FormalPara Key Summary Points

Why carry out this study?

There is a high prevalence of migraine in Japan that imposes a heavy social and economic burden on society.

There are still gaps in knowledge regarding patient preferences related to choices between acute and preventive treatment options.

This study aimed to quantify the preferences of patients with migraine in Japan when selecting preventive and acute treatments and to examine the impact of patient characteristics on treatment choice.

What was learned from the study?

“Method of delivery” was the most important attribute to patients and oral preventive treatments were preferred over injectable treatments.

Our results indicate the need for raising awareness of the availability of preventive medications among patients with migraine in Japan and the need for continued development of oral migraine treatment options that reflect treatment preferences of patients with migraine in Japan.

Introduction

Migraine is a chronic debilitating headache disorder that can be both painful and impact the quality of life of patients [1, 2]. The disease is particularly burdensome as the impact of migraines can extend beyond the pain from an event. For example, anxiety caused by the fear of having a migraine attack can negatively impact patients between migraine events [3]. The societal burden of migraine is further exacerbated by the high prevalence (estimated to be between 6.0% and 8.4%) in the Japanese population [4, 5]. These numbers are lower than in western countries such as the USA (age-adjusted prevalence of 15.9% in 2018 [6]) indicating the potential underreporting of cases in Japan. Regardless, the physical, psychological, and economic burdens caused by migraine are gaining recognition in Japan as shown in recent large-scale population-based studies [7].

In Japan, the treatment options for migraines in the past have focused on acute treatments rather than preventive care. These acute treatments include triptans, ergot alkaloids, ditans, and over-the-counter (OTC) medicines such as nonsteroidal anti-inflammatory drugs (NSAIDs) [8]. Despite the focus on acute treatments, the Clinical Practice Guideline for Headache Disorders 2021 [9] recommends both acute and preventive treatments. According to a recent population-based study, 80.4% of patients with migraine have tried an OTC and 20.1% have tried a triptan for acute treatment, and 10.2% have tried a preventive treatment [1].

Preventive treatment options for migraine have included medications used in other indications such as beta-blockers, antidepressants, anticonvulsants, and calcium channel blockers [8]. However, recent advances in calcitonin gene-related peptide (CGRP) inhibitors, including monoclonal antibodies and receptor antagonists (gepants), have resulted in the first category of pharmaceuticals developed specifically as a targeted therapy for migraine prevention. Moreover, there are multiple preventive migraine treatments that have recently become available or are currently under development or being studied in clinical trials [10]. Therefore, it is expected that new developments in migraine treatments in the coming years will improve the quality of life of patients suffering from migraines.

Given the development of novel preventive treatment options, which may reduce a patient’s need for acute treatment, it is important to understand patient preferences for the various acute and preventive treatment options. As patient-centric treatment gains further prominence in the medical field, patient preference is becoming an essential dimension of healthcare that should be incorporated into pharmaceutical decision-making. Although acute and preventive treatments differ in their benefits and burdens, patient preferences regarding these differences remain unclear. Acute treatments are meant to alleviate migraine attacks after onset. These treatments aim to reduce symptoms such as pain, nausea, and audiovisual sensitivity [11, 12]. These treatments differ in attributes related to medication class, efficacy, time-to-effect, administration type, side effects, and contraindications. Preventive treatments (beta blockers, antidepressants, anti-CGRPs) may be appropriate for patients who have frequent headache attacks (as determined by a physician) or who present with headache symptoms of moderate frequency but high severity and high impact on social function. Many of the attributes of preventive medications are similar to those of the acute treatments. However, preventive treatments may require lifestyle modifications due to the set intervals at which they are administered. It is also a common treatment strategy to combine preventive and acute treatments on an ad hoc basis for patients suffering from frequent migraines [13].

This study aimed to quantify the preferences of patients with migraine in Japan when selecting preventive and acute treatments and to examine the impact of patient characteristics on treatment choice.

Methods

Study Design

This study was a non-interventional, mixed-methods, cross-sectional virtually conducted survey using a discrete choice experiment (DCE) to quantify treatment preferences among patients receiving acute and/or preventive treatments in Japan. A DCE is a quantitative method widely used in healthcare to elicit preferences from participants [14]. In a DCE, participants are presented with a series of choice tasks in which they are asked to choose among two or more hypothetical profiles. Each profile is defined by several attributes, each of which has multiple levels (an example panel from this study is shown in Fig. 1). Prior to the DCE, a qualitative interview study was conducted to inform the attribute selection and questions for the DCE.

Fig. 1
figure 1

Example choice set from the DCE translated from the original Japanese into English. Two hypothetical options were shown, and respondents were asked to choose which they would prefer as a treatment. This process was repeated seven times for each participant

Full size image

Qualitative Interview Study

In order to select relevant attributes and formulate appropriate questions for the DCE, a virtual qualitative interview study was conducted with patients with migraine [15]. Qualitative data was collected from 10 patients using a semi-structured interview approach. In recruiting participants, several questions about clinical characteristics such as symptoms and treatment were asked at the same time as screening for inclusion criteria, to ensure that participants were not overly biased in terms of severity of migraine or type of treatment used. The question topic list is shown in Supplemental Table S1. Participants were interviewed on topics including their demographic characteristics, migraine frequency, migraine intensity, use of acute or preventive treatments, and the understandability of the DCE attributes, levels, and question format in a draft version of the main survey. The interview results were used to confirm the understandability of the DCE questions and informing the final attributes, levels, and choice questions. Only count data was collected during the qualitative interviews and no statistical analyses were performed.

Quantitative Survey (DCE)

The online survey including the DCE was informed by a literature search, the qualitative interview study, and two experts in the field (a headache specialist and a neurologist working at a headache center and in the pharmaceutical industry, respectively). The survey was developed and administered in the Japanese language. A description of the variables/questions in the survey is shown in Table 1. Furthermore, the attributes and levels for the DCE described in Table 1 were selected for the purpose of understanding patient preferences for both acute and preventive migraine medications. The attributes were selected on the basis of researching the relevant migraine literature [16,17,18,19,20,21], clinical relevance, importance to participants during the qualitative interviews, and relevance to the research question. The survey collected information related to participant demographics, severity of the condition (Migraine Disability Assessment (MIDAS) test and Migraine Interictal Burden Scale (MIBS-4)) and their treatment preferences. The question topic list is shown in Supplemental Table S2. These variables were chosen for the purpose of understanding the backgrounds and treatment preferences of patients suffering from migraines.

Table 1 Attribute descriptions and levels used for the DCE
Full size table

Participants

Qualitative Interview Study

Ten participants were recruited for the qualitative interview portion of this study. As Japan has a homogeneous society, 10 participants was sufficient to reach a saturation of ideas [22]. The patients for the qualitative interviews met the same inclusion criteria as the DCE participants below, and the two groups were mutually exclusive.

DCE

A total of 400 participants were recruited for the study (attrition shown in Fig. 2). The sample size needed for the DCE part of the study was calculated on the basis of the number of choice tasks (T), the number of alternatives (A), and the largest number of levels across all attributes (C). The sample size equation derived from the Qualtrics white paper [23] is as follows: N ≥ [(1000 × C)/(T × A)] with N being the required sample size. In this DCE, A and C are 2 and 3, respectively. As T should be a minimum of 6, the equation for N is 1000 × 3/(6 × 2) = 250 respondents. To ensure enough sample for model estimation, 400 participants were recruited for this study. Only valid responses were retained, and incomplete submissions were not retained or counted towards the 400 samples.

Fig. 2
figure 2

Participant attrition. 37,500 individuals were invited to the survey, from which 20,056 responses were received. If a response to screening questions or questionnaire did not meet the inclusion criteria, the data from the participant was excluded from the analysis. The survey did not specify any quotas for participant recruitment

Full size image

Patients with migraine were recruited to complete the DCE through the Rakuten Insight panel in October and November of 2023. The Rakuten Insight special panel, a subsidiary of Rakuten Insight Inc. (https://insight.rakuten.com [24]), is a research group that conducts surveys on specialized panels. Patients who self-reported having a physician diagnosis of migraine in the panel were notified of the survey through e-mail and participants were enrolled on a rolling basis. Participants were compensated for their participation in the study.

Participants who met all of the following inclusion criteria were included in the study:

  1. 1.

    Participants with a diagnosis of migraine from a physician (self-reported).

  2. 2.

    Signed and dated an informed consent document indicating that the participant (or a legally acceptable representative) had been informed of the nature, purpose, and duration of the study.

  3. 3.

    Participants who received their diagnosis within the past 10 years.

  4. 4.

    Participants older than 18 (inclusive) and younger than 65 (exclusive, age of retirement) years of age.

  5. 5.

    Participants with more than four (inclusive) migraines per month.

  6. 6.

    Participants who were receiving prescription medication for migraines at the time of the study.

Participants meeting the criteria below were not included in the study:

  1. 1.

    Participants who were unable to understand Japanese.

Statistical Analyses

Analyses were conducted using Qualtrics XM. Other outputs such as subgroup analysis were calculated using R version 4.3.2. Statistical analyses for the DCE were based on the Conjoint Analysis offering on the Qualtrics platform, specifically, the hierarchical multinomial logistic regression model [23]. This model uses hierarchical Bayesian modeling to estimate a preference weight for each attribute level [25]. Relative attribute importance (RAI) was calculated as the mean (for the sample) of the difference in preference weights of the highest and lowest level (for each patient). Standard errors for preference weights and RAI estimates were calculated as the square root of the variance of the estimate over the sample or subsample. The outputs generated include attribute importance (the relative impact that an attribute has on a patient’s preference, conditional on the range of levels of the attribute), and average attribute-level utility estimates (preference weights of selected subgroups). Descriptive statistics (n, mean, median, SD, range, min, max for continuous variables, and n and percentage for categorical variables) were used to summarize the sociodemographic and clinical characteristics of the sample.

Subgroup Analysis

Subgroups were selected on the basis of previous literature and clinical relevance suggested by two headache experts. The acute treatment usage days per month, monthly migraine days (MMDs), Migraine Interictal Burden Health (MIBS-4) score, and Migraine Disability Assessment (MIDAS) were determined prior to the execution of the study as specified below. The preventive versus preventive plus acute treatment subgroup was explored on the basis of the importance of the “method of delivery” attribute. To understand the differences between the preferences of patients using different methods of delivery (oral versus injectable), the preferences were also presented for more precise subgroups as follows:

Acute treatment usage days per month: Two subgroups defined by current acute treatment usage days per month were created based on the Medication Overuse Headache (MOH) criteria. MOH is a headache that can be caused by using medication too frequently, the cut-off for which is 15 or more days of medication use per month [26].

MMDs: Three subgroups defined by MMDs were created on the basis of ranges of MMDs from previous studies [27]. Typically, 7–8 MMDs and 8–14 MMDs per month are considered low- and high-frequency episodic migraines, respectively, while 15 or more MMDs is considered chronic.

MIBS-4: The MIBS-4 score is rated on a scale of 0–10, where a score greater than or equal to 5 is considered to be severe [28]. Therefore, the two subgroups were defined as non-severe and severe participants.

MIDAS: MIDAS score grades I and II (0–10) are considered minimal or mild with infrequent disability, whereas grades III and IV (11–20) are considered moderate to severe disabilities [28]. Two subgroups were defined on the basis of these groupings of MIDAS scores.

Acute treatment versus acute and preventive treatment users: Two subgroups were created on the basis of patients’ medication usage (acute only versus acute plus preventive). A separate subgroup for preventive only participants was not created as a result of limitations of sample size.

Average utility values of these subgroups were calculated for each attribute by subgroup, and the means and estimated 95% confidence intervals were used to compare the subgroups within each of the sets described above, while means and standard deviations were used to compare RAI. On the basis of the questionnaire and study design, there were no missing values, and no attempts at imputation were made.

Ethics

The study was conducted in accordance with legal and regulatory requirements in Japan, and in accordance with the Helsinki Declaration of 1964 and its later amendments. An electronic informed consent was obtained on the questionnaire prior to participants entering the survey where the nature, purpose, and duration of the study were explained to each participant. Each participant was informed that they could withdraw from the study at any time and for any reason. The final protocol, any amendments, and informed consent documentation were reviewed and approved by the Shiba Palace Clinic Ethics Review Committee SOUKEN for this study on June 29, 2023 (No. 152411_rn-35258).

Results

Qualitative Interviews

The 10 participants from the qualitative interviews ranged in age from 27 to 57 years and were mostly female (female to male ratio 7:3). Six of the participants reported using only acute medications, while four reported using both acute and preventive treatments. Of the six participants who only used acute medications, one was able to describe preventive medications, while four had heard of them but had little knowledge, and one had never heard of them at all. When asked about preferences on dosage form of a hypothetical migraine medication, 10 had positive views of oral medications, while four had positive views of injectable medications.

Background Demographics by Treatment Type

The results from this study are focused on participants using only acute treatments, or those who use both acute and preventive treatments as the sample of participants using only preventive treatments was very small (n = 8, 2.00%). Participants were mostly female (n = 267, 66.75%) with an age distribution as follows: 40–49 (34.75%), 30–39 (26.50%), 50–59 (20.75%), 18–29 (12.75%), ≥ 60 (5.25%). Around half of the participants were full-time employees (n = 211, 52.75%) while about a quarter were part time (n = 85, 21.25%). A summary of all demographic characteristics of participants in the full sample are presented in Table 2.

Table 2 Background demographics
Full size table

Clinical Characteristics

In the full sample, 49.50% (n = 198) of patients reported using only acute treatments, 48.50% (n = 194) used both acute and preventive treatments, and 2.00% (n = 8) used only preventive treatments (Tables 3, 4, and 5). The most common age ranges for the age of diagnosis of migraine by a physician were 40–49 (35.00%, n = 140) and 30–39 (28.00%, n = 112). Participants using both acute and preventive treatments tended to have more MMDs than those who used only acute treatments (mean/SD 10.64/6.56 versus 9.37/5.88). In the full sample, the most common age bracket when first experiencing a migraine was 18–29 (n = 132, 33.00%), while the most common age bracket for age of diagnosis was 40–49 (n = 140, 35.00%). Overall, the participants visited a healthcare facility for their migraines a median of five times in the past year, and participants using both acute and preventive treatments visited healthcare facilities more often (six times) than participants using only acute treatment medications (four times). Among participants using both acute and preventive medications, 86.10% (n = 167) used an oral method of delivery product, while 9.80% (n = 19) used an injectable, and 12.40% (n = 24) used another type of medication. The full clinical characteristics of the participants are presented in Table 3.

Table 3 Clinical characteristics
Full size table
Table 4 MIDAS and MIBS-4 scores
Full size table
Table 5 Subgroup analyses for the method of delivery attribute
Full size table

MIDAS and MIBS-4 Instruments

The average MIDAS score for all participants was 19.60 (SD = 33.78), while participants taking both preventive  and acute treatments had severe disability with an average score of 22.71 (SD = 37.89), whereas participants using only acute treatments had an average score of 16.42 (SD = 29.31). The complete description of the MIDAS and MIBS-4 scores for participants is presented in Table 4.

The MIBS-4 scores demonstrated that on average, participants suffered from severe disability regardless of whether they took only acute or both acute and preventive medications (full sample, 5.59 (SD = 3.30); preventive and acute, 6.01 (SD = 3.47); acute only, 5.12 (SD = 3.06)).

DCE Findings

The RAI estimates indicate the impact that an attribute has on treatment choice given the range of levels of each attribute in the DCE. “Method of delivery” had the highest RAI (51.92, SD = 10.20), indicating that this attribute had the greatest impact on treatment choice. In addition, acute treatment was consistently preferred to oral preventive treatment which, in turn, was consistently preferred to injectable preventive treatment. This was followed by “reduction of pain when experiencing a headache” (17.00, SD = 7.74) and “The impact of headache on your daily routines” (10.27, SD = 4.61). The adjustability of the dosage was the least important attribute (3.87, SD = 2.13). RAI estimates are presented in Fig. 3.

Fig. 3
figure 3

Relative attribute importance. *Error bars are standard deviation (SD). Relative attribute importance of treatment attributes for the full study sample (N = 400). The Y-axis shows the attributes, while the X-axis shows the percentage of attribute importance, which is 100% in total for all attributes

Full size image

Subgroup Analyses

RAI by Subgroup

The data from the DCE were used to compare preferences of subgroups of patients with different demographic and clinical characteristics (Fig. 4a, b; details of “method of delivery” can be found in Supplemental Table S3). Some of the trends observed for the subgroups included a decreased importance of “method of delivery” as MMDs increased. In contrast, as MMDs increased, the importance of “reduction of pain when experiencing a headache” also increased. Participants with 15 or more acute-medicine usage days per month placed more importance on the “method of delivery” but less importance on “reduction of pain when experiencing a headache” than did those who had 14 or fewer usage days.

Fig. 4
figure 4figure 4

Relative attribute importance of treatment attributes by subgroups related to a clinical features and b PRO instruments. Error bars are standard deviation (SD). The Y-axis shows the attributes, while the X-axis shows the percentage of attribute importance, which is 100% in total for all attributes. MIBS-4 Migraine Interictal Burden Scale, MIDAS Migraine Disability Assessment, MMDs monthly migraine days, OTC over the counter

Full size image

Subgroups for the Method of Delivery Attribute

When observing the confidence intervals for the subgroups of the “method of delivery”, for current acute medicine usage days per month (OTC), it was found that those using medications less than or equal to 14 days per month placed more importance on the “method of delivery” (mean 3.68, CI 3.50–3.87) compared to those who used OTCs greater than or equal to 15 days per month (mean 3.20, CI 2.89, 3.50). Preference for oral acute medication decreased and preference for self-administered injections increased as MMDs increased; however, all groups preferred oral modes to injectables. For MIBS-4, a numerical trend was observed (minor overlap in CI) suggesting, that on average, those suffering from milder symptoms between migraines (MIBS-4 score less than or equal to 4) placed more importance on “method of delivery” (mean 3.54, CI 3.40–3.68) than those with a score greater than or equal to 5 (mean 3.29, CI 3.17–3.41).

Discussion

This study aimed to quantify the preferences of patients with migraine for treatment. To the best of our knowledge, this study was the first in Japan to quantify patient preferences for migraine treatments including both acute and preventive therapeutics through a discrete choice experiment (DCE).

The demographic and clinical characteristics of patients in this sample were generally aligned with those from previous studies on patients with migraine. These trends included factors such as age distribution, gender ratio, and employment status [1, 29]. However, this present study found a higher proportion of patients with migraine taking preventive treatments compared to previous epidemiological studies mentioned by the Clinical Practice Guideline for Headache Disorders 2021.

The higher prevalence of participants using preventive treatments in this study may suggest that patients with migraine are shifting their preferences towards preventive treatment options. This trend is corroborated by the results from a recent database study that investigated treatment practices in Japan [29]. Therefore, comparing the clinical characteristics of patients using preventive treatments to those using only acute treatments may be important to understanding what may be driving the trend toward patients’ greater acceptance of preventive treatments in Japan. When comparing those using both acute and preventive treatments with those using only acute treatments, those using preventive treatments on average had higher MMDs, higher usage of triptans and ditans, more hospital visits, and higher MIDAS and MIBS-4 scores. These results suggest that preventive medication users may be suffering from more severe migraines which may be why these patients choose to be on a preventative migraine medication. One explanation for the comparatively low preventive medication usage may be the limitations on the prescription of preventive treatments in Japan as currently only headache specialists can prescribe preventive treatment drugs, especially injectables. Overall, the demographics of the participants from this present study were comparable to previous literature on patients with migraine in Japan.

Discrete Choice Experiment

The results from this study demonstrated that method of delivery was by far the most important attribute for patients selecting a migraine treatment, followed by efficacy. The importance of mode in this study is consistent with other migraine DCE studies featuring patients from the USA and Germany in that method of delivery is an important driver of treatment preference [18]. This present study also found that acute treatments were preferred over preventive, and that oral medications were preferred over injectables. One explanation for this finding may be a lack of awareness or knowledge regarding preventive treatments. Six participants in the qualitative phase of this study who were not taking a preventive treatment were asked about their knowledge of preventive medications. Among these participants, only four (66.67%) were aware of the availability of preventive treatments, and only one (16.67%) was able to identify a preventive treatment. Although recent studies indicate that usage of preventive treatments is increasing [29], the practice of taking medications for headache relief at regular intervals may not be well established in Japan.

Although this study found that acute medications were overwhelmingly preferred, there were also interesting differences identified between preventive oral and injectable treatments. When comparing oral and injectable preventive treatments in the DCE, oral treatments were preferred. During the qualitative interviews, when participants were asked about their impressions/preferences regarding the dosage form of a hypothetical migraine medication, all 10 viewed oral treatments positively, while four were hesitant about injectables. Their concerns about injectables included the handling of injectors, portability, and the frequency of hospital visits. It is important to note that although oral treatments were preferred for prevention, participants with busy or irregular schedules mentioned concerns regarding forgetting to take their medications or being unable to take them during a regular daily cycle. As a result of the overall preference of oral preventive options over injectable treatments, it would be prudent to continue the development of treatment options to match the treatment preferences of migraine patients in Japan.

The trends found in the overall population were consistent with the subgroups observed, although some differing trends were identified. First, the RAI of mode decreased, and the RAI of pain relief increased as MMD increased. In addition, the subgroups that used OTC acute medications more than 15 days per month (MOH group) tended to place more importance on method of delivery than pain reduction. Nevertheless, the overall trends were similar across all subgroups, suggesting that method of delivery was important to patients regardless of demographic or clinical characteristics.

Clinical Implications

The results of our patient preference study for migraine treatment indicate a high need for prompt pain relief in the acute setting. Contraindications and risks for patients with cardiovascular complications are challenges for existing treatments in the acute phase, and progression to medication overuse headache with frequent administration is a major concern. Therefore, new therapeutic agents with different mechanism of action and better safety profile than existing drugs are expected in Japan. Moreover, this study has shown that there is a need for oral rather than injectable prophylactic therapy. Thus, developing drugs that are both effective and safe in the acute phase and easy to administer for prophylaxis are important to develop. In addition to CGRP-related antibodies, gepants are already available in some countries, and CGRP-related drugs, including gepants, are recommended by experts as the first choice, especially for prophylaxis [30], which may also indicate the needs of patients with migraine.

Limitations

All statistical procedures including the primary and other analysis outputs were automatically conducted in the Qualtrics platform. The options for the analytical procedures are limited and may not meet all the demands of the analysis of discrete choice data. Details of the analytical procedures are described on the Qualtrics website. Although a hierarchical Bayesian model was used to estimate the parameters in this study, the 95% confidence intervals are approximations of credible intervals and are not strictly calculated confidence intervals. Furthermore, it was not possible to include all attributes that may be relevant to a patient’s treatment choice. However, these practices are aligned with the practices of other online patient surveys. Additionally, the options in the DCE of this study assume that acute and preventive medications are mutually exclusive for use, whereas in reality there are medications that can be used for both acute and preventive treatment. Finally, there may have been biases in our sample. First, participants were self-reported as having a diagnosis of migraine from a physician, which may be a reliability bias. Second, there may be a knowledge bias as our qualitative interview results indicate that the majority of patients did not have a strong understanding of preventive treatments. This, however, also reflects the real-world clinical situation in Japan. Third, as the participants were exclusively recruited through the Rakuten Insight panel, there may have been a selection bias which may limit the generalizability of the findings.

Conclusion

The most important treatment attribute to patients with migraine was the method of delivery, followed by efficacy. Comparing the results of this study to previous literature suggests that the usage of preventive treatments may be increasing. Moreover, oral treatments were preferred over injectable treatments in this study regardless of participant demographic or clinical characteristics. As the prevalence of headaches and migraines remains prominent in Japanese society, understanding patient preferences for migraine treatment remains critical for developing medications that meet patients’ needs. Increasing education and awareness of preventive treatments may enable patients with higher frequency migraines to increase their access to preventive treatments.