Abstract
Although the ID MigraineTM has been developed and used for adults, it can also be used in adolescents. We aimed to investigate the validity of the ID MigraineTM in the adolescent student population. The study was performed on 1064 adolescents aged 12-17 years. A face-to-face semistructured interview was conducted with students after they had answered a 17-item questionnaire on sociodemographics and headache. Students who had had two or more headache attacks during the last 3 months then answered the ID MigraineTM. The sensitivity and specificity of the ID MigraineTM were 62.1± and 71.1±, respectively. The sensitivity was higher for patients with migraine with aura (71.1±), for girls (71.3±) and for younger adolescents (69.1±). The specificity was higher for boys (77.9±) and for older adolescents (73.3±). The K value was 0.316 (P< 0.001). Although the K value of the ID MigraineTM is low, the test may be a useful tool in screening adolescent migraine after some changes are made.
Introduction
Migraine is a prevalent and disabling disorder characterized by severe headache attacks and associated symptoms. It has a 1-year prevalence of 10–12% and a life-time prevalence of 12–18%, which have been shown to be age and gender dependent in community-based studies from different parts of the world (1). The life-time prevalence of migraine in Turkey was reported to be 10.9% in men and 21.8% in women in a nationwide epidemiological survey (2). Migraine is one of the world's most 20 disabling medical illnesses according to the World Health Organization's Global Burden of Disease survey (3). Disability is the major determinant of the cost of migraine to society and the patient. The economic impact of migraine is mostly caused by lost work days and decreased productivity (4, 5). Despite its high frequency (1, 2, 6, 7), severity (8) and economic and social burden to society (6, 9–13), migraine is still an underdiagnosed and undertreated disorder (14–16). Less than half of migraine sufferers consult a doctor for their headache and slightly more than one-third of patients with tension-type headache are diagnosed with migraine (6). Lipton et al. have reported that only one-third of migraine patients take prescription drugs (15).
Many difficulties have been reported during the diagnosis and treatment of migraine (17). To overcome these, and for quick and appropriate screening, Lipton et al. have invented a self-administered migraine screener (ID MigraineTM) (ID MIGRAINE is a trademark of Pfizer Inc.) (18). They validated this newly invented screener in the primary care setting and found a sensitivity of 0.81 and a specificity of 0.75.
Migraine is also frequent during adolescence. A recent study conducted in Bursa province of Turkey has revealed that the estimated migraine prevalence was 16.4% among adolescents aged 12–17 years (19). Many migraine sufferers experience their first headache attack during adolescence or childhood. Diagnosis of adolescent migraine may be more difficult than that of adult migraine because to evolving and mixed headache characteristics during adolescence. In the second edition of International Classification of Headache Disorders (ICHD-2), the International Headache Society (IHS) notes two major differences between adolescent and adult migraine: (i) headache duration might be shorter in adolescent and childhood than in adulthood migraine (1–72 h); and (ii) headache is commonly bilateral in adolescents and children (20). Although the ID MigraineTM was developed for adults, it includes no questions regarding these major differences between adolescent and adult migraine. Therefore, we hypothesized that the ID MigraineTM could also be used in adolescents as a screening tool for migraine. In screen-positive adolescents, a face-to-face neurological evaluation might help the diagnosis of migraine in its early stages.
The aim was to investigate the validity of the ID MigraineTM in a large, school-based adolescent student population and to determine whether it could be used as a screening test in a large adolescent population.
Methods
This was a school-based, cross-sectional study, planned as part of a large student-based epidemiological headache study and conducted in the secondary schools of the central city of Bursa, Turkey, between October 2003 and June 2004, on adolescents aged between 12 and 17 years from 6th to 11th grades.
Headache prevalence among school children in Turkey has been reported to be 49.2% (21). The estimated sample size was 1270 for 12–14-year-old students and 1117 for 15–17-year-old students.
A multistep, stratified, cluster sampling method was used for subject selection. In the first step, schools were classified according to the region of the city. Of these schools, 17 were selected by random sampling. Our study sample included 6.5% of the schools and 1.8% of the total student population aged 12–17 years in the area. The gender ratio of the study population was also adjusted to the gender ratio of the student population of Bursa.
The study was planned in three phases. In phase I, a self-administered 17-item questionnaire was answered by all of the students. This questionnaire included 14 questions on the sociodemographic characteristics of the subjects and three questions on headache. Students were asked if they had had headache during the last year and, if so, whether it had recurred or not, and what was its frequency. In phase II, a face-to-face semistructured interview was conducted with students who had had a recurrent or non-recurrent headache. In phase III, students who had had two or more headache attacks during the last 3 months answered the Turkish version of the ID MigraineTM test as performed in the original study. The three questions of the ID MigraineTM test were: (i) do you feel nauseated or sick in your stomach during your headaches?; (ii) does light bother you during your headaches (more than when you do not have headaches)?; (iii) do your headaches limit your ability to work, study or perform necessary activities for at least 1 day (18)?
The Turkish version of the ID MigraineTM screening test has been validated in the workplace in a previous study, in which similar results with the original validation study had been obtained (22).
Parents or legal guardians were informed about the study by telephone and gave oral permission for inclusion of the subjects in the study. The local official authorities and school administrations gave written informed consent.
The headache criteria in the second edition of the ICHD-2 were used for the diagnoses (20). During the interview, students were questioned regarding all headache types and one or more headache diagnosis was made. The diagnosis of the neurologist was used as the gold standard. Eligibility criteria for this study were therefore different from those of the original study. The only inclusion criterion was two or more headache attacks during the last 3 months. Additional inclusion criteria of the original study that were the limitation of ability to work, study or enjoy life and willing to speak with a physician were not required. After the interview, all adolescents who met the inclusion criteria were given the Turkish version of the 3-item ID MigraineTM screening test. The test was explained to the students, who then completed the ID MigraineTM by themselves. Those who answered ‘yes’ to two out of three items of the test were accepted as screen-positive for migraine.
Statistical analysis was performed using SPSS for Windows (version 11.0; SPSS Inc., Chicago, IL, USA). The ID MigraineTM screening test results were analysed for different headache types—migraine, migraine without aura (MoA), migraine with aura (MA) or episodic tension-type headache (ETTH)—according to age group and gender by using Pearson's χ2 test and Fisher's exact test. The sensitivity, specificity, positive predictive value and negative predictive value of the test and of each of the three questions of the ID MigraineTM test were calculated for all of the headache types using the neurologist's diagnosis as the gold standard. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated. The statistical significance level was accepted as P < 0.05.
Results
A total of 1064 adolescent students [mean age (± SE) 14.67 ± 1.74] with recurrent headache were interviewed for this study, all of whom completed the ID MigraineTM screening test. Of those, 50 were excluded due to inadequate data quality. The screening test results of 1014 subjects were statistically analysed. The test results of patients from each headache subtype were compared with those of the other patients with headache. Of the analysed data, only the results of patients with migraine and ETTH were presented here, because patient numbers in other headache groups were small.
The sensitivity and specificity of the ID MigraineTM were 62.1% and 71.1%, respectively, for the migraine patients. The sensitivity was higher for patients with MA (71.1%), but specificity was higher for those with MoA (66.8%). For patients with ETTH, the sensitivity was 28.7% and specificity 44.7%. For migraine patients, of the three ID MigraineTM questions, disability showed the highest sensitivity (67.2%) and nausea the highest specificity (80.4%). The OR of the ID MigraineTM was 4.04 (95% CI 3.06, 5.32) and the negative predictive value was 79.2% for the migraine patients. The ID MigraineTM had an OR of 0.33 (0.25, 0.42) for ETTH. The OR, sensitivity, specificity, positive predictive value and negative predictive value of the ID MigraineTM according to the migraine types are given in Table 1.
The odds ratio, sensitivity, specificity, positive predictive value, and negative predictive value of the ID MigraineTM according to the migraine types
P < 0.001;
P < 0.05.
CI, Confidence interval; ETTH, episodic tension-type headache; MA, migraine with aura; MoA, migraine without aura; NPV, negative predictive value; OR, odds ratio; PPV, positive predictive value; Sn, sensitivity; Sp, specificity.
When the results according to age and gender were analysed, the sensitivity of the ID MigraineTM was found to be higher for girls than for boys (71.3% vs. 49.3%). Younger adolescents showed a higher sensitivity of the test than older ones (69.1% vs. 56.3%). On the other hand, the specificity of the ID MigraineTM was higher for boys than for girls (77.9% vs. 65.7%) and for older than for younger adolescents (73.3% vs. 69.0%) (Table 2).
The sensitivity, specificity, positive predictive value and negative predictive value of the ID MigraineTM according to age, gender and headache type
P < 0.001;
P < 0.01.
ETTH, Episodic tension type headache; MA, migraine with aura; MoA, migraine without aura; NPV, negative predictive value; PPV, positive predictive value; Sn, sensitivity; Sp, specificity.
The ID MigraineTM was positive in 80% of subjects diagnosed with migraine. The κ value for migraine was 0.316 (P < 0.001). This showed low agreement between the ID MigraineTM screening test and headache diagnosis by the neurologist. Agreement rates between the ID MigraineTM questions and the interview are given in Table 3.
The agreement rate between the interview and the ID MigraineTM test questions according to headache type
ETTH, Episodic tension type headache; MA, migraine with aura; MoA, migraine without aura.
Discussion
This study has shown that the ID MigraineTM is a valid screening test in adolescents aged 12–17 years. If properly and widely used, such a self-administered, easy-to-perform test might have important implications for adolescent migraine.
Migraine is still an underdiagnosed and undertreated disorder (14–16). Proper diagnosis and treatment of migraine decrease the economic burden of the disease to patients and society and improve the quality of life of patients. Central sensitization may occur during migraine evolution (23). Inadequate treatment might increase the risk of central sensitization, which might complicate treatment of migraine attacks and result in chronic migraine.
Migraine is frequently seen in adolescents, and diagnosis of adolescent migraine is more difficult than that of the adult form. Studies have shown lower sensitivity and specificity for IHS criteria in adolescent and childhood than in adult migraine (19, 21, 24, 25). The mixed characteristics of adolescent migraine might also be a challenge to diagnosis (24). During the first two phases of this study, the personal observations of the authors (M.Z. and N.K.) suggested that awareness and proper diagnosis and treatment of adolescent migraine were even lower than of adult migraine (24).
The sensitivity and specificity of the ID MigraineTM were 62.1% and 71.1%, respectively. As this is the first validation study in this age group, we could not compare our results with those of a similar study, but compared our results with those of Turkish and Italian ID MigraineTM validation studies and the original study, all of which were conducted in adults (18, 22, 26).
The specificity in our study (71.1%) was similar to values found in other studies (18, 22, 26). However, the sensitivity (62.1%) was lower than reported before (18, 22, 26), a finding attributable to different methodology and age distribution of the study population. Adolescent migraine shows mixed headache characteristics such as bilaterality, short duration (< 4 h) and infrequent association with nausea and vomiting (24). These characteristics of adolescent migraine may explain the relatively lower sensitivity and specificity rates of our study. The sensitivity and specificity for IHS criteria have also shown to be lower in adolescent and childhood than in adult migraine (19, 21, 24, 25). Therefore, it is not surprising that the ID MigraineTM has lower sensitivity in the adolescence period.
This study was conducted in a student population that is a representative cohort of the 12–17-year-old population of Bursa. In this regard, this study can be accepted as population based. None of the previous ID MigraineTM studies was population based (18, 22, 26). Our eligibility criteria were not the same as those of the original study (18), which might have resulted in the inclusion of less severe headache sufferers and caused lower sensitivity and specificity rates.
During adolescence, ETTH sufferers may show migraine characteristics (24). They might therefore receive a diagnosis of migraine during this period. Our study has shown that the sensitivity and specificity of the ID MigraineTM for ETTH are significantly different from those for migraine (sensitivity 28.7%, specificity 44.7% for ETTH). Migraine risk increased four times with a positive ID MigraineTM test.
Similar to the findings of other studies (18, 22), the sensitivity was higher for girls and specificity was higher for boys in our study. We also found that the sensitivity was higher for younger adolescents (12–14-year-old group) and the specificity was higher for older adolescents (15–17-year-old group). The specificity values of the ID MigraineTM for nausea, photophobia and disability were similar to those found in the original study (18).
The κ value showed low agreement between the screening test and the expert diagnosis based on IHS criteria. The above-mentioned reasons for the relatively low sensitivity and specificity are also valid for the low κ value. Although the κ value was low, the ID MigraineTM is still a valid screening test for adolescent migraine. When the agreement rates between the physician interview and the ID MigraineTM test questions were analysed, the agreement values were found to be lowest for disability. In migraine, overall agreement rates for nausea and photophobia were nearly 85%. However, the agreement for disability was unacceptably low (39.1%). Disability might have lowered the κ value. The disability question might have not been understood by the adolescents. Therefore, disability in the ID MigraineTM test should have been described in a more understandable way.
As a different application from other studies, we gave the ID MigraineTM test to subjects after the interview. This might have caused a selection bias. However, the test was given to all 1014 interviewed subjects regardless of their diagnosis. Therefore, application of the test after the interview had no impact on the selection of the subjects.
This study was performed as part of a large epidemiological study and included more subjects than previous studies (18, 22, 26). This large sample size increased the power of our study and helped to determine the validity of the test in a population-based setting. On the other hand, the large sample size probably decreased the sensitivity, specificity and κ values. These values might be expected to be higher in a primary care setting than in a population-based setting.
In conclusion, the ID MigraineTM may be a useful screening tool for adolescent migraine when appropriate changes are made to the definition of disability in the test. This screening test might help the diagnosis of adolescent migraine in the primary care setting. It can also be used as a screening tool for large adolescent student populations. Screening of large adolescent populations for migraine might help to increase awareness of the disease and to diagnose and treat it properly in its early stages.