A Review of Screening Tools for Psychiatric Comorbidity in Headache Patients


  • Morris Maizels MD,

  • Todd A. Smitherman PhD,

  • Donald B. Penzien PhD

  • From the Department of Family Medicine, Kaiser Permanente, Woodland Hills, CA (Dr. Maizels); and Head Pain Center, Department of Psychiatry and Human Behavior, University of Mississippi Medical Center, Jackson, MS (Drs. Smitherman and Penzien).

Address all correspondence to Morris Maizels, MD, Department of Family Medicine, Kaiser Permanente, 5601 De Soto Avenue, Woodland Hills, CA 91365.


Psychiatric comorbidity, especially depression and anxiety, has been well documented in patients with primary headache disorders. The presence of psychiatric comorbidity in headache patients is associated with decreased quality-of-life, poorer prognosis, chronification of disease, poorer response to treatment, and increased medical costs. Despite the prevalence and impact, screening for psychiatric disorders in headache patients is not systematically performed, either clinically or in research studies, and there are no guidelines to suggest which patients should be screened or in what manner. We review a variety of screening methods and instruments, focusing primarily on self-report measures and those available in the public domain. Informal verbal screening may be sufficient in a primary care setting, but should include screening for both anxiety and depression. Explicit screening for anxiety is important, as anxiety may have a more significant impact on headache than does depression and may occur in the absence of clinical depression. Formal screening with instruments that can identify a variety of psychiatric disorders is appropriate for patients with daily headache syndromes, patients who are refractory to usual care, and patients referred for specialty evaluation. Limitations of screening instruments include the influence of transdiagnostic symptoms and the need for confirmatory diagnostic interview. The following instruments appear most suitable for use in headache patients: for depression, the Patient Health Questionnaire Depression Module, the Beck Depression Inventory-II, or the Beck Depression Inventory-Primary Care; for anxiety, the Beck Anxiety Inventory and the Generalized Anxiety Disorder 7-item Scale; and for multidimensional psychiatric screening, the Patient Health Questionnaire or Primary Care Evaluation of Mental Disorders.


panic disorder


generalized anxiety disorder


obsessive-compulsive disorder


Diagnostic and Statistical Manual of Mental Disorders-4th ed.


Center for Epidemiological Studies Depression Scale


Beck Depression Inventory


Beck Depression Inventory-Primary Care


Patient Health Questionnaire Depression Module


Beck Anxiety Inventory


State-Trait Anxiety Inventory


Generalized Anxiety Disorder 7-item Scale


Primary Care Evaluation of Mental Disorders


Patient Health Questionnaire


Brief Symptom Inventory


Psychiatric Diagnostic Screening Questionnaire


Pain Patient Profile


Diagnostic Interview Schedule


Composite International Diagnostic Instrument


Mini-International Neuropsychiatric Interview


Epidemiological research indicates that headache patients are at much greater risk for comorbid psychiatric conditions compared to patients who do not experience recurrent headache. When compared to those with nonheadache, migraineurs are between 4 and 5 times more likely to suffer from major depression, dysthymia, and bipolar disorders; 3 to 10 times more likely to suffer from panic disorder (PD); 4 to 5 times more likely to suffer from generalized anxiety disorder (GAD); and 5 times more likely to suffer from obsessive-compulsive disorder (OCD).1,2 Even when controlling for depression, migraineurs attempt suicide at a higher rate than individuals without migraine.1 Migraineurs are also twice as likely to suffer from illicit drug abuse/dependence, alcohol abuse/dependence, and nicotine dependence.1,2 The lifetime prevalence rates of various psychiatric disorders among migraine patients are as follows: major depression (34%), dysthymia (9%), bipolar II (4%), PD (11%), GAD (10%), OCD (9%), and phobias (40%).3 Increased rates of psychiatric disorders have been observed in individuals suffering from chronic tension-type headache as well.4–6

Though more research is needed in these areas, increased prevalence of psychiatric disorders in headache populations is associated with chronification of headache, increased medical costs, and poorer prognosis for treatment. In terms of chronification of headache, at least one study has found that the majority (57%) of young adults with 2 or more psychiatric diagnoses showed no improvement or deterioration in headache at 8-year follow-up; 14% of these youth were headache free after 8 years.4 By comparison, of the youth with no psychiatric disorder or only one comorbid diagnosis at initial evaluation, no more than 15% had failed to improve and approximately 40% were headache free. There is also limited evidence to suggest that psychiatric comorbidity may play a significant role in the transformation of migraine to medication overuse headache6 and in the progression of headache frequency (eg, from episodic to chronic and daily).7,8 Specifically, chronic migraine patients endorse higher levels of depression and anxiety than do episodic migraine patients,9 and chronic nonmigrainous headache (eg, tension-type or cluster headache) patients endorse higher levels of depression and anxiety than do episodic nonmigraineurs.10

Migraineurs with a comorbid anxiety or depressive disorder accrue on average $4,000 to $5,500 more in medical costs each year compared with migraineurs without comorbid psychiatric diagnoses.11 Migraineurs with anxiety and depression also report lower satisfaction with headache treatment, increased headache-related disability, and reduced quality-of-life than do migraineurs with neither comorbid condition.12

An emerging body of literature suggests that psychiatric comorbidity also portends a poorer prognosis for response to headache treatment.2,13 Despite this evidence, there is a relative paucity of research investigating the effects of treating psychiatric disorders on headache or the converse, on the impact of treating headache on psychiatric disorders. Headache preventive studies that have assessed depression have produced mixed results. Some have found low-to-moderate correlations between depression and headache outcome, while most have not.14 Treatment studies assessing anxiety disorders are even more limited and prohibit us from drawing meaningful conclusions in this regard.

The literature reviewed above underscores the importance of early and accurate identification of comorbid psychiatric disorders in headache patients. Table 1 suggests additional reasons for screening based primarily on clinical experience. Table 2 considers potential pitfalls in screening. In this article, we review specific psychiatric screening tools and considerations unique to screening headache patients for psychiatric comorbidity, focusing predominantly on screening for depression and anxiety in adults presenting in primary care settings. We concentrate primarily on self-report screening tools, and wherever possible, on those that are available in the public domain. However, we also describe more comprehensive measures that may be used to confirm the presence of a diagnosis after a positive screen has been obtained. We close by providing clinical recommendations regarding screening for psychiatric comorbidity in headache patients.

Table 1.—.  Rationale for Psychiatric Screening in Headache Patients
1. The presence of anxiety and/or depression significantly impacts headache prognosis and satisfaction with headache treatment, and is associated with increased headache-related disability.
2. Anxiety and/or depression may yield differential response to headache prophylaxis. Anxiety and/or depression may suggest the preferential use of psychotropics to treat the comorbid disorder(s).
3. Anxiety and/or depression may influence compliance with medication and behavioral treatment, as well as the tendency to experience and report medication side effects.
4. Anxiety and/or depression have significant impact on quality-of-life and health care utilization, regardless of their impact on headache.
5. The use of antidepressants may trigger mania in a patient with unrecognized bipolar disorder.
6. Patients with bipolar disorder and/or a history of chemical dependency may have a tendency to medication overuse or drug-seeking behavior.
7. The recognition of psychiatric comorbidity may be a key component in developing a therapeutic doctor-patient relationship.
8. The use of screening tools may improve the patient's recognition of, and attention to, relevant psychiatric factors.
9. Screening tools may be useful in excluding a suspected psychiatric disorder whose presentation suggests a psychiatric basis for somatic complaints.
Table 2.—.  Possible Negative Consequences of Psychiatric Screening in Headache Patients
1. Patients may feel stigmatized.
2. Recognition of a psychiatric disorder may dissuade the clinician from adequately addressing the headache disorder.
3. Excessive costs of time and money (eg, purchasing screening measures) for patient and clinician.
4. Identification of psychiatric disorders may be harmful if appropriate follow-up treatment is not provided.
5. Clinicians may incorrectly diagnose a psychiatric disorder based on a positive screen, without appropriate confirmation.
6. Clinicians may prescribe unnecessary medications based on an unconfirmed positive screen.


When screening for psychiatric disorders in headache patients, several issues must be taken into consideration. Selection of an appropriate screening measure is of primary concern, as many different types of screening measures exist, varying in breadth and depth of coverage and in format. Disorder-specific measures are used to screen for particular disorders or obtain more information about the nature and severity of particular symptoms. Multidimensional screening measures emphasize breadth of coverage and are useful as initial screening measures, in cases where multiple psychiatric problems are of concern, or when the psychiatric presentation is unclear or complex.

An additional consideration is the role of transdiagnostic symptoms, or those symptoms that are characteristic of both headache and a psychiatric disorder. For example, sleep disturbances, fatigue, and difficulty concentrating are frequently observed in patients with depression or chronic headache. Dizziness, nausea, and paresthesias are common complaints of patients suffering with PD or headache. Likewise, irritability, muscle tension, and insomnia present in both headache and GAD patients.

The presence of transdiagnostic symptoms may result in a false positive screen, further complicating diagnosis and treatment. Many patients with a false positive screen for one psychiatric disorder meet criteria for another psychiatric disorder.15 Using a series of structured interviews based on the Diagnostic and Statistical Manual of Mental Disorders-4th ed. (DSM-IV)16 criteria, Leon et al found that a large proportion of primary care patients with a false positive screen for anxiety or depression met criteria for another anxiety or depressive disorder.


In a clinical setting, appropriate psychological inquiries may be inhibited by the difficulty of redirecting a patient's somatic preoccupation to a psychiatric consideration. In this context, we have found that asking about sleep and energy allows a comfortable transition to inquiring about mood (ie, “How is your sleep? How is your energy? How are your moods?”). The question “What do you do for fun?” is also useful in detecting anhedonia and at times may be elaborated upon to obtain information about social support, relationships, and other correlates of depression.

Direct Verbal Screening.— Two questions (“Over the past 2 weeks, have you felt down, depressed, or hopeless?” and “Over the past 2 weeks, have you felt little interest or pleasure in doing things?”) may be as effective as longer instruments to detect depression.17 The single question “Have you felt sad or depressed much of the time in the past year?” had sensitivity comparable to the Center for Epidemiological Studies Depression Screen (CES-D) (85% vs 88%) but was slightly less specific (66% vs 75%).18 No similar single question has been suggested for anxiety, but we have found the question “Are you a worrier?” to be a useful screening question for anxiety disorders that feature prominent cognitive components (eg, GAD) and as a lead-in to other, more specific questions about different anxiety symptoms (eg, panic attacks, obsessions/compulsions).

Other Clinical Cues.— Emotional responses to patients are important clues to patients' psychological states. Patients with depression or anxiety may trigger similar feelings in their clinician. Patients with obsessive concern about somatic symptoms or medication side effects may be displaying the chronic worry symptoms associated with GAD, contamination fears characteristic of some OCD patients, or symptoms of a somatoform disorder. A useful question is “Do you worry like this about things other than your health?” Clues to personality disorders include: particularly odd patient behavior, patients who arouse particularly strong feelings in the medical staff, patients who create discord among medical personnel, and patients who demand special treatment. However, it is our clinical experience that untreated pain may bring out behaviors that appear pathological but revert to normal when the pain is controlled.


Depression.— The US Preventive Services Task Force recently published guidelines for screening for depression.19 They concluded that questionnaire screening improves identification of adult primary care patients who are depressed. The Task Force concluded that more research was needed on the utility of screening in children and adolescents. They recommended that in order for screening to influence outcomes, there should be sufficient systems and resources to ensure accurate diagnosis, effective treatment, and appropriate follow-up. They also recommended “full diagnostic interviews” to confirm a positive screen.

In terms of specific screening instruments, Mulrow et al reviewed 9 case-finding instruments for depression suitable for primary care clinicians, and found no significant differences, with sensitivities for detecting major depression ranging from 67% to 99% and specificities from 40% to 95%.20 Certainly the most common and well-researched instrument used in the assessment of depression is the Beck Depression Inventory-II (BDI-II).21 The BDI-II consists of 21 groups of symptoms that respondents rate from 0 to 3, representing the severity of each symptom that they have experienced over the past 2 weeks. Symptom clusters include cognitive symptoms (eg, pessimism, self-dislike, anhedonia, suicidal ideation) and physical symptoms (eg, fatigue, changes in appetite or sleeping patterns), thus covering a broad range of depressive symptomatology. The BDI-Primary Care (BDI-PC), also known as the BDI-FastScreen, contains 7 items from the BDI-II, all of a psychological nature (sadness, pessimism, past failure, self-dislike, self-criticalness, suicidal thoughts, loss of interest).22,23 Another measure designed for use in primary care is the Patient Health Questionnaire Depression Module (PHQ-9),24 which is a subsection of the PHQ (described below). The PHQ-9 is based on the 9 DSM-IV criteria for major depressive disorder: 2 sentinel questions for depression (loss of interest and depressed mood), 4 additional psychological questions (including suicidal ideation), and 3 somatic items (sleep, energy, and psychomotor retardation or restlessness). Sensitivity and specificity of this measure for major depression have both been reported at 88%, based on a cutoff score of 10.24 Nease and Malouin concluded that PHQ-9, based on its brevity and high predictive value, is the best available screening tool for depression for use by primary care clinicians.25 The PHQ-2, a 2-item version similar to the PHQ-9 that inquires about the frequency of depressed mood and anhedonia over the past 2 weeks, is able to detect major depression with a sensitivity of 83% and a specificity of 92% (using a cutoff score of 3).26

Other self-report measures for depression include the Center for Epidemiological Studies Depression Scale (CES-D)27 and the Zung Self-Rating Depression Scale.28 The 20-item CES-D has demonstrated strong reliability and validity when used to assess depressive symptomatology in cancer patients29 and is used commonly with other medical populations. While it has good sensitivity, studies indicate that the CES-D has inadequate specificity and often results in a large number of false positive screens.30 The Zung Self-Rating Depression Scale includes 20 questions about the symptoms of depression as well as functional questions not included in the other screeners. Table 3 details the major screening measures for depression.

Table 3.—.  Measures for Assessing Mood Disorders and Depression in Headache Patients
Psychological Disorder/MeasureContent AreaTime to Complete
  1. Note: The interested reader is referred to Antony et al39 for descriptions and reprints of common measures used to assess social and specific phobias, as well as reprints of some of the disorder-specific anxiety measures described in this Table.

 Beck Depression Inventory-II (BDI-II)2121 cognitive and physiological symptoms of depression2-5 minutes
 BDI-Primary Care (BDI-PC) or BDI-FastScreen227 cognitive symptoms of depression2 minutes
 Patient Health Questionnaire Depression Module (PHQ-9)249 symptoms of depression (DSM-IV criteria)2 minutes
 PHQ-2262 core aspects of depression (depressed mood, anhedonia)<2 minutes
 Center for Epidemiological Studies Depression Scale (CES-D)2720 symptoms of depression2-5 minutes
 Zung Self-Rating Depression Scale2820 symptoms of depression, including functional impairment2-5 minutes
 Hospital Anxiety and Depression Scale (HADS)317 cognitive symptoms of depression and 7 of anxiety2-5 minutes
Bipolar symptoms
 Mood Disorder Questionnaire (MDQ)3313 items related to manic symptoms2-5 minutes
Broad anxiety symptoms (not disorder-specific)
 Beck Anxiety Inventory (BAI)3721 symptoms of anxiety (primarily panic/somatic symptoms)2-5 minutes
 State-Trait Anxiety Inventory (STAI)3820 symptoms of state anxiety and 20 symptoms of trait anxiety2-5 minutes
Specific anxiety disorders
Panic disorder (PD)
 Anxiety Sensitivity Index (ASI)41Fear of bodily sensations of arousal2-5 minutes
 Albany Panic and Phobia Questionnaire (APPQ)42Fear and avoidance of anxiety sensations2-5 minutes
Generalized anxiety disorder (GAD)
 Generalized Anxiety Disorder 7-item scale (GAD-7)43Symptoms of GAD2 minutes
 Generalized Anxiety Disorder Questionnaire-IV (GAD-Q-IV)44Diagnostic criteria and major themes of worry2-5 minutes
 Penn State Worry Questionnaire (PSWQ)45Severity and frequency of worry2-5 minutes
Obsessive-compulsive disorder (OCD)
 Maudsley Obsessional Compulsive Inventory (MOCI)46Prevalence of obsessions/compulsions related to certain themes5 minutes
 Padua Inventory47Obsessive-compulsive themes5 minutes
 Yale-Brown Obsessive Compulsive Scale (Y-BOCS)48“Gold standard” semistructured interview30-40 minutes
Posttraumatic stress disorder (PTSD)
 PTSD Checklist49 (PCL)Symptoms/diagnostic criteria of PTSD2-5 minutes
 Clinician-Administered PTSD Scale (CAPS)50“Gold standard” clinician-administered rating scale45-60 minutes

With respect to screening headache patients, the inclusion of somatic items (eg, BDI-II) may give a more comprehensive evaluation of the patient's overall condition, whereas a purely cognitive/emotive measure (BDI-PC) eliminates concern for the influence of overlapping somatic symptoms. Two depression-screening tools specifically address the problem of transdiagnostic symptoms by excluding somatic content, as most of the depression instruments previously described overlap significantly with anxiety. All 7 items on the BDI-PC refer to cognitive/psychological symptoms of depression. The Hospital Anxiety and Depression Scale (HADS) contains 7 items each for depression and anxiety, all of a cognitive/psychological rather than physical nature.31 The HADS is especially useful for headache patients because it has questions unique for anxiety and depression and eliminates somatic content.

Bipolar Disorder.— In a primary care population, 10% of patients screened positive for lifetime prevalence of a bipolar spectrum disorder; two-thirds of these had a current major depressive disorder, anxiety disorder, or substance use disorder.32 It is particularly important to screen headache patients for bipolar disorder, as antidepressants may trigger mania in a bipolar patient not on a mood stabilizer. Further, certain antiepileptic drugs may treat both migraine and bipolar disorder. Recognition of bipolar traits may also help to understand impulsive patient behaviors, including a tendency to uncontrolled medication use. The Mood Disorder Questionnaire (MDQ) is a brief and easy-to-use self-report inventory with 13 yes/no items. A cutoff point of 7 positive items has a sensitivity of 73% and specificity of 90%.33 Because bipolar patients may minimize their history of mania or hypomania, asking about energy and sleep may be helpful in eliciting a more accurate history of mood fluctuations (eg, “Have you ever had so much energy that you felt like you didn't need to sleep much at all?”). Confirming questionnaire and interview data with significant others may also be helpful in obtaining an accurate history of manic and hypomanic episodes (asking about spending sprees, irritability, impulsive decisions, need for little sleep, etc.).

Anxiety.— Compared to studies on assessment of depression in medical settings, assessment of comorbid anxiety has been given considerably less attention. Anxiety disorders are the most prevalent class of psychiatric disorder in the general population,34 and individuals with anxiety disorders present at primary care facilities as often as they do at mental health facilities.35 Problematically, clinicians often attribute the physical symptoms of anxiety to a medical condition rather than considering the presence of a specific anxiety disorder.36

Anxiety screening measures are numerous and can be broken down into those that focus on anxiety more generally and those that address specific anxiety subtypes (disorders). Two of the most-commonly used broad anxiety measures include the Beck Anxiety Inventory (BAI)37 and the State-Trait Anxiety Inventory (STAI).38 These measures have adequate psychometric properties39 and aim to identify the presence of anxiety symptoms rather than specific anxiety disorders. The BAI is one of the most frequently used and thoroughly studied anxiety measures for adults. The respondent rates 21 symptoms on a 0 to 4 Likert scale. Because the BAI was developed to assess symptoms of anxiety that are relatively distinct from depression, it is heavily loaded with somatic items that emphasize panic. The STAI consists of 2 20-item Likert-scale measures, one inquiring about current symptoms (state version) and the other inquiring about more longstanding symptoms (trait version). Some items on each version are reverse scored. One problem with the STAI is that it has repeatedly been found to correlate as much (or more) with depression as it does with anxiety.40 A patient who screens positive on a broad-based anxiety measure should be evaluated for specific anxiety diagnoses.

Disorder-specific anxiety measures focus on identifying the frequency and severity of symptoms that are characteristic of a particular anxiety disorder. Space does not allow a detailed review of the numerous disorder-specific screening instruments germane to the major anxiety disorders. Table 3 presents some of the most well-researched measures to assess the core features of the major anxiety disorders. Although these measures are not commonly used as routine screening instruments in primary care settings, they are particularly valuable when the clinician desires more specific information about a particular anxiety disorder (eg, to confirm a positive screen from a multidimensional measure) or in selecting participants for research studies on anxiety disorders. The interested reader is referred to the original citations to find out more about these valuable screening tools.

Somatoform Disorders.— Somatization refers to a physical symptom that is “unexplained” by a medical disorder and is thought to be the somatic expression of psychological distress. Somatization disorders may be associated with anxiety or depression or present independently of those diagnoses. The designation of somatization in headache patients is problematic because many headache patients experience numerous somatic symptoms and there is no objective manner to determine whether they are “unexplained.” Somatic counts have been suggested as a surrogate marker for somatoform disorder. Kroenke et al suggest that a threshold of 7 symptoms on the 15-item checklist of the PRIME-MD PQ should trigger screening for somatoform disorder, with a positive predictive value of 25%.51 A study of the PRIME-MD (described below) in a headache clinic population found that patients with chronic migraine and chronic daily headache syndromes had a mean somatic count of 6.5 and 7.3, respectively.52 Somatic counts increased with the frequency of severe headaches as well as with associated psychiatric disorders, suggesting that somatic prevalence might represent the spectrum of migraine disorder rather than a somatoform disorder.


The Primary Care Evaluation of Mental Disorders (PRIME-MD) is a multidimensional screening tool with 2 components: a 1-page self-report Patient Questionnaire (PQ) and a 12-page Clinician Evaluation Guide (CEG).53 The CEG provides a structured interview that the clinician uses to evaluate positive screening items from the PQ. The PHQ was developed from the original PRIME-MD in an effort to reduce clinician time required for evaluating positive answers on the PQ.54 The PHQ is an entirely self-reported version of the PQ and CEG, requiring less than 3 min of clinician time for 85% of patients. The PHQ has demonstrated similar sensitivity and specificity as the PRIME-MD,54 and is used frequently in primary care settings and clinical research studies due to its breadth of coverage and ease of administration and scoring. Like the PRIME-MD, the PHQ assesses 5 categories of disorders: depression, anxiety, somatoform, alcohol-use, and eating.

Other multidimensional measures include the Brief Symptom Inventory (BSI),55 the Psychiatric Diagnostic Screening Questionnaire (PDSQ),56,57 and the Pain Patient Profile (P-3).58 The BSI is a shortened version of the Symptom Checklist-90-Revised59,60 that assesses 9 domains: somatization, obsessive-compulsive, interpersonal sensitivity, depression, anxiety, hostility, phobic anxiety, paranoid ideation, and psychoticism. The PDSQ screens for 13 DSM-IV Axis I disorders. The P-3, a commercially available instrument from NCS, was designed to screen for psychiatric problems in pain patients; it contains subscales related to depression, anxiety, and somatization. The score for pain patients for depression was nearly 3 standard deviations above the mean for the community sample, 2 standard deviations for anxiety, and almost 4 standard deviations for somatization.58

The following semistructured and structured interviews, while requiring training and intended for experienced (or well-trained lay) clinicians, are included because they represent “gold standards” for comprehensive diagnosis of psychiatric disorders. The Structured Clinical Interview for DSM-IV (SCID-I) is a semistructured interview for diagnosis of mood, psychotic, substance abuse, anxiety, somatoform, eating, and adjustment disorders.61 The Diagnostic Interview Schedule-IV (DIS-IV)62 and the Composite International Diagnostic Instrument (CIDI)63 are fully structured diagnostic interviews, requiring up to 90 minutes to 2 hours to administer. The DIS includes 19 diagnostic modules, while the CIDI covers 11 diagnostic modules. The Mini-International Neuropsychiatric Interview (M.I.N.I.)64 evolved from the CIDI and uses screening questions for each of the psychiatric disorders to make screening more efficient yet comprehensive. It includes the core screening questions for the major DSM-IV Axis I diagnoses, as well as a follow-up module for each individual disorder. We have found the tool useful when trying to determine the presence of specific but less common axis I disorders.

The Minnesota Multiphasic Personality Inventory-II (MMPI-II) is a well-known multidimensional psychological instrument.65,66 Like other personality measures, the MMPI-II is time-intensive (60-90 minutes), expensive, and difficult to interpret without prior training, and thus is not a recommended tool in primary care settings. However, it may be useful in identifying personality styles that might influence treatment of comorbid psychiatric conditions and/or headache, including exaggeration (malingering) and minimization of psychopathology, tendency toward adopting “the sick role,” adoption of stereotypically male or female qualities, and introversion/extroversion.


The impact of anxiety and depression on migraine was highlighted by a recent landmark study of the French population by Lantéri-Minet et al.12 Using a low threshold on the Hospital Anxiety and Depression Scale (HADS), they found that 51% of subjects with active migraine were anxious or depressed, but importantly 28% had anxiety alone. Depression alone was not found to be more frequent among migraine than non-migraine subjects in this survey, and screening for depression alone would have missed over half of those with significant psychiatric comorbidity. These findings suggest that screening for depression alone is not sufficient, and that screening for anxiety must accompany screening for depression. In this study, coexistence of depression with anxiety was associated with more pronounced migraine-related disability and reduced health-related quality-of-life. Importantly, both perceived treatment efficacy and satisfaction with treatment were lower in subjects with anxiety (either alone or combined with depression). Fewer subjects with combined anxiety and depression used only a single medication to treat a migraine attack, possibly of significance in understanding the development of medication-overuse headache.

Despite the well-documented impact of psychiatric comorbidity on migraine, identification and treatment of psychiatric disorders in this population has not been systematically addressed by clinicians or researchers. Many screening tools with suitable operating characteristics are available which are clinically efficient to use. We recommend that all patients with migraine be screened for anxiety and depression. Simple verbal screening (especially about depressed mood, loss of interest/pleasure, and anxiety symptoms such as panic and chronic worry) may be sufficient at a primary care level, but more formalized and multidimensional screening is recommended at a referral level and in headache specialty clinics. Patients who are poorly responsive to usual care also merit formalized psychiatric screening. Similarly, all research evaluations of acute and preventive medications should incorporate documented screening for anxiety and depression. Table 4 summarizes our recommendations for psychiatric screening in headache patients.

Table 4.—.  Recommendations for Screening for Psychiatric Comorbidity in Headache Patients
Primary care settings
 1. Patients with infrequent or nondisabling headache in primary care should receive at least brief verbal screening for anxiety, depression, and other lifestyle issues. Positive replies should receive appropriate clinical evaluation.
 2. All patients with chronic or daily headache syndromes, and patients poorly responsive to standard headache management, should be formally screened for psychiatric disorders (at a minimum for anxiety and depression).
Headache specialty settings
 3. Patients referred for specialty evaluation of headache should have formal screening for depression and anxiety at a minimum (but preferably more comprehensive multidimensional psychiatric screening).
General guidelines
 4. No single instrument can be recommended as most suitable for screening headache patients. Those screening tools validated on medical populations are likely to be most appropriate for use with headache patients.
 5. Interpretation of scores on screening tools should consider the common prevalence of somatic items (eg, fatigue and insomnia) in headache and pain disorders, even in the absence of psychiatric disorders. When attempting to diagnose anxiety or depression, somatic items common to pain patients may need to be considered separately from cognitive/affective items.
 6. Referral for psychiatric evaluation and/or treatment is warranted in cases where psychiatric comorbidity is suspected but resources are not in place to provide sufficient assessment, treatment, and follow-up services.
 7. Diagnoses of anxiety and/or depression based on screening tools should ensure that the major diagnostic criteria for the disorder are satisfied (based on DSM-IV criteria published by the American Psychiatric Association61). This can often be accomplished by confirming positive screens by administration of clinical interviews.

Because of ease of use and interpretation, the following instruments appear most suitable for use in headache patients: for depression, the PHQ-9, the BDI-II, or the Beck Depression Inventory-Primary Care (BDI-PC); for anxiety, the Beck Anxiety Inventory (BAI) to assess somatic symptoms related to panic and the Generalized Anxiety Disorder 7-item Scale (GAD-7) to assess cognitive symptoms related to worry; and for multidimensional psychiatric screening, the PHQ or Primary Care Evaluation of Mental Disorders (PRIME-MD). Where comprehensive multidimensional psychiatric assessment is required (as in research or epidemiologic surveys), the M.I.N.I. accomplishes both screening and diagnosis. Table 5 provides a listing of these recommended measures as well as Internet links through which they may be obtained.

Table 5.—.  Recommended Instruments for Psychiatric Screening in Headache Patients
 • Beck Depression Inventory-II (BDI-II)21
  ○ Purchase online at http://www.psychcorp.com
 • Beck Depression Inventory-Primary Care (BDI-PC)22–23
  ○ Purchase online at http://www.psychcorp.com (listed as BDI-FastScreen)
 • Patient Health Questionnaire Depression Module (PHQ-9)24
  ○ Available online at no cost for clinical and research purposes at http://www.pfizer.com/pfizer/phq-9/index.jsp
Bipolar symptoms
 • Mood Disorder Questionnaire33
  ○ Reprinted in the original article and also available online at no cost at http://www.bipolar.com/mdq.htm
 • Beck Anxiety Inventory37 (BAI)
  ○ Purchase online at http://www.psychcorp.com
 • GAD-7 Generalized Anxiety Disorder 7-item Scale43
  ○ Obtained by contacting Dr. Kurt Kroenke, coauthor of the GAD-7, at kkroenke@regenstrief.org
Multidimensional measures
 • Patient Health Questionnaire (PHQ)54
 • Primary Care Evaluation of Mental Disorders (PRIME-MD)53
  ○ Both the PHQ and PRIME-MD assess symptoms related to depressive, anxiety, somatoform, alcohol, and eating disorders. Both may be obtained by contacting Dr. Kurt Kroenke, coauthor of these measures, at kkroenke@regenstrief.org
 • Mini-International Neuropsychiatric Interview (M.I.N.I.)64
  ○ Structured interview assessing major Axis I disorders. Also includes brief self-report screen (M.I.N.I. Screen) that may be administered prior to the relevant M.I.N.I. modules. Instructions for obtaining this measure at no cost for clinical and research purposes at http://medical-outcomes.com/Downloads/index.htm

Limitations of Screening.— Clinicians who use psychiatric screening instruments should be aware of their limitations. The problem of transdiagnostic symptoms has been previously mentioned. Anxiety disorders, in particular, are often difficult to assess using only one screening measure. As discussed above, broad anxiety measures such as the BAI that focus primarily on somatic symptoms often do not provide thorough coverage of cognitive aspects of anxiety. Broad anxiety measures emphasizing cognitive correlates of anxiety (eg, worry) often overlap significantly with depression. Further, the prevalence of somatic symptoms in headache patients suggests that somatic items common to these patients may need to be considered separately from cognitive/affective items when attempting to diagnose anxiety or depression.

Screening instruments are intended to be followed by a careful diagnostic interview, a process that may not occur in a general medical setting. Patients may thus receive an inappropriate psychiatric diagnosis. It is especially important for clinicians to remain alert to the core conditions required to diagnose depression (depressed mood and/or loss of pleasure or interest) and anxiety (feeling anxious/afraid, avoidance behaviors, or chronic worry). A patient may have a high score on a screening tool without satisfying the core diagnostic criteria. While many instruments may be used to monitor a patient's progress over time, a change in a score does not always equate with a change in a clinical condition (eg, a 50% decrease in a score does not necessarily indicate a 50% improvement in the condition). Finally, the presence of a psychiatric disorder should not deter the clinician from diagnosing or treating a medical disorder; the depressed patient with migraine merits migraine-specific therapy as well as treatment for depression.

Future Research.— The most important issues we identify to address in future research include: (1) To what extent do transdiagnostic symptoms need to be accounted for in diagnosing depression or anxiety in headache patients? (2) Does psychiatric comorbidity influence response to headache prophylactic medications or behavioral interventions? (3) What are the effects of treating one comorbid condition (either a psychiatric disorder or headache) on the other? and (4) To what extent do limbic factors other than psychiatric disorders influence headache outcomes (eg, lifestyle, “stress” and “hassles,” and subclinical anxiety and depression)?


The prevalence and impact of psychiatric disorders in headache patients suggest that all clinicians who treat headache patients should incorporate screening for psychiatric disorders as part of their routine clinical evaluation. Screening should address both anxiety and depression. Comprehensive evaluation of headache patients should incorporate formal and multidimensional psychiatric screening. Research studies, both of acute headache treatments and preventives, should also incorporate standardized psychological screening.

Conflict of Interest:  None