Functional improvement in children and adolescents with primary headache after an interdisciplinary multimodal therapy program: the DreKiP study

Patients in the program

Descriptive statistics and further results are reported for patients who provided outcome questionnaires for at least one follow up visit. This subcohort consists of 75 patients, 50/75 (67%) female, 9-19 years old (median = 14, IQR = 13:16; Supplementary Tables 1 and 2). All of the participants had at least one primary headache diagnosis according to the IHS criteria [22] (for details see Fig. 2).

Fig. 2figure 2

Primary headache diagnoses of patients in the program. 16 patients (21,3%) episodic Migraine without Aura (eMwoA), 9 patients (12%) episodic Migraine with Aura (eMwA), 8 patients (10,7%) chronic Migraine (cM), 15 patients (20%) episodic Tension Type Headache (eTTH), 3 patients (4%) chronic Tension Type Headache (cTTH), 9 patients (12%) chronic Tension Type Headache/episodic Migraine (cTTH/eM), 15 patients (20%) episodic Migraine/episodic Tension Type Headache (eM/eTTH)

In 45 (60%) participants, at least one accompanying chronic disease was diagnosed (Table 1).

Table 1 Frequency of additional diseasesHeadache days at baseline and after group therapy

At the beginning of the program (T0), headache frequency data were available for 72 participants (96%). Patients reported having had on average 43.74 (±33.44) headache days within the last 3 months, with median 30 days (IQR = 15:90). Girls presented more headache days than boys: mean 48.4 (± 34.3) and 34.9 (±30.5), medians 40 (IQR = 15:90) and 20 (IQR = 15:40) days, respectively (W = 706, p = 0.158). Younger participants (≤13) showed on average 22 days less headache in the last 3 months than patients ≥14 years (W = 312, p = 0.008).

At T0, headache days differed significantly depending on the headache diagnosis (Kruskal-Wallis test, χ2 (3) = 16.194, p = 0.001).

Six months after the program (T1), 65 (86.7%) patients provided information about their headache days. The mean of 35.65(±34.19) days was reported (median = 18), with girls presenting on average 20.7 more headache days than boys (W = 635.5, p = 0.024). The overall difference between headache diagnoses was still considerable (Kruskal-Wallis test, χ2 (3) = 17.096, p < 0.001), but the difference between chronic and episodic migraineurs was not significant anymore (W = 51, p-value = 0.112). Chronic migraineurs improved and the headache frequency became lower.

Twelve months after completing the program (T2), data were provided by 47 (62.7%) patients. The mean number of headache days in the last 3 month was 27.7 (± 30.9), median 13, and thus overall reduced compared to baseline and to T1 (Fig. 3). No major differences between headache diagnoses were observed (Kruskal-Wallis, χ2 (3) = 6.595, p = 0.086).

Fig. 3figure 3

Headache frequency. (Median headache days in the last 3 months) at baseline and follow up 6, 12 and 24 months after the program “DreKiP” (stands for Dresden children and adolescent headache program)

Two-year follow-up data (T3) were provided by 25 (33.3%) patients. They had, on average, 23.32 (± 26.04) days, median = 12, which may indicate that the reduction achieved by 12 months remained stable. No significant differences between headache diagnoses were observed (Kruskal-Wallis, χ2 (3) = 0.145, p = 0.986).

Time trends in headache days were investigated using a linear mixed-effects model with random intercept per patient and fixed effects for gender, time point, and main diagnosis. This analysis arrived to similar conclusions, indicating a strong reduction in headache days. As compared to the baseline, the average number of headache days decreased on average by 11.2 days per last 3 months T1 (p = 0.002) and by 17.3 days per last 3 months at T2 (p < 0.001). Finally, at T3, patients experienced on average 16.7 days less per last 3 months compared to the baseline (p = 0.001). The interaction between time and gender was found borderline significant (χ2 (3) = 7.975, p = 0.047). In particular, the estimated number of headache days of girls after 6 months was about 16.8 days per last 3 months more than those of boys at the same time point (p = 0.029). However, no difference could be inferred after 1 year (p = 0.057) or 2 years (p = 0.526; Fig. 3).

Headache intensity at baseline and after group therapy

At T0, 69 (92%) patients provided data describing their headache intensity. Seven of them (10.1%) reported having mild, 30 (43.5%) moderate, and 32 (46.4%) severe headache, and the intensity seems to be associated with headache diagnosis (χ2 (6) = 15.499, p = 0.017). While nobody reported having pain of a highest severity, patients with episodic migraine experienced severe headache more frequently than other patients (OR = 3.70, p = 0.026).

The intensity data of 57 (76%) patients were collected at T1. One adolescent (1.7%) did no longer complain about headache, 13 patients (22.8%) had mild, 24 (42.1%) moderate, 15 (26.3%) severe headache, and four (7.0%) reported the most severe headache. No association with headache diagnosis could be concluded (χ2 (12) = 12.152, p = 0.434), and the difference in intensity between genders was not significant (χ2 (4) = 8.183, p = 0.085).

At T2, information on pain intensity was obtained from 36 (48%) patients. One (2.8%) had no headache, 7 (19.4%) mild, 17 (47.2%) moderate, 10 (27.8%) severe, and one (2.8%) the most severe headache (Fig. 4). No difference due to headache type (χ2 (12) = 17.135, p = 0.145) or gender (χ2 (4) = 7.722, p = 0.102) could be inferred.

Fig. 4figure 4

Headache intensity. (proportion of patients with headache intensity classified as “most severe”, “severe”, “moderate”, “mild” and “no headache” in the last 3 months) at baseline and follow up 6, 12 and 24 months after the program “DreKiP” (stands for Dresden children and adolescent headache program)

Only 17 patients (22.7%) provided data on headache intensity at T3. Their data indicated no new trends as compared to T2.

The effects of time, headache diagnosis, and gender on headache intensity were explored using a linear mixed-effects model. As in the descriptive analysis, the effects of time and diagnosis were found to be significant (F = 4.089, p = 0.008, and F = 4.552, p = 0.006, respectively). A notable change in intensity was estimated after 12 months (about 30% decrease on average, p = 0.056). The estimated difference between baseline and 24 months was, on average, 40% (p = 0.025; Fig. 4).

Analgesic medication at baseline and after group therapy

Initially, information about analgesic intake was provided by 72 patients (96%). Among them, 16 (22.2%) reported no analgesic medication intake for headache, 36 (50%) used analgesics more than once a month and 20 (27.8%) more than once a week.

At T1, medication-intake data were available for 65 (86.7%) participants, of whom 19 (29.2%) did not need analgesics, 36 (56.9%) used medication more than once a month, and 9 (13.8%) more than once a week.

At T2, 45 patients (60%) provided information on analgesics. Of them, 18 (40%) stated that they were not taking any medication, 20 (44.4%) reported taking it more than once a month, and 7 (15.6%) more than once a week (Fig. 5). Finally, at T3, data on analgesics were provided by 23 patients (30.7%). Two of them (8.7%) were using analgesics more than once a week, 11 patients (47.8%) reported that they were taking none (Fig. 5).

Fig. 5figure 5

Acute Headache Medication intake at baseline and follow up, 6, 12 and 24 months after the program

No statistically significant changes in analgesic intake, compared to the baseline, were found at either 6 months or 12 months time point (McNemar–Bowker symmetry test, p = 0.105, p = 0.128, respectively), however, the two-year difference to baseline was significant (p < 0.001). The association of analgesic intake with headache diagnosis and with gender at either time points was not significant.

Headache related disability at baseline and after group therapy

At T0, PedMIDAS as measure of headache related disability of 75 children and adolescents were obtained. Scores ranged from 1 to 200, mean 40.51 (±39.25) and median 30, indicating outliers. Patients with chronic migraine presented higher than average scores, mean of 66.0 (±13.81). The second largest mean score of 40.0 (±7.97) was observed for the episodic migraine group. However, no overall association with headache diagnosis could be inferred (F(3,70) = 1.69, p = 0.177, log-scale).

At T1, PedMidas scores were obtained for 65 (86.7%) patients. The mean score 23.83 (± 24.93) was significantly lower than at the beginning of the program (V = 1465, p < 0.001). The maximum score of 120 and the median of 15 indicated a positive development as well. The difference in PedMidas due to headache diagnoses was not significant (F(3,61) = 1.167, p = 0.330, log-scale). As before, the chronic migraine group had the largest mean score of 36.6 (± 8.53), compared to other diagnosis groups.

At T2 data revealed a further reduction in PedMidas scores. The scores of 47 patients (62.7%) who returned for the follow-up ranged from 0 to 77 (median = 7), and their mean 16.89 (± 20.35), was significantly lower than the baseline value (V = 947.5, p < 0.001; Fig. 6). As before, no overall association with headache diagnosis was found (F(3,43) = 1.987, p = 0.130, log-scale). The mean score of participants with chronic migraine was 41.6 (± 8.51).

Fig. 6figure 6

PedMidas. (Median) at baseline and follow up 6, 12 and 24 months after the program “DreKiP” (stands for Dresden children and adolescent headache program)

At T3, PedMidas scores were obtained for 26 patients (34.7%). By then, the maximum score was 105, the mean was equal to 15.96 (± 26.16) and median = 3.5. Although these scores were not significantly different from the 12 months data (V = 89.5, p = 0.277), the improvement, compared to the start of the program, remained significant (V = 296, p < 0.001). As in the previous time points, no association with headache diagnosis could be inferred (F(3,22) = 1.196, p = 0.334, log-scale), with the mean score of participants with chronic migraine being the highest, 30.8 (± 10.04).

Further, the percentage change in PedMidas score over time was investigated using a linear mixed-effects model. According to the fit results, the scores decreased by 50% already in 6 months (p < 0.001), a decrease of about 70% was estimated to be achieved in 12 months (p < 0.001), and of almost 80% in 2 years (p < 0.001). Across all time points, the estimated scores for girls were on average about 90% higher than for boys (p = 0.024).

Pain related disability at baseline and after group therapy

At T0, PPDI scores were available for 63 (84%) children and adolescents. The minimum and maximum scores were 18 and 57, respectively, with mean = 33.13 (± 8.82), and median = 34.

At T1, PPDI scores were obtained from 60 (80%) patients. This time, the scores ranged from 12 to 57, with the mean of 29.93 (± 10.85) and median of 30.

At T2 scores were reported for 44 patients (58.7%). Although the score range stayed about the same, from 12 to 55, the new mean value of 27.91 (± 11.15) and median of 26 reflected a tendency of somewhat reduced pain disability (Fig. 7).

Fig. 7figure 7

PPDI. (Median) at baseline and follow up 6, 12 and 24 months after the program “DreKiP” (stands for Dresden children and adolescent headache program)

At T3 data were collected from 23 patients (30.7%). The scores, ranged from 12 to 47, with the mean of 23.61 (± 9.63) and median of 21 indicated a small reduction in pain disability (Fig. 7). At either time point, no significant differences in PPDI scores due to the headache diagnosis were found.

An additional analysis with linear mixed-effects models indicated an overall significant improvement in PPDI by about 25% starting from baseline to 24 months (p = 0.001), but implied a slightly different time dynamic in PPDI scores of boys and girls (χ2 (3) = 7.864, p = 0.049). Throughout the whole time, girls’ scores were estimated to be about 20% higher than those of boys (p = 0.022). By 6 months, boys’ scores decreased by 25% (p = 0.001), while girls’ scores stayed, on average, about the same (p = 0.3961). According to the model, a gradual decrease continued for both genders, leading to almost a 30% reduction in girls (p < 0.001) and 20% reduction in boys (p = 0.076) in 24 months after the program.

留言 (0)

沒有登入
gif