NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Archibald N, Lipscomb J, McCrory DC. Resource Utilization and Costs of Care for Treatment of Chronic Headache. Rockville (MD): Agency for Health Care Policy and Research (US); 1999 Feb. (Technical Reviews, No. 2.1.)

Cover of Resource Utilization and Costs of Care for Treatment of Chronic Headache

Resource Utilization and Costs of Care for Treatment of Chronic Headache.

Show details



Reliable, comprehensive, and generalizable information on the cost of headache has multiple important uses: for gauging the overall economic burden of headache in comparison with other problems; for use in cost-effectiveness analyses of preventive and therapeutic interventions; and for informing the development of clinical practice guidelines. Unfortunately, chronic pain syndromes such as headache are difficult to quantify in economic terms. Even with recent advances in the classification of headache syndromes, patients with chronic headaches remain difficult to diagnose, and data on the prevalence of the various types of headache are still lacking. Economic analysis of chronic headache is further complicated by the fact that many sufferers treat their headaches with over-the-counter medications and do not seek medical care.

For this report we surveyed the published literature on the cost of headache and summarized the key findings from papers reporting primary data that appeared to be both relevant and of adequate quality. Our intent was to address this key question: What are the direct medical and non-medical costs, as well as the indirect costs, associated with chronic headache care? For the purposes of this report, “chronic headache” includes migraine, tension-type headache, cluster headache, and mixed migraine and tension-type headache.


This review is based on a comprehensive search of the literature involving multiple, parallel strategies. The primary strategy involved a computerized search of the MEDLINE, HealthStar, and CINAHL databases for relevant studies published between January 1966 and December 1996, using a strategy combining the MeSH term “headache” (exploded) with terms and text words pertaining to cost and utilization. Other computerized bibliographic searches and databases, review articles textbooks, and experts were also utilized.

A total of 400 citations were identified from the primary search, and 24 citations were identified from other sources. These citations were then screened and included for further review if they were:

  1. Published in English;
  2. Full reports of original research (Le., not a letter, editorial, case report, review, or abstract);
  3. About an adult (non-pediatric) subject population; and
  4. About a subject population suffering from primary headache (Le., not traumatic, post-lumbar puncture, or other secondary headaches).

Of the 424 citations screened according to the above criteria, 127 (30%) were included for further review. We were not able to obtain a copy of one of these articles. The remaining 126 were read in full and subjected to an additional review of their content and methods. In order to be included at this stage, articles were required to:

  1. Meet the above screening criteria;
  2. Contain material relevant to the economic impact of chronic headache on medical resource utilization or work loss;
  3. Meet minimal methodological criteria related to the validity of cost or utilization measures.

Thirty-five of the 126 articles reviewed met these criteria. Data from the included articles were abstracted onto specially designed data collection forms.

The studies reviewed in this report used a variety of methods for gathering cost and utilization data. In some cases, epidemiological methods were used to assess utilization in a defined population. Other studies used before-after measures to determine the impact of an intervention on utilization or cost outcomes. While there have been numerous methodological standards published for evaluating the quality of cost-effectiveness or other economic analyses of clinical practices, we were not able to identify any widely accepted or particularly useful scales for grading the broad variety of studies included in this report; thus, we did not formally grade the quality of the evidence provided by the studies. We have, however, commented on the quality of this evidence in the text of the report and have also attempted to describe the potential biases of individual studies and the direction of these biases.


Utilization of Health Care Resources

Health Care Provider Consultation

Among people with migraine headache, health care provider consultation rates ranged between 56%–91 % in eight studies. Consultation rates for tension-type headache were generally lower, ranging from 16% to 71%.

The severity of headache appears to play a role in health care provider consultation; patients with more severe attacks are more likely to seek care from a provider, while patients who have never consulted a physician about their headaches report that their symptoms are not severe enough to warrant such a consultation and/or that they are able to treat their headaches satisfactorily with non-prescription medications.

Chronic headache sufferers who seek medical advice do not necessarily continue to consult a health care provider. In one interview survey, of 81 % of migraine sufferers who had sought care at any time previously, only 36% were receiving care at the time of the survey; similarly, of the 71 % of tension-type headache sufferers who had ever sought care, only 28% were still receiving care. Some respondents listed as their primary reasons for not continuing care the availability of non-prescription medications and the fact that they were not having headaches of sufficient frequency or severity to warrant continuing in care. More than 50%, however, said that they did not return because they were unhappy with their health care providers or had experienced negative side-effects with their medications.

As a group, patients with migraine have higher rates of health care provider consultation than does the general population. Furthermore, among migraineurs, a small number of patients have particularly high consultation rates. In one study, most migraineurs averaged 1–4 physician visits per year, while 7.6% of the subjects made more than 12 visits per year to primary care and specialty physicians for the care of their headaches.

Predictors of Receiving a Headache Diagnosis

Community-based studies suggest that a large number of people who satisfy widely accepted diagnostic criteria for migraine or tension-type headache never receive a diagnosis of their condition from a physician. Predictors associated with receiving a physician diagnosis of migraine include household income, sex, and older age. Among patients found to meet the International Headache Society (IHS) criteria for migraine, a physician diagnosis is more likely for persons whose household incomes are higher and for women.

Use of the Emergency Department (ED) for the Treatment of Migraine

Studies providing data on the use of the ED for the treatment of chronic headache show wide variation in ED utilization. While one study reported that a high percentage (48%) of migraine patients surveyed had visited the emergency department for treatment of an acute attack of migraine in the previous 12 months, other studies reviewed reported lower rates of ED use, ranging from 3% to 20% of migraineurs over periods of time ranging from 6 months to lifetime.

Prescription drug users had the highest lifetime rates of ED utilization (33% for men and 27% for women), and non-prescription drug users the lowest (one-third of the prescription drug user rates). Those migraineurs using no medications reported intermediate rates of ED utilization (20% for men and 15% for women). The investigators were not able to identify whether the high rate of ED use among prescription drug users was due to the severity of the subjects’ headaches, medication failure, or side-effects associated with their medication.

A substantial proportion of migraineurs present to the ED without having taken medication to treat their headaches. Medication failure is noted by another large proportion.


Between 6% and 11% of patients with migraine reported being hospitalized for treatment of their headaches at some point in their lifetime. Migraineurs have overall hospitalization rates nearly twice as high as those in a non-migraine comparison group.

Pharmaceutical Use

The discussion below focuses on the types of medications used by headache sufferers and the patterns with which they are used, rather than on summary measures of the economic cost of pharmaceutical use. Data on the average wholesale price of drugs are readily available from other sources.

Chronic headache sufferers have several options for the pharmacological treatment of their headaches: prescription versus non-prescription treatment and treatment for prevention versus treatment for the relief of acute episodes. Almost all headache sufferers report having used some sort of pharmaceutical agent to treat their headaches at some time. Overall rates of medication use are reported to be near 95%.

Prescription vs. Non-Prescription Drug Use

All the studies reviewed showed that non-prescription drugs are used more often than prescription drugs. Several clinical factors seem to be related to the use or non-use of prescription medications. Among patients with severe headache, those with migraine are more likely to use prescription drugs. Patients with higher headache-related disability, more frequent attacks, headache associated with vomiting, and migraine with aura were all more likely to use prescription medications. Use of prescription drugs increases with age and is higher among women than men; however, it is not related to income level or insurance coverage.

Black migraineurs had a much lower rate of prescription drug use than whites; however, blacks had a slightly higher rate of non-prescription drug use than whites.

In addition, it appears that about two-thirds of patients who try prescription medications do not continue to use them, citing the availability of effective non-prescription drugs, unwanted side-effects, and inadequate relief among reasons for discontinuing prescription drugs.

Use of Preventive Medications

Patients with recurrent headache may be treated with acute or preventive medication or with some combination of the two. It appears from four studies, however, that preventive drug treatments are used by only 3%–13% of migraineurs.

Frequency of Medication Use for Treatment of Acute Headache

Patterns of use of medication for acute headache episodes vary widely from patient to patient. In several studies, a small minority of headache sufferers were particularly heavy users. Factors associated with heavy use included increasing age and higher frequency of headaches.

Effect of Sumatriptan on Drug and Health Care Resources Utilization

Medication use can have an impact on other health service utilization. Two studies found similar decreases in health care provider (physician and emergency department) utilization in migraineurs after the introduction of the anti-migraine drug sumatriptan. Pharmaceutical costs were increased significantly in one study, but not available in the other. The routine use of sumatriptan in clinical practice may have a significant impact on practice patterns and the utilization of health care resources, but the precise nature and magnitude of this impact are as yet uncertain.

Non-Pharmacological Therapy

Several studies have attempted to estimate the economic impact of behavioral therapy for chronic headache. One study described a very large reduction in self-reported medical costs among patients undergoing relaxation and biofeedback training. Costs considered included medication, visits to family physicians and neurologists, hospital costs, and costs for alternative treatments such as acupuncture and chiropractic. Other studies have described the duration of patient-therapist contact as a proxy for costs in comparisons of home- versus clinic-based training in relaxation and biofeedback or relaxation and/or cognitive therapy. Because no differences have been observed in the efficacy of the less intensive (and less costly) home-based training and more traditional office-based treatment, researchers have concluded that the home-based treatments may be more cost-effective.

Work Loss

Unlike the costs described in previous sections which result in direct expenditures, work loss due to headache is an indirect cost. A dollar figure for lost productivity can be estimated from rates of absenteeism or reduced effectiveness.

Studies describing rates of absenteeism have been conducted in the workplace, community, and clinical settings. Rates of absenteeism due to headache of 2 to 4.3 days per year were reported in employee and community studies, while higher rates, ranging from 14.3 to 61 days per year, were observed in two studies conducted in clinical settings. These average figures obscure the fact that while many to most patients miss no days of work or school, a relatively small proportion of subjects accounts for many missed days.

While headache does cause a large number of absences from work, it appears that many headache sufferers stay on the job, functioning with significantly reduced effectiveness, when they have a headache. Most subjects reported a slight to moderate effect of migraine on their work; men and women experience similar reductions in effectiveness with their headaches. Studies have found no differences in the frequency of missed work days by race, but black women and men were less likely to report reduced efficiency at work than either whites or Asian-Americans.

There have been few efforts to estimate the economic impact of this lost productivity. The cost of errors or inaccuracies in work performed during a headache attack may never be known.

Future Research

Most studies of the cost of care for headache, and of the clinical management of headache, pertain to migraine; the costs of tension-type, cluster, and other chronic primary headache disorders remain to be estimated. Low rates of recognition and diagnosis of chronic headache conditions and the large proportion of chronic headache sufferers who receive no treatment or who self-treat with over-the-counter medications make it difficult to estimate the scope of chronic headache in economic or human terms. Epidemiological data collection methods have identified and begun to quantify this problem. In clinical and administrative data, ascertainment of chronic headache is also difficult, as there is evidence of under-diagnosis, and, furthermore, administrative coding methods do not correspond with current headache disease classification systems. Until these and other problems can be resolved, it will remain difficult to estimate the aggregate economic burden of headache or to estimate the costs of headache for cost-effectiveness analysis or cost-benefit analysis.

Large gaps exist in the literature. We do not, for example, know the average number of health care provider visits it takes to reach a diagnosis of chronic headache, how many and what kinds of tests are performed to make a diagnosis, what kinds of treatment options patients are offered, or how many medications are tried before relief is obtained or the patient gives up treatment. We need to better understand why patients with chronic headache are admitted to the hospital and what is done for them while they are there. We also do not know all there is to know about the impact of chronic headache on work—whether chronic headache prevents people from obtaining or retaining jobs or has an effect on their ability to advance in their job. We must also determine from whose perspective the economic impact of headache should be viewed. Insurers, for example, may not have an interest in work loss or absenteeism. Employers may be concerned about medical costs and work loss but not the impact on family life or social role functioning.

Finally, several studies have documented the magnitude of work loss and reduced productivity due to headache. While the mean number of days absent from work annually appears to vary roughly from about 2 to 14 for chronic headache sufferers, there has been little systematic effort to translate these missed days into corresponding estimates of economic loss from either the societal or individual perspective. Also requiring much more study is the impact on productivity of headaches that impair worker performance, but do not lead to work loss days.

We find that each paper reviewed here provides some useful information about the economic cost of headache. But no study attempts to estimate the total cost for a nationally representative population of sufferers. To produce reliable, comprehensive, and generalizable estimates of the economic cost of headache, good quality resource utilization information from multiple, complementary data sources almost certainly will be required. An analytical framework for identifying the required data elements and carrying out the calculations can be built around standard economic models of disease cost.


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...