Abstract
Objective
To examine acute and preventive migraine prescription treatment needs among participants with active migraine and identify factors associated with needing treatment optimization.
Background
A high unmet treatment need exists for individuals with migraine.
Methods
ObserVational survey of the Epidemiology, tReatment and Care Of MigrainE (United States) is a prospective, multi‐cohort, longitudinal, web‐based survey in a demographically representative sample of US adults with active migraine in the preceding year. The current study is a pooled analysis of the baseline surveys of the 2018, 2019, and 2020 migraine cohorts. Criteria for determining need for migraine treatment initiation/modification were based on the American Headache Society 2021 consensus statement. For preventive treatment, need was defined based on the number of headache days per month and Migraine Disability Assessment Scale scores. For acute treatment, need was defined using the Migraine Treatment Optimization Questionnaire scores or based on the presence of two or more disability days per month. Participants were categorized into four groups based on their need for initiation/modification of acute or preventive treatment or both. The categories were “not a candidate (no need for treatment initiation/modification),” “acute need only,” “preventive need only,” or “preventive + acute needs.” Interictal burden was measured using the Migraine Interictal Burden Scale‐4.
Results
Among 59,001 participants with migraine, 77.1% of the total population were candidates for migraine prescription treatment initiation/modification. Among those, 39.4% had unmet acute needs only, 2.8% had unmet preventive needs only, and 34.9% met criteria for both. Up to one‐quarter of participants in each group were taking preventive medication. Acute prescription medication use was reported by 55.7% of those with preventive + acute needs, 46.9% of those with acute needs only, 36.9% of those with preventive needs only, and 27.1% of those who were not candidates for treatment initiation/modification. The highest rate of acute medication overuse was seen in the preventive + acute needs group (37.0%). Over half of the participants in the preventive + acute (69.9%) and acute (58.7%) needs groups reported experiencing moderate‐to‐severe interictal burden on the Migraine Interictal Burden Scale‐4. Migraine‐related stigma was highest among those with acute + preventive treatment needs, with 46.4% experiencing stigma often or very often. Experiencing severe interictal burden was most associated with having acute treatment needs (odds ratio [OR], 2.66; 95% confidence interval [CI], 2.48–2.86), whereas overusing acute medication was most associated with having preventive treatment needs (OR, 10.04; 95% CI, 8.64–11.65) and preventive + acute treatment needs (OR, 10.19; 95% CI, 9.09–11.43).
Conclusion
Over 75% of participants with migraine in this population sample were candidates for initiation or modification of migraine prescription treatment. These findings highlight the opportunity for optimizing treatment and improving outcomes for patients with migraine.
Keywords: acute, medication, migraine, preventive, unmet treatment need
Plain Language Summary
This study evaluated the need for prescription migraine treatment in people with active migraine by examining headache frequency, migraine disability scores, and treatment optimization questionnaire results. It found that many participants had unmet treatment needs, with acute medication overuse, burden between attacks, and seeking higher levels of care being the strongest factors linked to the need for treatment optimization. These findings highlight important opportunities to improve care and outcomes for people living with migraine.
Abbreviations
- AHS
American Headache Society
- AMO
acute medication overuse
- AMPP
American Migraine Prevalence and Prevention Study
- CaMEO
Chronic Migraine Epidemiology and Outcomes
- CI
confidence interval
- ED
emergency department
- ICHD‐3
International Classification of Headache Disorders, 3rd edition
- LASSO
least absolute shrinkage and selection operator
- MAST
Migraine in America Symptoms and Treatment
- MIBS‐4
Four‐Item Migraine Interictal Burden Scale
- MiRS
Migraine‐Related Stigma questionnaire
- MSQ‐RFR
Migraine‐Specific Quality of Life Questionnaire v2.1–Role Function‐Restrictive
- MSSS
Migraine Symptom Severity Score
- mTOQ
Migraine Treatment Optimization Questionnaire
- OR
odds ratio
- OVERCOME
ObserVational survey of the Epidemiology, tReatment and Care Of MigrainE
- PHQ‐4
Four‐Item Patient Health Questionnaire
- RC
retail community
- RF
random forest
- SD
standard deviation
- UC
urgent care
- US
United States
INTRODUCTION
Migraine is a complex neurologic disease that is highly prevalent 1 and affects as many as one in every six Americans. 2 It ranks as the world's second leading cause of years lived with disability, 3 , 4 and many experience migraine‐related stigma . 5 , 6 , 7 , 8 Effective migraine treatment can reduce disability and improve functioning. 9 , 10 Preventive treatment can reduce the frequency and burden of migraine attacks and decrease overall healthcare costs. 11
Individuals with migraine often use over‐the‐counter acute migraine treatments 12 ; however, potentially effective acute prescription treatments are underutilized. 13 , 14 The underutilization of preventive treatment for migraine is even more pronounced. 1 , 15 Together, this highlights the large unmet treatment need among patients with migraine. Guidelines and consensus statements on acute and preventive migraine treatment indicate a number of available medication classes beneficial for preventing and/or treating migraine. 16 , 17 The 2021 consensus statement from the American Headache Society (AHS) outlines criteria for identifying patients for whom preventive treatment should be “offered” or “considered.” 18 This statement expands on previous recommendations developed by the American Migraine Prevalence and Prevention (AMPP) study advisors to “consider” or “offer” preventive treatment based on monthly headache days and attack‐related impairment. 1 Given these recommendations, it is important to understand how candidates for migraine treatment optimization differ across the population, and which sociodemographic, clinical, and migraine characteristics are associated with being a candidate for treatment. This information will enable clinicians to better tailor treatments to the individual needs of people with migraine.
The ObserVational survey of the Epidemiology, tReatment and Care Of MigrainE (OVERCOME [US]) study characterized sociodemographic and migraine‐related characteristics, along with medication use, in individuals with migraine. One key objective of the current analysis was to characterize unmet preventive and acute treatment needs in the OVERCOME study population using the AHS recommendations. The second objective was to use machine‐learning methods to identify sociodemographic, clinical, and migraine‐related factors associated with the need for treatment optimization. We hypothesized that there would be a high unmet need for preventive and acute treatment optimization among participants and that several clinical and migraine‐related factors would be associated with the need for better treatment.
METHODS
Ethics approval and consent to participate
Individuals who were interested in participating in the survey voluntarily provided electronic informed consent. Approval for this study was provided by the Sterling Institutional Review Board (ID #6425‐001).
Study design
OVERCOME (US) is a prospective, multi‐cohort, longitudinal, web‐based survey that annually (2018–2020) recruited a demographically representative adult sample in the United States.
Quota sampling was used to establish a demographically representative sample (by geographic region, age, race, and sex). Additional details regarding the recruitment and fielding process used in the OVERCOME study have been previously published. 14 , 19 The current analysis focuses on the baseline survey of the 2018, 2019, and 2020 migraine cohorts.
Participants
Individuals who passed the demographic screener were identified with active migraine if they reported ≥1 headache in the previous 12 months (not caused by illness, injury, or hangover) and met criteria for migraine per the validated American Migraine Study/AMPP study migraine diagnostic questionnaire, 1 , 20 which uses the ICHD‐3 criteria, 21 , 22 or self‐reported a medical diagnosis of migraine. Individuals with active migraine then completed the full migraine survey. Each question in the survey required an answer, thus eliminating missing data. For the current analysis, only those meeting migraine criteria assessed via the diagnostic questionnaire were included (Figure S1). The sample size required was calculated based on the level of precision obtained on the proportion of patients with a specific attribute (e.g., percent qualifying for preventive migraine treatment); a sample of 40,000 individuals with migraine imparted a minimum precision of 0.75%. An additional cohort of ~20,000 individuals was later added to capture patients on newly approved migraine‐specific treatments.
Determining treatment needs
Participants were categorized by candidacy for initiating or modifying prescription migraine treatment(s) (Table 1A). The four groups were: (1) not a candidate (those who were not candidates for treatment initiation/modification), (2) acute (candidate for acute prescription treatment initiation/modification only), (3) preventive (candidate for preventive prescription treatment initiation/modification only), and (4) preventive + acute (candidate for both preventive and acute prescription treatment initiation/modification). Groups were mutually exclusive. These categories were based on defined treatment need regardless of whether medication was currently being used. When analyzing current medication use, “recommended” preventive and acute medications were defined as those with established/probable efficacy as identified by the American Academy of Neurology/AHS guidelines and the AHS consensus statement regarding new migraine treatments. 16 , 18
TABLE 1.
Definitions of candidacy for initiating or modifying migraine treatment and categories of treatment need: OVERCOME baseline sample (N = 59,001).
| A. Treatment need categories by acute and preventive candidacy | |||
|---|---|---|---|
| Preventive | |||
| No | Yes | ||
| Acute | Yes | 39.4% (acute) | 34.9% (preventive + acute) |
| No | 22.9% (not a candidate) | 2.8% (preventive) | |
| B. Candidate for initiating or modifying preventive prescription treatment a | ||
|---|---|---|
| Headache days per month | MIDAS | |
| ≥6 | and | Any |
| 4 or 5 | and |
Moderate (11–20) or Severe (≥21) |
| 3 | and | Severe (≥21) |
| C. Candidate for initiating or modifying acute prescription treatment | ||
|---|---|---|
| mTOQ‐4 b | MIDAS | |
| Poor/very poor (0–5) | or |
Mild (6–10) Moderate (11–20) Severe (≥21) |
Abbreviations: mTOQ‐4, Migraine Treatment Optimization Questionnaire‐4; OVERCOME, Observational Survey of the Epidemiology, Treatment, and Care of Migraine.
Recommendations from the American Headache Society consensus statement were used. 18
The mTOQ‐4 was not applicable for some respondents who were not taking acute medication.
Criteria for determining the need for initiating or modifying prescription migraine treatment were based on the AHS 2021 consensus statement. 18 To determine candidacy for initiating or modifying preventive treatment, we applied the criteria from the AHS consensus statement for identifying when preventive treatment should be “offered.” 18 Respondents needed to meet one of the following criteria: six or more headache days per month, four to five headache days per month with moderate‐to‐severe disability, or three headache days per month with severe disability (Table 1B).
The AHS consensus does not specify criteria for determining the adequacy of acute treatment or the need for additional optimization, in part because conventional wisdom states that everyone with migraine needs some form of acute treatment. Accordingly, the OVERCOME medical advisory group developed criteria using a consensus process. We determined that acute treatment should be modified for those with poor optimization of current acute treatment and/or those experiencing disability from migraine on an average of 2 or more days per month. We operationalized these criteria based on poor or very poor treatment efficacy per the Migraine Treatment Optimization Questionnaire (mTOQ‐4) (defined by scores 0–5) or mild, moderate, or severe disability based on a MIDAS score ≥6 (Table 1C). This requires an average of 2 disability days per month for 3 months.
Variables and measures of interest
The need for preventive treatment was based on the number of headache days per month and the MIDAS score. 22 , 23 The need for acute treatment was based on the MIDAS or the mTOQ‐4 24 score. The MIDAS 22 , 23 quantifies the number of days an individual missed or had reduced productivity at work/home/social events over the preceding 3 months. Disability was categorized as 0–5 (little/none), 6–10 (mild), 11–20 (moderate), or ≥ 21 (severe). The mTOQ‐4 24 assessed acute treatment optimization via four items each scored 0–2, generating total scores ranging from 0 to 8 (0 = very poor treatment efficacy, 1–5 = poor treatment efficacy, 6–7 = moderate treatment efficacy, and 8 = maximum treatment efficacy).
Sociodemographic characteristics included age, natal sex, race, ethnicity, income, health insurance, urban residence, marital status, employment status, and educational attainment (Table 2). Clinical characteristics included years with migraine, migraine symptoms interfering with sleep or mood, headache pain intensity, acute medication overuse (AMO) for migraine medications (defined according to medication use thresholds in the ICHD‐3 diagnostic criteria for medication‐overuse headache and assessed via survey questions asking about frequency of use for each acute medication), 21 the Four‐Item Patient Health Questionnaire (PHQ‐4), 25 which assesses depressive and anxious symptomology, and the most specialized level of care sought for migraine/severe headache in the previous 12 months. For level of care, response options were grouped into “primary care” (primary care/internal medicine/family medicine), “specialist” (neurology/headache specialist/pain specialist), “emergency department/urgent care/retail care” (includes retail community/pharmacy walk‐in/convenience care clinic), or “other” (other specialist). Other measures examined included the following:
Migraine‐Related Stigma questionnaire (MiRS): The frequency of experiencing migraine‐related stigma was assessed using a 12‐item questionnaire. 26 Items assess two main domains: “feeling that others viewed migraine was being used for secondary gain” and “feeling that others were minimizing the burden of migraine.” Responses were classified into three groups by frequency of experiencing secondary gain and/or minimizing burden: never, rarely/sometimes, or often/very often. 26
Four‐Item Migraine Interictal Burden Scale (MIBS‐4): Burden of migraine between attacks during the previous 4 weeks was measured using the MIBS‐4. 27 , 28
Migraine‐Specific Quality of Life Questionnaire v2.1–Role Function‐Restrictive (MSQ‐RFR): The functional impact of migraine on work‐related and social activities over the previous 4 weeks was measured using the seven‐item MSQ‐RFR subscale. 29 , 30 , 31
Migraine Symptom Severity Score (MSSS): The frequency of experiencing certain migraine‐related symptoms was measured using the MSSS. 32 , 33 Full details regarding these clinical characteristics can be found in Table 3.
TABLE 2.
Sociodemographic characteristics by treatment need group.
| Not a candidate (N = 13,535 [22.9%]) | Acute needs only (N = 23,217 [39.4%]) | Preventive needs only (N = 1675 [2.8%]) | Preventive + acute needs (N = 20,574 [34.9%]) | Total (N = 59,001) | |
|---|---|---|---|---|---|
| Age, years; mean (SD) | 44.0 (15.9) | 39.9 (14.1) | 46.3 (16.2) | 40.8 (13.7) | 41.3 (14.6) |
| Female, n (%) | 9680 (71.5) | 16,847 (72.6) | 1234 (73.7) | 16,417 (79.8) | 44,178 (74.9) |
| Race, n (%) a | |||||
| White | 9769 (72.2) | 15,180 (65.4) | 1317 (78.6) | 15,121 (73.5) | 41,387 (70.1) |
| Black | 1059 (7.8) | 2222 (9.6) | 82 (4.9) | 1197 (5.8) | 4560 (7.7) |
| Other | 2707 (20.0) | 5815 (25.0) | 276 (16.5) | 4256 (20.7) | 13,054 (22.1) |
| Hispanic, n (%) | |||||
| Yes | 1238 (9.1) | 2957 (12.7) | 134 (8.0) | 2082 (10.1) | 6411 (10.9) |
| No | 11,888 (87.8) | 19,346 (83.3) | 1486 (88.7) | 17,787 (86.5) | 50,507 (85.6) |
| Prefer not to answer | 409 (3.0) | 914 (3.9) | 55 (3.3) | 705 (3.4) | 2083 (3.5) |
| Income categories, n (%) | |||||
| Prefer not to answer | 560 (4.1) | 603 (2.6) | 62 (3.7) | 461 (2.2) | 1686 (2.9) |
| Less than $49,999 | 5853 (43.2) | 11,760 (50.7) | 770 (46.0) | 11,183 (54.4) | 29,566 (50.1) |
| $50,000–$99,999 | 4489 (33.2) | 7139 (30.7) | 535 (31.9) | 6255 (30.4) | 18,418 (31.2) |
| $100,000 or over | 2633 (19.5) | 3715 (16.0) | 308 (18.4) | 2675 (13.0) | 9331 (15.8) |
| Urban residence, n (%) | 11,454 (84.6) | 19,606 (84.4) | 1404 (83.8) | 16,523 (80.3) | 48,987 (83.0) |
| Health insurance, n (%) | |||||
| Yes | 11,686 (86.3) | 19,594 (84.4) | 1453 (86.7) | 17,503 (85.1) | 50,236 (85.1) |
| College graduate or above, n (%) | |||||
| Yes b | 5393 (39.8) | 8097 (34.9) | 671 (40.1) | 6289 (30.6) | 20,450 (34.7) |
| Married, n (%) | |||||
| Yes | 7634 (56.4) | 12,526 (54.0) | 952 (56.8) | 11,386 (55.3) | 32,498 (55.1) |
| Employed, n (%) | |||||
| Yes c | 7928 (58.6) | 13,931 (60.0) | 919 (54.9) | 10,728 (52.1) | 33,506 (56.8) |
Abbreviations: n, sample population; N, total population; SD, standard deviation.
Respondents selected all that applied among the following races: American Indian or Alaska Native; Asian or Asian American (including Asian Indian, Chinese, Filipino, Japanese, Korean, and Vietnamese); Black or African American; Native Hawaiian or Pacific Islander; White or Caucasian; and other.
College graduates: 0.4% to 0.6% of participants among the groups preferred not to answer the college graduate question.
Employment: 0.8% to 1.1% of participants among the groups preferred not to answer the employment question.
TABLE 3.
Clinical characteristics by treatment need group.
| Not a candidate (N = 13,535) | Acute needs only (N = 23,217) | Preventive needs only (N = 1675) | Preventive + acute needs (N = 20,574) | Total (N = 59,001) | |
|---|---|---|---|---|---|
| Migraine diagnosis class, n (%) | |||||
| Migraine only | 2145 (15.8) | 4223 (18.2) | 300 (17.9) | 3506 (17.0) | 10,174 (17.2) |
| No migraine or headache diagnosis | 5102 (37.7) | 4727 (20.4) | 427 (25.5) | 2438 (11.8) | 12,694 (21.5) |
| Migraine plus other headache diagnosis | 2921 (21.6) | 8499 (36.6) | 542 (32.4) | 10,773 (52.4) | 22,735 (38.5) |
| Other headache diagnosis only | 3367 (24.9) | 5768 (24.8) | 406 (24.2) | 3857 (18.7) | 13,398 (22.7) |
| MSSS, mean (SD) a | 15.8 (3.0) | 17.1 (2.9) | 16.4 (3.0) | 18.3 (2.5) | 17.2 (3.0) |
| PHQ‐4, mean (SD) b | 2.8 (3.0) | 4.4 (3.3) | 3.4 (3.3) | 5.6 (3.5) | 4.4 (3.5) |
| Years with migraine, mean (SD) | 20.4 (15.8) | 17.3 (14.2) | 21.2 (16.5) | 19.0 (14.0) | 18.7 (14.6) |
| Acute medication overuse – any, n (%) c | |||||
| Yes | 430 (3.2) | 1729 (7.4) | 462 (27.6) | 7610 (37.0) | 10,231 (17.3) |
| MIBS‐4, n (%) d | |||||
| None/mild interictal burden | 10,295 (76.1) | 9592 (41.3) | 1190 (71.0) | 6194 (30.1) | 27,271 (46.2) |
| Moderate/severe interictal burden | 3240 (23.9) | 13,625 (58.7) | 485 (29.0) | 14,380 (69.9) | 31,730 (53.8) |
| Most specialized level of headache care sought, n (%) e | |||||
| None | 10,012 (74.0) | 11,041 (47.6) | 1005 (60.0) | 6677 (32.5) | 28,735 (48.7) |
| ED/UC/RC | 417 (3.1) | 1343 (5.8) | 52 (3.1) | 1040 (5.1) | 2852 (4.8) |
| Primary care | 2015 (14.9) | 6019 (25.9) | 341 (20.4) | 6222 (30.2) | 14,597 (24.7) |
| Specialist | 904 (6.7) | 4485 (19.3) | 242 (14.4) | 6357 (30.9) | 11,988 (20.3) |
| Other | 187 (1.4) | 329 (1.4) | 35 (2.1) | 278 (1.4) | 829 (1.4) |
| MiRS, n (%) f | |||||
| Never | 3193 (23.6) | 2036 (8.8) | 345 (20.6) | 1141 (5.5) | 6715 (11.4) |
| Rarely or sometimes | 8793 (65.0) | 13,814 (59.5) | 1087 (64.9) | 9884 (48.0) | 33,578 (56.9) |
| Often or very often | 1549 (11.4) | 7367 (31.7) | 243 (14.5) | 9549 (46.4) | 18,708 (31.7) |
| MSQ‐RFR, mean (SD) g | 72.8 (19.4) | 53.0 (21.5) | 67.9 (18.7) | 41.3 (20.8) | 53.9 (23.9) |
| Migraine symptoms interfere with sleep, n (%) h | |||||
| A good bit of time or more | 2801 (20.7) | 11,150 (48.0) | 444 (26.5) | 12,860 (62.5) | 27,255 (46.2) |
| Migraine symptoms interfere with mood, n (%) i | |||||
| A good bit of time or more | 4582 (33.9) | 14,787 (63.7) | 643 (38.4) | 16,168 (78.6) | 36,180 (61.3) |
| Pain has moderate/severe intensity, n (%) j | |||||
| Half of the time or more | 8421 (62.2) | 16,748 (72.1) | 1255 (74.9) | 17,803 (86.5) | 44,227 (75.0) |
Abbreviations: ED, emergency department; ICHD‐3, International Classification of Headache Disorders, 3rd edition; MIBS‐4, Four‐Item Migraine Interictal Burden Scale; MSQ‐RFR; MiRS, Migraine‐Related Stigma questionnaire; Migraine‐Specific Quality of Life Questionnaire v2.1–Role Function‐Restrictive; MSSS, Migraine Symptom Severity Score; n, sample population; N, total population; PHQ‐4, Four‐Item Patient Health Questionnaire; RC, retail community; SD, standard deviation; UC, urgent care.
Assesses seven headache features with five response options: 0 = never, 0 = rarely, 1 = less than half the time, 2 = half the time or more, and 3 = all or nearly all the time. Scores range from 0 to 21.
Measures how frequently key symptoms of depression and anxiety were experienced during the previous 2 weeks. Scores range from 0 to 12, and higher scores indicate more frequent symptoms.
Respondents indicated which medications they used to relieve or treat migraine or severe headache attack and the number of days using each of those medications to treat migraine in the past 30 days. Medication overuse was defined according to medication use thresholds included in the ICHD‐3 diagnostic criteria for medication‐overuse headache.
The interictal burden of migraine (i.e., burden of migraine between attacks) was measured using the MIBS‐4. Each item in the MIBS‐4 contains item responses ranging from “never” to “most or all of the time,” with total scores ranging from 0 to 12. For the current study, we utilized the following MIBS‐4 interictal burden categories: 0 = none, 1–2 = mild, 3–4 = moderate, and ≥5 = severe.
Respondents indicated whether they sought care in any of the following settings: primary care (primary care/internal medicine/family medicine), neurology, headache specialist, pain specialist, ED, UC, and RC/pharmacy walk‐in or convenience care clinic. Responses were grouped as none, ED/UC/RC, primary care, specialty care (neurology, headache specialist, pain specialist), and other.
Frequency of experiencing migraine‐related stigma was measured using the 12‐item MiRS questionnaire. Items assess two main domains: “feeling that others viewed migraine was being used for secondary gain” and “feeling that others were minimizing the burden of migraine,” with responses ranging from “never” to “very often.” Scores in each domain determined assignment into one of three overall groups: never, rarely or sometimes, and often/very often.
The functional impact of migraine on social and work‐related activities over the previous 4 weeks was measured using the seven‐item MSQ‐RFR. Each item contains six response options ranging from “none of the time” to “all of the time.” The raw score is transformed to a score of 0 to 100, with higher scores indicating better role function.
Respondents were asked how often the pain and symptoms accompanying their typical migraine or severe headache attack interfered with their sleep. Response options included “none of the time,” “a little bit of the time,” “some of the time,” “a good bit of the time,” “most of the time,” “all of the time,” and “not applicable.” Table depicts a combination of “a good bit of the time,” “most of the time,” and “all of the time.”
Respondents were asked how often the pain and symptoms accompanying their typical migraine or severe headache attack interfered with their mood. Response options included “none of the time,” “a little bit of the time,” “some of the time,” “a good bit of the time,” “most of the time,” “all of the time,” and “not applicable.” Table depicts a combination of “a good bit of the time,” “most of the time,” and “all of the time.”
Respondents were asked how often they experienced pain with moderate or severe intensity with their most severe headache. Response options included “never,” “rarely,” “less than half the time,” “half the time or more,” and “all or nearly all of the time.” Table depicts a combination of “half the time or more” and “all or nearly all of the time.”
Statistical analyses
Sociodemographic, clinical, and migraine‐related variables were summarized using means with standard deviations (SD) for continuous variables, percentages for dichotomous variables, and proportions for ordinal categorical variables. Using the definitions specified above, each eligible participant with migraine was classified into one of the four treatment need groups. The analyses focused on identifying which of the 55 predictor variables (Table S1) were associated with the need for treatment initiation/modification (variable selection) and quantifying the level of association for those variables. Two methods were used for variable selection: (1) a random forest (RF) consisting of 1000 trees, and (2) a linear model structure with main effects for all 55 variables with a least absolute shrinkage and selection operator (LASSO) algorithm. SAS Enterprise Guide 8.2 (SAS Institute Inc., Cary, NC, USA) was used for all analyses.
An RF is an efficient approach to prediction algorithms because it handles nonlinearities in response and accounts for all possible interactions. It also handles missing data through surrogate splits. RFs with bootstrap aggregation (bagging) generalize single tree prediction ability and provide variable importance measures (using the RF algorithm in SAS) that are used to help understand and rank variables in their order of predictive importance. The out‐of‐bag Gini measure of variable importance was used to assess variables for further inspection. Variables with a Gini index near 0 were not selected for further evaluation because they indicate variables of low importance.
An additional method of linear regression with the LASSO algorithm (using the implementation in SAS) was used to conduct variable selection in predicting the need for treatment initiation/modification by shrinking as many regression coefficients to zero as possible, so the terms remaining have the most important predictive value and were subjected to further inspection. This was used as a supporting analysis to the RF under a different model structure of the response and could be described and predicted as a linear function of a subset of the 55 variables.
The variables selected from the RF and the set resulting from the LASSO regression were combined and used in a logistic regression for treatment need to evaluate the odds ratio (OR) for each of the variables in the regression. The logistic model for computing ORs included all associated factors; therefore, all ORs are adjusted. All continuous variables were standardized and are shown in Figure 1. The 95% confidence interval (CI) was computed for each OR. ORs were considered statistically significant at the 5% level of significance if the 95% CI did not include the value "1" using a 2‐tailed hypothesis test.
FIGURE 1.
Factors associated with treatment need by (A) acute treatment need only, (B) preventive treatment need only, and (C) preventive + acute treatment need: Results of a standardized multivariable logistic regression model (OR estimates). Logistic regressions are shown with standardized data for continuous variables. Standardized variables included: age, PHQ‐4, years with migraine, MSSS, Headache Pain Intensity Score, MSQ‐RFR. OR and 95% CI are presented for each variable. CI, confidence interval; ED, emergency department; MIBS‐4, Four‐Item Migraine Interictal Burden Scale; MSQ‐RFR, Migraine‐Specific Quality of Life Questionnaire v2.1–Role Function‐Restrictive; MSSS, Migraine Symptom Severity Score; OR, odds ratio; PHQ‐4, Four‐item Patient Health Questionnaire; RC, retail community; UC, urgent care. [Color figure can be viewed at wileyonlinelibrary.com]
RESULTS
Candidacy for prescription pharmacologic migraine treatment initiation/modification
Within the OVERCOME (US) baseline cohort of individuals meeting diagnostic criteria for migraine (N = 59,001) (Figure S1), 77.1% were candidates for migraine prescription treatment initiation/modification; 39.4% met criteria for acute prescription medication initiation/modification only, 2.8% for preventive prescription medication initiation/modification only, and 34.9% for both preventive and acute prescription medication initiation/modification (Table 1A). In total, 74.2% had unmet acute treatment needs and 37.7% had unmet preventive treatment needs. Overall, 22.9% of respondents did not need initiation or modification of pharmacologic treatment (Table 1A).
Sociodemographic and clinical characteristics
Sociodemographic characteristics of the study population are shown in Table 2. Overall, sociodemographic characteristics across the treatment groups were similar. The mean age of participants ranged from 39.9 to 46.3 years; 71.5% to 79.8% identified as female; 65.4% to 78.6% identified as White; and 8.0% to 12.7% identified as Hispanic. The percentage of individuals who were college graduates or above ranged from 30.6% to 40.1%, and about half, overall, were employed (52.1% to 60.0%).
Clinical characteristics of the study population are shown in Table 3. The mean number of years with migraine among respondents ranged from 17 to 21 years and was longest in the preventive need group. Not having a self‐reported medical diagnosis of migraine or other headache types was more common in people without a treatment need. Rates of self‐reported medical diagnosis of migraine and/or headache were highest in those with both acute and preventive needs (88.2%), lower in those with only preventive needs (74.5%) or only acute needs (79.6%), and lowest in those with no treatment need (62.3%). AMO was evident in all groups, with the highest percentage in the preventive + acute need group (37.0%). Moderate‐to‐severe interictal burden was reported by more than half of respondents in the preventive + acute need group (69.9%) and acute‐only group (58.7%), lower in those with preventive treatment only (29.0%), and lowest in those without need for acute or preventive treatment (23.9%) (Table 3). All groups reported experiencing migraine‐related stigma, with 46.4% of the preventive + acute group experiencing stigma often or very often (not a candidate, 11.4%; acute, 31.7%; preventive, 14.5%).
Factors associated with candidacy for treatment initiation/modification
Factors associated with candidacy for migraine prescription treatment initiation/modification were assessed using both RF and LASSO analyses. The top three variables identified by the RF were MSQ‐RFR, AMO, and interictal burden (MIBS‐4) (Table S2). The LASSO analyses identified six variables associated with the need for treatment initiation/modification (Table S2). There were 13 variables in total identified by either RF or LASSO with three variables identified by both methods: MSQ‐RFR, average headache pain intensity, and PHQ‐4 score.
The results of the multivariable logistic regression analysis revealed the factors most associated with candidacy for treatment initiation/modification in each group (Figure 1). The factors strongly associated with acute treatment needs were moderate‐to‐severe interictal burden (MIBS‐4 moderate vs. none: OR, 1.93 [95% CI, 1.78–2.08]; severe vs. none: OR, 2.66 [95% CI, 2.48–2.86]), seeing a specialist (OR, 1.99 [95% CI, 1.83–2.17]), and AMO (OR, 1.70 [95% CI, 1.51–1.92]). The factors most associated with preventive treatment needs were AMO (OR, 10.04 [95% CI, 8.64–11.65]) and seeing a specialist (OR, 1.84 [95% CI, 1.55–2.18]). In the preventive + acute treatment need group, the factors strongly associated with candidacy were AMO (OR, 10.19 [95% CI, 9.09–11.43]), seeing a specialist (OR, 2.81 [95% CI, 2.56–3.08]), mild‐to‐severe interictal burden (MIBS‐4 mild vs. none: OR, 1.81 [95% CI, 1.67–1.96]; moderate vs. none: OR, 2.38 [95% CI, 2.18–2.60]; severe vs. none: OR, 2.73 [95% CI, 2.51–2.96]), migraine‐related stigma (OR, 1.65 [95% CI, 1.53–1.78]), and greater functional impact of migraine (MSQ‐RFR: OR, 0.39 [95% CI, 0.37–0.40]).
Medication use by treatment group
Use of acute and preventive migraine medications was examined by treatment need groups (Figure 2). In Figure 2A, use of acute prescription migraine medication was reported by about half of the preventive + acute (55.7%) and acute (46.9%) groups, 36.9% of the preventive group, and 27.1% of the not a candidate group. Of those taking acute prescription medication, the majority were taking a recommended acute prescription medication.
FIGURE 2.
Use of (A) acute prescription and (B) preventive migraine medications by treatment need group (Note: scales differ) aRecommended medications refer to those with established/probable efficacy as identified by the American Academy of Neurology/AHS guidelines and the AHS consensus statement regarding new migraine treatments. AHS, American Headache Society. [Color figure can be viewed at wileyonlinelibrary.com]
In all groups, approximately one‐quarter or less of participants were taking preventive medication (not a candidate, 5.9%; acute, 18.8%; preventive, 11.6%; preventive + acute, 25.3%). Of those taking preventive medication, the majority in all groups were taking a recommended preventive medication (Figure 2B).
In respondents whose use of acute migraine medications met criteria for AMO, 21 the majority were not taking a preventive therapy (not a candidate, 85.1%; acute, 64.0%; preventive, 84.2%; preventive + acute, 67.6%).
DISCUSSION
This analysis from OVERCOME evaluated acute and preventive prescription treatment needs of individuals with migraine in the United States. The data collected in OVERCOME, which is the largest population‐based survey with collection of person‐level data to support a diagnosis and assess patterns of care, 1 , 14 , 32 , 34 provide a view of the treatment needs among the population with migraine when newer classes of preventive and acute therapies started to become available.
Our approach provides an opportunity to identify individuals with migraine who have unmet medication needs for acute treatment, preventive treatment, or both. To understand the needs for preventive migraine medication, we applied the AHS consensus statement criteria to this population. 18 In the absence of guidelines or recommendations to determine acute prescription migraine medication needs, our OVERCOME medical advisory committee made recommendations based on previous literature and clinical experience. 21 , 35 , 36 Treatment satisfaction or moderate‐to‐severe migraine‐related disability (MIDAS) were previously used in the AMPP study to identify unmet treatment needs. 37 Most recently, the Chronic Migraine Epidemiology and Outcomes (CaMEO) study defined six acute treatment management gap domains. 38 We operationalized unmet acute treatment needs in this study using two of these gaps: mild‐to‐severe disability per MIDAS and lack of current acute treatment optimization on the mTOQ because the prior study showed that the majority of individuals fell into these gaps.
These analyses revealed that 77.1% of the individuals with migraine were candidates for initiating or modifying migraine treatment, and 37.7% were candidates for initiating/modifying preventive migraine treatment. The AMPP study found that 38.8% of people with migraine in the population needed preventive treatment, with 25.7% in the offer prevention category and 13.1% in the consider prevention category. 1 Similarly, the more recent analysis from the CaMEO study (2012–2013) showed that 39% of participants surveyed were potentially eligible for preventive treatment. 15 Though the results are similar across the US studies, there are differences among studies. It is important to note that whereas prior studies evaluated the need to initiate preventive treatment, the current analysis also includes the need to modify a potential existing treatment. This is an important part of individualizing preventive treatment for patients, who may still experience the need for additional clinical improvement even while on preventive medication.
In the current study, we estimated that 39.4% of participants with migraine met criteria for acute treatment modification/initiation. Other studies have not used the same definition of unmet acute treatment needs. However, 37.4% of participants with migraine in the AMPP study were dissatisfied with acute treatment, whereas 37.7% of participants from the Migraine in America Symptoms and Treatment (MAST) study experienced poor/very poor treatment optimization. 35 , 37 Despite differences in definitions, results are similar among US population studies.
The group we categorized as not a candidate for initiation/modification of acute treatment had no/little migraine‐related disability, a lower number of monthly headache days, and good efficacy from their acute treatments (Table S3). Though this group was doing well by many standards, one‐third had not received a medical diagnosis. Awareness of a migraine diagnosis could facilitate better outcomes through targeted education or behavioral interventions. In addition, some individuals reported interictal burden and migraine symptoms interfering with mood and sleep (Table 3). Only 5.9% of this population were taking preventive medication, and this group was presumably well controlled on their current treatment. Approximately one‐quarter were taking a prescription acute medication for migraine; similarly, most likely treatment was meeting their needs. In general, membership in this group with no need to change treatment reflects treatment success. It would be interesting to follow those taking medications to examine if and how their needs change over time. As a corollary, nearly two‐thirds of individuals with migraine might benefit from treatment changes.
The group in need of better preventive treatment only was defined to exclude those who also needed changes to acute treatment. We did this to create mutually exclusive groups. Everyone with mild, moderate, or severe MIDAS disability was ineligible for this group by virtue of being candidates for modification of their acute treatments (Table S3). Those in the preventive group reported more monthly headache days because six or more monthly headache days made them eligible for membership in this group in accordance with AHS guidance. They were more likely to see a specialist for migraine and to report overuse of acute medications. Though acute treatment is effective as a group, more than one‐quarter of the participants (27.6%) were overusing acute medications. This, along with a higher number of headache days per month, emphasize the need to offer preventive treatment in this group because the majority (88.4%) were not taking preventive medication.
Most participants in the acute needs group had little‐to‐mild disability and a lower number of monthly headache days, and they qualified for treatment initiation/modification mainly based on poor efficacy of acute treatment (Table 1C; Table S3). However, delving further into measures of migraine impact such as MIBS and MSQ, over 50% of participants experienced moderate‐to‐severe interictal burden and, overall, higher functional impact. Like the preventive needs group, the use of MIDAS alone does not capture the full impact of migraine. Of note, the results for both the acute and preventive needs groups are distorted because both groups exclude those individuals who have both preventive and acute treatment needs (N = 20,574).
Use of recommended prescription migraine medication, both acute and preventive, is low in the acute needs group. If someone is taking a guideline‐/consensus‐recommended acute prescription medication and still needs treatment modification, this suggests that the acute treatment is not delivering sufficient benefits. Even though acute treatment efficacy is poor in this group, only a small percentage were overusing acute medications. Half of these participants sought some level of healthcare in the previous 12 months, providing an opportunity to initiate conversations to determine their medication needs. Based on our criteria, these participants were not candidates for preventive medications, but some could be “considered” for preventive treatment, as recommended in the AHS consensus statement, 18 based on a combination of disability and headache days per month.
The preventive + acute needs group experienced a high degree of impact from migraine, with the majority experiencing severe disability, moderate‐to‐severe interictal burden, and increased functional impact (MSQ‐RFR), all of which speak to the need for more effective preventive treatment. Over one‐third were overusing acute medications, and the majority had poor efficacy from acute treatment; together, these features suggest the need for more effective acute treatment. However, the answer to AMO is not simply more effective acute medications. To qualify for AMO, respondents had to report using acute medications on ≥10 or ≥ 15 days per month, depending on the medication. 21 Incorporating an effective preventive medication, with the goals of improving responsiveness to acute treatment and avoiding escalation of acute treatments and reducing reliance on acute treatment, is warranted. 18 Because fewer than one in five of these participants were on a recommended preventive medication and with most seeking some level of healthcare, there are opportunities for clinicians to improve care. If someone is taking a recommended preventive medication and needs preventive treatment modification, that indicates there may be room for improvement in preventive treatment.
Previous results from OVERCOME have shown that individuals with migraine often hesitate to seek care for reasons such as believing their attacks are not serious enough or concerns that their healthcare professionals would not take their migraine seriously. 39 Additionally, previous research suggests that clinician–patient conversations about the impact of migraine happen infrequently 40 , 41 , 42 but are associated with more appropriate treatment. 28 , 43 The results of our analyses emphasize the opportunity for clinicians to initiate a broader conversation with patients about the full impact of migraine, both during and between attacks, to elucidate the need to initiate or modify treatment.
The major strength of this study is its large scale, with a population (N = 59,001) that is three times larger than any other population study in individuals with migraine. 1 , 14 , 32 , 34 In addition, it was conducted from 2018 to 2020, which allows a glimpse at the new era in migraine treatment as new treatments were becoming available during this time. Statistical analyses utilized two methods to select predictors for needing treatment initiation/modification, which included supervised machine learning.
In terms of limitations, our categorization of treatment need did not distinguish between those needing modification of current treatment and those needing treatments initiated; thus, the unique needs of these distinct populations were not captured. This study analyzed factors that predicted membership into any of the four groups versus treatment pairs. If RF analysis was utilized on treatment pairs, the factors associated could differ. Additionally, some predictors are correlated with other factors used to categorize participants; thus, some estimates of the associations could be distorted (e.g., stigma and interictal burden are highly correlated). 44 , 45 This is a cross‐sectional study, so causation could not be determined. This study is based on self‐reported data that were not validated by a medical professional, healthcare claims, or electronic health records, and thus could be susceptible to recall bias. Responses were required for all survey questions, which could affect the data quality due to response fatigue. Because the survey was web‐based, the respondents may not be fully representative; those without internet access were not able to participate. Although the candidates in this study were recruited during the 2018 to 2020 timeframe, the criteria for assessing preventive treatment needs were from the AHS 2021 consensus statement. 18 If clinicians were to rigorously implement the current AHS guidelines, and considering the availability of newer migraine treatments, fewer unmet treatment needs might be found in a future population‐based survey. To evaluate preventive medication needs, we utilized current guidelines/consensus statements, but there are currently no published guidelines that define acute prescription medication needs. This is an area that needs continued research, and for which published guidelines would be beneficial.
CONCLUSIONS
Our analysis of a large population of individuals with migraine has shown that, despite the resources and focus on migraine education for patients and healthcare professionals prior to and during the time period of the study, there continued to be high unmet treatment needs for individuals with migraine. Applying the AHS consensus statement and our acute medication needs criteria to people with migraine in OVERCOME has revealed that more than three‐quarters of the participants are candidates for initiation or modification of their current migraine prescription therapy. Machine learning identified AMO, interictal burden, and seeking higher levels of care to be associated with the need to initiate or modify treatment. Initiating a discussion about the full impact of migraine, especially relating to function and burden between attacks, can help clinicians elucidate the need for optimizing preventive therapy beyond the frequency of headache days.
FUNDING INFORMATION
This study was funded by Eli Lilly and Company.
CONFLICT OF INTEREST STATEMENT
Susan Hutchinson has served as a consultant, speaker, and/or received honoraria from AbbVie/Allergan, Alder BioPharmaceuticals/Lundbeck, Amgen, Biohaven, Currax Pharmaceuticals, electroCore, Eli Lilly and Company, Impel Pharmaceuticals, Novartis, Teva, Theranica, and Upsher‐Smith.
Elizabeth Johnston, Robert A. Nicholson, Anthony Zagar, Gilwan Kim, E. Jolanda Muenzel, and Eric M. Pearlman are employees and minor stockholders of Eli Lilly and Company.
Michael L. Reed has received research support from the National Headache Foundation. He serves as a consultant, advisory board member, and/or has received honoraria or research support from AbbVie/Allergan, Amgen, Dr. Reddy's Laboratories (Promius), and Eli Lilly and Company.
Dawn C. Buse has received research support from the Food and Drug Administration (FDA) and the National Headache Foundation. She serves as consultant, advisory board member, and/or has received honoraria or research support from AbbVie/Allergan, Amgen, Biohaven, Eli Lilly and Company, Lundbeck, Novartis, Teva, and Theranica.
Richard B. Lipton has received research support from the National Institutes of Health, the FDA, and the National Headache Foundation. He serves as consultant, advisory board member, and/or has received honoraria or research support from AbbVie/Allergan, Amgen, Biohaven, Dr. Reddy's Laboratories (Promius), electroCore, Eli Lilly and Company, GlaxoSmithKline, Lundbeck, Merck, Novartis, Pfizer, Teva, Vector Pharma, and Vedanta Research. He receives royalties from Wolff's Headache and Other Head Pain, 8th edition (Oxford University Press, 2007) and Informa. He holds stock/stock options in Axon Pharma, Biohaven, CoolTech, and Mainistee Therapeutics.
Supporting information
Figure S1. Participant flow chart for the OVERCOME (US) study from 2018 to 2020 (all cohorts).
Table S1. Sociodemographic, clinical, and migraine‐related variables included in the LASSO and random forest analyses.
Table S2. Factors associated with treatment need as identified by RF (in order of association) and LASSO.
Table S3. Characteristics used to determine treatment need group.
ACKNOWLEDGMENTS
The authors would like to thank Deborah Dsouza from Syneos Health for writing support, Alyssa Luck from Syneos Health for editorial support, and Armen Zakharyan from TechData Service Company for statistical support.
DATA AVAILABILITY STATEMENT
Lilly will provide access to anonymized individual participant data collected during the study. The data will be available to request on vivli.org after the study team has completed analyses and publications. Access will be provided after a proposal has been approved by an independent review committee identified for this purpose and after receipt of a signed data sharing agreement. After a proposal is approved, data and documents, including the study protocol, will need to be provided in a secure data sharing environment. For details on submitting a request, see the instructions provided at www.vivli.org.
REFERENCES
- 1. Lipton RB, Bigal ME, Diamond M, et al. Migraine prevalence, disease burden, and the need for preventive therapy. Neurology. 2007;68:343‐349. [DOI] [PubMed] [Google Scholar]
- 2. Burch R, Rizzoli P, Loder E. The prevalence and impact of migraine and severe headache in the United States: figures and trends from government health studies. Headache. 2018;58:496‐505. [DOI] [PubMed] [Google Scholar]
- 3. Steiner TJ, Stovner LJ, Jensen R, Uluduz D, Katsarava Z, Lifting The Burden: the Global Campaign against Headache . Migraine remains second among the world's causes of disability, and first among young women: findings from GBD2019. J Headache Pain. 2020;21:137. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. GBD 2016 Headache Collaborators . Global, regional, and national burden of migraine and tension‐type headache, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet Neurol. 2018;17:954‐976. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Parikh SK, Kempner J, Young WB. Stigma and migraine: developing effective interventions. Curr Pain Headache Rep. 2021;25:75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Young WB, Park JE, Tian IX, Kempner J. The stigma of migraine. PLoS One. 2013;8:e54074. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Buse DC, Fanning KM, Reed ML, et al. Life with migraine: effects on relationships, career, and finances from the Chronic Migraine Epidemiology and Outcomes (CaMEO) Study. Headache. 2019;59:1286‐1299. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Seng EK, Shapiro RE, Buse DC, Robbins MS, Lipton RB, Parker A. The unique role of stigma in migraine‐related disability and quality of life. Headache. 2022;62:1354‐1364. [DOI] [PubMed] [Google Scholar]
- 9. Bordini CA, da Mariano Silva H, Garbelini RP, Teixeira SO, Speciali JG. Effect of preventive treatment on health‐related quality of life in episodic migraine. J Headache Pain. 2005;6:387‐391. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Grosu O, Nooshin Y, Nathaniel S, et al. Effects of anti‐CGRP monoclonal antibodies for episodic and chronic migraine on migraine characteristics, disability, impact and quality of life beyond 3 months of treatment: a systematic review and meta‐analysis. Headach Med Con. 2021;1:23‐30. [Google Scholar]
- 11. Lipton RB, Lee L, Saikali NP, Bell J, Cohen JM. Effect of headache‐free days on disability, productivity, quality of life, and costs among individuals with migraine. J Manag Care Spec Pharm. 2020;26:1344‐1352. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Lambru G, Andreou AP, Guglielmetti M, Martelletti P. Emerging drugs for migraine treatment: an update. Expert Opin Emerg Drugs. 2018;23:301‐318. [DOI] [PubMed] [Google Scholar]
- 13. Hutchinson S, Lipton RB, Ailani J, et al. Characterization of acute prescription migraine medication use: results from the CaMEO study. Mayo Clin Proc. 2020;95:709‐718. [DOI] [PubMed] [Google Scholar]
- 14. Lipton RB, Nicholson RA, Reed ML, et al. Diagnosis, consultation, treatment, and impact of migraine in the US: results of the OVERCOME (US) study. Headache. 2022;62:122‐140. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Nahas SJ, Buse DC, Hutchinson S, et al. Characterizing preventive treatment gaps in migraine: results from the CaMEO study [abstract]. Neurology. 2021;96:1514. [Google Scholar]
- 16. Silberstein SD, Holland S, Freitag F, et al. Evidence‐based guideline update: pharmacologic treatment for episodic migraine prevention in adults: report of the Quality Standards Subcommittee of the American Academy of Neurology and the American Headache Society. Neurology. 2012;78:1337‐1345. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Marmura MJ, Silberstein SD, Schwedt TJ. The acute treatment of migraine in adults: the American Headache Society evidence assessment of migraine pharmacotherapies. Headache. 2015;55:3‐20. [DOI] [PubMed] [Google Scholar]
- 18. Ailani J, Burch RC, Robbins MS, Board of Directors of the American Headache Society . The American Headache Society Consensus Statement: update on integrating new migraine treatments into clinical practice. Headache. 2021;61:1021‐1039. [DOI] [PubMed] [Google Scholar]
- 19. Ashina S, Muenzel EJ, Nicholson RA, et al. Machine learning identifies factors most associated with seeking medical care for migraine: results of the OVERCOME (US) study. Headache. 2024;64:1027‐1039. [DOI] [PubMed] [Google Scholar]
- 20. Bigal ME, Serrano D, Buse D, Scher A, Stewart WF, Lipton RB. Acute migraine medications and evolution from episodic to chronic migraine: a longitudinal population‐based study. Headache. 2008;48:1157‐1168. [DOI] [PubMed] [Google Scholar]
- 21. Headache Classification Committee of the International Headache Society (IHS) . The International Classification of Headache Disorders, 3rd edition. Cephalalgia. 2018;38:1‐211. [DOI] [PubMed] [Google Scholar]
- 22. Lipton RB, Stewart WF, Sawyer J, Edmeads JG. Clinical utility of an instrument assessing migraine disability: the Migraine Disability Assessment (MIDAS) Questionnaire. Headache. 2001;41:854‐861. [PubMed] [Google Scholar]
- 23. Stewart WF, Lipton RB, Dowson AJ, Sawyer J. Development and testing of the Migraine Disability Assessment (MIDAS) Questionnaire to assess headache‐related disability. Neurology. 2001;56(suppl 1):S20‐S28. [DOI] [PubMed] [Google Scholar]
- 24. Lipton RB, Fanning KM, Serrano D, Reed ML, Cady R, Buse DC. Ineffective acute treatment of episodic migraine is associated with new‐onset chronic migraine. Neurology. 2015;84:688‐695. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Löwe B, Wahl I, Rose M, et al. A 4‐item measure of depression and anxiety: validation and standardization of the Patient Health Questionnaire‐4 (PHQ‐4) in the general population. J Affect Disord. 2010;122:86‐95. [DOI] [PubMed] [Google Scholar]
- 26. Shapiro RE, Nicholson RA, Seng EK, et al. Migraine‐related stigma and its relationship to disability, interictal burden, and quality of life: results of the OVERCOME (US) study. Neurology. 2024;102:e208074. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Buse DC, Bigal MB, Rupnow M, Reed M, Serrano D, Lipton RB. Development and validation of the Migraine Interictal Burden Scale (MIBS): a self‐administered instrument for measuring the burden of migraine between attacks [abstract]. Neurology. 2007;68:A89. [Google Scholar]
- 28. Buse DC, Rupnow MFT, Lipton RB. Assessing and managing all aspects of migraine: migraine attacks, migraine‐related functional impairment, common comorbidities, and quality of life. Mayo Clin Proc. 2009;84:422‐435. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Cole JC, Lin P, Rupnow MFT. Validation of the migraine‐specific quality of life questionnaire version 2.1 (MSQ v. 2.1) for patients undergoing prophylactic migraine treatment. Qual Life Res. 2007;16:1231‐1237. [DOI] [PubMed] [Google Scholar]
- 30. Jhingran P, Osterhaus JT, Miller DW, Lee JT, Kirchdoerfer L. Development and validation of the migraine‐specific quality of life questionnaire. Headache. 1998;38:295‐302. [DOI] [PubMed] [Google Scholar]
- 31. Rendas‐Baum R, Bloudek LM, Maglinte GA, Varon SF. The psychometric properties of the Migraine‐Specific Quality of Life Questionnaire version 2.1 (MSQ) in chronic migraine patients. Qual Life Res. 2013;22:1123‐1133. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Lipton RB, Munjal S, Alam A, et al. Migraine in America Symptoms and Treatment (MAST) Study: baseline study methods, treatment patterns, and gender differences. Headache. 2018;58:1408‐1426. [DOI] [PubMed] [Google Scholar]
- 33. Serrano D, Buse DC, Reed ML, Runken MC, Lipton RB. Development of the Migraine Symptom Severity Score (MSSS): a latent variable model for migraine definition [abstract]. Headache. 2010;50:S40. [Google Scholar]
- 34. Adams AM, Serrano D, Buse DC, et al. The impact of chronic migraine: the Chronic Migraine Epidemiology and Outcomes (CaMEO) study methods and baseline results. Cephalalgia. 2015;35:563‐578. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Lipton RB, Munjal S, Buse DC, et al. Unmet acute treatment needs from the 2017 Migraine in America Symptoms and Treatment Study. Headache. 2019;59:1310‐1323. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Lipton RB, Buse DC, Dodick DW, et al. Burden of increasing opioid use in the treatment of migraine: results from the Migraine in America Symptoms and Treatment Study. Headache. 2021;61:103‐116. [DOI] [PubMed] [Google Scholar]
- 37. Lipton RB, Buse DC, Serrano D, Holland S, Reed ML. Examination of unmet treatment needs among persons with episodic migraine: results of the American Migraine Prevalence and Prevention (AMPP) Study. Headache. 2013;53:1300‐1311. [DOI] [PubMed] [Google Scholar]
- 38. Buse DC, Nahas SJ, Schwedt TJ, et al. Acute treatment management gaps in people with migraine: results of the CaMEO study [abstract]. Neurology. 2020;94:105. [Google Scholar]
- 39. Shapiro RE, Nicholson RA, Zagar AJ, et al. Reasons for hesitating to consult for migraine care: results of the OVERCOME (US) study [abstract]. Headache. 2021;61:10‐11. [Google Scholar]
- 40. Hahn SR, Lipton RB, Sheftell FD, et al. Healthcare provider‐patient communication and migraine assessment: results of the American Migraine Communication Study, phase II. Curr Med Res Opin. 2008;24:1711‐1718. [DOI] [PubMed] [Google Scholar]
- 41. Lipton RB, Hahn SR, Cady RK, et al. In‐office discussions of migraine: results from the American Migraine Communication Study. J Gen Intern Med. 2008;23:1145‐1151. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Grosberg BM, Nicholson RA, Vallarino C, et al. What patients report discussing during their most recent visit for headache with a neurologist or headache specialist: association with being diagnosed with migraine [abstract]. Headache. 2022;62:170. [Google Scholar]
- 43. Holmes WF, MacGregor EA, Sawyer JP, Lipton RB. Information about migraine disability influences physicians' perceptions of illness severity and treatment needs. Headache. 2001;41:343‐350. [DOI] [PubMed] [Google Scholar]
- 44. Shapiro R, Nicholson R, Zagar AJ, et al. Identifying factors pertaining to migraine‐related stigma from a novel question set: results of the OVERCOME (US) study [abstract]. Neurology. 2022;98:2997. [Google Scholar]
- 45. Vincent M, Viktrup L, Nicholson RA, Ossipov MH, Vargas BB. The not so hidden impact of interictal burden in migraine: a narrative review. Front Neurol. 2022;13:1032103. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1. Participant flow chart for the OVERCOME (US) study from 2018 to 2020 (all cohorts).
Table S1. Sociodemographic, clinical, and migraine‐related variables included in the LASSO and random forest analyses.
Table S2. Factors associated with treatment need as identified by RF (in order of association) and LASSO.
Table S3. Characteristics used to determine treatment need group.
Data Availability Statement
Lilly will provide access to anonymized individual participant data collected during the study. The data will be available to request on vivli.org after the study team has completed analyses and publications. Access will be provided after a proposal has been approved by an independent review committee identified for this purpose and after receipt of a signed data sharing agreement. After a proposal is approved, data and documents, including the study protocol, will need to be provided in a secure data sharing environment. For details on submitting a request, see the instructions provided at www.vivli.org.