From the Health Economics and Outcomes Research, AstraZeneca, LP, Wilmington, DE; HealthMetrics Outcomes Research LLC, Groton, CT; Department of Health Administration, School of Business, Quinnipiac University, Hamden, CT.
The Medical Costs of Migraine and Comorbid Anxiety and Depression
Version of Record online: 4 JUN 2004
Headache: The Journal of Head and Face Pain
Volume 44, Issue 6, pages 562–570, June 2004
How to Cite
Pesa, J. and Lage, M. J. (2004), The Medical Costs of Migraine and Comorbid Anxiety and Depression. Headache: The Journal of Head and Face Pain, 44: 562–570. doi: 10.1111/j.1526-4610.2004.446004.x
- Issue online: 4 JUN 2004
- Version of Record online: 4 JUN 2004
- Accepted for publication February 6, 2004.
Objective.—To examine the direct medical costs associated with migraine, when diagnosed alone and in conjunction with anxiety and/or depression in adults and children.
Background.—Migraine is a common disorder that can often be accompanied by comorbid anxiety and/or depression. Given the prevalence of migraine and the likelihood for comorbid conditions, it is not surprising that migraine is extremely costly for society.
Methods.—Migraine cohorts were identified in a 1999-2000 database capturing inpatient, outpatient, and prescription drug services from approximately 45 large employers. Four cohorts of adults (migraine only, migraine and anxiety, migraine and depression, migraine and both conditions), and two cohorts of children (migraine only, migraine and anxiety and/or depression), were compared to respective “healthy” cohorts. t-statistics were used to capture differences in costs between the migraine cohorts and the healthy cohorts whereas ANOVA was used to test for differences in costs between subgroups of migraine sufferers.
Results.—Compared to nonmigraineurs, adults and children with migraine had significantly higher total direct medical costs in all examined categories (P < .0001) ($7,089 vs $2,923 adults; $4,272 vs $1,400 children). For adults, the presence of depression and/or anxiety along with migraine equated to significantly greater total direct medical costs when compared to their matched healthy cohorts (P < .0001) ($12,642 vs $5,179 anxiety; $11,290 vs $3,135 depression). Children with migraine and either anxiety or depression (or both) incurred an average of $9,875 in total direct medical costs as compared with only $1,165 for healthy comparators. For children and adults, the presence of comorbid anxiety or depression was associated with significantly higher medical costs when compared to migraine alone (P < .0001).
Conclusions.—This analysis quantifies the economic impact of a migraine diagnosis for both adults and children. The results of this analysis demonstrate that individuals identified as migraineurs have significantly higher medical costs than healthy comparators, with or without comorbid anxiety and/or depression. This study also suggests that clinicians should be aware that while proper treatment of migraine with effective acute and prophylactic therapy is important, attention must also be directed to comorbid conditions.
Health Maintenance Organization
Commercial Claims and Encounters
Health and Productivity Management
Food and Drug Administration
International Headache Society
Migraine is a common and disabling disorder, affecting approximately 23 million Americans annually.1 Estimates of the 1-year prevalence of migraine range from 1.2% to 3.2% in children aged 7 years,2–4 4% to 11% at age 15 years,2–5 and 7.5% to 25% in adults.6 As early as 1937, Wolff described the “migraine personality,” as characterized by a set of psychological features including anxiety, depression, and social phobias.7 Evidence from a large number of methodologically sound population studies has appeared to verify this hypothesis by supporting the association between migraine and comorbid mental health conditions.8–13 One such population-based study among young adult enrollees in a Health Maintenance Organization (HMO) found that migraineurs had an increase in lifetime rates of major depression and anxiety disorders.9 Although the relationship between migraine and anxiety and depression has not been extensively studied in children and adolescents, the body of work in this area points to a tendency for these disorders to coexist.14–16
Given the high prevalence of migraine and the likelihood for comorbid conditions, it is not surprising that migraine is extremely costly for society, with estimates of the annual total costs of migraines to be between $13 billion and $17.2 billion.17,18 These costs consist of both direct and indirect costs associated with lost work and diminished productivity. For example, research has found that migraineurs have approximately twice as many medical claims as nonmigraineurs, 2.5 times as many pharmaceutical claims, and more usage of emergency room services.19,20 Whereas there has not been any study focused specifically on resource utilization among pediatric migraineurs, headache is a common reason for visits to pediatric emergency departments.21
The purpose of this study was to examine the costs associated with a diagnosis of migraine in adults and children. In addition, since anxiety and depression disproportionately affect migraineurs, this research also examined the costs associated with dual diagnoses of migraine and anxiety or migraine and depression. Using a case-control methodology, the analyses allowed for an examination of the direct costs associated with migraine, migraine and comorbid anxiety, and migraine and comorbid depression compared to a “healthy” cohort. The study therefore provides data that will help managed care decision makers and employers gain a better understanding of the economic impact of migraine, as well as the impact of migraine accompanied by comorbid anxiety or depression.
MedStat's MarketScan Commercial Claims and Encounter (CCE) and Health and Productivity Management (HPM) databases for the years 1999 and 2000 provided the data for these analyses. These databases include private sector health data from approximately 100 payers and contain data on clinical utilization, expenditures, and enrollment across inpatient, outpatient, prescription drug, and carve-out services. The databases contain data on healthcare service use of more than 3.5 million privately insured individuals. The individuals in the database are covered under fee-for-service, fully capitated, and partially capitated health plans.
Adults who were diagnosed with migraine (based upon ICD-9 codes of 346.xx) or received a migraine medication (ergot or triptan) were included in the migraine cohort provided they were continuously insured (including outpatient prescription drug coverage) and had complete data on medical services during the calendar year of diagnosis. A total of 5997 adults met these criteria. In addition, comorbidities were identified in the migraine sample in order to gain a comprehensive view of the medical conditions for adults diagnosed with migraine. Given that research has shown that migraine tends to coexist with anxiety and/or depression,22 analyses also focused on the subgroup of individuals in the migraine cohort who were diagnosed with anxiety (based upon ICD-9 codes of 293.84, 300.00, 300.01, 300.09, 300.21, 300.22, 300.23, 300.3, or 309.81)23 or depression (based upon ICD-9 codes of 296.2, 296.3, 300.4, 309.0, 309.1, or 311).24
To examine the incremental costs associated with migraine, adults in the migraine cohort were matched to individuals without migraine. Given the emphasis on the comorbidities associated with migraine, the nonmigraine cohort was also required to have no diagnosis or medication for anxiety or depression. In addition, the nonmigraine cohort was required to have continuous insurance coverage and complete data on medical services during the year of analysis. The migraine cohort was matched to the nonmigraine, or “healthy,” cohort at a 1:1 ratio based upon age, sex, and metropolitan statistical area. Based on the match criteria, 5666 individuals diagnosed with migraine were matched to an individual in the control group, for a successful match rate of 94.48%.
Separate analyses were carried out to examine the medical costs associated with migraine among children. Given that there are currently no prescription migraine medications approved by the Food and Drug Administration (FDA) for use in children or adolescents, the migraine cohort was identified based solely on a diagnosis of migraine. Both the migraine cohort and their controls were required to have continuous insurance coverage and complete medical data during the year of analysis. There were 473 individuals aged 18 or younger with a diagnosis of migraine, complete medical data, and continuous insurance coverage during the calendar year of diagnosis, of which 445 were successfully matched by age, sex, and metropolitan statistical area, for a successful match rate of 94.08%. As with the adults, the analysis also examined the impact of a comorbid diagnosis of depression or anxiety. However, unlike the adult population, the smaller sample precluded a separation of the individuals diagnosed with depression or anxiety.
Costs are defined for these analyses as direct medical costs, including inpatient, outpatient, and prescription drug costs. Costs of outpatient prescription drugs were calculated using the average wholesale price of the drugs, while inpatient and outpatient costs were calculated as the total gross payments to all providers associated with an admission or service. When comparing the migraine cohort with the control group, t-statistics were used to compare differences between cohorts, while ANOVAs controlling for age, sex, and region were used to examine differences within migraine cohorts. Findings with a P value less than or equal to .05 were considered to indicate statistically significant differences between the groups. All analyses were conducted using SAS Version 8.1.25
Adults.— The characteristics of the 11,332 adults included in the analyses are presented in Table 1. The average age was 41 years and all individuals were of traditional working age (19-65). The population consisted of more females than males (82% females), and the majority of individuals resided in the South (39%), North Central (31%), and Northeast (17%) regions of the United States. Note that since the individuals in the migraine cohort were matched to their healthy comparators based upon age, sex, and metropolitan statistical area, these characteristics are identical in the migraine and nonmigraine cohorts.
|Variables||Mean (SD)||Min and Max|
Total medical costs for adults on the migraine cohort were significantly higher than for the control subjects ($7,089 vs $2,923; P < .0001) (Table 2). These higher costs consisted of significantly higher outpatient costs ($3,179 vs $1,603; P < .0001), inpatient costs ($942 vs $636; P= .0031), and outpatient prescription drug costs ($2,968 vs 685; P < .0001).Table 2 also displays how migraine combined with comorbid depression, anxiety, or both depression and anxiety affects medical costs. Within the adult migraine cohort, 219 individuals were classified as having anxiety, 692 individuals were classified as depressed, and 190 individuals were classified as both anxious and depressed. Similar to the results for the entire migraine cohort, individuals in the migraine cohort who were anxious, depressed, or both anxious and depressed, experienced significantly higher medical costs. These higher medical costs consisted of significantly higher outpatient costs, inpatient costs, and prescription drug costs.
|Migraine Compared to Control Group|
|Components||Mean Migraine||Mean Control Group||P Value|
|Migraine and Anxiety Cohort Compared to Control Group|
|Total medical costs||12642.0||5178.6||<.0001|
|Migraine and Depression Cohort Compared to Control Group|
|Total medical costs||11290.0||3134.8||<.0001|
|Migraine, Depression, and Anxiety Cohort Compared|
|to Control Group|
|Total medical costs||13943.0||2858.8||<.0001|
While Table 2 compares the migraine cohort to the healthy comparator group, Table 3 examines the differences within the migraine cohort. Individuals diagnosed with migraine without any comorbid anxiety or depression had significantly lower inpatient costs compared with individuals with comorbid anxiety and/or depression. However, there were no significant differences between the migraine and anxiety, migraine and depression, and migraine, anxiety, and depression cohorts when considering inpatient costs. In contrast, for outpatient costs, prescription drug costs, and total medical costs, individuals with migraine and anxiety or migraine and depression showed significant differences in costs compared to the health control group. As the number of comorbidities increased, these cost components increased significantly. For example, total medical costs for individuals with migraine and no comorbid anxiety or depression were $5,590. These costs were significantly lower than the total medical costs for individuals with migraine and either comorbid anxiety ($10,223) or depression ($10,582). Furthermore, individuals with both comorbid anxiety and depression experienced significantly higher costs still ($13,442).
|Components||Migraine||Migraine and Anxiety||Migraine and Depression||Migraine, Anxiety, and Depression||P Value|
|Total medical costs||5589.5||10223.4||10582.2||13441.7||A|
Children and Adolescents.— The study also considered the medical costs associated with a diagnosis of migraine among children and adolescents. Table 1 presents the demographic characteristics of the 890 children included in this subanalysis. The average age of these individuals was approximately 15 years and, as in the adult population, there were more females than males (63% females). The majority of the children resided in the South (38%), North Central (32%), and Northeast (18%) regions of the United States.
Table 4 presents medical costs comparing the children diagnosed with migraine to the control group. Similar to the results in the adult population, total medical costs for children in the migraine sample were significantly higher than for the control subjects ($4,272 vs $1,400; P < .0001). These higher costs for children diagnosed with migraine consisted of significantly higher outpatient costs ($2,460 vs $987; P < .0001) and prescription drug costs ($1,363 vs $206; P < .0001). However, in contrast to the adult population, there was no significant difference in inpatient costs between the migraine cohort and the healthy comparator group ($436 vs $207; P= 0.4663).
|Migraine Compared to Control Group|
|Components||Mean Migraine||Mean Control Group||P Value|
|Migraine and (Depression or Anxiety) Compared|
|to Control Group|
|Total medical costs||9875.4||1165.4||<.0001|
Table 4 also examines how migraine combined with comorbid depression and/or anxiety affects medical costs. Of the 445 children diagnosed with migraine, 60 were also identified as anxious and/or depressed. Compared to their “healthy” comparators, children with migraine and comorbid anxiety and/or depression had significantly higher total medical costs ($9,875 vs $1,165; P < .0001). These higher total medical costs consisted of higher outpatient costs ($5,045 vs $995; P < .0001) and higher prescription drug costs ($1,989 vs $162; P < .0001).
Finally, Table 5 examines differences among the cohorts. Children who were diagnosed with migraine alone had significantly lower total medical, outpatient, and inpatient costs compared to individuals diagnosed with migraine in conjunction with comorbid anxiety and/or depression. However, while children diagnosed with migraine along with anxiety and/or depression had higher prescription drug costs than individuals without such comorbidities, this difference was not statistically significant ($1,948 vs $1,095; P= .0553).
|Components||Migraine Only||Migraine and Depression or Anxiety||P Value|
|Total medical costs||1638.1||8256.5||<.0001|
The results of this study offer further support that migraine is burdensome to health care insurers and employers. Moreover, migraineurs who suffer from common mental health comorbidities exert an even greater economic impact. The high prevalence of migraine and the tendency for migraineurs also to suffer from depression and anxiety at a higher rate than the general population reinforce the need to understand better what is driving the costs of treating this patient population. To our knowledge, this study is the first actually to quantify the direct medical costs for individuals with migraine and common comorbid conditions among both adults and pediatric subjects.
Much has been published on the economic burden of migraine, both in terms of direct medical costs and indirect costs related to lost workdays and reduced performance while at work during attacks. Studies representing a range of designs all provide evidence that migraine is a costly disorder. Compared to nonmigraineurs, migraineurs are more likely to use emergency room services26–29 or to visit their physician.27,29 In addition, research has found higher prescription drug costs27 and higher direct medical costs19,26 associated with migraine. The results of this study support the literature, with adult migraineurs experiencing 2.4 times greater direct medical costs than matched controls. The category with the greatest differential was annual average prescription drug charges, with migraineurs incurring costs over four times higher than nonmigraineurs ($2,968 vs $685). Charges for outpatient services were roughly double ($3,179 vs $1,603) for migraineurs, as were charges for inpatient services ($942 vs $636).
The findings regarding direct medical costs for children diagnosed with migraine are similar to the findings in the adult population. Compared with their healthy counterparts, children and adolescents in this study who were diagnosed with migraine incurred significantly higher annual total health care costs ($4,272 vs $1,400). While there were no differences between groups with respect to inpatient treatment costs, children and adolescents in the migraine cohort had significantly higher outpatient and prescription drug costs. The finding of significantly higher outpatient costs is consistent with the Bulloch and Tenenbein (2000) analysis, which found that there was a tendency for children and adolescents with migraine attacks to seek emergency room care.21 Furthermore, laboratory tests may be ordered to rule out more severe pathology, with brain-imaging studies almost always performed in children under 6 years of age.30
In this study, prescription drug costs were found to be almost seven times higher among children suffering from migraine than among the control group. This result is similar to the finding for the adult migraine population compared to their controls, despite the fact that there is currently no FDA-approved migraine medication for children. Although there are few published data on prescription drug patterns in children with migraine, clinicians consider conventional oral, rectal, or intravenous analgesics as the basis for achieving symptomatic pain relief in their patients.30 Bearing in mind that many of these agents have not been extensively tested in clinical trials with pediatric populations,31 clinicians do rely on a wide arsenal of medications to treat their pediatric patients. Some examples of medications used in practice are acetylsalicylic acid, naproxen, acetaminophen, ibuprofen, and ergots. More severe and chronic patients may require prophylactic treatment, such as propranolol, flunarizine, sodium valproate, amitriptyline, and trazodone. Emergency relief of pediatric migraine can be attained via administration of oral dihydroergotamine (DHE) or a triptan (eg, zolmitriptan, sumatriptan).32
Despite the paucity of data on the economic impact of migraine combined with comorbid mental health conditions, many clinical and epidemiologic studies have provided strong evidence supporting their tendency to coexist.8,10,33–36 There is an abundance of data indicating that migraineurs are more likely to suffer from panic disorder,8,33,34 major depression,10,35 or any anxiety disorder.36 For example, Stewart et al (1992) conducted a population-based telephone interview survey of 10,169 young adults and found that among those who reported a headache in the previous year and sought treatment, a high proportion had a history of panic disorder.33 In a prospective study of young adults, a history of migraine was associated with a fourfold increased likelihood of major depressive disorder compared with nonmigraine controls.10 Not only were similar results found in a random sample of young adult HMO enrollees with respect to the association between migraine and affective and anxiety disorders, but also persons with migraine and mental health comorbidities were found to be three times more likely to attempt suicide than persons without migraine.36
In this analysis, individuals diagnosed with migraine and comorbid anxiety or depression were significantly more costly than individuals diagnosed with migraine alone. After adjusting for age, sex, and regional differences, adults diagnosed with migraine alone had average total direct medical costs of $5,590, while individuals diagnosed with migraine and anxiety or depression had approximately twice as high direct total medical costs, and adults diagnosed with all three conditions had average direct total medical costs of $13,442. These results are consistent with Fishmen et al (1997), who looked at health care cost data for adults enrolled in large-staff-model HMOs to estimate the impact of chronic conditions, including anxiety, depression, and migraine, on costs.37 Enrollees with one chronic health condition were twice as costly to the insurer as those without any chronic conditions, while those two or more chronic conditions were at least three times as costly.
It is apparent from our study that children with migraine and anxiety or depression are much heavier users of the health care system than children with migraine alone or healthy children. Outpatient costs were over five times higher ($5,045 vs $945) and prescription drug costs were over ten times higher among children with multiple conditions compared with matched controls. Although inpatient costs for this group were not statistically different than for controls, they were substantially higher ($2,838 vs $0), indicating that children with these multiple diagnoses were more likely to be hospitalized. These findings are consistent with evidence that these disorders have a tendency to coexist.15,38–40 For example, Braccili et al (1999) studied 73 children with diagnosed migraine (mean age 11.2 years) and administered various instruments to assess the presence of anxiety and depressive disorders. Based on the results of these tests, 53.4% of the subjects presented with depressive symptomology, 21.9% presented with anxiety-depressive symptomology, and 24.6% were negative for any of these conditions.15
These findings should be interpreted in the context of the limitations of the study design. First, the sample consisted of individuals with continuous insurance coverage for at least 1 year. These individuals may not be similar to individuals who are not employed or continuously employed for at least 1 year. Therefore, results from this study may not be generalizeable to other populations. Second, the use of diagnostic codes and the use of medications to identify adults in the migraine cohort are not as rigorous as formal diagnostic assessments for identifying individuals with migraine. Furthermore, identification of children and adolescents to be included in the migraine cohort depended entirely upon diagnostic codes. However, a migraine diagnosis in a child may follow a “looser” definition than in an adult due to fervent disagreement within the headache community on the International Headache Society's (IHS) criteria for migraine in this population.31 In addition, diagnosis of migraine in children is challenging, as other symptoms (nausea, abdominal pain, vomiting) may mask the underlying problem.41 Third, the use of medical claims data precludes the use of patient assessments, and, as a result, the analysis cannot examine quality of life, functioning, or any clinical outcomes. Finally, it is important to note that the analysis focuses exclusively on the direct medical costs and hence does not include other costs associated with migraine, such as productivity costs and caregiver burden.
While the existence of a relationship between migraine and mental health comorbidities has been well established in the literature,33,34 we know of no other study that examined the concomitant economic impact of these disorders. Using data from over 100 payers, this study quantifies the direct medical costs incurred by both children and adult migraineurs with and without anxiety and depression compared with relevant control subjects who were free from these conditions. In addition, this analysis compares the incremental costs of anxiety or depression comorbidities by comparing costs within the migraine cohort. These results therefore quantify both the total direct medical costs for individuals identified with migraine compared to a healthy cohort, as well as the incremental medical costs for migraineurs associated with comorbid anxiety and/or depression.
Acknowledgments: Financial support for this study was provided by AstraZeneca.
- 5Headache in children. In: OlesenJ, ed. Headache Classification and Epidemiology. New York : Raven Press; 1994: 273-281..
- 7Personality features and reactions of subjects with migraine. Arch Neur and Psychiatry. 1937;37: 895-921..
- 15Evaluation of anxiety and depression in childhood migraine. Eur Rev Med and Pharmacol Sci. 1999;3: 37-39., , , , , , , .
- 23A review of the diagnosis, pharmacologic treatment, and economic aspects of anxiety disorder. J Clin Psychiatry. 2001;3: 110-117., .
- 25SAS/STAT User's Guide: Version 8. NC : SAS Institute ; 1999.
- 27Pharmacoeconomic impact of injectable sumatriptan in migraine-associated healthcare costs. Am J Manage Care. 1996;2: 139-143., .