FormalPara Key Points

The OVERCOME (Japan) survey asked Japanese people with migraine about the impact of migraine on their daily lives and the effectiveness of their current migraine medication.

Patients using triptans reported high migraine burden, and many reported poor treatment effectiveness.

Regardless of the current treatment used, the survey respondents who experienced severe burden between migraine attacks were less likely to report moderate-to-maximum treatment effectiveness.

1 Introduction

Migraine is a debilitating neurological disorder characterized by severe pulsating headaches, frequently accompanied by vomiting, nausea, and/or sensitivity to sound and light. Migraine is the second most common cause of years lived with disability worldwide [1] and has substantial impacts on day-to-day functioning and health-related quality of life (HRQoL) [2]. To reduce the impacts of migraine on daily functioning and HRQoL, pharmacotherapy for migraine includes both preventive and acute treatments. Preventive treatment aims to reduce frequency and severity of migraine attacks [3], while acute treatment during attacks aims to resolve the attack and allow return to normal function as soon as possible [4].

In Japan, migraine affects around 8.5% of the population [5, 6], and Japanese people with migraine can experience significant disability, reduced HRQoL, and reduced productivity [7,8,9,10]. Japanese guidelines for the acute treatment of migraine recommend stratified care, in which drugs are selected to meet the degree of disability during migraine attacks [11]. Under these guidelines, nonsteroidal anti-inflammatory drugs (NSAIDs) are recommended for mild-to-moderate severity headaches, and triptans are recommended for moderate-to-severe headaches and mild-to-moderate headaches if NSAIDs were previously ineffective [11]. Antiemetic medications may also be prescribed in combination with either NSAIDs or triptans. Many people with migraine also take over-the-counter (OTC) pain medications, whether alone or in combination with prescription medications [5, 6]. However, there is a lack of up-to-date knowledge of factors and outcomes associated with the acute therapy options available in Japan. For example, the background characteristics, physician consultation patterns, migraine burden, and treatment outcomes of patients using only OTC medications, patients prescribed NSAIDs, and patients prescribed triptans are not currently known.

The ObserVational survey of the Epidemiology, tReatment, and Care Of MigrainE (OVERCOME) (Japan) study assessed symptomatology, burden, and care patterns in a representative sample of Japanese people [5]. Previous articles in the OVERCOME (Japan) series have addressed diagnosis, healthcare practitioner (HCP) consultation patterns, and medication use [5]; burden of migraine [10]; and potential unmet needs for acute treatment of migraine [12]. The objectives of this analysis of OVERCOME (Japan) study data were (1) to describe the background characteristics, HCP consultation patterns, and migraine burden among three separate acute treatment groups and (2) to examine the relationship between burden and self-reported treatment effectiveness in each of these groups. The acute treatment groups were OTC treatment only, prescription (Rx) NSAIDs/acetaminophen (ACE) only, and triptans; these groups represent the three most common types of acute treatment used by the OVERCOME (Japan) respondents [12]. We hypothesized that patients with mild symptoms may not seek care and may be poorly informed about their condition, and thus may be more likely to use only OTC treatment. We also expected that triptan users may report greater treatment effectiveness than the other acute treatment groups, given that triptans are migraine-specific medications.

2 Materials and Methods

2.1 Study Design and Study Participants

The OVERCOME (Japan) cross-sectional, population-based web survey was conducted from July 28 to September 24, 2020. Ethical approval for the study was granted by the Research Institute of Healthcare Data Science Ethical Review Board (ID: RI2020003). All survey respondents provided electronic informed consent, and all data were anonymized before analysis.

Eligible participants were adults; residents in Japan; able to understand Japanese; able to access the internet; members of the Kantar Profiles (Lightspeed) online global survey panel or its panel partners led by Cerner Enviza, an Oracle company (formerly Kantar Health); and able to provide electronic informed consent. As previously described [5], OVERCOME (Japan) consisted of three stages. In Stage I, a sample population representative of the Japanese population was identified. This population-based sampling method selected panel participants to match preset quotas for sex, age, and geographic region, based on Japanese population data [5]. In Stage II, people with migraine and severe headache were identified: people with migraine met modified International Classification of Headache Disorders, 3rd edition (ICHD-3) criteria [13] as assessed by a validated American Migraine Study/American Migraine Prevalence and Prevention Study migraine diagnostic questionnaire [14, 15] or had a self-reported physician diagnosis of migraine. In Stage III, migraine symptoms, impact and burden, and healthcare patterns of people with migraine were assessed. An overview of the OVERCOME (Japan) methodology is provided in Fig. S1 of the Electronic Supplementary Material (ESM).

2.1.1 Acute Treatment Categories

People with migraine identified in Stage II of OVERCOME (Japan) were categorized according to their current use of acute treatment(s) for migraine attacks or severe headaches. Current use was defined as “currently using or typically keep on hand” [5]. In this analysis, the current acute treatment groups were defined based on data from OVERCOME (Japan) respondents [12]. Group 1 (OTC only) consisted of people who reported using only OTC medications for acute treatment of their migraine attacks; OTC medications included OTC painkillers. Group 2 (Rx-NSAIDs/ACE only) consisted of people who reported using only prescription NSAIDs and/or prescription ACE for acute treatment (i.e., no other prescriptions and no OTC medications). Group 3 (triptans) consisted of users of oral, injectable, and/or nasal triptans. In line with Japanese guidelines for management of migraine [4], people in Group 3 could also be taking other acute medications, including NSAIDs (either OTC or prescription). Triptans are only available with a prescription in Japan.

2.2 Measures

2.2.1 Demographics and Clinical Features of Migraine

Demographic data, health status data, medication use, and clinical features of migraine are reported in this article for each of the current acute treatment groups. Demographic and baseline health data include age, sex, marital status, education level, employment status, household income, and current smoking habit. Medication use and experience data reported include current use of preventive medication and experience with triptan medication.

Clinical data reported for each of the current acute treatment groups include monthly headache days category (0–3 days, 4–7 days, 8–14 days, or ≥ 15 days per month), migraine attack pain severity (0–10 scale), cutaneous allodynia category (none, mild, moderate, or severe), and acute medication effectiveness. Cutaneous allodynia symptoms were measured using the 12-item Allodynia Symptom Checklist (ASC-12) [16]. Effectiveness of current acute medication was measured using the validated 4-Item Migraine Treatment Optimization Questionnaire (mTOQ-4) [17]. Treatment optimization categories by total mTOQ-4 score were as follows: 0 = very poor treatment effectiveness, 1–5 = poor treatment effectiveness, 6–7 = moderate treatment effectiveness, and 8 = maximum treatment effectiveness [17].

2.2.2 Healthcare Behaviors

Respondents were asked both whether they had ever sought medical care for their migraines/severe headaches and whether they had sought care in the previous 12 months. Those who reported seeking medical care in the previous 12 months were asked to identify all types of healthcare practioners (HCPs) they had consulted for their migraines. The list of possible HCPs included primary care physician/internist, neurosurgeon, general neurologist, pain specialist, emergency room physician, pharmacist at pharmacy, obstetrician-gynecologist, industrial physician, or other type of doctor/specialist. Respondents could select multiple options if they had consulted multiple HCPs. Respondents were also asked whether they had ever hesitated to seek medical care for their migraines/severe headaches.

2.2.3 Burden and Impact of Migraine

Four separate patient-reported outcome measures were used to assess the burden and impact of migraine for respondents in each of the acute treatment groups. Quality of life (QoL) was assessed via the Migraine-specific Quality of Life Questionnaire v.2.1 (MSQ), a well-established, reliable, and validated measurement tool [18,19,20,21]. The MSQ measures the functional impact of migraine through limitations on daily social and work activities (Role-Function-Restrictive [RFR]), prevention of daily social and work activities (Role-Function-Preventive [RFP]), and emotional impact (Emotional Function [EF]) over the previous 4 weeks [18]. Raw scores in each domain are standardized to a 0–100 scale, with higher scores indicating better QoL [18, 21]. The Japanese-language version of the MSQ has been validated [22, 23].

Migraine-related disability was measured using the Migraine Disability Assessment (MIDAS), a reliable and validated scale that correlates with physicians’ assessments regarding the need for medical care [24,25,26]. The MIDAS tool assesses disability in the 3 months prior to the survey. Higher MIDAS scores indicate more days with disability. Respondents were grouped into one of four categorical grades based on their total MIDAS score: Grade I: 0–5 = little or no disability; Grade II: 6–10 = mild disability; Grade III: 11–20 = moderate disability; and Grade IV: ≥ 21 = severe disability [24, 25]. The Japanese-language version of the MIDAS has been validated [26].

The burden of migraine between migraine attacks (the interictal period) was assessed using the Migraine Interictal Burden Scale (MIBS-4), a validated four-item measurement tool [27, 28]. MIBS-4 questions address disruption at work and school, diminished family and social life, difficulty planning, and emotional difficulty, during days without a headache attack (recall period: previous 4 weeks). Responses are summed across all four items to give a total score ranging from 0 to 12. Respondents were categorized into one of four interictal burden groups: none (score of 0), mild (score of 1–2), moderate (score of 3–4), and severe (score ≥ 5) [27, 28].

The impact of migraine on workplace and general productivity was assessed using the Work Productivity and Activity Impairment Questionnaire: Migraine (WPAI-M) [29]. The WPAI-M measures impacts of migraine on work productivity and regular activities over the previous 7 days. WPAI item responses are used to calculate four WPAI-M scores: absenteeism, presenteeism, total work productivity impairment, and total activity impairment. Scores are calculated as impairment percentages, with higher numbers indicating worse outcomes (greater impairment and less productivity) [29].

2.3 Statistical Analysis

Demographic characteristics, migraine-related clinical characteristics, and patient-reported outcomes (MSQ, MIDAS, MIBS, and WPAI-M) were compared by pairwise comparison of each acute treatment group. Comparisons were by t-test or F-test for continuous variables and chi-squared or Fisher exact test for categorical variables, with a significance level of 0.05. Supervised machine learning was used to identify factors associated with sufficient effectiveness (mTOQ-4 categories: moderate/maximum) versus insufficient effectiveness (mTOQ-4 categories: poor/very poor) in each acute treatment group. A random forest (RF) with 1000 trees was used to access the relative importance of predictor variables. In addition, a Least Absolute Shrinkage and Selection Operator (LASSO) linear regression model was employed to identify relevant variables in a linear regression structure. The selected variables from the RF and LASSO were used in the final logistic model. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated for those factors. The candidate factors included demographic variables (age, sex, employment status, marital status, and smoking status), general health variables (anxiety, depression, and dizziness/vertigo/balance problems in the past 12 months), migraine-related variables (migraine by ICHD-3 criteria, monthly headache days, headache severity, presence/absence of aura, ASC-12 category [mild/moderate/severe], MSQ-RFR score, MIDAS category [mild/moderate/severe], and MIBS-4 category [mild/moderate/severe]), and healthcare-related variables (sought medical care in the past 1 year, ever hesitated to seek medical care, and current preventive medication use). SAS Enterprise Guide 7.15 (SAS Institute Inc., Cary, NC, USA) was used for all statistical analyses.

3 Results

3.1 Patient Distribution and Demographic Characteristics

The OVERCOME (Japan) migraine survey was completed by 17,071 respondents (Fig. 1). There were 5791 current users of OTC treatments only (Group 1), 751 current users of Rx-NSAIDs/ACE only (Group 2), and 2533 current users of triptans (Group 3).

Fig. 1
figure 1

Survey respondent flow diagram. Note that the three current acute treatment groups (Group 1, Group 2, and Group 3) exclude those that were not current users of acute treatment and those who received other combinations of acute treatments. aTargeted sampling to represent the Japanese adult population in terms of key demographic characteristics (age, sex, and geography) was applied. bInvolved participants who represented the census adult population based on the Statistics Bureau, Ministry of Internal Affairs & Communications of Japan (April 2019). ACE acetaminophen, NSAID nonsteroidal anti-inflammatory drug, OTC over-the-counter, Rx prescription

Full size image

The triptan group was demographically distinct from the other acute treatment groups (Table 1), with a significantly lower proportion of female respondents (58.2%) and a significantly higher proportion of current smokers (34.3%). Triptan users also had a higher proportion of full-time workers (63.0%), respondents with at least some college education (75.1%), and respondents with a household income of ≥ 5 million yen (54.6%).

Table 1 Demographic features of the acute treatment groups
Full size table

3.2 Clinical Features of Migraine and Healthcare Behaviors

There were significant differences between the acute treatment groups with respect to clinical features of migraine (Table 2). The OTC-only group had the highest proportion of people with low-frequency episodic migraine (0–3 headache days per month; 71.3%). Although the triptan group had the highest proportion of people with high-frequency episodic migraine (8–14 headache days per month; 14.4%), the proportion of people with chronic migraine (≥ 15 headache days per month) was similar in the triptan and Rx-NSAIDs/ACE-only groups (10.4% and 10.1%, respectively). The proportion of respondents who experienced severe cutaneous allodynia was the highest among triptan users (17.2%) and reported by similar proportions of the other two acute treatment groups (7.4% of OTC users and 6.8% of Rx-NSAIDs/ACE users, respectively). Attack pain severity was highest in the triptan group. Substantial proportions of all the acute treatment groups reported poor or very poor effectiveness of their current treatment (mTOQ-4 scores ≤ 5). This proportion was significantly higher in the triptan group (60.9%) than in the OTC-only group (43.1%) and the Rx-NSAIDs/ACE-only group (47.6%). Triptan users reported significantly poorer effectiveness than the other acute treatment groups for each individual question in the mTOQ-4 questionnaire (Table S1 of the ESM).

Table 2 Migraine-related clinical characteristics of the acute treatment groups
Full size table

The proportion of current users of acute treatments who had sought medical care for their migraines or severe headaches in the previous year ranged from 19.2% of the OTC-only group to 83.9% of the triptan group (Table 2). People in all of the current acute treatment user groups were most likely to consult their primary care physician/internist (56.2% of the OTC-only group, 69.8% of the Rx-NSAIDs/ACE-only group, and 59.6% of the triptan group; Fig. 2). Neurosurgeons, general neurologists, headache specialists, and pain specialists were consulted by higher proportions of the triptan group (~10–32%) than the OTC-only and Rx-NSAIDs/ACE-only groups (~2–26% and ~2–19%, respectively). Obstetrician-gynecologists were consulted by higher proportions of the OTC-only and Rx-NSAIDs/ACE-only groups (8.5% and 8.3%, respectively) than the triptan group (4.4%).

Fig. 2
figure 2

Types of HCPs visited for migraine treatment in the previous 12 months, by current acute treatment type. Respondents could report seeking care in multiple settings. ACE acetaminophen, HCP healthcare practitioner, NSAID nonsteroidal anti-inflammatory drug, OB-GYN obstetrician-gynecologist, OTC over-the-counter, Rx prescription

Full size image

3.3 Burden and Impact of Migraine

Across all the patient-reported outcomes, the triptan group reported higher burden and greater impact of their migraine/severe headaches (Table 3) compared with the other acute treatment groups. All three MSQ domain scores were significantly lower in the triptan group compared with the other acute treatment groups. A higher proportion of triptan users than other acute treatment users reported moderate-to-severe disability and severe interictal burden. Triptan users reported that 11.4% of their work time was lost to absenteeism, compared with 2.0–2.7% of work time lost in the OTC-only and Rx-NSAIDs/ACE-only groups.

Table 3 Patient outcomes
Full size table

3.4 Factors Associated with Insufficient Treatment Effectiveness

The combined RF and LASSO methods were used to identify factors to be included in a logistic model of factors associated with insufficient treatment effectiveness (mTOQ-4 categories: moderate/maximum) in each acute treatment group. In the OTC-only group, the variables included in the final logistic model were age, pain severity score, MSQ-RFR score, MIBS score, MIDAS score, use of preventive medication, migraine by ICHD-3 criteria, HCP visits in the past year, anxiety in the past year, depression in the past year, dizziness or vertigo in the past year, and presence of aura. The ORs (with 95% CIs) for all the variables in the model are listed in Table S2 of the ESM. Insufficient effectiveness of OTC treatment was associated with higher pain severity (OR for 1 unit increase: 0.94 [95% CI, 0.91–0.96]), presence of aura (OR versus no aura: 0.85 [0.76–0.96]), dizziness/vertigo (OR versus no dizziness/vertigo: 0.84 [0.74–0.96]), higher MSQ-RFR score (OR for 1 unit change: 1.02 [1.02–1.02]), and severe interictal burden (OR versus no burden: 0.47 [0.40–0.54]) (Fig. 3a). However, the differences in effectiveness associated with pain severity score and MSQ-RFR were small and unlikely to be clinically relevant (OR range: 0.94–1.02; Fig. 3a).

Fig. 3
figure 3

Forest plot of migraine-related clinical factors associated with insufficient effectiveness of acute treatments (mTOQ score < 6) in each of the treatment groups. Group 1, OTC only: Covariates in the model not shown in the figure are age, use of preventive medication, anxiety, depression, HCP visits, and migraine by ICHD-3 criteria (Table S2 of the ESM) (a). Group 2, Rx-NSAIDs/ACE only: Covariates in the model not shown in the figure are age, MSQ-RFR score, use of preventive medication, anxiety, depression, tobacco use, sex, and migraine by ICHD-3 criteria (Table S3 of the ESM) (b). Group 3, triptans: Covariates in the model not shown in the figure are age, MSQ-RFR score, use of preventive medication, depression, HCP visits, migraine by ICHD-3 criteria, and employment status (Table S4 of the ESM) (c). Non-migraine-related factors of tobacco use and anxiety were also significant for Group 3 and are included in this figure. aContinuous variable, range 0–10. bContinuous variable, range 0–100. cUpper 95% CI was 0.995 before rounding. ACE acetaminophen, ESM Electronic Supplementary Material, HCP healthcare practitioner, ICHD-3 International Classification of Headache Disorders, 3rd edition, MIBS Migraine Interictal Burden Survey, MIDAS Migraine Disability Assessment, MSQ-RFR Migraine-specific Quality of Life Questionnaire v.21, Role-Function-Restrictive domain, mTOQ Migraine Treatment Optimization Questionnaire, NSAID nonsteroidal anti-inflammatory drug, OR odds ratio, OTC over-the-counter, Rx prescription

Full size image

In the Rx-NSAIDs/ACE-only group, the variables included in the final logistic model were age, sex, tobacco use, pain severity score, MSQ-RFR score, MIBS score, MIDAS score, use of preventive medication, migraine by ICHD-3 criteria, anxiety in the past year, and depression in the past year (Table S3 of the ESM). Insufficient effectiveness of Rx-NSAID/ACE treatment was significantly associated with higher pain severity (OR for 1 unit increase: 0.82 [0.75–0.89]) and severe interictal burden (OR versus no burden: 0.56 [0.35–0.89]) (Fig. 3b).

In the triptan group, the variables included in the final model were age, employment status, tobacco use, pain severity score, MSQ-RFR score, MIBS score, MIDAS score, use of preventive medication, migraine by ICHD-3 criteria, HCP visits in the past year, anxiety in the past year, depression in the past year, and presence of aura (Table S4 of the ESM). Insufficient effectiveness of triptan treatment was significantly associated with presence of aura (OR versus no aura: 0.73 [0.60–0.88]), experience of anxiety (OR versus no anxiety: 0.60 [0.48–0.75]), tobacco use (OR versus nonsmokers: 0.81 [0.66–0.98]), moderate or severe interictal burden (OR versus no burden: 0.53 [0.41–0.70] and 0.41 [0.32–0.52], respectively), and severe disability (OR versus little/no disability: 0.61 [0.46–0.81]) (Fig. 3c). The association of pain severity score with effectiveness appeared to be small.

4 Discussion

This analysis of data from OVERCOME (Japan), a population-based, demographically representative survey of people with migraine in Japan, provides the first detailed description of migraine burden and treatment outcomes for users who treated migraine attacks with OTC medications only, Rx-NSAIDs/ACE only, and triptans in Japan. As expected, based on Japanese treatment guidelines, triptan users experienced more severe headaches. Triptan users also reported higher burden and greater impact of their migraine/severe headaches compared with the other acute treatment groups. Within each of the three treatment groups, survey respondents with severe interictal burden were less likely to experience moderate/maximum treatment effectiveness. Among triptan users, treatment effectiveness was also negatively associated with smoking, aura, anxiety, moderate interictal burden, and severe disability.

Of the three acute treatment groups examined in the OVERCOME (Japan) study, the OTC-only group had the highest proportion of people with low-frequency episodic migraine (0–3 headache days per month; 71.3%). This group also had the lowest percentage who had visited an HCP in the previous 12 months (19.2%). However, reporting of severe allodynia was similar in the OTC-only and Rx-NSAIDs/ACE-only groups (7.4% and 6.8%, respectively). These results were generally consistent with our hypothesis that people with milder symptoms would be less likely to see HCPs and more likely to rely on OTC medications. However, there was a proportion of the OTC-only group who reported severe burden of migraine (6.3% reported severe disability, and 21.2% reported severe interictal burden). Our results suggest that not all patients with severe migraine burden are seeing HCPs, indicating the potential for underdiagnosis and undertreatment of migraine in Japan.

A greater proportion of triptan users than other acute treatment users experienced higher headache frequencies and more severe cutaneous allodynia. Given the clinical guidelines for prescription of triptans based on severity of clinical features and/or levels of disability experienced [11], such differences were expected. For example, patients with higher headache frequencies, with more severe headache pain, and experiencing severe cutaneous allodynia as a symptom of migraine may be more likely to seek HCP assistance and be prescribed triptans. However, because triptans are acute therapy, they will not reduce the number of migraine attacks experienced by patients. The amount of reduction of allodynia and headache pain may be functions of the timing of dosing, as it is recommended to take triptans at the beginning of a migraine attack. As previously reported for the OVERCOME (Japan) study, only 30.7% of current triptan users took their medication at the first sign of an attack [12].

Triptan users reported the highest attack pain severity scores and higher burden and greater impact of their migraine/severe headaches compared with the other acute treatment groups. This was consistent across all the patient-reported outcomes, including disability, HRQoL, interictal burden, and workplace and activity impairment. One reason for the higher burden seen in the triptan group may be the high proportion (> 60%) of triptan users who reported poor or very poor effectiveness of their current treatment. In both global and Japanese real-world assessments, patients who responded insufficiently to triptans had poorer HRQoL, higher levels of disability, and higher levels of workplace impairment than patients for whom triptans were effective [30, 31]. MSQ domain scores in the triptan group of the current analysis (RFR: 65.1; RFP: 72.7; EF: 68.3) were consistent with those of triptan insufficient responders in a real-world study of Japanese patients with migraine (RFR: 68.5; RFP: 75.9; EF: 70.3) [31]. Similarly, the WPAI scores for triptan users in the current analysis were of the same order of magnitude as those of triptan insufficient responders in the earlier study (eg, absenteeism scores of 11.4% and 6.4%, respectively) [31]. The triptan users in the OVERCOME study appear to have unmet needs in the acute treatment of their migraine, leading to an overall high burden.

The proportion of triptan users reporting insufficient effectiveness in this study (60.9%) is higher than has been reported previously: 34.2% and 44.0% of the global and Japanese study groups, respectively, were triptan insufficient responders [30, 31]. In these previous studies, insufficient response was defined as patients who achieved freedom from headache pain within 2 h of taking triptan acute treatment in no more than three of five migraine attacks [30, 31]. The mTOQ-4 questionnaire used in the current study assesses not only 2-h pain freedom, but also relief from pain for 24 h, patients’ willingness to make plans, and patients’ perceived control of their migraine or severe headaches [17]. The additional criteria included in the mTOQ-4 score may partially explain the higher proportion of triptan users reporting overall poor/very poor effectiveness compared with previous studies. Triptan users reported significantly poorer response to treatment than the other acute treatment groups for each individual question in the mTOQ-4 questionnaire, including the questions about 24-h pain relief, plans, and perceived control.

In addition to the impact on patient-reported outcomes, insufficient effectiveness of acute treatment for migraine has been associated with an increase in headache frequency and development of chronic migraine at 1-year follow-up [17]. Identifying factors associated with insufficient effectiveness of acute treatment provides a valuable resource to potentially identify subpopulations who may be more likely to experience insufficient treatment effectiveness. In the current analysis of OVERCOME (Japan) data, severe interictal burden was associated with insufficient effectiveness of treatment in all the acute treatment groups. People with severe interictal burden experience significant impairment between their migraine attacks, likely in all four domains of the MIBS-4 (work/school, family/social, making plans, and emotional/cognitive distress) [28]. The relationship identified here between severe interictal burden and reduced treatment effectiveness highlights the importance of this aspect of migraine burden to developing an optimal treatment plan. As noted by Buse et al. physicians should routinely ask their patients about the interictal burden of their headaches [27, 28].

Treatment effectiveness in the triptan group was also negatively associated with smoking, the presence of aura, anxiety, moderate interictal burden, and severe disability. Severe disability and psychiatric conditions such as anxiety have previously been reported as factors associated with difficult-to-treat migraine [32]. Current smokers were a significantly greater percentage of the triptan group than the other treatment groups (34.3% versus 18.8% and 21.4% in the OTC-only and Rx-NSAIDs/ACE-only groups, respectively). The association of smoking with reduced treatment effectiveness may be due to related lifestyle factors among smokers, such as poorer general health [33]. The factors identified in the current analysis are different from those previously reported for migraine patients who experienced insufficient triptan response: combining OTC medications with triptans, timing of triptan administration, greater migraine severity, and low general HRQoL [31]. In the earlier study, interictal burden and smoking status were not included in the model. In addition, patients had to be receiving a triptan as their sole acute prescription medication [31], whereas in the current study some participants may have received combinations of prescription medications. These differences may partially explain the differences in the associated factors identified.

Of the three acute treatment groups in our analysis, the triptan group had the greatest proportion (83.9%) who sought medical care for their migraines in the previous year. Regular medical care may also partially explain the higher frequency of preventive treatment use in the triptan group (37.6%), as both triptans and preventive medications require a prescription. Triptan users most commonly consulted primary care/internist physicians, but also sought care from specialists, including neurosurgeons, general neurologists, headache specialists, and pain specialists. Consultation with primary care physicians was most common in the Rx-NSAIDs/ACE-only group; this group were also unlikely to consult neurosurgeons, general neurologists, or headache specialists. These data suggest that disease-specific migraine treatments may be less likely to be prescribed by primary care physicians and highlight the importance of primary care physicians in the care of Japanese people with migraine. It may therefore be important to target education about headache disorders, including new medications and treatment approaches, toward primary care physicians. This approach is aligned with the structured headache services model of Tinelli et al., who propose that most patients with migraine should be under the care of primary physicians well educated in the treatment of headache disorders [34,35,36].

OVERCOME (Japan) is the largest-scale survey of migraine ever conducted in Japan. Strengths of the study included the use of several validated patient-reported outcome scales to assess treatment outcomes and burden of migraine and the selection of a population-based sample. OVERCOME also included respondents with a large variation in headache frequency and migraine burden and people with migraine who were taking only OTC medications and may not be under the care of a physician for migraine treatment. Limitations of the study primarily include those common to survey studies. All self-reported data can be subject to recall bias, and there was a risk of participation bias because respondents self-selected their participation in the study (they opted in to both the original healthcare panel and the OVERCOME survey, with an ~7% response rate to the survey invitation [5]). However, participation bias was reduced by using a population-based quota sampling method. An additional limitation of this analysis was the use of exclusionary categories for some of the acute treatment groups; for example, we do not know what outcomes are experienced by people with migraine who combine OTC medications with prescription NSAIDs. In addition, the category of triptan users was broad, and thus any influence of medications used in combination with triptans was not examined. For example, the combination of a triptan with an NSAID is a treatment strategy recommended in Canadian, French, German, and European Headache Federation treatment guidelines for patients who report insufficient effectiveness of triptans [37,38,39,40]. The analysis presented here cannot provide any insight into outcomes of this treatment protocol. Because of the cross-sectional study design, we are unable to determine if the demographic differences between the three groups influenced any of the results reported. Finally, although we conducted a logistic model analysis, OVERCOME (Japan) was a cross-sectional study. Thus, causal relationships cannot be established between variables.

5 Conclusion

Among Japanese people with migraine, burden and treatment outcomes differed across users of OTC medications only, Rx-NSAIDs/ACE only, and triptans. People with migraine who used triptans experienced more severe migraine-related burden, and many triptan users self-reported insufficient treatment effectiveness. Although people using only OTC medications tended to report less frequent headaches, a proportion of this group had severe disease or severe interictal burden, which suggests that OTC-only treatment is inadequate for some OTC users. Improved patient and physician education may be required to promote better treatment of migraine, including stratified care, earlier introduction of disease-specific acute medications, and preventive treatment, to support better effectiveness of acute treatments.