Migraine Diagnosis and Treatment: A Knowledge and Needs Assessment of Women's Healthcare Providers

Allison M. S. Verhaak PhD; Anne Williamson PhD; Amy Johnson MD; Andrea Murphy APRN; Matthew Saidel MD; Abigail L. Chua DO; Mia Minen MD, MPH; Brian M. Grosberg MD


Headache. 2021;61(1):69-79. 

In This Article


Characteristics of Respondents

Of the approximate 402 providers invited to participate in the web-based survey, 115 completed the survey over the 12 weeks in which it was available (response rate 28.6%). Table 1 reveals the demographics of the participants. Of the respondents, 80.0% (92/115) were female, and 88.7% (102/115) were physicians (of which 28.7%, or 33/115, were obstetrics and gynecology residents or fellows). Other participants included Advanced Practice Providers (APP), including advanced practice registered nurses, physician assistants, and certified nurse midwives (n = 13). Response numbers by question varied due to differing practice patterns (e.g., some respondents indicated that they do not prescribe and/or are not comfortable prescribing certain medications for migraine, and therefore would not complete additional medication-specific questions). The number of responses for each question are indicated in referenced tables. There were no missing data.


Table 2 displays women's healthcare providers' knowledge of migraine prevalence, comorbidity prevalence, and other knowledge considerations pertaining to treatment of migraine. Respondents reported that 18.1% ± 14.9% of the population has migraine, and that nearly one-third of people with migraine have depression (32.2%) and anxiety (31.6%). Regarding evidence-based non-pharmacological treatments for migraine, 56.5% (65/115), 50.4% (58/115), and 40.9% (47/115) of respondents identified biofeedback, cognitive behavioral therapy (CBT), and relaxation, respectively, as treatments known to them, as opposed to the other non-pharmacological treatments.

A small percentage of the respondents were familiar with the American Academy of Neurology guidelines for prescribing preventive medications (6.3%; 7/111) or with the Choosing Wisely campaign for guidance with medication management (17.3%; 19/110). Providers were also queried about the potential risks of medication-overuse headache (MOH) secondary to migraine abortive medications. The vast majority (70.4%; 81/115) of respondents correctly identified barbiturate-containing combination analgesics as being linked with medication overuse headaches, but were less knowledgeable about other medications that can lead to medication overuse headaches, including opioids (43.5%; 50/115), NSAIDs (40.9%; 47/115), and triptans (20.9%; 24/115).

Practice Patterns

Survey respondents reported that they served as the PCP for approximately one-third (31.8%) of their patient population, and believe that 23.4% of their patient population experiences headache (Table 3). Approximately half of providers (57.9%; 66/114) routinely ask patients about headaches during the annual visit; roughly two-thirds (67.8%; 78/115) of respondents have made the diagnosis of migraine, and nearly half (45.9%; 50/109) have made the diagnosis of medication overuse headache. The majority (82.6%; 95/115) stated that they were somewhat comfortable or very comfortable diagnosing migraine. When diagnosing migraine, 69.2% (54/78) assess for disability, 59.0% (46/78) assess sleep patterns, and approximately half of respondents assess for depression (51.3%; 40/78) and anxiety (46.2%; 36/78) symptoms.

Headache with neurologic symptoms was the most common reason for ordering magnetic resonance imaging (MRI) (47.8%; 55/115), followed by headache not responding to treatment (41.7%; 48/115), worsening headache (35.7%; 41/115), and new type of headache (24.3%; 28/115) (Table 3). Nearly half of survey respondents reported that they did not feel comfortable ordering an MRI for headache (45.2%; 52/115). A far greater percentage of respondents reported that they would send a patient to the emergency department (ED) for headache with neurologic symptoms (85.2%; 98/115), as well as worsening headache (51.3%; 59/115) and headache not responding to treatment (41.7%; 48/115).

Treatment and Referrals

Nearly 60% (68/114) of respondents were comfortable prescribing medication for migraine (Table 4). NSAIDs were the most commonly prescribed drug (78.1%; 89/114), followed by acetaminophen (63.2%; 72/114). More than half of providers (58.8%; 67/114) prescribed barbiturate-containing combination analgesics, and 37.7% (43/114) prescribed triptans. Only one provider (0.9%; 1/114) reported prescribing opioids for migraine; those who responded indicated that they only used them in the context of pregnancy. Although a stratified approach is the current recommendation, 39.5% (45/114) reported creating step-wise treatment plans with abortive medications, and only 13% (15/115) reported prescribing preventative medications. Of those who do not prescribe preventative medications, respondents primarily cited a lack of comfort with prescribing preventives (76.0%; 76/100); 9.0% (9/100) had not considered the use of preventives, and 16% (16/100) deferred to the patient's PCP or neurologist to manage headache medications. Respondents reported prescribing anti-emetic medication for less than one-fifth of their patients with migraine (18.6%).

Regarding referral patterns, respondents reported referring nearly half (48.4%) of their patients with migraine to an internist, nearly one-third (31.3%) to a neurologist, and very few patients (1.3%) are referred for CBT or biofeedback therapies (Table 4). The primary obstacles to recommending and referring patients for non-pharmacological treatments (such as CBT and biofeedback) include not knowing to whom they can refer for these services (65.1%; 69/106), issues with insurance covering these treatments (42.5%; 45/106), lack of availability of providers for these services (34.9%; 37/106), and patient skepticism (33.0%; 35/106).

Use of Hormonal Therapies in the Context of Migraine

A portion of the survey was specific to women's health care and prescribing patterns for hormonal therapy in patients with migraine (Table 5). The vast majority of respondents (91.9%; 102/111) reported awareness of American College of Obstetrician and Gynecologists (ACOG) guidelines concerning migraine with aura as a contraindication for use of estrogen-containing contraception. Most providers reported feeling comfortable with prescribing hormonal contraception to patients with migraine with aura (72.5%; 79/109) and without aura (80.9%; 89/110). With regard to contraception-prescribing patterns for patient with migraine, based on presence of aura, 6.3% or less of respondents would prescribe estrogen-containing contraception in the context of migraine with aura. A wider variety of contraceptives, including those containing estrogen, were prescribed to patients with migraine without aura (see Table 5). With regard to hormone replacement therapy, the majority of providers reported they would prescribe to patients with migraine without aura (82.4%; 89/108), while only one-fourth (26.9%; 29/108) would prescribe to those with migraine with aura. The majority of respondents (85.1%; 97/114) correctly identified migraine with aura as a risk factor for ischemic stroke in women.

Migraine Education

The final section of the survey questioned provider's past education around migraine management and the types of education that they felt they needed in order to better serve their patients. Only 37.3% (41/110) of the providers reported having received headache/migraine-specific education, and many indicated that they would welcome educational opportunities in this area. A large majority (96.3%; 105/109) reported that receiving education around migraine prevention and appropriate medication use would improve their patient care, 74.3% (81/109) endorsed wanting to learn more about the degree of disability associated with migraine, and 59.6% (65/109) reported wanting to know more about diagnostic testing.

Respondents Serving as PCP

There was a significant difference between level of training (attending vs. resident/fellow vs. other) and percentage of patients for whom the provider assumed they were the PCP, as determined by one-way ANOVA, F(2,109) =9.87, p < 0.001. Post hoc testing (Tukey) revealed that fellows/residents reported serving as PCP for a higher percentage of their patients (M = 46.3, SD = 25.9) than did attending providers (M = 24.7, SD = 20.5). There were no differences observed between other groups (attending vs. other; resident/fellow vs. other). In addition, providers who reported feeling comfortable with prescribing medications for migraine also reported serving as PCPs for a higher percentage of their patients (M = 38.5, SD = 24.5) than providers who reported that they were not comfortable with prescribing medications for migraine (M = 22.1, SD = 21.3), F(1,108) = 12.87, p = 0.001. There were no other significant differences in knowledge or practice patterns based on the frequency by which providers act as PCPs for their patients.