Anxiety Disorders: Efficient Screening Is the First Step in Improving Outcomes Free

Anxiety disorders have been the neglected stepchild of primary care–based mental health care. Compared with the extensive research in primary care on the adverse effects of depression on somatic symptom burden, decrements in function, and medical utilization and costs (1, 2), far less research has been completed on anxiety disorders. Yet, the National Comorbidity Survey (3) has shown that anxiety disorders are the most frequent disorders in the general population and are associated with substantial social and vocational impairment. Primary care patients with anxiety disorders are high users of primary care services and have many medically unexplained symptoms (that is, chest pain and rapid heart rate), clinically significant decrements in function, and high medical costs (1, 4–6). They also have high levels of psychiatric comorbidity, particularly major depression. Meeting Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV), criteria for 1 or more anxiety disorder markedly increases the risk for major depression over the following year (7). Anxiety disorders are also associated with adverse health behaviors, such as smoking and sedentary lifestyle (8, 9), which may contribute to the high levels of medical comorbidity found in adults with anxiety disorders.

In this issue, Kroenke and colleagues describe 4 important findings (10) about anxiety disorders in primary care: 1) the high prevalence (19.5%) of the 4 most common anxiety disorders in primary care (generalized anxiety disorder, panic disorder, social anxiety disorder, and posttraumatic stress disorder [PTSD]); 2) the association of these 4 disorders with high levels of comorbid depression, somatic symptom burden, and functional impairment, as well as high medical utilization; 3) 2 brief screening questionnaires (Generalized Anxiety Disorder [GAD]-2 and GAD-7) can efficiently screen for all 4 disorders; and 4) more than 40% of patients with an anxiety disorder reported that they were not receiving any current mental health treatment.

Unlike the extensive research on the use of depression scales in primary care, far fewer reports have focused on the development of screening scales for anxiety. Kroenke and colleagues’ paper shows that 2 new screening questionnaires, the GAD-2 and the longer GAD-7, not only perform well (as measured by receiver-operating characteristic [ROC] curve analysis) for detecting generalized anxiety disorder, but also performed almost as well for detecting panic disorder, social anxiety disorder, and PTSD (10). This result is important because using a different screening test for each disorder would be inefficient and might discourage screening in primary care, especially given the extensive time that is already required for screening for multiple medical disorders. However, we wonder whether the responses to the questionnaire may only reflect anxiety. Given the relatively high depression scores in Kroenke and colleagues’ study population, the high predictive validity of the GAD-7 may reflect a high level of nonspecific emotional distress rather than something specifically unique about anxiety. Comparing the ROC curves for depression screening measures, such as the Patient Health Questionnaire (PHQ)-9 or PHQ-2 (11), with the ROC curve for GAD-7 would be a first step in testing this hypothesis.

Screening for anxiety disorders is a necessary first step, but the evidence suggests that making a diagnosis is not sufficient to improve the quality of mental health care and outcomes among primary care patients with anxiety disorders (12). Studies have shown that primary care patients who are accurately diagnosed with anxiety disorders have important gaps in the mental health care that they receive, including poor adherence to psychoactive medications, poor follow-up by primary care physicians, and infrequent exposure to evidence-based psychotherapies (such as cognitive–behavioral therapy) (13). These findings, together with the high prevalence of anxiety disorders, indicate a serious failure in the management of anxiety in the primary care setting.

Research on primary care–based care of depression suggests that changing the system for anxiety disorders may help to close the gap. These studies have shown that system-of-care changes that linked depression screening programs to exposure to evidence-based depression treatments had the largest effect on the quality of care for depression and improvement in depression outcomes (14). According to very strong evidence, quality of care and outcomes improve when screening is coupled with multimodel intervention programs, such as collaborative care (15, 16).

Collaborative care models are an important element in improving the quality of mental health care. They use allied health professionals, such as nurses, to support primary care physicians. These care managers provide patient education about depression, closely follow patients, track adverse effects of care, monitor adherence and depression outcomes (with standard depression scales, such as the PHQ-9) (11), and facilitate return visits with the primary care physician if depression is not improving (15). In the most successful collaborative care programs, a psychiatrist supervises the care manager and recommends medication adjustments that the allied health professional communicates to the primary care physician.

These collaborative care models are very effective in improving depression care. A recent meta-analysis reviewed 37 trials that compared the effect of collaborative care models for depression with that of usual primary care (16). The meta-analysis showed 2-fold increases in adherence to antidepressants and improvements in depression outcomes that lasted for 2 to 5 years (16). Three recent studies that tested the effectiveness of collaborative care models for patients with generalized anxiety disorder or panic disorder also found statistically significant improvements in the quality of care and anxiety outcomes compared with usual care (17–19).

A potentially key element of the most successful collaborative care programs is patient self-activation, which may be particularly important for patients with anxiety because they actively avoid specific anxiety-provoking situations and behaviors. Self-activation is an important element in overcoming avoidance, which is necessary for anxiety syndromes to fully remit. As Kroenke and colleagues note, this aspect of anxiety disorders also suggests that cognitive–behavioral treatments may be particularly important for patients with anxiety.

Kroenke and colleagues have performed a large, descriptive study that shines a spotlight on a largely neglected disorder. We point out 2 minor concerns. The authors show that people with any of the 4 anxiety disorders have poorer function than people without these disorders. In the analysis that led to their conclusion, the authors did not make statistical adjustments to control for the higher prevalence of comorbid major depression and medical disorders in patients with anxiety disorders versus those without. However, other researchers have found that anxiety disorders are associated with statistically significant decrements in function after statistically controlling for depression and medical comorbidity (6). A second concern is the method for forming the study sample. The authors recruited the sample from patients attending clinic and did not select the patients randomly, which may explain the slightly higher rates of anxiety disorders compared with those of other recent studies.

Accurate recognition of anxiety disorders in primary care is the first step in providing effective treatment and potentially preventing development of major depression and social and vocational impairment. Kroenke and colleagues have developed and validated a brief anxiety questionnaire that efficiently screens for the 4 most common anxiety disorders. Hopefully, this contribution will help stimulate the continued development of effective primary care–based models to treat these highly prevalent disorders.