A total of 400 physician members of the JHS responded to the survey, and the responses of 397 of the physicians were included in the analyses; the other three physicians' answers were deemed inappropriate for the analyses. Another nine responses were partly excluded from the analysis since they included inexplicable responses regarding the number of patients.

The average length of the first appointments for the headache patients was 23.5 ± 9.5 (20 [15, 30]) min, and the average length of the follow-up appointments was 9.0 ± 3.9 (10 [5, 10]) min (Table 1). As depicted in Fig. 1, while 91% of them spent ≥ 15 min for the first appointment (15–19 min: 19%, 20–29 min: 31%, 30–39 min: 34%, ≥ 40: 7%), 85% of them spent < 15 min for the follow-up appointments (1–4 min: 4%, 5–9 min: 33%, 10–14 min: 48%). The physicians regularly followed up 121.8 ± 263.7 (30 [15, 100]) patients with migraines (0–19: 27%, 20–49: 29%, 50–99: 17%, 100–199: 9%, 200–499: 12%, ≥ 500: 6%), of whom 77.1 ± 165.7 (20 [9, 67.5]) patients had ever used a migraine preventive treatment and 25.5 ± 45.9 (10 [4, 25]) had ever been treated with a CGRPmAb (Table 1). The responses also revealed that 80.6% (n = 320) of the physicians had experience prescribing CGRPmAbs. The ratio of patients who had received a migraine preventive treatment over the total number of migraine patients was 64.9 ± 23.2%. The ratio of patients who had received a CGRPmAb as a migraine preventive treatment over the total number of migraine patients was 23.9 ± 17.1%. These data also revealed that the ratio of patients treated with a CGRPmAb over the total number of patients who received a migraine preventive treatment was 37.6 ± 26.1%.

Appointment length and number of patients who are followed up regularly. a Length of the headache patients' first appointments (min). b Length of the headache patients' follow-up appointments (min). c No. of patients with migraine who are followed up regularly

The types of facilities where the physician respondents were working were mainly community hospitals (n = 194, 48.9%), followed by clinics (n = 126, 31.7%) and university hospitals (n = 77, 19.4%) (Table 2). Most of the respondents worked in either neurology (n = 170, 42.8%) or neurosurgery (n = 185, 46.6%) services, but there were also physicians in the fields of anesthesiology/pain clinic (n = 16, 4.0%), internal medicine (n = 9, 2.3%), pediatrics (n = 9, 2.3%), and others. Regarding age, most of the respondents were > 40 years old (n = 362, 91.2%). They were also well experienced in migraine treatment, with only 24 (6.0%) of the respondents having < 5 years of experience treating headaches. Moreover, about half of the respondents (n = 202, 50.9%) had ≥ 10 years of JHS membership. Only 23 respondents (5.8%) did not have any board certifications at the time of the survey (Table 2).

In terms of CGRPmAb usage, the most widely available CGRPmAb was galcanezumab (n = 337, 84.9%), followed by fremanezumab (n = 272, 68.5%) and erenumab (n = 244, 61.5%) (Table 2). The number and types of CGRPmAbs available at each facility differed among the community hospitals, clinics, and university hospitals (Suppl. Table S2). All three CGRPmAbs, i.e., galcanezumab, fremanezumab, and erenumab, were available in more than half of the facilities. The percentage of facilities where all three CGRPmAbs were available was highest in the clinics (n = 85, 67.5%) (Suppl. Table S2).

The threshold for the number of MMDs for recommending a CGRPmAb was 10 for 81 (26.0%) of the respondents, and the number of migraine preventives usually tried before prescribing a CGRPmAb was two for 170 respondents (54.5%) followed by one for 64 respondents (20.5%) (Table 3). When the physician respondents were asked about the appropriate time point at which to assess the response to a CGRPmAb in patients with EM or CM, the response of 217 (69.6%) was after 3 months for EM patients; 188 (60.3%) specified after 3 months for CM patients. For patients with CMs, 55 (17.6%) of the physicians waited for 4–6 months before the assessment, whereas 19 (6.1%) of the physicians waited for 4–6 months in cases of EM (Table 3).

Regarding efficacy, we asked the percentage of patients whose MMDs had decreased by ≥ 50% and defined it as the responder rate (RR). For EM patients, the RR was ≥ 80% as answered by 107 (34.3%) respondents. The RR in CM patients was ' ≥ 40% and < 60%,' with 104 (33.3%) responses. The most frequently reported reasons for CGRPmAb responders to discontinue treatment with a CGRPmAb were: (i) sufficient improvement of migraines (n = 169, 54.2%), (ii) high cost (n = 76, 24.4%), and (iii) adverse effects (n = 39, 12.5%) (Table 3). Our survey also asked whether the physicians had any requests about CGRPmAbs, and "cost of CGRPmAbs" was the most frequently mentioned request in their answers (n = 110, 27.7%), followed by "establishment of detailed treatment plans" (n = 13, 3.3%).

A comparison of the survey item responses between the physicians with experience prescribing CGRPmAbs and those without such experience revealed that in the group with experience, the length of follow-up appointments was slightly shorter (p = 0.012) and the number of patients who were followed up was significantly higher (p < 0.0001) (Table 4). The age, years of headache practice, and years of JHS membership also showed significantly different distributions between the experienced and not-experienced groups. The physicians with CGRPmAb experience also had a significantly higher rate of having at least one board certification (p < 0.0001) (Table 4).

Focusing on how the physicians decided whether or not to prescribe a CGRPmAb to patients, we determined the MMD threshold for prescribing CGRPmAbs and the number of migraine preventives tried before CGRPmAbs (Suppl. Table S3). Although approximately half of the respondents answered 'two' for the number of migraine preventives attempted before CGRPmAbs (n = 170, 54.5%) followed by 'one' (n = 64, 20.5%), the MMD threshold for prescribing CGRPmAbs was rather dispersed among ' ≥ 4' (n = 71, 22.8%), ' ≥ 6' (n = 68, 21.8%), ' ≥ 8' (n = 76, 24.4%), and ' ≥ 10' (n = 81, 26.0%). Regarding the combination of these parameters, the MMDs of ' ≥ 10' and 'two' migraine preventives before the use of a CGRPmAb accounted for the highest proportion (n = 48, 15.4%), followed by ≥ 8 MMDs and two preventives (n = 40, 12.8%), ≥ 6 MMDs and two preventives (n = 39, 12.5%), and ≥ 4 MMDs and two preventives (n = 36, 11.5%) (Fig. 2).

Number of migraine preventives before an anti-calcitonin gene-related peptide monoclonal antibody (CGRPmAb) is prescribed, and the monthly migraine days (MMD) threshold for prescribing a CGRPmAb

We also analyzed the between-subgroup differences in the MMD threshold for prescribing a CGRPmAb, and we observed that the number of migraine patients who were followed up regularly (p = 0.0026) and the number of migraine patients who had ever been treated with a CGRPmAb (p = 0.0109) were related to the MMD threshold. Both the number of patients with migraine who were followed up regularly and the number of patients with migraine who were followed up regularly and had ever been treated with a CGRPmAb were the highest among the physician respondents who used the threshold of ≥ 8 MMDs (Table 5).