FormalPara Key Points for Decision Makers

People with migraine who participated in focus groups consistently ranked reducing the frequency and severity of migraine attacks as the most important aspect of treatment, and also ranked avoiding adverse events as important.

People with migraine are dissatisfied with currently available preventive oral (non-calcitonin gene-related peptide targeting) medications due to a lack of efficacy and tolerability concerns; most of the focus group participants would be willing to self-inject a preventive treatment that is effective and tolerable.

People with migraine have different preferences for device characteristics that may affect the treatment burden of self-injectable treatments, such as the injection time, method of dose confirmation, mechanism of needle insertion and retraction, and storage requirements.

1 Introduction

Migraine affects more than 1 billion people worldwide [1], most of whom are working-age adults [2]. Among neurological disorders, it is the second-most common cause of disability [3]. Migraine can be episodic or chronic [4,5,6], and common symptoms include throbbing pain, photo- or phonophobia, nausea, and aura [7]. Many people with migraine experience a negative impact on functioning, including decreased productivity at work, school, and home, restricting family and social activities and resulting in an overall decrease in health-related quality of life [8,9,10]. In addition to the personal impact, the direct and indirect costs of migraine are high [11].

Individuals who experience at least four migraine headache days per month should be considered for preventive treatment [12]. Until recently, the main preventive treatment options were daily oral agents, such as beta-blockers, antidepressants, and antiepileptics. However, adherence and persistence to these treatments is low [13], primarily due to adverse events and a perceived lack of efficacy [14]. Monoclonal antibodies (mAbs) targeting calcitonin gene-related peptide (CGRP) that bind either CGRP itself or the CGRP receptor provide a novel alternative for people with migraine, especially for those in whom standard (non-CGRP-targeting) oral agents have failed. Erenumab, galcanezumab, and fremanezumab are self-injectable CGRP-targeting mAbs indicated for migraine prevention [15]. The CGRP-targeting mAbs appear to have similar efficacy and safety [16], although they have not been compared in controlled head-to-head studies.

Patients are expected to prefer treatments with favorable clinical properties; however, their preferences may be influenced by other factors, such as frequency of administration, ease of use, and storage, especially if different treatment options have comparable efficacy and safety [17]. Understanding the experiences and perspectives of people with migraine pertaining to preventive treatment characteristics, including delivery mechanisms (e.g., injection devices), is needed to help guide decision-making about preventive treatments and, for injectable treatments, the devices for administering them [18, 19]. This study aimed to inform the design of a preference-elicitation instrument, which is being developed to quantify preventive treatment preferences of people with migraine [20]. To this end, we explored what people with episodic or chronic migraine value as key characteristics of preventive treatments, including attributes of devices for subcutaneously administering self-injectable treatments.

2 Methods

2.1 Study Design

An in-person focus group study was conducted to identify patient-relevant treatment aspects. The focus groups used a discussion guide, whose design was informed by a targeted literature review (Online Resource 1, see the electronic supplementary material). The targeted literature review was conducted in May 2019 using the Ovid platform (which includes key medical literature databases such as EMBASE and MEDLINE). It yielded a total of 14 studies to be reviewed (six qualitative and eight quantitative). Data for outcomes and results were extracted, including patient and clinician experiences and treatment perception; themes and subthemes from patient-reported outcomes for qualitative studies; and preference-elicitation methods, attributes and levels, and trade-off measures for quantitative studies. A review of the extracted data identified 28 potentially relevant characteristics of migraine treatments: thirteen benefits (mostly related to symptom relief), eleven risks (e.g., loss of function, fatigue, drowsiness, and weight gain), and four regimen-related measures (e.g., mode and frequency of administration and ease of use). Only two out of the 14 studies considered preventive treatments for migraine [21, 22]; other studies addressed acute treatments or did not specify a treatment of interest. Manuscripts describing clinical trials of self-injectable preventive treatments for migraine were also reviewed to augment the targeted literature review. Additional risk measures pertinent to CGRP-targeting mAbs that were identified from clinical trials included nausea and injection site reaction [23,24,25].

Development of the focus group discussion guide was also based on concepts covered in the Subcutaneous Administration Assessment Questionnaire, a self-administered instrument that evaluates patient experiences of using devices such as a prefilled syringe or autoinjector [26]. The discussion guide (Online Resource 2, see the electronic supplementary material) included questions to elicit participant feedback about device characteristics.

Nine focus groups were conducted at five locations across the United States (US), the United Kingdom (UK), and Germany in October and November 2019, with the aim of having three focus groups per country, each with approximately seven participants (total n = 63). These countries were targeted to recruit individuals who, given the approval status in the countries and current treatment guidelines/recommendations from payers or physician societies, would be eligible for a prescription of a CGRP-targeting mAb.

A previous study on health-seeking behaviors found that 80% of the themes emerged within the first two to three focus groups, and that 90% emerged within the first three to six focus groups [27]. Nine focus groups conducted across three countries were therefore considered as sufficient for the purpose of this study. Saturation grids were used to analyze theme saturation within the target sample rather than determine sample size. The study also aimed to recruit an approximately equal number of participants with episodic and chronic migraine. The overall study design and format of the focus group discussions are illustrated in Fig. 1.

Fig. 1
figure 1

Overall study design and format of the focus group discussions

Full size image

2.2 Participants

The study included adults (≥ 18 years) with a self-reported diagnosis of migraine by a clinician, at least 4 headache days a month, and a Migraine Symptom Severity Score (MSSS) [28] indicating at least moderate severity. Participants were also currently using or had used a prescription preventive treatment for migraine within the last 5 years. Participants were considered to have episodic migraine if they had up to 14 headache days per month during the previous 3 months and chronic migraine if they had ≥ 8 days with the features of migraine out of a total of ≥ 15 headache days per month for > 3 months [6]. Eligibility was ascertained using a self-completed online screening questionnaire, which included questions about age, country of residence, health conditions, number of headache days, and treatment history, as well as the MSSS. To reach the maximum number of eligible participants at the desired locations within a recruitment period lasting 1-month, potential participants were recruited by convenience sampling through referrals from healthcare professionals, databases, social media, and patient association groups and were approached via email or telephone.

2.3 Focus Groups

Focus groups were held in conference rooms or market research facilities. Each focus group discussion lasted approximately 90 min and was conducted by a moderator [SM and ZB (both female) in the US; AH (female) in the UK; NP (female) in Germany] and co-moderator [JS (female) in the US; CT (female) and CAS (male) in the UK; TB (male) in Germany] in the country’s local language with the help of the discussion guide. The moderators and co-moderators were patient-centered outcome research professionals. They were trained in qualitative interview techniques, held a minimum of a master’s degree, and were fluent in the local language. Participants were given the moderator’s name and employer, as well as their reason for doing the research. Before the focus groups, there had been no contact between the moderators/co-moderators and the participants.

The focus groups began with an open discussion about experiences with migraine and preventive treatments (Fig. 1). Participants were encouraged to share their expectations and concerns about preventive treatments for migraine, to help inform development of benefit and risk attributes for the planned preference study. The moderator then presented the participants with five different demonstration single-use medical devices for subcutaneously administering CGRP-targeting mAbs (Fig. 2) and demonstrated how they should be used per the package insert instructions for use. The devices comprised two prefilled syringes and three autoinjectors that were unbranded, had already been discharged, and had their needles removed. The devices were presented to different focus groups in different orders to eliminate the possibility of ordering effects. The participants were then encouraged to simulate self-injection with each medical device. During and after handling and simulating self-injection using the five devices, participants were asked to share their perspectives on the aspects of the medical devices most important to them in making treatment decisions. This discussion identified device characteristics that mattered to patients.

Fig. 2
figure 2

Medical devices presented to the focus groups

Full size image

Next, participants were asked to rank treatment attributes identified in previous discussions by importance in a live polling exercise. This exercise was designed to help select candidate attributes for inclusion in the planned preference study [29]. The attributes included in the exercise were selected by the co-moderator and were ranked by participants anonymously using their cell phones. The number of attributes varied between six and 11 for the different focus groups. The polling results were shared and discussed within the focus group.

The focus group discussions were audio recorded digitally and then transcribed. The transcripts from the focus groups in Germany were subsequently translated into English. In addition, the co-moderator took field notes. Confidentiality was ensured by assigning each participant a unique participant number at enrolment and by removing any personally identifiable information from the transcripts before coding. Transcripts were not returned to participants for comment or correction.

2.4 Data Analysis

The analysis focused on identifying the themes that explain patients’ valuation of key characteristics of preventive treatments for migraine, as well as their perceptions of differences between treatments and the medical devices for administering them [30,31,32,33]. A combined approach consisting of inductive (bottom-up) and deductive (top-down) methods was used to capture data-driven themes while ensuring that specific topics of interest included in the semi-structured discussion guide were covered. This approach initially involved two researchers (CAS and EB [female]) independently double-coding the same two transcripts inductively using the qualitative data analysis program ATLAS.ti version 8.4.22 [34]. Next, a wider team meeting was held to compare and discuss the coding by reviewing the transcripts in conjunction with the initial codes; identifying, modifying, and merging codes of similar meaning; and identifying discrepancies and resolving them through discussion. In addition, the aims of the study were reviewed to inform the inclusion of additional codes to deductively capture data of interest that were not already obtained from the initial transcripts. This process of coding the first two transcripts yielded a coding framework (Online Resource 3, see the electronic supplementary material), which was used by one researcher (CAS) to code the remaining transcripts. The coding framework was updated iteratively to reflect new themes that emerged during coding and regular discussions with the research team. Focus group participants were not invited to provide feedback on the coding.

A conceptual map was produced to illustrate the main outcomes of the focus group study in a hierarchical framework [35]. Saturation grids were produced to evaluate whether new themes emerged toward the end of data collection [27].

3 Results

3.1 Participants

Forty-seven participants took part in nine focus groups (14 participants in the US, 17 in the UK, and 16 in Germany). Saturation analysis indicated that many of the higher-level themes reached saturation (i.e., no substantially new information would be expected to emerge from additional focus groups) (Online Resource 4, see the electronic supplementary material).

Of the 47 participants, 28 (59.6%) had episodic migraine and 19 (40.4%) had chronic migraine (Table 1). Most participants (n = 40; 85.1%) were female, and the median age was 50.0 years (interquartile range 37.5–56.0). Most participants (n = 28; 59.6%) scored ≥ 41 on the Migraine Disability Assessment, indicating very severe disability due to migraine [8]. The mean MSSS was 28.6, indicating that, on average, participants experienced migraine-related symptoms half the time or more during their most severe type of headache [28]. The most commonly used type of preventive treatment for migraine over the previous 5 years was antiepileptics (n = 20; 42.6%), followed by antidepressants and beta-blockers (n = 17 [36.2%] each). Twelve participants (25.5%) had experience using self-injectables (for migraine or another indication), and five participants (10.6%) were currently using self-injectable CGRP-targeting mAbs for migraine prevention. Most participants (n = 34; 72.3%) had no or a low-level fear of needles.

Table 1 Participant demographic and clinical characteristics
Full size table

3.2 Themes Arising from Focus Group Discussions

Fourteen key themes with 43 subthemes were identified across four different areas of discussion. These are illustrated as a conceptual map in Fig. 3.

Fig. 3
figure 3

Conceptual map summarizing identified themes and subthemes

Full size image

3.2.1 Migraine Experience

Themes and illustrative quotations related to migraine symptoms are provided in Online Resource 5 (see the electronic supplementary material). Migraine affected participants’ ability to work or attend school, causing them to “have many times off school,” feel “half-functioning,” switch from full-time to part-time, change to a day or night shift, be simply “not able to work,” or “quit” their online job. Severe migraine left some participants struggling to get out of bed due to pain, vomiting, or suicidal feelings and limited one participant to function just enough to “go to the toilet.” Participants also reported that migraine made it hard for them to care for their children. Some, particularly those from more suburban areas in the US, expressed an inability or unwillingness to drive because they would “lose vision” while driving.

3.3 Treatment Expectations and Concerns

Themes and illustrative quotations related to preventive treatments for migraine are provided in Table 2. The most discussed expectation of any preventive treatment for migraine was reduced severity and frequency of migraine. Participants in five of the nine focus groups indicated that they would like a preventive treatment that allows them to have “normal functions” and live a “normal life,” enabling them to work, leave the house, and perform daily activities that they were otherwise unable to do during a migraine attack. Participants anticipated that this would be achieved by the preventive treatment substantially reducing the frequency and severity of migraine attacks.

Table 2 Themes and illustrative quotes related to expectations and concerns about preventive treatments for migraine
Full size table

Participants from all focus groups had tried multiple daily (non-CGRP-targeting) oral preventive treatments for migraine. They expressed dissatisfaction with these treatments because of ineffectiveness or intolerable adverse events, which in some cases had caused them to switch or discontinue treatment. Participants wanted a preventive treatment to be completely free of adverse events, although discussions revealed that they would be willing to make trade-offs between effectiveness and adverse events. The most frequently discussed physical adverse event associated with preventive treatment (emerging in six focus groups) was weight change, with which participants from all three countries expressed their dissatisfaction. One participant who had used a self-injectable CGRP-targeting mAb reported constipation, which led to them switching to another CGRP-targeting mAb.

Participants across four focus groups reported experiencing depression and mood-related adverse events associated with preventive treatment. In four focus groups, participants discussed cognitive adverse events that they attributed to preventive treatments, such as impaired memory and concentration, which had interfered with work and school and made it difficult for them to communicate with others.

Treatment-associated adverse events that were discussed included concerns about withdrawal effects (with antidepressants), dependence, contraindications, and drug–drug interactions. One participant with past addiction behavior expressed a general wish for treatments not to be habit forming. Preventive treatment use also raised concerns about contraindication in patients with glaucoma and about drug–drug interactions with acute treatments for migraine.

3.4 Review of Medical Devices

Themes and illustrative quotations related to the devices are provided in Table 3. After simulating self-injection with the five different medical devices, participants generally preferred autoinjectors over prefilled syringes, describing them with words such as “robust,” “stable,” and “easy to use.” Some participants from three of the focus groups preferred prefilled syringes because they expected them to provide greater “control” of the administration process than autoinjectors. These participants generally preferred a large thumb rest and flange for providing a good grip of the prefilled syringe. Participants who had experience with self-injecting insulin pens were not opposed to using prefilled syringes because they would “get used to it” despite any fear of needles. However, most participants disliked prefilled syringes, describing them as “old-fashioned,” “creepy,” “scary,” “clumsy,” and “cheap and nasty.” Some felt that they were fragile and unreliable, believing that they might break during use and would leave “room for error.”

Table 3 Themes and illustrative quotes related to impressions of the five medical devices
Full size table

Many participants who were asked if they would be willing to try any of the self-injectable treatments they had handled indicated their willingness to try one that was effective and tolerable. In discussions of dosing frequency, willingness to self-inject depended on the effectiveness of the preventive treatment. As long as a treatment worked, some participants would be willing to self-inject it “every day” or “daily,” including participants who would otherwise be unhappy to have daily injections. One participant declared that they would be willing to administer an effective medicine “twice a day just to be able to function properly” and another “three times a day would be the maximal number that is still good to handle when [the medicine was] working”; other participants indicated that monthly administration would be ideal.

Device characteristics discussed by participants could be divided into six main themes: ease of handling, administration/preparation, needle, injection time, portability/storage, and dose confirmation. Device characteristics often judged to contribute to a stable grip included size, material, and shape, with participants across five focus groups expressing concern that they may experience an accident (i.e., the devices might “slip”) during administration. In particular, the bases of certain autoinjectors mattered, especially the large flat surface of device E. This was largely because it allowed the device to be held in place at the correct angle, thereby ensuring stable administration. Size and shape were felt to be particularly helpful for “the elderly” and people with “shaky” hands or “arthritis,” who would want a device that provides good grip and control.

Participants disliked overcomplicated or multistep administration and appreciated features that, in their view, made the process easier, reducing the risk of mistakes and accidents. Some participants specifically liked the automatic injection of autoinjectors, which was perceived to give “less room for error” and to avoid concerns about manual needle insertion and retraction. On the other hand, one participant was not convinced that the injection would be initiated when pressing a button on an autoinjector. The need to unlock autoinjectors was perceived as “more secure if you have children in your house” by one participant and a useful way of preventing accidental skin penetration or device activation by another. Some participants expressed concern that the requirement for a 45° angle with the prefilled syringes would make them “see the needle” against their will and might result in administration errors.

Although most participants preferred not having to pinch the skin before needle insertion, opinions about skin pinching were diverse, with some participants favoring pinching and others expressing indifference. Advantages of pinching the skin included feeling less anxious about accidentally injecting at the wrong site. Discomfort was mentioned as a disadvantage.

Participants considered prefilled syringes “not a good fit” and “dangerous,” especially for children, because of the exposed needle. Needle protection features of autoinjectors, which include auto-retraction or a needle shield/guard, were appreciated by participants, especially those averse to needles, because they contributed to safe, patient-friendly administration with the needle hidden. Needle auto-retraction offered by one of the autoinjectors was valued because people do not have to manually retract the needle from their skin and because it avoids errors and accidents caused by moving the device with the needle inserted or incorrectly retracting the needle from the skin. Similarly, some participants were worried about the possibility of such accidents if they were to move the device while the needle is inserted.

The respective injection times of the three autoinjectors were 3, 10, and 15 s. A shorter duration of injection was predominantly preferred. Several participants were not comfortable with an injection lasting 15 s or longer, especially one participant who disliked injections. Notably, one participant was concerned that a long injection time might cause them to miss the dose confirmation.

Although participants expressed differing views about whether control of injection speed was better achieved with prefilled syringes or autoinjectors, one participant indicated that using autoinjectors would eliminate concerns about how quickly they were injecting the drug.

Autoinjectors have two mechanisms to confirm dose completion: by making an audible click sound and visually via a confirmation window, which can be either a narrow window that changes color on dose completion or transparent housing providing a 360-degree view. Some autoinjectors click shortly before the full dose is administered, which concerned some participants who thought they might “pull it out sooner than the time.” However, one autoinjector clicks after dose completion; this was valued because it removed the need to count how many seconds had elapsed to confirm dose completion.

Although a narrow confirmation window that changes color was found to be reassuring, some participants felt it could be better positioned because it was obstructed by their breasts or adipose tissue or because they did not want to see the injection. A transparent housing was preferred by some because it allowed dose completion to be observed from any angle and because it could serve as a back-up confirmation if audible clicking did not work or if they missed the click sound. One participant considered visual confirmation more important than audible confirmation.

Portability was important for frequent travelers, especially for one participant who felt embarrassed about having to fly with a large box of medicines. However, portability was not a concern for everyone.

Storage requirements were also considered, because all the prefilled syringes and autoinjectors must be stored in a refrigerator prior to use. One participant explained that the size of the device and its packaging was important to them because of having limited refrigerator space.

Concerns about the impact of single-use devices on the environment were raised, particularly by participants from the UK and Germany, due to the size of certain devices and the amount of non-recyclable plastic they contain.

3.5 Ranking of Treatment Characteristics

Across focus groups, reducing the frequency of migraine and reducing the severity of migraine were consistently ranked as the most important characteristics (Online Resource 6, see the electronic supplementary material). However, some differences emerged among countries. Adverse events were deemed important by four of the nine focus groups (two in the UK and two in Germany). Two of the three US focus groups highlighted speed of injection as being important. Also, vision problems due to migraine were ranked as the most important attribute by participants of the focus group conducted in a US suburb, where it is necessary to drive. Environmental impact was a notable device concern among German focus group participants, although they ranked it as less important than other attributes.

3.6 Preference-Elicitation Instrument

Insights gained from the focus groups were used to develop a preliminary set of attributes and levels for a preference-elicitation instrument (Table 4). The levels of the device attributes were derived from the characteristics of the autoinjectors used to administer CGRP-targeting mAbs (Emgality® [galcanezumab], Aimovig® [erenumab], and Ajovy® [fremanezumab]), which are approved in the target countries [36,37,38,39,40,41].

Table 4. Attributes and levels included in the preliminary preference-elicitation instrument
Full size table

4 Discussion

This focus group study examined which characteristics of preventive treatments for migraine matter to people with migraine and why, with a focus on gaining insight on preferences for devices used to inject CGRP-targeting mAbs indicated for migraine prevention. During the focus group discussions, participants were given the opportunity to handle and simulate self-injection using several devices and were queried about which characteristics mattered to them. The study confirmed that reducing the frequency and severity of migraine attacks was a critical benefit expected from preventive treatments and that migraine symptoms greatly interfere with normal daily activities, including the ability to work [42, 43]. This is consistent with clinical guidelines that recommend change in migraine headache days per month as a primary outcome measure in clinical trials [44,45,46]. The focus group study also identified a number of adverse events of concern that are linked to oral and self-injectable preventive migraine treatments, including weight change, cognitive problems, and psychological problems; these were anticipated based on the results of the targeted literature review. The study also revealed adverse events and lack of effectiveness to be principal reasons for changing or discontinuing previous preventive treatments. This aligns with other studies showing that effectiveness and tolerability matter greatly to people with migraine and substantially affect adherence and persistence [13, 47, 48]. The focus group discussions identified only two new themes relating to treatment expectations and concerns (nausea and injection site pain) that had not been identified from the targeted literature review but from supplemental review with CGRP-targeting mAb clinical trial studies. By contrast, various device characteristics beyond mode and frequency of administration and out-of-pocket costs were highlighted as important, including ease of holding, specific injection steps (twisting to unlock and pressing the button), number of injection steps, shape, size and weight of the device, needle protection or visibility, disposal, portability, storage, injection time, dose confirmation, and pinching. Comparisons of the clinical findings were limited by the fact that only two papers identified during the targeted literature review considered preventive treatment for migraine. It is also possible that some participants may not have been fully able to disentangle treatment-related adverse events from migraine symptoms.

Patients’ valuation of self-injectable preventive treatments for migraine may be driven by the characteristics of the device used to administer them [49]. The current study discovered additional attributes explaining ease of use of devices that were not identified from the targeted literature review. The participants in this study generally found autoinjectors to be easy and safe to use, and they valued characteristics such as the unlocking step, automated insertion and retraction, and needle protection. Audible and visual dose confirmation also increased confidence in administering the full dose with an autoinjector. Although instructions for using autoinjectors were initially overwhelming to some participants, they generally believed this would not be a barrier once they became accustomed to using the devices.

The ranking exercise results further helped confirm the potential of the identified aspects to be attributes in the planned preference study. While ranked benefit attributes did not differ across the focus groups, it was difficult to consolidate risk attributes, as many discrete side effects were included in the ranking exercise. Another challenge was to materialize device characteristics into attributes. If only plausible pairs of device characteristics are to be considered based on actual devices, some characteristics can form a single attribute rather than being separate attributes. For instance, base of autoinjectors and skin pinching can be a single attribute for having three plausible pairs together: an autoinjector that has a narrow base and requires skin pinching, an autoinjector that has a narrow base and does not require pinching, and an autoinjector that has a wide base and does not require pinching. Furthermore, some device characteristics may overlap, such as size of the device, portability, storage, and disposal. This issue may be overcome by considering whether levels of attributes can be differentiated and represent a reasonable improvement from a base level. For instance, variation in storage requirements may include different numbers of days that the device can be stored at room temperature for the size of the device would be different dimensions of the package or the device itself.

This study had some limitations that should be considered when interpreting the results. Because this study had a relatively small sample size and was conducted in a limited number of geographic regions, its generalizability may be limited. Also, although the study included people with both episodic and chronic migraine spanning a wide age range (20–69 years) across different countries, the sample may not have been large or diverse enough to fully understand the treatment preferences of people with migraine or to explore any potential differences based on participant characteristics [19]. Notably, the study did not capture the perspectives of people with undiagnosed migraine or who have not previously received preventive treatments. However, the study included some participants who had experienced self-injecting treatments, which helped broaden the range of experiences and perspectives captured. A further potential limitation is that it was difficult to determine whether perceptions or experiences differed between participants with chronic and episodic migraine because all the UK focus groups and two of the German focus groups included both types of patients.

Describing attributes and levels can pose a challenge when transferring the results of a focus group study to a subsequent preference study. How to communicate device characteristics to participants in the planned preference study has to be carefully considered because, unlike the focus group participants, participants in the preference study will not be given the opportunity to physically handle and simulate device usability. Without tangible experience of self-injectable devices, people might find it difficult to understand how the devices look and function. We plan to leverage the focus group findings to develop visual aids such as illustrations and videos to help participants make informed choice decisions in the preference study.

5 Conclusion

In conclusion, this focus group study identified relevant attributes that might influence the decisions people with migraine make when selecting a preventive treatment for migraine. The study suggests that many people with migraine are willing to take a self-injectable preventive treatment. Faced with multiple self-injectable treatments with comparable efficacy and safety, patient preferences for injection time, dose confirmation, needle protection, storage requirements, and other factors that may affect the treatment burden can help guide treatment selection. This study will serve as a foundation for future preference studies that aim to quantify the trade-offs that people with migraine are willing to make when choosing between different preventive treatments for migraine, including self-injectables.