 620 patients with 5378 controls, the patients reported a diminished functioning and well-being on all health status domains. They had also a poorer health status than patients with asthma, but a better one than those with chronic musculoskeletal pain (4). Passchier et al. (5, 6) found that health status was more reduced in migraine patients who consult their physician for their headaches and in those who belong to the Dutch Society of Headache Patients than in patients from the general population. Patients with chronic tension headache appeared to be similar or even more impaired than those with migraine (7, 8). Health status measures have also shown to reflect the positive effects of imigran on migraine (9).
Several standardized measures specific to chronic headache patients have been developed. Because of the relevance of their content to patients, such specific instruments generally have higher reliability and validity than the older generic health status measures (10). The Migraine Quality of Life Questionnaire was designed to assess the impact of a migraine headache over the 24 h following its onset (11, 12), focusing on impairments rather than quality of life. The Migraine-Specific Quality of Life Questionnaire (MSQLQ) was developed for the measurement of the migraineur's quality of life between attacks. It was empirically based, but lacks a firm theoretical basis (9).
Quality of life can be conceived of as the extent to which an individual's needs are met (13). Interviews conducted with migraineurs indicated that they described their experience of migraine in terms of needs that were, or were not, being met—rather than in functional terms (14). The content of the Migraine-Specific Quality of Life instrument (MSQOL) consists of statements derived from these interviews and the questionnaire has been shown to have excellent internal consistency and test-retest reliability (14, 15). Scores on the MSQOL were also found to be related as expected to those on a comparator measure of well-being and to perceived severity of migraine and disruption to life caused by the illness.
The way in which a patient copes with pain has been shown to be related both to the severity of the pain and to the resultant problems caused to the patient's life, as a number of studies show. In patients with chronic tension-type headache, self-efficacy regarding headache management predicted a lower headache index and less headache disability (16). Patients with fibromyalgia who used active and positive strategies for controlling their pain showed better functioning and fewer symptoms, while those who used emotion-focused strategies (such as catastrophizing and depression) had more pain and worse sleep and quality of life (17). Catastrophizing was also found to be positively associated with pain levels, migraine frequency and pain problems in migraine patients and patients with low back pain (18). A negative relation has also been found between active coping and disability in children and adolescents with headache (19). Passive methods of coping with stress in adolescents were positively related to headache intensity and duration (20).
To date, studies on the association between headache coping and quality of life in migraineurs have not employed validated migraine-specific quality of life measures. It also remains to be established whether the link between migraine and quality of life can be explained fully by pain parameters. A further question is whether or not the presence of a headache during the completion of a quality of life questionnaire influences this measurement.
The present paper reports on the adaptation of the MSQOL for use in the Netherlands and on its use in investigating the relation between headache coping and quality of life. On the basis of previous research, two hypotheses were generated:
Active headache coping strategies are positively and passive headache coping strategies negatively associated with headache intensity and quality of life.
The relation between headache coping strategies and quality of life is partly independent of the headache intensity.
Method
Patient groups and procedure
The development of the Dutch version of the MSQOL followed the guidelines of the European Group for Quality of Life Assessment and Health Measurement (21). This involved the use of two translation panels (bilingual and lay). The bilingual panel (which consisted of five Dutch people with high proficiency in English) produced translations of the items and instructions in the MSQOL. This translated version was then considered by a lay translation panel consisting of five persons with average to low educational level, working only in Dutch. This second panel was employed to ensure that the adapted questionnaire had an appropriate level of language and that it would feel ‘natural’ to future Dutch respondents.
The translated version was then tested for face and content validity with 16 patients who had received a diagnosis of migraine from their GP. Fifteen of these were recruited from a general practice and one from the Dutch Society of Headache Patients. Comments from these interviews led to further minor changes in wording being made to the Dutch MSQOL. The final phase of the adaptation involved testing the reliability and construct validity of the instrument. For this phase, patients aged 18 years and above were recruited from the Dutch Society of Headache Patients by an announcement in its monthly journal ‘Hoofdzaken’. Those patients who agreed to participate were sent a booklet containing the Dutch MSQOL and a questionnaire concerning their demographic and headache characteristics. The participants returned the completed booklet in a reply paid envelope. Two weeks later they were sent a second booklet containing the Dutch MSQOL and questionnaires on psychological well-being, level of role functioning and headache coping.
Questionnaires
Headache characteristics
Headache characteristics were collected using a questionnaire about their headaches and associated symptoms. It has been shown that migraine can be satisfactorily classified at a group level by the use of questionnaires (22).
Quality of life
Quality of life was assessed with the 20-item Dutch MSQOL. Each item has four possible responses (scored 1–4) allowing total scores to range from 20 (lowest quality of life) to 80 (highest quality of life) (14). To facilitate interpretation and comparison, we also calculated a standard score according to the procedure of Patrick et al. (23).
Well-being
Well-being was measured by the Psychological General Well-being Index (PGWB). A previous study had demonstrated that this measure is able to detect impaired well-being between migraine headaches, underlining the construct validity of the scale for this patient group (24). Scores on the PGWB can range from 22 to 110, with a high score indicating good well-being. Since the underlying concept of this scale is related to the concept of quality of life and a validated Dutch version is available, the PGWB was used as the main comparator for testing the construct validity of the Dutch MSQOL.
Headache-related disability
Headache-related disability was assessed by the Migraine Disability Assessment Score (MIDAS). The MIDAS score is derived from five questions concerning time lost from or reduced productivity at work/school and in the home. The internal consistency and test-retest reliability and validity of the original MIDAS score are high and its validity has been supported (25). Standard scores on the measure can range from 1 to 4, with a high score indicating high disability. While the MIDAS has been translated into Dutch and used in previous studies, this version has not been formally validated. The measure is therefore used in the present study for illustrative purposes and associations between the MSQOL and individual MIDAS items are also reported.
Headache coping
Headache coping was measured by the Pain Coping Inventory (PCI). This is a 33-item questionnaire, designed to measure both cognitive and behavioural pain coping strategies. It has six subscales classified under two coping dimensions: ‘active coping’ (‘pain transformation’, ‘distraction’ and ‘reducing demands’) and ‘passive coping’ (‘retreating’, ‘worrying’ and ‘resting’). Previous research has demonstrated that the PCI has good reliability and validity (26). ‘Pain’ was replaced with ‘headache’ in the PCI in order to focus on coping with the pain of headache.
Data-analysis
The internal consistency of the Dutch MSQOL was assessed by Cronbach's alpha and its test-retest reliability by the Spearman rank correlation coefficient. Spearman correlation coefficients were also calculated to explore the level of association between the main outcome variables in the assessment of convergent validity. Regression analyses were performed on the Dutch MSQOL and PGWB scores. Predictors were the migraine parameters and the pain-coping strategies that showed significant correlations with these outcome variables. Mann–Whitney U-tests were carried out to test differences in score on the MSQOL between patients who completed the tests during a headache and those who were headache-free.
Results
Patient characteristics
Of the 107 patients who agreed to participate in response to the advertisement, 96 returned the booklets on both occasions (a response rate of 89.7%). The answers of six participants were omitted from the analyses. Three of these reported headache characteristics that failed to meet the migraine criteria of the International Headache Society (27); two had nausea nor phonophobia, in one patient the attacks had a duration less than one hour. Two respondents did not answer questions regarding their headache characteristics and one participant was less than 18 years of age. Details of the sample are shown in Table 1.
Demographic details of the postal survey sample
∗Standard deviation
A large majority(77.8%) of the sample was married or living together with a spouse. Forty-two patients(47%) had aura and 48(53%) no aura. The majority of the patients had migraine only. The median attack frequency was ‘a few times a month’, lasting an average of 36.5 h (
Descriptive statistics for main outcome measures
∗IQR, Inter quartile range
†M = 51.5 and
‡M = 3.0 and
§Pain intensity measured with VAS(0–10)
Psychometric characteristics of the Dutch MSQOL
No significant differences in MSQOL scores were found associated with age, gender or comorbidity of tension-type headache on either administration. The presence of a headache during the completion of the Dutch MSQOL did not affect scores significantly.
The internal consistency of the Dutch MSQOL was 0.92 on the first administration and 0.91 on the second. Test-retest reliability was 0.90. Correlations between the Dutch MSQOL scores and the other outcome measures are shown in Table 3. These values were each significant and in the expected direction.
Spearman rank correlation coefficients between the Dutch MSQOL and the other outcome measures
Headache coping and quality of life
Three headache coping strategies were significantly related to headache intensity and to scores on the Dutch MSQOL, PGWB, and MIDAS. Patients who coped with their pain by ‘retreating’, ‘worrying’ or ‘resting’ had more intense headaches and diminished quality of life, well-being and role functioning (see Table 4).
Correlations between methods of coping with headaches, migraine parameters and outcome measures
∗ P < 0.05;
∗∗ P< 0.01;
∗∗∗ P< 0.005;
∗∗∗∗ P< 0.001.
As scores on the MIDAS were not normally distributed, the regression analyses were restricted to scores on the Dutch MSQOL and the PGWB. Each coping score was separately entered with headache intensity.
The analyses showed that the relationship between ‘retreating’ and ‘resting’ and each of the outcome variables disappeared after correction for headache intensity. However, following this correction, the relation between ‘worrying’ and scores on Dutch MSQOL remained statistically significant (see Table 5). Age was not a confounder as it appeared not significantly associated with the MSQOL (r = 0.09) and PGWB (r =− 0.07) scores.
Regression coefficients of the headache coping scores on the Dutch MSQOL and PGWB
Finally, when coping strategies and headache intensity were entered simultaneously into the regression analysis as predictors of the Dutch MSQOL score, only the effect of ‘worrying’ remained statistically significant (P = 0.02).
Discussion
Psychometric characteristics of the Dutch MSQOL
The postal survey sample consisted mainly of females with relatively severe migraine, as indicated by a long migraine history, the large number with aura and a high proportion of triptan users. The response rate was high for a postal questionnaire and suggests that this group had more severe migraine than patients in general. We had no control group, but when we compare their standardized MSQOL score with that of an US migraine sample that had participated in placebo-controlled trials (23), they scored somewhat lower (0.3
The Dutch MSQOL was found to have very good reliability and validity. Test characteristics were similar to those found for the original English version: the internal consistency and reliability coefficients of both language versions was 0.90 or above and correlations with the PGWB were 0.50 for the Dutch version and 0.38 for the English. These comparable results suggest that the underlying construct of ‘quality of life’ is assessed in a similar way by both language versions. Confidence in the validity of the Dutch MSQOL was further strengthened by its moderate correlations with the individual questions that make up MIDAS and with the total score on that measure. Moderate correlations were expected as the MIDAS assesses disability, which influences quality of life. The association between the Dutch MSQOL and the headache parameters was only positive for intensity, suggesting that intensity is more disruptive to the life of the patient than history, duration or frequency of headache. However, other studies with a larger sample size found also evidence for the relevance of headache frequency for a low quality of life (23) and health status (3, 4).
The absence of differences in Dutch MSQOL scores associated with gender and age underlines its discriminative validity. Furthermore, as scores on the measure are not affected by whether or not the respondent is experiencing a headache, the Dutch MSQOL reflects the quality of life of the patients between attacks, as was intended. The similarity of scores on the measure for migraine patients with and without tension headache suggests that the test is mainly sensitive to migrainous headache.
The next step in the validation process is to investigate the responsiveness of this version to changes in quality of life resulting from effective treatment of the migraine.
Headache coping and quality of life
In contradiction with our first hypothesis, active coping strategies of headache were not associated with a better quality of life. We had expected on basis of previous research on other patients (17, 19) that active pain coping may have led to less pain and a maintenance of a normal life pattern. On the other hand, given the specific character of migraine pain as becoming more during exertion, these coping stategies can also have increased the pain and impeded the relaxation needed. The first effect might have compensated the second, resulting in absence of the association predicted.
The more the patients applied ‘retreating’, ‘resting’ and ‘worrying’ coping strategies, the greater their headache intensity and the lower their quality of life. This outcome sustained the other part of our first hypothesis. We also hypothesized that these associations with quality of life were partly independent from headache intensity. This was only found for coping with headache by ‘worrying’. The importance of ‘worrying’ for the life of the patient is in line with previous studies that have found ‘catastrophizing’ thoughts to be directly related to problems, pain and dysfunctioning in children and adolescents with migraine (19), in fibromyalgia (17) and in low back pain (18). The associations of the passive coping strategies with the PGWB were lower than with the MSQOL and the MIDAS. This is probably due to the headache-specificity of the coping, MSQOL and the MIDAS measures, while the PGWB covers the general well-being of the patient.
A methodological point is that we tested the hypotheses on the associations of headache coping and quality of life on the same sample which was used for the psychometric part of the study. However, as our predictor variables (the coping scores) played no role in the validation process, these tests can be considered as independent from those of the construct validity of MSQOL.
While a prospective study is required before definite conclusions could be drawn, it can be hypothesized that behavioural interventions should more focus on decreasing passive headache coping than increasing active headache coping in migraine patients. In particular, cognitive behavioural therapy that is directed to a reduction of worrying about the headache, might both improve headaches and quality of life.
Footnotes
Acknowledgements
This study was funded by AstraZeneca. Appreciation is also expressed to Mrs L. Zeebregts-Kibbel of the Dutch Society of Headache Patients for her administrative support.