Ms B is a 42-year-old computer programmer with a history of irritable bowel syndrome who presents to her primary care physician for a blood pressure check. Six months ago, she began caring for her chronically ill mother, and she reports increased stress. You note that she had a visit to urgent care after having transient chest pain, shortness of breath, and palpitations. Myocardial ischemia was ruled out without requiring hospital admission. Female sex, stressful life events, and chronic medical illness place her at increased risk for an anxiety disorder. What tools could be used by the physician or nurse to determine whether Ms B's symptoms and related behaviors indicate an anxiety disorder?
Why Is This Question Important?
Anxiety disorders are prevalent, are often chronic, cause important functional impairment, and are associated with increased health care use.1,2 Two of the more common anxiety disorders are generalized anxiety disorder (GAD) and panic disorder. In community samples, the estimated lifetime prevalence rates for GAD and panic disorder are 5.1% and 3.5%, respectively, and 12-month rates (experienced anytime within the last 12 months, including currently) are 3.1% and 2.3%, respectively.3 Primary care patients have higher rates of both GAD (8%) and panic disorder (6.8%), and the prevalence rate of GAD increases to 22% among those with anxiety problems as the presenting concern.4,5 Many patients with anxiety disorders present to their primary care physician with somatic symptoms, which contributes to underrecognition of these conditions and can result in unnecessary and costly diagnostic testing.6 When diagnosed, both GAD and panic disorder can be treated successfully with medication and/or psychotherapy. Furthermore, care management trials have shown that screening, coupled with effective primary care treatment, improves outcomes for patients with anxiety disorders.7
How to Diagnose GAD and Panic Disorder
Anxiety symptoms such as worry or physical tension are experienced nearly universally in response to stressful or threatening situations. Anxiety may be an adaptive emotional experience that helps a person to avoid or prepare for future challenges. In contrast, anxiety disorders cause severe and persistent symptoms that impair functioning. The criterion standards for GAD and panic disorder are summarized in Table 95-2. Generalized anxiety disorder is characterized by at least 6 months of persistent, excessive anxiety or worry that is difficult to control and causes significant distress or impairment. The diagnosis requires at least 3 of 6 additional symptoms: restlessness, fatigue, irritability, decreased concentration, muscle tension, and sleep disturbance.8 Panic disorder is characterized by frequent and unexpected panic attacks, and individuals with this disorder exhibit intense worry about having them. Panic attacks are periods of intense fear or terror associated with autonomic arousal, and typical symptoms include palpitations; sweating; trembling or shaking; shortness of breath; feeling of choking; chest pain or discomfort; nausea or abdominal distress; feeling dizzy, unsteady, lightheaded, or faint; paresthesias; chills; or hot flushes.8 Although agoraphobia was previously considered to be a subtype within the panic disorder diagnosis, in the Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition) (DSM-5) it is now classified as a discrete disorder characterized by avoidance of public spaces for fear of having a panic attack.
A clinical evaluation of anxiety disorders can begin with an open-ended question such as "Tell me about your worries, fears, concerns, and stresses, and how they affect you."9 When GAD is inquired about specifically, a question such as "Would you say that you have been bothered by 'nerves' or feeling anxious or on edge?" can elicit symptoms of the disorder. When inquiring about panic disorder specifically, the clinician can ask a question such as "Did you ever have a spell or an attack when all of a sudden you felt frightened, anxious, or very uneasy?"10
Another approach to the diagnosis of GAD and panic disorder in primary care clinics is to ask all patients, or those with risk factors, to complete a self-report screening instrument. Depending on the prevalence of the disease, the physician may want to optimize the positive likelihood ratio (LR+) to avoid unnecessary additional testing or the negative likelihood ratio (LR-) to be confident that anxiety disorders do not require additional consideration. An alternative as part of the initial diagnostic assessment would be to evaluate only patients who present with symptoms that raise suspicion of an anxiety disorder. For routine use in primary care settings, the ideal instrument should be brief, accurate, easy to score and interpret, and studied in mixed populations of patients. For patients with a positive screening result, a careful clinical interview coupled with a targeted physical examination and any indicated diagnostic testing to evaluate for an underlying explanatory medical illness is required for a definitive diagnosis. To inform decision making regarding a standard instrument to assess primary care patients for anxiety disorders, we conducted a systematic review of the literature to evaluate the performance of self-report instruments used to diagnose GAD and panic disorder in primary care settings.
Search Strategy and Study Selection
We searched MEDLINE, PsycINFO, and the Cochrane Library from January 1980 through April 2014 for studies conducted in general medical settings that compared a self-report instrument with an acceptable criterion standard. The search strategy included the terms generalized anxiety disorder and panic disorder, the names of anxiety instruments, and a validated search filter for retrieving articles on the diagnosis of health disorders (Supplemental Appendix 1).11,12 Electronic searches were supplemented by examining the bibliographies of systematic reviews, a recent technical report, and the studies we ultimately included in the technical report.13
We included studies that were conducted with patients aged at least 18 years who were treated in general medical settings (ie, general internal medicine, family medicine, geriatrics, emergency department, and women's health clinic); compared self-report questionnaires for GAD or panic disorder with diagnostic interviews, using criteria from either the Diagnostic and Statistical Manual of Mental Disorders (Third Edition) (DSM-III) or International Classification of Diseases, Ninth Revision, or more recent editions of these publications; and were peer-reviewed, English-language publications from North America, western Europe, New Zealand, or Australia. Geographic and language limitations were designed to identify studies with the highest applicability to US populations. Two reviewers independently examined each abstract for relevance. Next, full-text articles identified by either reviewer as potentially relevant were examined by 2 reviewers, who evaluated the articles' eligibility according to predetermined criteria (Supplemental Appendix 2). Disagreements were resolved by discussion or a third reviewer.
Data Abstraction and Quality Ratings
We extracted selected data elements informed by the principles outlined by the Standards for Reporting of Diagnostic Accuracy.14 These elements included descriptors to assess applicability (eg, setting, sample characteristics, anxiety disorder prevalence), test performance, and quality (eg, recruitment method, blinding, reference standard, sample size) of each study. When provided, raw data for the 2 × 2 table were extracted, and when not provided, data were derived from other performance characteristics such as sensitivity and specificity. When results were adjusted for the sampling design (eg, partial verification of the criterion-based diagnosis), we use the adjusted results. A second reviewer verified all data abstractions, and disagreements were resolved by reviewer discussion or by obtaining a third reviewer's opinion.
For each selected study, raters completed the Quality Assessment of Diagnostic Accuracy Studies, a 14-item tool that assesses study quality (Supplemental Appendixes 3-4). We followed recommendations from The Rational Clinical Examination series15 by assigning a level of evidence for each study, ranging from I (high quality) to V (low quality).
Sensitivity, specificity, and likelihood ratios (LRs) were calculated with CIs for instruments evaluated in each study. An LR+ is the ratio of the likelihood of a positive test result in an individual with the condition to the likelihood of a positive test result in an individual without it. An LR- is the ratio of the likelihood of a negative test result in an individual with the condition to the likelihood of a negative test result in an individual without it. Tests with higher specificity generally have higher LRs, and positive results are most useful for identifying patients with an anxiety disorder, whereas tests with higher sensitivity generally have lower LRs, and negative results are most useful for ruling out patients who do not have an anxiety disorder. If an LR+ is 2, a positive test result (in this case, a positive score on an anxiety questionnaire) is twice as likely to occur in an individual with an anxiety disorder as opposed to an individual without one. An LR- of 0.2 means that a negative screening result is one-fifth as likely to occur in an individual with an anxiety disorder as opposed to an individual without one. Because GAD and panic disorder are 2 distinct clinical entities, we calculated summary estimates separately for studies on GAD-specific instruments and panic disorder–specific instruments.
To estimate the prior probability of GAD and panic disorder, we calculated a random-effects summary measure from the included studies. The Symptom Driven Diagnostic System for Primary Care (SDDS-PC) instrument was evaluated in 3 studies, which allowed us to calculate separate summary measures for the sensitivity, specificity, and LR with 95% CI. All other instruments were evaluated in only 1 study, for which we show the test's point estimate and 95% CI. We explored heterogeneity among the studies with Cochran Q and I2, which describe the percentage of total variation across studies due to heterogeneity rather than chance, and we used meta-regression to evaluate the effect of age and sex on the LRs. Heterogeneity was categorized as low, moderate, or high according to I2 values of 25%, 50%, and 75%, respectively. We used Comprehensive Meta-Analysis (Biostat version 2.2.064) for all meta-analyses.
We identified 3605 unique citations from a combined electronic search of MEDLINE via PubMed (n = 1167), PsycINFO (n = 1810), and the Cochrane Library (n = 605) and from a manual examination of references (n = 23). After inclusion and exclusion criteria were applied, 3529 articles were excluded at the title and abstract level. We retrieved 76 articles for full-text review and excluded 63. For data abstraction and evidence synthesis, we retained a total of 13 articles representing 10 unique studies.16,17,18,19,20,21,22,23,24,25 Because some studies included more than 1 sample or evaluated more than 1 instrument, we included 14 unique evaluations of anxiety instruments. The Supplemental Figure illustrates the literature search process.
Of 13 articles describing 10 studies, 9 different instruments were evaluated (Table 95-3). Across all studies, diagnostic interviews determined that 257 of 2785 patients assessed had a diagnosis of GAD while 224 of 2637 patients assessed had a diagnosis of panic disorder. The average age of patients in studies of GAD (n = 6) (Table 95-4) was similar across 5 of the samples18,21,23,25 (range, 38-47 years), whereas 1 study20 contained older patients (mean age, 73 years). The studies were similar in sex, with 64% to 85% women. The studies of panic disorder (n = 6) (Table 95-4) among unselected patients included participants with a more homogeneous age (mean range, 39-54 years), with a similar distribution of women (66%-72%). A study of patients presenting with palpitations included similarly aged participants (mean age, 47 years), with a slightly smaller proportion of women (57%).16
Table 95-3.Characteristics of 8 Self-report Measures for Generalized Anxiety and Panic Disorder |Favorite Table|Download (.pdf) Table 95-3. Characteristics of 8 Self-report Measures for Generalized Anxiety and Panic Disorder
|Instrument ||No. of Items ||Response Format ||Time Frame ||Score Range ||Usual Cut Point ||Literacy Levelsa ||Completion Time ||Tracking of Symptoms |
|ADS-GA26 ||11 ||Yes or no ||Unknown ||0-11 ||4-5 ||Easy ||Unknown ||Unknown |
|GAD-723 ||7 ||4 Frequency ratings: not at all, several days, more than half the days, nearly every day ||2 wk ||0-21 || |
5 = mild
10 = moderate
15 = severe
|Average ||Unknown ||Unknown |
|GAD-Q-IV25 ||9 ||5 Yes or no; 2 Likert (9 response choices); 1 count of worries; 1 physical symptom checklist ||6 mo ||0-12 ||≥5.7 ||Average ||Unknown ||Unknown |
|SDDS-PC18 ||4 GAD ||Yes or no ||6 mo ||0-5 ||Unclear ||Easy ||<2 min ||Yes (scale has separate longitudinal tracking module) |
|Panic Disorder |
|BPDS27 ||4 ||Symptom severity: very little, a little, some, much, very much ||None ||0-16 ||≥11 ||Average ||Unknown ||Unknown |
|PHQ22 (panic module from 3-page diagnostic form)b ||5 ||Yes or no ||4 wk ||0-5 ||Yes on all 5 questions ||Easy || |
<1 min for 42%
1-2 min for 43%
3-5 min for 13%
>5 min for 3%
|SDDS-PC18 ||5 Panic ||Yes or no ||Past mo ||0-5 ||Unclear ||Easy ||<2 min || |
|Unnamed 10-item scale16 ||10 ||Symptom severity: not at all, a little bit, moderately, quite a bit, a great deal ||Unknown ||0-50 ||>21 ||Average ||Unknown ||Unknown |
|GAD or Panic Disorder |
|BAI-PC17 (GAD and panic) ||7 ||4 Items of symptom severity ||Past 2 wk to today ||0-21 ||≥5 ||Easy ||≈1 min ||Unknown |
|PRIME-MD24 (Multiple components with GAD and panic) ||3 ||Yes or no ||Past mo ||0-3 ||≥1 ||Easy ||<1 min ||No |
Table 95-4.Performance Characteristics of Self-report Instruments |Favorite Table|Download (.pdf) Table 95-4. Performance Characteristics of Self-report Instruments
|Instrument ||Study ||No. (% Prevalence)a ||Age, Mean (SD), y ||Females, % ||(95% CI) ||Quality Rating |
|Sensitivity ||Specificity ||LR+ ||LR− |
|GAD-7 ||Spitzer et al,23 2006 ||965 (7.6) ||47 (16) ||65 ||0.89 (0.82-0.96) ||0.83 (0.80-0.85) ||5.1 (4.3-6.0) ||0.13 (0.07-0.26) ||I |
|GAD-Q-IV ||Moore et al,25 2014 ||99 (27) ||39 (13) ||85 ||0.89 (0.77-1.0) ||0.63 (0.51-0.74) ||2.4 (1.7-3.3) ||0.18 (0.1-0.5) ||III |
|ADS-GA ||Krasucki et al,20 1999 ||88 (15) ||73 ||64 ||0.39 (0.12-0.65) ||0.88 (0.81-0.95) ||3.2 (1.3-8.0) ||0.70 (0.45-1.08) ||III |
|SDDS-PC ||Leon et al,21 1996 ||501 (16) ||49 (13) ||66 ||0.74 (0.64-0.83) ||0.82 (0.78-0.86) ||4.1 (3.2-5.2) ||0.32 (0.22-0.46) ||I |
| ||Broadhead et al,18 1995 ||257 (5.4) ||40 (13) ||79 ||0.92 (0.76-1.00) ||0.54 (0.49-0.59) ||2.0 (1.6-2.4) ||0.15 (0.02-1.01) ||I |
| ||Broadhead et al,18 1995 ||388 (3.1) ||39 (12) ||73 ||0.86 (0.67-1.00) ||0.60 (0.53-0.66) ||2.1 (1.6-2.8) ||0.24 (0.07-0.87) ||I |
| ||Summary SDDS-PC || || || ||0.78 (0.66-0.87) ||0.67 (0.47-0.82) ||2.6 (1.6-4.1) ||0.31 (0.22-0.43) || |
|Panic Disorder |
|PHQ ||Spitzer et al,22 1999 ||585 (6.0) ||46 (17) ||66 ||0.81 (0.68-0.93) ||0.99 (0.98-1.00) ||78 (29-210) ||0.20 (0.11-0.37) ||I |
|SDDS-PC ||Leon et al,21 1996 ||501 (8.0) ||49 (13) ||66 ||0.70 (0.56-0.84) ||0.91 (0.88-0.93) ||7.9 (5.5-11) ||0.33 (0.20-0.53) || |
| ||Broadhead et al,18 1995 ||257 (6.2) ||40 (13) ||79 ||0.78 (0.62-0.94) ||0.80 (0.76-0.84) ||3.9 (2.9-5.2) ||0.28 (0.14-0.56) ||I |
| ||Broadhead et al,18 1995 ||388 (7.0) ||39 (12) ||73 ||0.63 (0.39-0.86) ||0.83 (0.78- 0.88) ||3.8 (2.3-6.0) ||0.45 (0.23-0.70) ||I |
| ||Summary SDDS-PC || || || ||0.71 (0.60-0.80) ||0.86 (0.77-0.91) ||4.9 (3.0-7.9) ||0.35 (0.25-0.48) || |
|10-Item scale ||Barsky et al,16 1997 ||124 (26) ||47 ||57 ||0.72 (0.56-0.88) ||0.71 (0.60-0.80) ||2.4 (1.7-3.6) ||0.40 (0.22-0.70) ||II |
|BPDS ||Johnson et al,19 2007 ||295 (14) ||54 (11) ||66 ||0.61 (0.46-0.76) ||0.29 (0.23-0.35) ||0.86 (0.66-1.1) ||1.36 (0.88-2.08) ||I |
|GAD or Panic Disorder |
|BAI-PC ||Beck et al,17 1997 ||56 (23) ||49 (16) ||73 ||0.85 (0.65-1.00) ||0.81 (0.67-0.92) ||4.6 (2.3-8.9) ||0.19 (0.05-0.68) ||III |
|PRIME-MD ||Spitzer et al,24 1994 ||431 (18) ||55 (16) ||60 ||0.93 (0.88-0.99) ||0.53 (0.48-0.58) ||2.0 (1.8-2.3) ||0.12 (0.05-0.29) ||I |
Most studies were rated low risk of bias (Table 95-4, Supplemental Appendixes 3-4). All of the questionnaires were self-administered and did not require specialized equipment or trained personnel, making them suitable for patients to complete in a variety of settings. Based on diagnostic interviews, the random-effects summary estimate for prevalence of GAD was 10.1% (95% CI, 5.7%-17%), whereas prevalence of panic disorder was 8.8% (95% CI, 6.6%-12%). The panic disorder range does not include the results of Barsky et al,16 which found a panic disorder prevalence of 26% among patients presenting with a complaint of heart palpitations.
Performance Characteristics of Self-report Screening Instruments
Generalized Anxiety Disorder
The Generalized Anxiety Disorder Scale 7 Item (GAD-7), using a cut point of greater than or equal to 10, had good sensitivity (89%) and specificity (83%), had the highest LR+ (5.1; 95% CI, 4.3-6.0), and is also the only measure that reported test-retest reliability (intraclass correlation, 0.83). A GAD-7 score less than 10 had an LR- (0.13; 95% CI, 0.07-0.26) similar to that of the Generalized Anxiety Disorder Questionnaire Fourth Edition at a threshold less than 5.7 (LR-, 0.18; 95% CI, 0.06-0.52; P = .65 for the comparison). The SDDS-PC takes less than 2 minutes for completion and has an "easy" literacy level with 3 different formulations evaluated in differing populations (summary LR+, 2.6 [95% CI, 1.6-4.1]; LR-, 0.31 [95% CI, 0.22-0.43]). The Anxiety Disorder Scale–Generalized Anxiety was studied in older patients (mean age 72 years) and had the least useful LR-, 0.70 (95% CI, 0.45-1.1).
The instruments18,20,21,23,25 showed high heterogeneity (LR+: I2 = 93%, P < .001; LR−: I2 = 76%, P = .001) among studies conducted in primary care with unselected patients. Although meta-regression revealed that the LR+ did not vary by the mean age in the study samples (P = .23), older mean age was strongly associated with the LR-, accounting for 94% of the heterogeneity (P < .001). Studies of GAD with a higher frequency of younger patients found a lower LR- (easier to rule out GAD) compared with studies with older patients. Sex accounted for only 29% of the heterogeneity in the summary LR+ (P < .22) and only 5% of the heterogeneity in the summary LR- (P = .43). Thus, these screening instruments for GAD yielded similar diagnostic accuracy results across the sex distribution of the studies we evaluated (range female, 64% to 85%).
We assessed the heterogeneity of 4 of the 6 studies for identifying patients with panic disorder. One study16 was not included because it assessed patients with palpitations who presented to specialists rather than unselected patients presenting to a primary care provider. A second study19 was not included because it had no diagnostic utility (both LR CIs included 1), so it could not classify the presence or absence of panic disorder.
The Patient Health Questionnaire (PHQ), using a positive response to all 5 questions, had good sensitivity (81%) and specificity (99%), the best LR+ (78; 95% CI, 29-210), and the best LR- (0.20; 95% CI, 0.11-0.37). The PHQ requires less than 1 minute for completion and has an easy literacy level. The SDDS-PC is also efficient, with a summary LR+ of 4.9 (95% CI, 3.0-7.9) and summary LR- of 0.35 (95% CI, 0.25-0.48). An additional advantage of the PHQ is that it includes a brief depression module previously found to have high sensitivity and specificity for diagnosing depression.28
The 4 instruments had high heterogeneity for the LR+ (I2, 92%; P < .001), but the LR- showed low heterogeneity (I2, 14%; P = .32). In a meta-regression, age was not associated with the summary LR+ (R2, 0), suggesting that the results are similar in the age range we evaluated (mean age range 39 to 54 years). The meta-regression showed that the summary LR+ accounted for a small amount of the variability in the LR+ (R2, 15%; P = .03).
Combined Screening for GAD and Panic Disorder
For identifying patients who may have either GAD or panic disorder, the Beck Anxiety Inventory–Primary Care performed well compared with other instruments, with an LR+ of 4.6 (95% CI, 2.3-8.9) and an LR- of 0.19 (95% CI, 0.05-0.68). The instrument has an easy literacy and can be completed quickly (approximately 1 minute). An alternative combined instrument, the Primary Care Evaluation of Mental Disorders, has the fewest number of questions for the patient (3), short completion time (1 minute), and easy literacy level. At a threshold score of less than or equal to 1 question with a positive response, individuals with no positive responses have the lowest LR- with the narrowest CI for either anxiety disorder (LR-, 0.12; 95% CI, 0.05-0.29).
We found that 2 screening instruments, GAD-7 for GAD and the PHQ for panic disorder, have good performance characteristics and are feasible for use in primary care. Further validation of these instruments is needed because neither instrument was replicated in more than 1 primary care population.
This study was a highly structured and systematic review of the extant evidence. Our evidence synthesis was guided by a carefully designed standardized protocol, including a systematic search of research databases and relevant bibliographies, double data abstraction, and use of validated criteria to assess the quality of identified studies. Our multidisciplinary team included expertise in internal medicine, primary care, psychiatry, and psychology. Our search identified a large number of anxiety screening instruments, but few had been studied in primary care populations. These instruments had moderate to good operating characteristics, but unlike instruments used in the detection of other common mental illnesses such as depression or dementia, the operating characteristics have not been replicated in multiple samples. Even for the SDDS-PC—the only instrument evaluated in multiple studies—the versions studied were different, which might change the test performance.
In most studies, threshold values for the screening instrument were specified after analysis of results instead of before. Thus, replication is needed to validate the cutoffs recommended in these studies. Additionally, many of the studies did not confirm the diagnosis with the reference standard in all patients, or in a random sample of them, which could introduce partial verification bias. A further limitation is the lack of studies reporting on patient outcomes and societal influence. This lack of important patient outcomes has been recognized as a challenge in systematic reviews of diagnostic tests.29 Because our eligibility criteria were designed to exclude poor-quality studies (ie, studies in which the same person conducted the screening and criterion standard), we may have excluded studies that could provide low-level evidence on the topic. Furthermore, one of the better-performing measures, the Beck Anxiety Inventory–Primary Care, was tested in a very small sample (n = 56) and that study17 was rated as having a higher risk of bias (quality rating = III). A solution to these issues is to encourage future high-quality validation studies, which are notably absent despite that many of them were published almost 20 years ago. The criterion standard for GAD and panic disorder has not changed appreciably in that time, and thus the performance characteristics of these measures remain applicable to current diagnoses. Finally, these studies were not designed to address differing performance in subgroups, so our evaluation of age and sex as explanations for varying performance is based on a small number of studies, uses indirect comparisons, and should be considered exploratory. Indeed, future studies would benefit from the inclusion of older patients (>65 years) and more ethnically diverse samples to better determine how these screening measures perform in different subgroups.
How to Learn a Method for Diagnosing GAD and Panic Disorder
Both the GAD-7 and PHQ screening instruments are available online (www.phqscreeners.com) and have been translated into many languages. Because both of these instruments are self-administered, minimal clinician training is needed to administer them. Additional advantages of GAD-7 are that it has good operating characteristics in a 2-item abbreviated version (the GAD-2) and in screening for anxiety disorders other than GAD.4 A manual for scoring both instruments is also available online. All of the instruments included in this review are for screening or case-finding purposes and do not diagnose GAD or panic disorder. Although these instruments may be used as part of the initial diagnostic evaluation, a criterion-based diagnosis must be established through further evaluation by a primary care physician or by a mental health professional to whom the patient is referred. Such confirmation should be determined by follow-up questions based on the DSM-5 (outlined in Table 95-2) and should rule out psychiatric disorders with related symptoms (eg, posttraumatic stress disorder, depression) and medical causes of symptoms suggestive of anxiety. The studies we reviewed used DSM-III or DSM-IV diagnostic criteria for GAD or panic disorder; no significant changes in these criteria were introduced in DSM-5.
Screening alone is not sufficient to ensure that patients with anxiety disorders in the primary care setting receive appropriate treatment. Although referring a patient for a psychiatric evaluation is one option, primary care physicians should also familiarize themselves with the diagnostic criteria for GAD or panic disorder, as well as with pharmacologic and other treatments for these disorders that are appropriate for primary care. Collaborative care models integrating psychiatric treatment in the primary care setting have also been shown to be effective for anxiety disorders.7 Furthermore, because there is symptom overlap between GAD or panic disorder and other psychiatric diagnoses, false-positive results on any of these screening instruments may be not only "true" false-positives (ie, when the patient meets the criteria for no related diagnoses) but also due to the presence of a related psychiatric disorder. As such, a positive screening result, even if it is a false-positive for GAD or panic disorder, may indicate the need for further evaluation of the patient.
You observe that Ms B has important risk factors for an anxiety disorder, and her trip to urgent care suggests a possible panic attack. You decide that in addition to checking her blood pressure, you will conduct case-finding for GAD or panic disorders. You administer the GAD-7 and PHQ, wherein she scores 12 on the GAD-7 and answers no to the PHQ item about anxiety attacks. With a pretest probability of 20% for GAD (based on an estimate of twice the prevalence in unselected primary care patients) and a GAD-7 LR+ of 5.1, Ms B. has a 59% probability of having GAD. After discussing options for evaluation and treatment, you refer her for a psychiatric evaluation in which her condition may be diagnosed and treated with empirically supported treatments such as cognitive behavioral therapy or an appropriate pharmacotherapy.
There are several promising case-finding instruments with good performance characteristics for GAD or panic disorder in primary care populations. In particular, the GAD-7 and PHQ stand out as the most efficient instruments, whereas the SDDS-PC may be an adequate alternative when a fast screen is desired because it assesses both conditions with relatively few questions. For clinical practices that opt for patient-completed screening instruments (eg, in the waiting room), the Primary Care Evaluation of Mental Disorders shows promise for identifying anxiety that might prompt additional questions during an examination. Further replication of these initial validation studies, in particular with samples of older and more ethnically diverse patients, is needed in primary care settings.
The following disclosures were reported at the time this original article was first published in JAMA.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.
et al.. The functioning and well-being of patients with unrecognized anxiety disorders and major depressive disorder
. J Affect Disord
. 1997;43(2):105–119. Medline:9165380
B. Anxiety disorders in primary care: prevalence
, impairment, comorbidity
, and detection. Ann Intern Med
. 2007;146(5):317–325. Medline:17339617
et al.. Comparative effectiveness of collaborative chronic care models for mental health conditions across primary, specialty, and behavioral health care settings: systematic review
. Am J Psychiatry
. 2012;169(8):790–804. Medline:22772364
VP. Anxiety. In: Henderson
GW, eds. The Complete Patient History: An Evidence-Based Approach to Differential Diagnosis. 2nd ed. Lange Medical Books/McGraw Hill; 2012.
MB. Using five questions to screen for five common mental disorders in primary care: diagnostic accuracy of the Anxiety and Depression
Detector. Gen Hosp Psychiatry
. 2006;28(2):108–118. Medline:16516060
RB; Hedges Team. EMBASE search strategies for identifying methodologically sound diagnostic studies for use by clinicians and researchers. BMC Med
. 2005;3:7. Medline:15796772
NL. Optimal search strategies for retrieving scientifically strong studies of diagnosis from Medline: analytical survey
. 2004;328(7447):1040. Medline:15073027
et al.; Standards for Reporting of Diagnostic Accuracy. Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD Initiative. Ann Intern Med
. 2003;138(1):40–44. Medline:12513043
DL. Update: primer on precision and accuracy. In: Simel
SA, eds. The Rational Clinical Examination: Evidence-Based Clinical Diagnosis. New York, NY: McGraw Hill; 2009:9–16.
M. Use of the Beck Anxiety and Depression
Inventories for primary care with medical outpatients. Assessment
et al.. Development and validation of the SDDS-PC screen for multiple mental disorders in primary care. Arch Fam Med
. 1995;4(3):211–219. Medline:7881602
A. The FEAR: a rapid screening
instrument for generalized anxiety in elderly primary care attenders. Int J Geriatr Psychiatry
. 1999;14(1):60–68. Medline:10029937
et al.. Brief screens for mental disorders in primary care. J Gen Intern Med
. 1996;11(7):426–430. Medline:8842936
DM. Using the GAD-Q-IV to identify generalized anxiety disorder
in psychiatric treatment seeking and primary care medical samples. J Anxiety Disord
. 2014;28(1):25–30. Medline:24334213
E. The Guy's/Age Concern survey
rates of cognitive impairment, depression
and anxiety in an urban elderly community. Br J Psychiatry
. 1989;155:317–329. Medline:2611541
L. A brief screen for panic disorder
. J Anxiety Disord
et al.. Challenges in systematic reviews of diagnostic technologies. Ann Intern Med
. 2005;142(12 pt 2):1048–1055. Medline:15968029