Abstract
The objective of the present study was to investigate the influence of headache-related disability on the recognition and management of migraine by French general practitioners (GPs). Forty-nine teaching GPs at the Faculty of Medicine in the Nice-Sophia-Antipolis University were involved in this study. On one day, each patient who presented during the surgery hours of these GPs was invited to complete a questionnaire aimed at identifying if he/she was a headache sufferer and, if so, whether the headache corresponded to migraine and had an impact on his/her functional ability. Functional disability was measured by the short-form of the Headache Impact Test (HIT-6). Being blind to the patients' responses, the GPs completed a questionnaire for each patient aimed at identifying if he/she considered the patient to suffer from migraine and, if so, whether he/she managed the patient for migraine. A total of 696 patients were included in this study and 289 (41.52%) of them had episodic headache. According to the new International Headache Society (IHS) criteria, 113 (16.24%) patients suffered from headache without migrainous features and 176 (25.29%) patients were migraine sufferers (migraine according to IHS categories 1.1 and 1.2.1: 11.21%, and probable migraine according to IHS categories 1.6.1 and 1.6.2: 14.08%). The mean HIT score of these migraine sufferers was 59.1 + 8.8 and 50% of them presented with a very severe impact score (HIT score > 60). Among the 176 migraine sufferers, 105 (59.7%) were not recognized as having migraine, 21 (11.9%) were recognized as having migraine but without migraine management and 50 (28.4%) were recognized as having migraine with migraine management. Recognition of migraine by GPs was statistically associated with the HIT score (OR = 1.105, 95% CI: 1.056-1.157, P < 0.001) and with the 1.1 and 1.2.1 IHS diagnostic categories (OR = 2.942, 95% CI: 1.286-5.025, P = 0.0107) whereas management of patients recognized as having migraine was only associated with the patient's age (OR = 1.051, 95% CI: 1.000-1.104, P = 0.0486). These results indicate that the continuing medical education of GPs should focus on the diagnosis of migraine and its impact on the lifestyle of the patient.
Introduction
The detrimental effects of migraine on patients’ functional ability and health-related quality of life are well-established (1, 2) and clinical evaluation of migraine-related disability plays an important role in migraine management. In particular, clinical evaluation of migraine-related disability can facilitate treatment decisions. The Disability and Strategies of Care (DISC) study demonstrated that, when treatment was selected based on migraine severity, clinical outcomes improved relative to those of conventional step-care strategies (3). Nonetheless, a recent population-based study demonstrated that the therapeutic attitudes of migraine sufferers (use of non specific analgesics by the great majority of them) were poorly influenced by headache-related disability (4). Considering the high level of migraine self-awareness observed in this study (close to 80%), the inadequate migraine management did not appear to result from a deficit in diagnosis itself but instead from a deficit of treatment related to ignorance of the relationship between headache-related disability and treatment need. This deficit of treatment could explain why, whereas more than 80% of migraine sufferers were aware of the disease and had consulted previously for migraine, only 20% of them were currently followed up by a physician (4). Thus, the point at which the disability of an individual migraine sufferer influences his recognition and management by physicians should be understood in order to reduce the burden of migraine. Such data are especially important for primary care patients, since the great majority of migraine patients should be well managed in general practice.
The objective of the present study was to investigate the influence of headache-related functional impact on the recognition and management of migraine by general practitioners (GPs). In order to achieve this objective, we used the Headache Impact Test to measure migraine-induced disability. The Headache Impact Test (HIT) was created by applying Item Response Theory and other psychometric techniques to widely used questionnaires of headache impact (5). HIT was initially programmed for computerized adaptative assessments of headache impact (6), and then a short paper-based form (HIT-6) was developed and validated (7). The HIT-6 questionnaire measures the impact of headache using only 6 questions broadening it to include social-role functioning, pain, emotional distress and well-being, cognitive functioning, and vitality. Thus, HIT-6 is distinguished from other paper-based headache impact assessments by coupling precision with brevity, characteristics that render it ideal for use in busy clinical practices.
Methods
Selection of general practitioners
The general practitioners involved in the study were selected from the 75 teaching GPs at the Faculty of Medicine in the Nice-Sophia-Antipolis University. These GPs teach primary care which is considered as a specialized area of medicine in France and they play an essential role in the tutorial system for students finishing their medical degree course. They were contacted by phone or e-mail during the 11–19 December, 2002 and asked to participate in the study.
Conduct of the study
The study took place on one day, January 7, 2003. Each patient who presented during surgery hours to the GPs who had agreed to participate in the study was invited to complete a structured questionnaire aiming at identifying if he/she was a headache sufferer and, if so, whether the headache corresponded to migraine and had an impact on his/her functional ability. Patients completed the questionnaire in the waiting room before the consultation. At the end of the consultation, being blind of the patients’ responses, the GPs completed a brief questionnaire aimed at identifying if he/she considered the patient to have migraine and, if so, whether he/she managed the patient for migraine. Statistical analyses were then performed to establish the degree of migraine recognition by GPs and to measure the influence of the patients’ migraine-related disability on that recognition and an eventual management for migraine.
Patient questionnaire
The patient questionnaire was completed by each patient in the waiting room before the consultation. The initial question was ‘do you have headache attacks between which you do not suffer from headache?’. If the patient answered ‘yes’ he/she continued to complete the patient questionnaire.
The first part of patient questionnaire was based on the migraine diagnosis criteria defined and recently modified by the International Headache Society (IHS) (8). It comprised questions which were administered according to a validated algorithm (9) used in the main French studies on migraine epidemiology (4, 10). This algorithm for which sensitivity and specificity have been estimated at 95% and 78%, respectively (9) allowed the exclusion of headache patients without migrainous features and the assignment of headache patients with migrainous features to either migraine (IHS categories 1.1 and 1.2.1) or probable migraine (IHS categories1.6.1 and 1.6.2) groups. The occurrence of aura was not looked for and patients belonging to the migraine group might suffer from migraine without aura (IHS category 1.1) or from typical aura with migraine headache (IHS category 1.2.1) if their aura was followed by a headache with the features of migraine without aura.
The second part of the patient questionnaire comprised the headache impact test (HIT-6). HIT-6 is the short paper-based version of the HIT which was developed by applying Item Response Theory and other psychometric techniques (5) to widely used questionnaires of headache impact, the Headache Disability Inventory (11), the Headache Impact Questionnaire (12), the Migraine Disability Assessment questionnaire (13) and the Migraine Specific Questionnaire (14). HIT-6 is a questionnaire that measures the impact of headache using only 6 questions and which broadens the measurement of headache impact to include social-role functioning, pain, emotional distress and well-being, cognitive functioning, and vitality. HIT-6 is easily interpreted by clinicians and patients because it quantifies headache impact by an easily understood score which varies from 36 to 78. This score relates to the degree of headache impact: little or no impact from 36 to 49, some impact from 50 to 55, substantial impact from 56 to 59 and very severe impact above 60. A French version was available since HIT-6 has been translated into 28 languages (7), including French, according to a method developed for the International Quality of Life Assessment (IQOLA) project (15).
GP questionnaire
The GP questionnaire was completed for each patient at the end of the consultation while the patient was still present in the office. The GPs were blinded to the patients’ responses. The GP questionnaire was brief and comprised only two questions: ‘Do you regard this patient as having migraine?’ and ‘If so, do you manage this patient for his/her migraine?’. The latter question referred to specific treatments (triptans, ergot derivatives, prophylactic drugs) and also to any treatment including analgesics and non steroidal anti-inflammatory drugs if they were given specifically for the treatment of migraine.
Data analysis
Categorical variables were summarized by the corresponding percentages and continuous ones were generally summarized by descriptive statistics (mean with standard deviation [SD]). Between-group differences were evaluated using χ2 test or Fischer's test for categorical variables whereas Student's t-test or Mann–Whitney's test were used for continuous variables. To determine the factors associated with both recognition and management of migraine a logistic regression model was used. All analyses were performed with Staview (SAS Institute). A probability level of P < 0.05 was considered as statistically significant.
Results
GP and patient groups
Of the 75 teaching GPs contacted, 63 gave their consent to participate in the study but only 49 participated effectively. These 49 GPs recruited 696 patients all of whom agreed to take part in the survey. Four hundred and sixteen (59.77%) patients were women and 280 (40.23%) were men. Their mean age was 52.8 ± 20.3 years (52.1 ± 20.9 for women; 53.9 ± 19.6 for men).
Identification of migraine sufferers in the patient group
In the initial patient group, all patients completed correctly the first part of patient questionnaire that identified whether they suffered from migraine. Two hundred and eighty-nine (41.5%) patients reported they had episodic headaches. According to the IHS criteria, 113 patients (16.2%) suffered from headaches without migrainous features and 176 (25.3%) were migraine sufferers (Fig. 1).
Venn diagram illustrating the numbers of patients in the initial sample, in the headache group and in the migraine sufferers group.
The characteristics of the migraine sufferers are presented in Table 1. The migraine sufferers group was composed of 126 (71.6%) women and 50 (28.4%) men, with a mean age of 44.2 ± 16.8 years. In this group, 162 completed the second part of the patient questionnaire that comprised the HIT (14 missing data). The mean HIT score of these patients was 59.1 ± 8.8 and 50% of them presented with a very severe functional impact (HIT score > 60). Migraine sufferers were represented by 78 migraine patients (IHS categories 1.1 and 1.2.1, 11.21%) and 98 patients with probable migraine (IHS categories 1.6.1 and 1.6.2, 14.08%). Patients with probable migraine were older (mean age, 46.3 ± 16.8 years vs. 41.5 ± 16.6 for migraine patients) and the sex ratio in favour of women was less marked (sex ratio, 1.88 vs. 3.87 for migraine patients), but neither of these characteristics was significantly different between the groups. The functional impact of the disease in patients with probable migraine was significantly less important than for migraine patients, with regard to both mean HIT score (57.5 ± 8.3 vs. 61.1 ± 8.3, P = 0.0085) and HIT grade distribution (P = 0.0329) (HIT data were obtained for 91 patients with probable migraine and 71 migraine patients).
The characteristics of the migraine sufferers
GPs’ recognition and management of migraine
Among the 176 migraine sufferers, 105 (59.7%) were not considered by their GP to have migraine, 21 (11.9%) were considered by their GP as having migraine but without migraine management and 50 (28.4%) were considered by their GP as having migraine with migraine management (Fig. 2).
Venn diagram illustrating the numbers of total migraine sufferers, those recognized by their GP as having migraine (recognition) and those with migraine management (management).
Influence of migraine-related functional impact on GPs’ recognition and management of migraine
A comparison of migraine sufferers not considered by their GP as having migraine and migraine sufferers considered by their GP as having migraine is presented in Table 2. Patients in whom migraine was recognized by their GP had a higher mean HIT score than patients in whom migraine was not recognized by their GP (63.3 ± 8.2 vs. 56.3 ± 8.6, P < 0.001). Similarly, comparison showed a significant difference in distribution of HIT grades (P = 0.002) with a larger proportion of patients with severe impact being in the group in whom migraine was recognized by their GP (70.8% vs. 36.1% in patients in whom migraine was not recognized by their GP). Nevertheless, 70.4% of patients with substantial functional impact (HIT score between 56 and 59) and 43.2% of patients with very severe functional impact (HIT above 60) were not recognized by their GPs as migraine sufferers. The percentage of patients with substantial or very severe functional impact not recognized by their GP as migraine sufferers was higher in patients with probable migraine (84.6% and 53.8%) than in patients suffering from migraine according to 1.1 and 1.2.1 IHS category criteria (57.1% and 33.3%).
Comparison of migraine sufferers not considered by their GP as having migraine (no recognition) with migraine sufferers considered by their GP as having migraine (recognition)
The mean HIT score (63.1 ± 7.8 vs. 63.4 ± 8.5, P = 0.8808) and HIT grade distribution (P = 0.6714) (Table 3) were not significantly different between migraine sufferers considered by their GP as having migraine but without migraine management and migraine sufferers considered by their GP as having migraine and with migraine management.
Comparison of migraine sufferers considered by their GP as having migraine but without migraine management (recognition without management) and migraine sufferers considered by their GP as having migraine and with migraine management (recognition with management)
Using a logistic regression model, recognition of migraine by GPs was statistically associated with the score HIT (odds-ratio = 1.105, 95% CI: 1.056–1.157, P < 0.001) and with the 1.1 and 1.2.1 IHS diagnostic categories (odds-ratio = 2.942, 95% CI: 1.286–5.025, P = 0.0107), whereas management of patients considered by their GPs as having migraine was only associated with the patients’ age (odds-ratio = 1.051, 95% CI: 1.000–1.104, P = 0.0486).
Discussion
The present study revealed that around a quarter of patients attending French GPs’ surgeries are migraine sufferers and that, despite a high level of migraine-related disability, the majority of them were not recognized by their GPs as having migraine and under a third of them were recognized as having migraine and managed for this disease.
Considering the mean age of the patient group in the present study, the frequency of migraine sufferers observed is in agreement with a recent French nationwide population-based study in which migraine prevalence was estimated to be 17% with highest values in the third, fourth and fifth decades (10). The similar frequencies of migraine sufferers according to 1.1 and 1.2.1 IHS categories criteria and patients with probable migraine according to 1.6.1 and 1.6.2 IHS categories criteria are also in agreement with the study of Henry and colleagues in which prevalences of strict migraine and migraine disorder according to the first IHS classification were equivalent (10). In the present study, the functional impact for patients with probable migraine was significantly less important than that for migraine patients, with regard to both mean HIT score (57.5 ± 8.3 vs. 61.1 ± 8.3, P = 0.0085) and HIT grade distribution (P = 0.0329). These data confirm that, in terms of functional impact, probable migraine occupies an intermediate position between strict migraine that is the most severe headache and other episodic headaches that are the least severe ones (10). Nevertheless the lower disability of probable migraine is relative and more than half of migraine sufferers according to 1.6.1 and 1.6.2 IHS criteria categories included in our study described a substantial (14.3% with HIT score between 56 and 59) or a very severe functional impact of their disease (42.9% with HIT score above 60). These results support the guidelines on migraine management recently drawn up in France at the request of the French Society for the Study of Migraine and Headache (16). These guidelines state that critical analysis of diagnostic criteria for IHS 1.1 and 1.2 categories demonstrates an acceptable level of interobserver variability (17) and good specificity, but unsatisfactory sensitivity (9, 18). The criteria are therefore restrictive and cannot provide a diagnosis for all patients with migraine. To avoid this problem in routine practice and thus prevent depriving some patients of appropriate management, the French guidelines recommended using the probable migraine IHS diagnosis (16). Implementation of this recommendation in primary care is essential since the logistic regression analysis performed in our study revealed that patients with probable migraine were statistically associated with lack of recognition of their migraine by GPs.
The high proportion of migraine sufferers not recognized as having migraine by their GPs in the present study is related to the fact that fewer than half of individuals with active migraine actually see a doctor each year for headache (19). Comparison of GPs’ migraine recognition rate with patients’ self-awareness of migraine would be informative but the question ‘You are a migraine sufferer. Did you know this?’ was not included in the patient questionnaire in the present study. In a previous study we reported that 80% of migraine sufferers selected in the French general population were aware of their migraine (4) but Lipton et al. (20) reported that the rate of self-awareness of migraine was only of 53.4% in United States. The self-awareness of migraine in our previous study was perhaps overestimated considering that there was a concern that some people might be embarrassed to confess their ignorance and might therefore report that they knew they had migraine, even if their diagnosis was in doubt. Nevertheless, our previous study concerned migraine sufferers with ‘active migraine’ (at least one attack in the previous three months) and it is possible that the high rate of self-awareness was related to the greater disability experienced by these ‘active’ migraine sufferers.
Irrespective of migraine categories defined by the IHS classification, logistic regression analysis used in the present study revealed that recognition of migraine by GPs was associated with high disability as measured by the HIT score whereas management of migraine in patients considered as having migraine by their GP was not associated with such a disability. The relationship between disability and migraine recognition has already been noted in studies performed to analyse the use of the health care system by migraine sufferers (21). The absence of a relationship between disability and management could explain our previous results showing that the therapeutic attitudes of migraine sufferers (use of non specific analgesics by the great majority of them) were poorly influenced by headache-related disability (4). More than 40% of patients with very severe impact (HIT score above 60) in our migraine sufferers’ group were not even considered as having migraine by their GP whereas, extrapolating the results of the DISC study (3), these patients should ideally have been treated by specific drugs such as triptans. The continuing medical education of GPs should include the impact-based management of migraine in order to reduce the burden of migraine.
Questionnaires designed to elicit data regarding headache-related disability may provide a basis to improve patient-physician communication and allow doctors to appreciate the medical need and design the most effective treatment plan. No consensus exists as to whether any of these questionnaires comes near to meeting a theoretical ideal. While appropriate for population-based studies, validated tools that measure headache disability often lack the accuracy needed to make clinical decisions for individual patients, especially in primary care, whereas short forms designed to allow quick administration may sacrifice comprehensiveness and precision across the full range of headache impact. To facilitate clinical decision-making, physicians and patients would benefit from a tool that is quick and easy to administer yet provides a precise and interpreted score. The HIT-6 questionnaire is short and measures the impact of headache using only 6 questions broadening it to include social-role functioning, pain, emotional distress and well-being, cognitive functioning, and vitality. These characteristics result from its development using the Item Response Theory model to adapt items of four published and widely used headache and migraine questionnaires: the Headache Disability Inventory (11), the Headache Impact Questionnaire (12), the Migraine Disability Assessment questionnaire (13) and the Migraine Specific Questionnaire (14). Thus, HIT-6 is distinguished from other short questionnaires by coupling precision with brevity and this renders it ideal for use in primary care. The clinically useful HIT-6 questionnaire was completed without any difficulty by more than 90% of patients included in our study.