Factors and Reasons Associated with Hesitating to Seek Care for Migraine: Results of the OVERCOME (US) Study

Abstract

Introduction

Despite a variety of available treatment options for migraine, many people with migraine do not seek medical care, thereby reducing opportunities for diagnosis and effective treatment and potentially leading to missed opportunities to reduce the burden of disease. Understanding why people hesitate to seek care for migraine may help healthcare professionals and advocates address barriers and improve outcomes. The aim of this study, in a large adult population sample in the United States (US), was to identify factors associated with and reasons for hesitating to seek healthcare for migraine.

Methods

The web-based OVERCOME (US) survey study identified adults with active migraine in a demographically representative US sample who answered questions about hesitating to seek care from a healthcare provider for migraine and reasons for hesitating. Supervised machine learning (random forest, least absolute shrinkage and selection operator) identified factors associated with hesitation; logistic regression models assessed association of factors on hesitation.

Results

The study results show that of the 58,403 participants with active migraine who completed the OVERCOME (US) baseline survey and provided responses to the question on hesitating to seek care for migraine, 45.1% (n = 26,330/58,403) with migraine indicated that they had ever hesitated to seek care for migraine. Factors most associated with hesitating to seek care were hiding migraine (odds ratio [OR] = 2.69; 95% confidence interval [CI]: 2.50, 2.89), experiencing migraine-related stigma (OR = 2.13; 95% CI 1.95, 2.33), higher migraine-related disability (OR = 1.30; 95% CI 1.23, 1.38), and higher ictal cutaneous allodynia (OR = 1.26; 95% CI 1.19, 1.35). The most common reasons participants stated for hesitating included (1) 44.2% wanting to try and take care of migraine on their own, (2) 33.8% feeling that their migraine or headache would not be taken seriously, (3) 29.2% thinking that their migraine was not serious/painful enough, and (4) 27.4% not being able to afford it or not wanting to spend the money. The main limitation of the study includes the requirement for respondents to have internet, access which may have reflected cohort bias, and the quota sampling rather than random sampling to create a demographically representative sample.

Conclusions

Hesitating to seek migraine care is common and is most strongly associated with hiding the disease and migraine-related stigma. Those experiencing higher migraine-related burden are more hesitant to seek the care that might alleviate the burden. These findings suggest that migraine’s social context (e.g., stigma) is a major determinant of hesitance to seek migraine care.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40120-024-00668-9.

Key Summary Points

Why carry out this study?

Approximately half of Americans with migraine do not seek medical care for their headaches, presenting a large population burden.

The aim of the current analysis was to understand reasons why some individuals hesitate to seek medical care for migraine and to explore the sociodemographic, clinical, and migraine-related characteristics that may be associated with this hesitancy.

The current study utilized data from the Observational Survey of the Epidemiology, Treatment and Care of Migraine (OVERCOME) study, the largest web-based survey to date, with >60,000 adult participants with migraine.

What was learned from the study?

The results demonstrate that hesitating is very common (45% of adults with migraine report ever having hesitated to talk to a healthcare provider about their headache symptoms) and most frequently associated with hiding the disease, perceived external stigma, and experiencing higher migraine-related burden.

These findings suggest that migraine’s social context is a major determinant of hesitancy and may provide a springboard for the development of strategies and possible interventions at the patient level.

Introduction

Migraine is a highly prevalent primary headache disorder associated with substantial disability, burden, and socioeconomic impact [1]. Despite a range of safe and effective treatments, a large percentage of people with migraine in the United States (US) do not seek medical care for migraine, reducing opportunities for diagnosis and effective treatment and potentially leading to missed opportunities to reduce the burden of disease.

Although rates of consultation and diagnosis have improved over the past three decades, the overall pattern of undertreatment and high disability persists for people living with migraine [2–5]. Understanding why people hesitate to seek care for migraine may help healthcare professionals and advocates address barriers and improve outcomes. The ObserVational survey of the Epidemiology, tReatment and Care Of MigrainE (OVERCOME) (US) study is a population-based web survey that recruited demographically representative samples of US adults with migraine to better understand patterns of healthcare use for migraine [6].

The current analysis aims to identify the sociodemographic, clinical, and migraine-related factors associated with hesitating to seek medical care for migraine and to further explore the reasons respondents provided for their hesitation. Understanding the prevalence of patient hesitancy to seek migraine care, and the factors and reasons associated with it, may provide bases for strategies to reduce hesitance, thereby improving health outcomes.

Methods

Study Design and Data Source

OVERCOME (US) is a prospective, cross-sectional and longitudinal, multi-cohort, web-based survey conducted annually between 2018 and 2020 in a representative sample of US adults (by geographic region age, race, and sex). The study design and population have been detailed previously [6, 7]. Briefly, people with “active” migraine who reported having ≥ 1 headache(s) in the previous 12 months, met criteria for migraine [8, 9], and/or self-reported having received a diagnosis of migraine from a healthcare professional were invited to complete the full migraine survey to collect information related to characteristics, beliefs, behaviors, and outcomes. The current analysis is a cross-sectional analysis of the baseline survey data.

The Sterling Institutional Review Board (IRB ID #6425–001) provided approval for this observational study. All respondents provided electronic informed consent to participate in a general health-related survey. Supplemental Fig. 1 displays the overall flow of all participants.

Fig. 1.

Variables most associated with hesitating to seek care for migraine. The top 10 factors identified by random forest analysis show that the strongest associations are hiding and MiRS (panel A). Variables in pattern fill indicate those factors that were also identified by LASSO analysis. Variables identified by LASSO but not random forest analysis include MSSS and years with migraine. Multivariate logistic regression analysis (panel B) revealed that odds of hesitating to seek care were nearly double for the top two factors of hiding and stigma. Continuous variables were standardized (for raw data see Supplemental Fig. 2). The standardized odds ratios for continuous variables reflect a one-standard-deviation change in the modeled variable. Note: Panel B—values for odds ratios and confidence intervals were rounded up to one decimal place. For factors with an odds ratio = 1, the 95% confidence intervals rounded to 1.0, and thus do not show error bars. ASC-12 Allodynia Symptom Checklist-12, LASSO least absolute shrinkage and selection operator, MiRS Migraine-Related Stigma questionnaire, MiRS-MB Migraine-Related Stigma–Minimizing Burden; MIBS-4 Migraine Interictal Burden Scale-4, MIDAS Migraine Disability Assessment Questionnaire, MSQ-RFR Migraine-Specific Quality of Life Questionnaire v2.1–Role Function Restrictive, PHQ-4 Patient Health Questionnaire-4, MiRS-SG Migraine-Related Stigma–Secondary Gain, MSSS Migraine Symptom Severity Score

Question Generation

For the generation of the questions around hesitating to seek medical care for migraine, seven focus groups with a total of 48 people living with migraine were conducted to understand whether or not they had sought medical care for their symptoms. Hesitating to seek care emerged as an important factor. A panel of subject matter experts then developed the questions around hesitating based on their clinical experience and the feedback from the patient focus groups. To ensure questions were easy to understand, cognitive interviews were conducted with a separate group of people who were recruited from local consumer databases and who met the International Classification of Headache Disorders 3rd edition criteria [9] for migraine.

Patient Outcomes

Assessing Hesitancy

The participants of interest for this analysis were those who provided a response to the survey question “Have you ever hesitated to seek care from (or talk to) a doctor or healthcare provider for your migraine or severe headache attacks?” Respondents who answered “yes” were included in the hesitation to seek care “yes” group.

Those who had hesitated to seek care were then asked “What are the reasons you hesitated to seek care from (or talk to) a doctor or healthcare provider for your migraine or severe headache attacks?” and could select all that applied from a list of 14 potential reasons (presented in random order from survey to survey to minimize order effect). This list of reasons (shown as part of Fig. 1, panel A) was developed from qualitative interviews among persons with migraine and by the OVERCOME advisor group based on clinical knowledge and expert perspective. The reasons, presented in random order, were (1) I did not think they were serious/painful enough; (2) I was too busy; (3) I tried and could not get an appointment; (4) I did not have health insurance or my health insurance did not cover the cost; (5) I did not know where to find a doctor or healthcare provider who treated them; (6) I wanted to try and take care of/deal with them on my own; (7) I do not like seeing a doctor/do not think they are helpful; (8) The medications or other treatments that do not need a doctor’s prescription worked well enough; (9) I did not want to have medical treatment done; (10) I was afraid they might diagnose me with something serious; (11) I could not afford it or did not want to spend the money; (12) I did not want to be labeled as a person with headache or migraine; (13) I felt my migraine or headache would not be taken seriously; and 14) other reasons. The group that hesitated to seek care was then stratified into subgroups of whether they actually sought care for migraine.

Assessing Hiding

To assess “hiding migraine from others,” respondents answered the question “How often do you hide your migraine or severe headache from: (1) family?; (2) friends?; (3) co-workers?; 4) a supervisor, boss or employer?” Similarly, for the outcome of “hiding migraine at work,” scores for each question of hiding from coworkers or hiding from supervisor (maximum score = 5 per question) were added. For both hiding from others and from work, response options were “never” (score = 1), “rarely” (score = 2), “sometimes” (score = 3), “often” (score = 4), “very often” (score = 5), and “not applicable.” In this manuscript we focus on hiding from family and friends only, as only employed participants answered the questions about work. To assess “hiding migraine from family and friends,” scores for each question (hiding from family/hiding from friends, maximum score = 5 per question) were added and respondents were assigned to a category of “hiding from friends and family” based on the total score (1,2 = never, 3,4 = rarely, 5,6 = sometimes, 7,8 = often, and 9,10 = very often). Similarly, for the outcome of “hiding migraine at work,” scores for each question (hiding from coworkers/hiding from supervisor, maximum score = 5 per question) were added and respondents were assigned to a category of “hiding at work” based on the total score (1,2 = never, 3,4 = rarely, 5,6 = sometimes, 7,8 = often, and 9,10 = very often). As “hiding at work” was only assessed among the respondents who had indicated that they were employed, this variable was not included in the modeling.

Assessing Migraine Stigma

The 12-item Migraine-Related Stigma (MiRS) questionnaire assessed how frequently respondents experienced migraine-related stigma [10, 11]. MiRS includes two factors: (1) feeling that others viewed migraine as being used for secondary gain (MiRS-SG) and (2) feeling that others were minimizing disease burden (MiRS-MB). These factors were combined for identifying an individual’s level of migraine-related stigma: (1) MiRS-both = experiencing secondary gain and minimizing migraine burden often/very often; (2) MiRS-SG = experiencing secondary gain often/very often; (3) MiRS-MB = experiencing minimizing burden often/very often; (4) MiRS-rarely/sometimes = experiencing secondary gain and minimizing burden rarely/sometimes; and (5) MiRS-never = experiencing secondary gain and minimizing burden never.

Other Assessments

Participants were asked questions about socioeconomic and demographic characteristics and were asked to complete several patient-reported outcome measures [6] not yet discussed. Migraine-related disability, burden, impact, frequency, and allodynia were assessed with the following measures previously detailed [6, 7]: five-item Migraine Disability Assessment Scale (MIDAS) [12, 13], four-item Migraine Interictal Burden Scale (MIBS-4) [14, 15], seven-item Migraine-Specific Quality of Life Questionnaire v2.1–Role Function Restrictive (MSQ-RFR) [16–19], Migraine Symptom Severity Score (MSSS) [20], 12-item Allodynia Symptom Checklist (ASC-12) [21], and the four-item Patient Health Questionnaire (PHQ-4) [22]. Assessment scales were publicly available for use, with the exception of the MSQ-RFR, in which licensing for use was obtained.

Participants were asked to self-report having received medical diagnoses from a list of comorbidities, including cardiovascular diseases (including prediabetes, diabetes, high cholesterol, and hypertension), joint or pain conditions (chronic back pain, fibromyalgia, osteoarthritis, and rheumatoid arthritis), and psychiatric conditions (anxiety, depression, and panic disorder). Additionally, self-report of medical diagnoses that are contraindicated for triptan use were queried (i.e., aneurysm, angina, cerebral hemorrhage, claudication, myocardial infarction, stroke, transient ischemic attack, and blood clots in legs/lungs).

Statistical Methods

This current analysis of pooled data from the OVERCOME (US) baseline cohorts utilized supervised machine learning to determine sociodemographic, clinical, and migraine-related factors associated with respondents who hesitated to seek care for migraine. Machine learning is widely used in population-based surveys across many disease areas but not commonly employed in population-based studies in migraine [23–25]. The advantage of machine learning is that large numbers of variables can be evaluated and weighed simultaneously, and complex relationships can thereby be identified.

Sociodemographic, clinical, and migraine-related factors were summarized with means ± standard deviations (SD) for continuous variables and with percentages for dichotomous or ordinal categorical variables. Statistical significance was assessed using analysis of variance (ANOVA) for continuous variables or chi-square test for categorical variables; all tests were two-sided, and P-values < 0.05 were considered statistically significant.

To identify the sociodemographic and disease-related factors most associated with hesitating to seek care, we used two selection approaches: a random forest algorithm [26] and a least absolute shrinkage and selection operator (LASSO) algorithm [27]. SAS^® software was used to implement the random forest and LASSO algorithms. Supplemental Table 1 details the variables that were chosen for inclusion in the original survey, and their response types. The variables were chosen based on potential association with the response variable based on the authors’ clinical experience. The random forest, consisting of 1000 trees, assessed the importance of 56 variables in predicting whether or not a respondent hesitated to seek care for migraine and reasons given for hesitating to seek care (selected by more than 25% of participants). The out-of-bag (OOB) Gini index for the 1000 trees measured the relative importance of the variables; those with the largest Gini indices were selected for further evaluation [28, 29]. A logistic regression (main effects, no interactions) with the LASSO algorithm (using the implementation in SAS^®) was used as an additional method for variable selection to potentially identify additional variables that the random forest model may have thought were less important than others it selected. LASSO is commonly used for variable selection; it has a simpler structure based on a linear model and effectively handles multicollinearity in a sample. A similar approach was used in a prior analysis of the OVERCOME (US) dataset [7]. A simultaneous multivariable logistic regression with main effects was then used to estimate the odds ratios (OR) and 95% confidence intervals (CI) for each of the variables identified by the random forest model and/or LASSO. Logistic regressions are shown with standardized continuous variables and raw data. SAS^®  Enterprise Guide software (version 7.15) was used for all analyses and summaries.

Table 1.

Demographics and clinical characteristics of participants who responded to the question regarding hesitation to seek care for migraine

	Total	Ever hesitated to seek care for migraine
	N = 58,403	Yes (N = 26,330)	No (N = 32,073)
Sociodemographic characteristics
Age in years, mean (SD)*	41.8 (14.7)	39.3 (13.8)	43.7 (15.2)
Sex at birth, n (%)*
Female, n	43,530 (74.5)	19,420 (73.8)	24,110 (75.2)
Male, n	14,873 (25.5)	6910 (26.2)	7963 (24.8)
Racea*
White, n	46,125	20,366	25,759
Column %	79.0	77.3	80.3
Row %	100	44.2	55.8
Black, n	5140	2406	2734
Column %	8.8	9.1	8.5
Row %	100	46.8	53.2
Asian or Asian Americanb, n	1896	798	1098
Column %	3.2	3.0	3.4
Row %	100	42.1	57.9
Otherc, n	5242	2760	2482
Column %	9.0	10.5	7.7
Row %	100	52.7	47.3
Hispanic origind: Yes, n*	6261	3302	2959
Column %	10.7	12.5	9.2
Row %	100	52.7	47.3
Regione, n*
Northeast, n	10,291	4200	6091
Column %	17.6	16.0	19.0
Row %	100	40.8	59.2
Midwest, n	13,255	5943	7312
Column %	22.7	22.6	22.8
Row %	100	44.8	55.2
South, n	23,103	10,695	12,408
Column %	39.6	40.6	38.7
Row %	100	46.3	53.7
West, n	11,754	5492	6292
Column %	20.1	20.9	19.5
Row %	100	46.7	53.3
Employed: Yes, n (%)*	33,335 (57.1)	15,216 (57.8)	18,119 (56.5)
Health insurance: Yes, n (%)*	50,131 (85.8)	21,945 (83.3)	28,186 (87.9)
College graduate or above: Yes, n (%)*	20,817 (35.6)	8680 (33.0)	12,137 (37.8)
Annual household income, n (%)*
< $50,000	28,776 (49.3)	13,885 (52.7)	14,891 (46.4)
$50,000–$99,999	18,493 (31.7)	8072 (30.7)	10,421 (32.5)
$100,000+	9540 (16.3)	3822 (14.5)	5718 (17.8)
Preferred not to answer	1594 (2.7)	551 (2.1)	1043 (3.3)
Self-reported medical diagnosis for, n (%)*
Migraine only	10,717 (18.4)	4280 (16.3)	6437 (20.1)
Other headache type onlyf	12,588 (21.6)	5907 (22.4)	6681 (20.8)
Migraine + other headache	23,515 (40.3)	11,442 (43.5)	12,073 (37.6)
Clinical characteristics and selected comorbidities
Number of self-reported joint or pain comorbiditiesg, n (%)*
None	39,958 (68.4)	18,050 (68.6)	21,908 (68.3)
1	11,973 (20.5)	5421 (20.6)	6552 (20.4)
2 or more	6472 (11.1)	2859 (10.9)	3613 (11.3)
Number of self-reported cardiovascular comorbiditiesh, n (%)*
None	30,938 (53.0)	13,806 (52.4)	17,132 (53.4)
1	13,131 (22.5)	5813 (22.1)	7318 (22.8)
2 or more	14,334 (24.5)	6711 (25.5)	7623 (23.8)
Self-reported comorbidity contraindicated for triptan usei:  Yes, n (%)*	10,148 (17.4)	5419 (20.6)	4729 (14.7)
Number of self-reported psychiatric comorbiditiesj, n (%)*
None	25,548 (43.7)	9925 (37.7)	15,623 (48.7)
1	10,907 (18.7)	5069 (19.3)	5838 (18.2)
2	12,088 (20.7)	5960 (22.6)	6128 (19.1)
3 or more	9860 (16.9)	5376 (20.4)	4484 (14.0)
PHQ-4 total score, mean (SD)*	4.3 (3.5)	5.1 (3.5)	3.7 (3.4)
Cannabis use: Yes, n (%)*	15,613 (26.7)	8346 (31.7)	7267 (22.7)
OTC medication overuse: Yes, n (%)*	3119 (5.3)	1702 (6.5)	1417 (4.4)
Migraine-associated features
Years with migraine, mean (SD)*	18.9 (14.8)	17.5 (13.9)	20.1 (15.5)
Monthly headache days, n (%)*
0–3	34,461 (59.0)	14,177 (53.8)	20,284 (63.2)
4–7	11,376 (19.5)	5594 (21.2)	5782 (18.0)
8–14	6193 (10.6)	3163 (12.0)	3030 (9.4)
15 or more	6373 (10.9)	3396 (12.9)	2977 (9.3)
MSSSk, mean (SD)*	16.9 (3.4)	17.4 (3.1)	16.5 (3.5)
Headache pain intensity score: 0–10, mean (SD)*	6.9 (1.9)	7.2 (1.8)	6.7 (2.0)
ASC-12, n (%)*
None (0–2)	27,488 (47.1)	9870 (37.5)	17,618 (54.9)
Mild (3–5)	14,041 (24.0)	6786 (25.8)	7255 (22.6)
Moderate (6–8)	8388 (14.4)	4605 (17.5)	3783 (11.8)
Severe (9 or more)	8486 (14.5)	5069 (19.3)	3417 (10.7)
Migraine-related outcomes
MSQ-RFR, mean (SD)*	54.4 (24.1)	48.6 (22.6)	59.1 (24.3)
MIBS-4 score, n (%)*
None (0)	17,734 (30.4)	5653 (21.5)	12,081 (37.7)
Mild (1–2)	9712 (16.6)	4054 (15.4)	5658 (17.6)
Moderate (3–4)	8414 (14.4)	4011 (15.2)	4403 (13.7)
Severe (5 or more)	22,543 (38.6)	12,612 (47.9)	9931 (31.0)
MIDAS score, n (%)*
Little to none (0–5)	23,076 (39.5)	7426 (28.2)	15,650 (48.8)
Mild (6–10)	9723 (16.6)	4581 (17.4)	5142 (16.0)
Moderate (11–20)	1,0257 (17.6)	5470 (20.8)	4787 (14.9)
Severe (21 or more)	15,347 (26.3)	8853 (33.6)	6494 (20.2)
MiRS, n (%)*
Never MiRS-SG/MiRS-MB	6870 (11.8)	1550 (5.9)	5320 (16.6)
Rarely or sometimes MiRS-SG/MiRS-MB	33,260 (56.9)	13,525 (51.4)	19,735 (61.5)
Often/very often MiRS-SG	1493 (2.6)	878 (3.3)	615 (1.9)
Often/very often MiRS-MB	9733 (16.7)	5630 (21.4)	4103 (12.8)
Often/very often both MiRS-SG/MiRS-MB	7047 (12.1)	4747 (18.0)	2300 (7.2)
Hiding migraine from family and friends often/very often, n (%)*	15,230 (26.1)	9487 (36.0)	5743 (17.9)
Hiding migraine from work often/very often (among those employed), Yes, n/N (%)*	11,325/33,335 (34.0)	6755/15,216 (44.4)	45,70/18,119 (25.2)

Percentages are reflective of row percent unless otherwise indicated

ASC-12 12-item Allodynia Symptom Checklist; ICHD-3 International Classification of Headache Disorders, 3rd edition; MiRS Migraine-Related Stigma questionnaire; MiRS-MB Migraine-Related Stigma–Minimizing Burden; MiRS-SG Migraine-Related Stigma–Secondary Gain; MIBS-4 Migraine Interictal Burden Scale-4; MIDAS Migraine Disability Assessment; MSQ-RFR Migraine-Specific Quality of Life Questionnaire v2.1–Role Function Restrictive; MSSS Migraine Severity Symptom Score; NSAID = nonsteroidal anti-inflammatory drugs; OTC over-the-counter; PHQ-4 Patient Health Questionnaire-4; SD standard deviation

*All comparisons were significant, P < 0.001

^aRace was queried by instructing the respondent to select any that applied from responses of American Indian/Alaska Native, Asian/Asian American, Black/African American, Native Hawaiian/Asian or Pacific Islander, White or Caucasian, Other, or Prefer not to answer. Those who selected “White or Caucasian” were included in the group “White,” those who selected “Black/African American” were included in the group “Black,” and all other races selected were grouped into “Other.”

^bCategory includes Asian, Asian American, Asian Indian, Chinese, Filipino, Japanese, Korean, and Vietnamese

^cOther includes American Indian only, Native Hawaiian only, and two or more races

^dInformation on ethnicity was queried as “Hispanic” with responses of “yes,” “no,” or “prefer not to answer.”

^eRegions: Northeast (Maine, New Hampshire, Vermont, Massachusetts, Rhode Island, Connecticut, New York, New Jersey, and Pennsylvania); Midwest (Wisconsin, Illinois, Michigan, Indiana, Ohio, North Dakota, South Dakota, Nebraska, Kansas, Minnesota, Iowa, and Missouri); South (Kentucky, Tennessee, Mississippi, Alabama, Florida, Georgia, South Carolina, North Carolina, Virginia, West Virginia, District of Columbia, Maryland, Delaware, Texas, Oklahoma, Arkansas, and Louisiana); and West (Montana, Idaho, Wyoming, Nevada, Utah, Colorado, Arizona, New Mexico, Washington, Oregon, California, Alaska, and Hawaii)

^fOther headache types included cervicogenic headache, chronic daily headache, cluster headache, new daily persistent headache, medication overuse headache or rebound headache, menstrual headache or menstrual migraine, post-traumatic headache or post-concussion headache, sinus headache, stress headache, and tension-type headache or tension headache

^gJoint or pain comorbidities included chronic back pain, fibromyalgia, osteoarthritis, and rheumatoid arthritis

^hCardiovascular comorbidity included self-reported medical diagnosis of pre-diabetes, diabetes, high cholesterol, and hypertension

ⁱAmong aneurysm, angina, cerebral hemorrhage, claudication, myocardial infarction, stroke, transient ischemic attack, and blood clots in legs/lungs

^jPsychiatric comorbidities included self-reported anxiety, depression, and panic disorder

^kThe MSSS composite index of seven ICHD-3 headache criteria of unilateral pain, pulsatile pain, moderate or severe pain intensity, routine activities worsen pain, nausea, photophobia, and phonophobia

Results

Sociodemographic and Migraine-Specific Variables

A total of 58,403 participants with active migraine completed the baseline survey and provided responses to the question on hesitating to seek care for migraine. Among this group, 45.1% (n = 26,330) had hesitated to seek care for migraine and 54.9% (n = 32,073) had not hesitated (Table 1).

Respondents who hesitated to seek care, relative to those that did not hesitate, were younger (age mean ± SD 39.3 ± 13.8 years vs. 43.7 ± 15.2 years) and less likely to be female at birth (73.8% vs. 75.2%). Hesitating to seek care was most common among Hispanic ethnicity (52.7%), followed by racial groups of Black (46.8%), White (44.2%), and Asian (42.1%). Respondents that hesitated were also less likely to be college graduates (33.0% vs. 37.8%), report an annual household income of ≥ $50,000 (45.2% vs. 50.3%), and have health insurance (83.3% vs. 87.9%) (Table 1).

Respondents who hesitated to seek care, relative to those that did not hesitate, had lived with migraine for fewer years (mean ± SD 17.5 ± 13.9 years vs. 20.1 ± 15.5 years), but reported much higher burden of migraine across all patient-reported outcome measures. Specifically, they were more likely to have ≥ 8 headache days per month (24.9% vs. 18.7%), a higher MSSS (mean ± SD; 17.4 ± 3.1 vs. 16.5 ± 3.5), moderate/severe migraine-related disability (MIDAS ≥ 11; 54.4% vs. 35.1%), severe ictal cutaneous allodynia (ASC-12 ≥ 9; 19.3% vs. 10.7%), lower migraine-specific quality of life (MSQ-RFR mean ± SD; 48.6 ± 22.6 vs. 59.1 ± 24.3), severe migraine interictal burden (Migraine Interictal Burden Scale-4 [MIBS-4] ≥ 5; 47.9% vs. 31.0%), and severe migraine-related stigma (MiRS, often/very often; 42.7% vs. 21.9%). They were also more likely to hide their migraine from family or friends (36.0% vs. 17.9%), hide their migraine at work (44.4% vs. 25.2%), and have higher anxiety/depression symptoms (PHQ-4 mean ± SD; 5.1 ± 3.5 vs. 3.7 ± 3.4). Differences among those who hesitated to seek care relative to those who did not hesitate were statistically significant among all sociodemographic and migraine-specific characteristics and outcome variables evaluated (Table 1).

Factors Associated with Hesitating to Seek Care

The top 10 factors associated with hesitating to seek care for migraine, as identified by random forest, are shown in Fig. 1, panel A. The top two factors were (1) higher levels of hiding migraine from family and friends and (2) higher levels of migraine-related stigma. The other factors associated with hesitancy (in decreasing order of association) included (3) lower migraine-specific quality of life; (4) higher migraine-related disability; (5) experiencing symptoms of anxiety and depression; (6) higher levels of ictal cutaneous allodynia; (7) higher burden of migraine between attacks; (8) migraine impact on mood, (9) younger age at the time of the survey; and (10) migraine impact on sleep. In addition to random forest analysis, LASSO identified five variables associated with hesitancy; three overlapped with those identified by random forest analysis (migraine-specific quality of life, experiencing symptoms of anxiety and depression, and younger age), whereas two were only identified by LASSO (total migraine symptom severity and years with migraine).

One multivariate logistic regression model was run simultaneously examining all 12 factors identified by random forest and LASSO (Fig. 1, panel B; raw data in Supplemental Fig. 2). This model revealed that the odds of hesitating to seek care for migraine more than doubled for those individuals who hid migraine from family and friends often/very often (OR = 2.69; 95% CI 2.50, 2.89) or often (OR = 2.27; 95% CI 2.13, 2.41) in comparison with those who never hid migraine. The odds of hesitating were also more than double for those who often/very often experienced migraine-related stigma (OR = 2.13; 95% CI  1.95, 2.33) in comparison with those who never experienced migraine-related stigma. The odds of hesitating were also greater among those with higher migraine-related disability and interictal burden, as well as more ictal cutaneous allodynia and higher PHQ-4 and MSSS scores.

Fig. 2.

Reasons for hesitating to seek care for migraine among those who report hesitating (n = 26,330). Panel A shows the reasons participants reported for ever hesitating to seek care for migraine. Panel B shows further stratification of the top four reasons (given by > 25% of the participants) by whether they did (Yes) or did not seek medical care (No) for migraine

Reasons for Hesitating to Seek Care

Among the 14 options for having hesitated to seek care for migraine, the four most common reasons, identified by > 25% of respondents, were as follows: reason #1 “I wanted to try and take care of/deal with them on my own” (44.0%); reason #2 “I felt my migraine or headache wouldn’t be taken seriously” (33.6%); reason #3 “I didn’t think they were serious/painful enough” (29.1%); and reason #4 “I couldn’t afford it or didn’t want to spend the money” (27.2%) (Fig. 2, panel A). Given that more than one reason could be selected, further analysis showed that 50.2% of participants selected one of these reasons, 32.1% selected two reasons, 14.6% selected three reasons, and 3.1% selected all four reasons of these top four reasons for hesitating to seek care for migraine. Of note, 6.7% of respondents selected “other reasons” as to why they hesitated (2.7% of respondents selected this as their only reason), thus confirming that the list of options offered covered most of the reasons for having hesitated to seek care for migraine.

Overall, among those respondents that hesitated to seek care (n = 26,330), 57.1% (n = 15,024) eventually consulted for care. Respondents who selected reason #2 for hesitating (“I felt my migraine or headache wouldn’t be taken seriously”) were most likely to seek care (61.5%), while those who selected reason #3 (“I didn’t think they were serious/painful enough”) were least likely to do so (47.9%) (Fig. 2, panel B).

The characteristics that drove hesitation to seek care among the participants who selected the four most cited reasons for hesitating were identified by machine learning analysis among 56 sociodemographic, clinical, and migraine-related factors (Fig. 3, Supplemental Tables 1 and 2). For each of the four reasons, one overarching factor was evident. For those who selected reason #1 (“I wanted to try and take care of/deal with them on my own”), the dominant associated factor was lower interictal burden (measured by MIBS-4); fewer people citing this reason experienced at least moderate interictal burden (56.4%) compared to those who did not cite this reason for hesitating (68.4%). Among respondents that selected reason #2 (“I felt my migraine or headache wouldn’t be taken seriously”), experience of migraine-related stigma (measured by MiRS) was the top factor; 56.8% of those who cited reason #2 experienced migraine-related stigma often/very often compared to 35.6% who did not cite this reason. For reason #3 (“I didn’t think they were serious/painful enough”), lower mean headache pain intensity was the top factor. The mean ± SD intensity among those who cited reason #3 was 6.7 ± 1.7, whereas the mean ± SD intensity among those who did not cite this reason was 7.4 ± 1.7. For reason #4 (“I couldn’t afford it or didn’t want to spend the money”), having health insurance overwhelmingly dominated the remaining factors, with fewer respondents who cited this reason having health insurance (70.3%) compared to those who did not cite this reason (88.2%).

Fig. 3.

Five factors associated with reason #1 (panel A), reason #2 (panel B), reason #3 (panel C), and reason #4 (panel D) given by > 25% people with migraine for hesitating to seek care for migraine as identified by random forest analysis. Shown in the black bar is the predominant factor associated with each of the top four reasons for hesitating to seek care. For each reason, factors of interictal burden (reason #1), severe stigma (reason #2), headache pain intensity (reason #3), and lack of health insurance (reason #4) dominated over remaining factors identified. Note: Migraine interictal burden was assessed using the MIBS-4, migraine-associated disability was assessed by using MIDAS, and migraine-related stigma was assessed by using MiRS. MIBS-4 4-item Migraine Interictal Burden Scale, MIDAS Migraine Disability Assessment Questionnaire, MiRS Migraine-Related Stigma questionnaire, PHQ-4 Patient Health Questionnaire-4. Panel A Between groups of those who cited the following factors and those who did not cite the factor, respectively, as reasons for hesitating to seek care: interictal burden, 56.4% reported experiencing moderate/severe migraine burden between attacks vs. 68.4%; behavioral management, 22.0% were currently using behavioral treatments to manage migraine vs. 14.8%; vomiting with headache, 13.6% reported vomiting half the time or more with migraine attack vs. 19.1%; health insurance, 86.5% had healthcare insurance vs 80.9%; and cannabis use, 28.0% reported current cannabis use vs. 34.6%. Panel B Between groups of those who cited the following factors and those who did not cite the factor, respectively, as reasons for hesitating to seek care: migraine-related stigma, 56.8% reported experiencing migraine-related stigma often/very often vs. 35.6%; frequency of migraine headache days, 33.3% reported having ≥ 8 monthly headache days vs. 20.7%; psychiatric comorbidities, 70.8% self-reported medical diagnosis of at least one of anxiety, depression, and/or panic attacks vs. 58.0%; migraine-associated disabilities, 43.5% reported severe migraine-related disability vs. 28.6%; and interference with mood, 57.9% reported pain and migraine symptoms interfering with mood most/all of the time vs. 41.9%. Panel C Between groups of those who cited the following factors and those who did not cite the factor, respectively, as reasons for hesitating to seek care: headache pain intensity, mean pain intensity was 6.7/10 vs. 7.4/10; vomiting with headache, 9.7% reported vomiting half the time or more with a migraine attack vs. 19.5%; interictal burden, 52.4% reported experiencing moderate/severe migraine burden between attacks vs. 67.5%; current age, mean age was 37.4 years vs. 40.1 years; and nausea with headache, 42.5% reported nausea half the time or more with a migraine attack vs. 56.2%. Panel D Between groups of those who cited the following factors and those who did not cite the factor, respectively, as reasons for hesitating to seek care: health insurance, 70.3% had health insurance vs. 88.2%; region, 11.0% lived in the Northeast and 46.0% lived in the South vs. 17.8% lived in the Northeast and 38.6% lived in the South; frequency of migraine headache days, 28.5% reported having ≥ 8 monthly headache days vs. 23.6%; PHQ-4, mean score of 5.7 for key symptoms of anxiety and depression over previous 2 weeks vs. mean score of 5.0; and psychiatric comorbidities, 66.5% self-reported medical diagnosis of at least one of anxiety, depression, and/or panic attacks vs. 60.8%

Discussion

Hesitation to seek healthcare for migraine may likely lead to poorer health outcomes by delaying time to receiving accurate diagnoses and effective therapies. In this analysis, hesitating to seek migraine care was highly prevalent; 45.1% of respondents with active migraine “ever hesitated to seek care for migraine.” Using machine learning (random forest and LASSO), this analysis identified two factors (among 56) most associated with migraine care hesitancy: (1) hiding of migraine from friends and family (often/very often; OR = 2.7) and (2) migraine-related stigma (often/very often; OR = 2.1). Other variables significantly associated with hesitating to seek care, by multivariate logistic regression model, included higher disability, higher ictal cutaneous allodynia, and higher interictal burden.

It is not surprising that hiding migraine is the factor identified as most strongly associated with hesitating to seek care, since both hiding and hesitating behaviors entail active nondisclosure of the disease. Our finding that more than a quarter of respondents with active migraine hid it from family and friends aligns with a recent study reporting that about a third of people with migraine avoided telling others about it [30]. Hiding migraine at the workplace was not included in our current study, as this variable can be assessed only among those employed. Future analyses will include this subgroup. However, our current study showed that 44.4% of employed respondents who hesitated to seek care also often/very often hid migraine in the workplace.

The strong association of migraine-related stigma with hesitating to seek care suggests that both stigma and hiding contribute to hesitancy to seek medical care. The temporal associations and causal sequences cannot be unpacked in these cross-sectional analyses. While our machine learning models identified stigma as a strong factor associated with hesitating, they did not identify either headache frequency or headache severity as strongly associated with hesitating, despite the fact that the likelihood of hesitating to seek migraine care increased with a higher number of monthly headache days and migraine symptom severity score. This apparent disparity illustrates that in performing multiple simultaneous analyses of large datasets using machine learning, factors that may show statistically significant effects, when taken individually, may not emerge as key variables when compared to other factors. Comparably, we had previously shown that while migraine-related stigma increases directly as headache frequency increases, severe migraine-related stigma is more strongly associated with diminished migraine-specific quality of life than with headache occurring at any frequency [30]. Taken together, these findings suggest that the social context of migraine (i.e., migraine-related stigma) may have a significantly greater impact for many people living with the disease than migraine frequency or symptoms (e.g., headache).

Additional migraine-related factors identified in our analyses as most strongly associated with hesitating to seek migraine care also reflected higher migraine burden (i.e., lower quality of life, higher disability, higher anxiety/depression, higher ictal cutaneous allodynia, higher interictal burden). It may be that hesitating to seek care delays seeking care and leads to more severe disease, although our cross-sectional analyses cannot clarify whether hesitating to seek care precedes development of higher migraine burden. Those who hesitate to seek care have had migraine for a shorter period than those who do not hesitate (17.5 vs. 20.1 years). Those who hesitate may also have more severe migraine even after they seek care, because they were reluctant to implement aspects of treatment for migraine. Stigma may also play a role. Our previous study found that for any frequency of monthly headache days, the presence of migraine-related stigma was associated with significantly higher disability, higher interictal burden, and lower quality of life [30]. Stigma may mediate the association between hesitating to seek care and severity. That is, people who experience a higher symptomatic burden may consequently experience greater migraine-related stigma and greater hesitancy to seek care by virtue of stigma. Concealing a stigmatized medical identity such as migraine might be associated with some advantages (e.g., avoiding discrimination) that may outweigh considerable potential costs (e.g., reduced disease management) in some circumstances [31]. Further studies are needed to understand whether stigma could be a mediator of hesitating to seek care for migraine.

Our analyses identified four, non-mutually exclusive, reasons endorsed by at least 25% of people for why they hesitated to seek migraine care. As the factors associated with the various reasons for hesitation differ, alternative interventions may be considered. For respondents citing reason #1 (“I wanted to try and take care of/deal with them on my own”), the variables most highly associated suggest that these respondents may be experiencing relatively lower migraine burden (lower MIBS-4, less vomiting with headache) for whom self-care may be an appropriate option. Moreover, these respondents reported relatively better access to effective migraine education and care (using behavioral treatments, available insurance), compared to others that hesitated to seek care; self-efficacy is reported to be a mediator between health literacy and health behavior [32]. For a subgroup with treatment-responsive disease, self-care may be an appropriate option. Education may be directed towards knowing when to seek medical care. By contrast, for respondents citing reason #2 (“I felt my migraine or headache wouldn’t be taken seriously”), the most highly associated variables suggest that they are experiencing significantly higher migraine-related stigma, greater migraine-related disability, more psychiatric comorbidities, and higher headache frequency, compared to others that hesitated to seek care. For this group, emphasizing efforts to reduce disability and destigmatize migraine might be helpful. For respondents citing reason #3 (“I didn’t think they were serious/painful enough”), associated variables suggest these respondents were experiencing relatively lower migraine symptom burden (i.e., lower headache severity, less nausea, less vomiting with headache, lower interictal burden) and were younger, relative to others that hesitated to seek care. Some individuals in this group may not have needed medical care, and again, education focused on how to know when one needs a doctor may be helpful. Finally, for respondents citing reason #4 (“I couldn’t afford it or didn’t want to spend the money”), the strongest associated variable was lower rates of healthcare insurance, where improving access to high-quality medical care might be crucial.

It is notable that 57.1% of people with active migraine who hesitated to seek care eventually did so, suggesting that hesitating is not an absolute determinant of consulting. Moreover, the likelihood of further consulting was related to the particular reasons that individuals may give for hesitating. For example, respondents that indicated “I felt my migraine or headache wouldn’t be taken seriously” (associated with higher stigma burden) as a reason for hesitating were more likely to consult than those that indicated “I didn’t think they were serious/painful enough” (associated with lower disease burden) (i.e., 61.5% vs. 47.9%). Notably, more than a quarter of respondents (27.4%) who hesitated cited “I couldn’t afford it or didn’t want to spend the money,” which is overwhelmingly associated with lack of access to health insurance. Further, only about half (51.8%) of those respondents consulted. Collectively, this finding suggests that approximately one in eight Americans with active migraine (27.1% of the 45.1% of those with active migraine who ever hesitated to seek care) are hesitating, in part, for financial reasons, and nearly half had not yet sought treatment at the time of the survey. This finding underscores the serious barrier that high healthcare costs and/or poor health insurance may present for equitable access to migraine care.

In summary, those who hesitated because they preferred self-care (reason #1) or whose headaches were perceived as “not that bad” (reason #3) may have lower perceived needs, less severe migraine presentations, and/or more effective migraine self-care. However, those who feared disrespect (reason #2) or experienced financial barriers (reason #4) have evidence of significant burden, high stigma, lack of health insurance, psychiatric comorbidities, and residency in the South. There is a divergence in reasons for hesitating to seek care between the factors underlying reasons #1 and #3 versus reasons #2 and #4. In other words, those experiencing mild disease burden may not perceive the relative need for care whereas those experiencing severe disease burden may confront real or perceived access barriers to receiving appropriate and timely care.

Heightened migraine social burden (stigma), migraine symptomatic burden, and psychiatric comorbidities are each significant factors associated with these hesitating behaviors in our overall population of respondents, and successful efforts to improve migraine care-seeking will likely require addressing all of them. However, our cross-sectional survey analyses do not allow us to identify causal factors that may mediate or moderate hesitating behaviors.

Finally, while machine learning approaches are helpful in guiding interpretation of multiple simultaneous factors in large datasets at a whole population level, these approaches may also obscure factors of importance in smaller subpopulations. For example, hesitancy to seek migraine care appears to be particularly prevalent among Hispanic (52.7%) and Black (46.8%) respondents, relative to White (44.2%) and Asian (42.1%) respondents, as well as among respondents in Western (46.7%) and Southern (46.3%) regions, relative to those in the Midwest (44.8%) or Northeast (40.8%). Understanding the basis for these disparities may help mitigate health inequities.

The clinical implications of the findings in this manuscript are manifold. Healthcare professionals often see patients hesitant to seek medical care. These individuals often have a long gap from migraine onset to initial diagnosis. Understanding that these delays are common and understanding the reasons for hesitating can improve clinician–patient communication and help clinicians to sensitively address patient needs and concerns. Population-based surveys, such as OVERCOME (US), provide a method for understanding public health needs that includes those individuals whose engagement with the healthcare system is delayed or does not occur. This includes individuals who hesitate to seek medical care. Developing strategies to address these barriers can be implemented on both the patient–healthcare provider level and the local, delivery system, or national levels. Our analysis showed that high healthcare costs and/or poor health insurance were associated with hesitation to seek care for migraine. Hence, public health interventions to mitigate these aspects may be able to improve equitable access to migraine care. Further, heightened awareness of migraine as a primary neurological disease, through local/national campaigns as well as conversations between individuals and their healthcare provider, may be able to improve acceptance of the disease, reduce migraine-related stigma and hiding of the disease, and promote the seeking of treatment for their disease.

Strengths and Limitations

This study has significant strengths. OVERCOME (US) is the largest population-based web-based survey addressing people with migraine and assessing hesitancy in seeking healthcare. The machine learning approaches utilized in this study are capable of simultaneously analyzing large sets of variables and detecting differences with greater relative effect sizes beyond simply determining statistical significance for individually identified factors. Finally, we employed two machine learning algorithms (random forest and LASSO), reflecting differing statistical approaches and with overlapping but nonidentical results. We see this as a methodological strength rather than a weakness, since there is no consensus surrounding the application of any particular machine learning approach as most appropriate for this type of study.

The study also has limitations. The OVERCOME (US) cohorts may reflect multiple biases including the requirement for respondents to have internet access, quota sampling rather than random sampling to create a demographically representative sample, the necessity for respondents to complete the survey to be counted, and self-reporting of data by respondents. These factors may have systematically limited inclusion of some individuals, such as those with severe disabilities, limited financial resources, or time constraints. Another important limitation is that the study relied on self-reported symptoms to establish a diagnosis of migraine and self-assessment of the extent of the burden of the migraine. Examination of participants by headache or migraine specialists would likely strengthen the results of this study. With that said, this study design allowed for a very robust cross-sectional sample size that would not have been feasible if conducted in a traditional clinical trial setting. While this analysis was a comprehensive evaluation of multiple variables, we may have failed to include all important reasons for hesitating to seek to care. Finally, recognizing that the OVERCOME (US) survey included questions around treatment interventions for migraine (including pharmacological therapies) and that the study was sponsored by a pharmaceutical company, the survey was executed through a third-party vendor. Individuals completing the questionnaire did not have line of sight regarding sponsorship of the study, which may also have inherent biases. In addition, recruiting from the population circumvents the biases inherent in clinic-based studies.

Conclusion

We found that nearly half of the participants with active migraine in OVERCOME (US) had ever hesitated to seek care for it, although most eventually did seek care. The key factors associated with hesitation were hiding migraine and experiencing migraine-related stigma. Further, greater migraine-related symptomatic burden (e.g., disability, low quality of life, severe symptoms) may be associated with greater hesitancy to seek care to mitigate migraine burden. Disease progression might therefore be consequent to delays in seeking such care in the face of stigma towards the disease and financial disparities.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Robert E. Shapiro, Eva Jolanda Muenzel, Robert A. Nicholson, Anthony A. Zagar, Michael L. Reed, Dawn C. Buse, Susan Hutchinson, Sait Ashina, Eric M. Pearlman and Richard B. Lipton contributed to the study conception and design. Statistical analysis was performed by A. Zagar. The first draft of the manuscript was written by R. Shapiro and J. Muenzel; all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

Sponsorship for this study and the journal’s Rapid Service Fee were funded by Eli Lilly and Company.

Data Availability

Lilly will provide access to anonymized individual participant data collected during the study. The data will be available to request on www.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.

Declarations

Conflict of Interest

Robert E. Shapiro: Consulting: Eli Lilly and Company, Lundbeck, AbbVie/Allergan, Theranica. Michael L. Reed: Consulting, Research Grants: AbbVie/Allergan, Amgen, Eli Lilly and Company, GlaxoSmithKline, National Headache Foundation, Promius. Dawn C. Buse: Consulting, Honoraria, Research Grants: AbbVie/Allergan, Amgen, Biohaven, Collegium, Eli Lilly and Company, Teva. Susan Hutchinson: Consulting, Speaking, Honoraria: AbbVie/Allergan, Amgen, Biohaven, Currax, electroCore, Eli Lilly and Company, Impel, Lundbeck, Teva, Theranica, Upsher-Smith. Sait Ashina: Consulting, Teaching, Honoraria: AbbVie/Allergan, Biohaven Pharmaceuticals, Eli Lilly and Company, Impel NeuroPharma, Lundbeck, Satsuma, Percept, Teva, Theranica; Consulting and Advisory Board: Pfizer. Richard B. Lipton: Stock or options: Manistee Health; Consultant, Advisory Board Member, Honoraria from or research support: Abbvie (Allergan), American Academy of Neurology, American Headache Society, Aeon, Amgen, Biohaven, Biovision, Boston, Dr. Reddy’s (Promius), electroCore, Eli Lilly and Company, Equinox, GlaxoSmithKline, Grifols, Lundbeck, Merck, Pernix, Pfizer, Teva, Vector, Vedanta. E. Jolanda Muenzel, Robert A. Nicholson, Anthony J. Zagar, and Eric M. Pearlman: Employee/Shareholder: Eli Lilly and Company.

Ethical Approval

The Sterling Institutional Review Board (IRB ID #6425–001) provided approval for this non-interventional, observational survey study. All respondents provided electronic informed consent to participate in a general health-related survey. This was a non-interventional, observational study and not registered with a clinical trials registry.