Generalized Anxiety Disorder After 65: What the Screening Evidence Does and Does Not Cover

Anxiety Screening Recommendations Do Not Extend Beyond 65

In 2023, the US Preventive Services Task Force recommended screening for anxiety disorders in adults, finding a moderate net benefit based on moderate-certainty evidence. For adults aged 65 years or older, the Task Force instead issued an “I” statement, concluding that the evidence was insufficient to assess the balance of benefits and harms. Specifically, the evidence was insufficient regarding the accuracy of screening tools, and on the benefits and harms of screening and treatment in older adults.¹

Prevalence falls with age, but isolated cases are rare

A Wave 2 analysis of 12,312 adults aged 55 and older from the National Epidemiologic Survey on Alcohol and Related Conditions found a past-year GAD prevalence of 2.80%. The prevalence declined with age: 3.73% among adults aged 55 to 64 years, 2.28% among those aged 65 to 74 years, and 1.88% among those aged 75 years and older.² Among adults aged 55 years and older with past-year GAD, only 0.53% had GAD without another Axis I or II disorder, so the diagnosis in later life is rarely isolated. Among those with past-year GAD and no comorbid Axis I disorder, only 18% sought professional help in the past year, compared with 28.3% of all past-year GAD cases.² These DSM-IV diagnoses were based on data collected in 2004 and 2005, limiting their applicability to current practice.

Lower GAD Screening Cut Points in Older Adults

The Task Force review found that few studies reported the accuracy of anxiety screening tools in older adults.¹ In a German population-based study of 438 community-dwelling adults aged 58 to 82, with a blinded structured clinical interview as the criterion standard, 6.2% (95% CI, 3.9%-8.2%) met DSM-IV criteria for current GAD. The optimal GAD-7 cut point in this sample was 5 or greater, with a sensitivity of 0.63 and a specificity of 0.90. The GAD-2 performed similarly at a cut point of 2 or greater with a sensitivity of 0.67, and a specificity of 0.90.³ The authors concluded that lower cut points may be appropriate in the general older population, although this conclusion was based on a single cohort.

Beers Criteria Recommend Avoiding Benzodiazepines After 65

The 2023 American Geriatrics Society Beers Criteria recommends avoiding benzodiazepines in adults 65 and older, based on moderate-quality evidence. The panel cites increased sensitivity, decreased metabolism of long-acting agents, physical dependence, and increased risk of cognitive impairment, delirium, falls, fractures, and motor vehicle crashes.⁴ For older adults with a history of falls or fractures, the criteria state shorter-acting agents are not safer than long-acting ones, and it lists severe generalized anxiety disorder among the circumstances in which these agents may be appropriate. The panel separately advises considering total anticholinergic burden during medication review because cumulative exposure is associated with increased risk of falls, delirium, and dementia.⁴

Treatment Evidence Is Limited After 65

The Task Force review identified the accuracy of GAD screening tools and the effectiveness of GAD treatment in older adults as research priorities. The review located only 7 randomized trials of psychological treatment in this population, enrolling 215 participants, with a pooled standardized mean difference of -0.66 (95% CI, -0.94 to -0.38).¹ Taken together, the available evidence remains limited for both GAD screening and treatment in adults aged 65 and older.

Financial support was provided by Definium Therapeutics. Psychiatrist.com independently developed the content and maintained final editorial control.

Prevalence and Impact

Generalized anxiety disorder (GAD) ranks among the most prevalent psychiatric conditions. The Mental and Substance Use Disorders Prevalence Study, using clinician-administered interviews from 2020 to 2022, estimated a past-year GAD prevalence of 10.0% among US adults aged 18 to 65, representing approximately 20.2 million people.¹ GAD is chronic and recurrent, and the functional toll is substantial, with impairment across work, social, and physical domains; reductions in quality of life comparable to those seen in major depression; and higher medical costs than in patients without the disorder.² Comorbidity can further magnify the burden, with mood and other anxiety disorders frequently accompanying the diagnosis.³

Comorbidity and Cost Add to the Burden

GAD rarely occurs on its own. In the World Health Organization World Mental Health Surveys, 81.9% of people with GAD met criteria for another lifetime disorder, including a mood disorder in 63.0% and another anxiety disorder in 51.7%.³ GAD also carries higher healthcare costs than anxiety-free populations: pooled controlled cost-of-illness data found direct medical costs were 2.60 times higher (95% CI, 2.01-3.36).⁴

Standard Treatments With Incomplete Results

SSRIs and SNRIs remain the foundation of pharmacologic care, but their efficacy is less than ideal. In a network meta-analysis of double-blind randomized trials, first-line agents such as duloxetine, escitalopram, paroxetine, and venlafaxine roughly doubled the odds of remission compared with placebo, with odds ratios ranging from 1.7 to 2.3.⁵ Despite these odds, only about one-third of patients experience remission during acute first-line treatment, leaving most still symptomatic.⁶ Treatment options have also remained limited, with no pharmacotherapy carrying a new mechanism of action approved for GAD in the United States since 2007.^7,8

Balancing Efficacy and Tolerability

Response is a more attainable target than remission, but it is not guaranteed. An estimated 30% to 60% of patients do not experience any response to first-line treatment,⁶ and among those who do respond, the benefit is modest. A 2025 Cochrane review of 37 randomized trials involving 12,226 adults with GAD found that antidepressants improved response compared with placebo (risk ratio, 1.41; 95% CI, 1.29-1.55), with a number needed to treat of 7.⁹ At the same time, discontinuation because of adverse effects was more than twice as likely as with placebo (risk ratio, 2.18; 95% CI, 1.81-2.61), corresponding to a number needed to harm of 17.⁹ Overall dropout rates were similar between the groups.

Where Standard Care Falls Short

Standard pharmacotherapy helps many patients but leaves many others symptomatic, and the absence of a mechanistically new approval for nearly two decades limits what clinicians can offer those who do not respond adequately. For a condition this common and this persistent, closing that gap remains an unmet need in GAD care.

Recognizing GAD In Clinical Practice

Generalized anxiety disorder (GAD) is usually described through its diagnostic criteria and its prevalence, but diagnosing anxiety can be challenging. Many patients report trouble sleeping, muscle tension, fatigue, or pain, but the underlying worry may not be mentioned. How GAD presents in practice, how often it is missed, and its impact on a patient’s daily life are important factors for understanding the disorder.

How GAD Presents

The hallmark of GAD is persistent, hard-to-control worry, but patients often reach out to the clinician when their symptoms present in physical form. Somatic symptoms frequently serve as the presenting complaint, and GAD may be unrecognized or mistaken for a physical condition.¹ Painful physical symptoms are especially common. In a study of 981 patients with GAD, 59.0% of those without comorbid depression and 78.0% of those with it reported painful physical symptoms, compared with 28.3% of controls.² A patient who presents with headaches, back pain, or insomnia does not, at first glance, look like a patient with an anxiety disorder.

Why it goes unrecognized

GAD is common in primary care, where roughly 8% of patients meet criteria,³ yet it often goes undetected. In one primary care survey, GAD was correctly diagnosed in only 34% of the patients who had it, and just 20% were receiving treatment.⁴ Treatment is often delayed. Among people with GAD, an estimated 86.1% eventually reach treatment, but the median delay from the onset of the disorder to first contact is 9 years.⁵ The mismatch is between how patients describe their distress, often in terms of physical symptoms, and how the disorder’s diagnostic criteria. As a result, many patients go undiagnosed and undertreated for reasons that have little to do with how severe their anxiety is.

A burden the numbers understate

The impact of GAD on functioning is significant. In the Primary Care Anxiety Project, patients with GAD (n=128) showed worse psychosocial functioning than patients with type 2 diabetes, hypertension, recent myocardial infarction, or congestive heart failure.⁶ Among patients with pure GAD (n=20) who were free of any comorbid medical or psychiatric illness, 75% scored below the 25th percentile of the general population on mental functioning, so the burden could not be attributed to another condition.⁶ Physical symptoms resulted in poorer functioning. When patients had painful physical symptoms, disability scores were more than one and a half times worse.² Everyday symptoms such as disrupted sleep, difficulty concentrating, irritability, and muscle tension can affect work and relationships in ways that a symptom count does not capture.

Recognition as the first intervention

For many patients, the unmet need in GAD begins before treatment does. GAD can present through physical symptoms and have a substantial impact on functioning, making recognition an important part of care. Identifying GAD in a patient who reports only physical complaints is an important first step towards better care.

Beyond Persistent Worry

Generalized anxiety disorder is more than persistent worry. Evidence from neuroimaging and preclinical work increasingly describe it as a disorder of circuit regulation, in which the brain’s threat-detection system may be insufficiently regulated by the regions involved in controlling it. In particular, altered communication between the amygdala and prefrontal cortex may affect how the brain responds to perceived threats and regulates anxious states.

A circuit that loses its brake

In GAD, the balance between the amygdala and the prefrontal cortex appears to be altered. In drug-naive patients, resting-state imaging shows weaker functional connectivity between the left basolateral amygdala and the anterior cingulate and medial prefrontal cortex, with connectivity differences associated with state and trait anxiety.¹ At the same time, the right basolateral amygdala  demonstrated increased functional connectivity with the left superior temporal gyrus and insula, and these connectivity differences were also associated with anxiety symptoms.¹  The findings suggest altered communication between threat-processing and regulatory circuits, including differences in the connectivity of regions involved in vigilance and emotional regulation.

How chronic stress remodels the architecture

Animal models show that sustained stress can produce contrasting structural changes across brain regions. In the amygdala, chronic stress drives dendritic expansion, strengthening the threat circuitry. In the hippocampus and prefrontal cortex it does the opposite, producing dendritic retraction and spine loss.^2,3 In the prefrontal cortex, chronic uncontrollable stress removes spines and dendrites through sustained calcium signaling, falling BDNF support, and the active pruning of synapses by glial cells, weakening the region that provides top-down regulation of thought, action, and emotion.⁴ In these animal models, chronic stress can increase amygdala reactivity while weakening prefrontal regulation. Whether the same structural sequence occurs in patients with generalized anxiety disorder has not been established.

Partial reversibility after stress

Glucocorticoid stress hormones are among the signals involved in stress-related structural remodeling.² This may help explain how chronic stress affects the course of anxiety. A sensitized amygdala and a thinned prefrontal cortex form a self-reinforcing loop that can keep anxiety running and slow the pace of recovery. Stress-induced remodeling in the hippocampus and prefrontal cortex shows at least partial reversibility after the stressor resolves, while amygdala hypertrophy has been reported to persist.² This raises the possibility that restoring synaptic architecture may be relevant to recovery, rather than only rebalancing chemistry.

Where standard treatment acts

First-line pharmacotherapy may act along this same axis. Beyond blocking monoamine reuptake, common antidepressants also bind the neurotrophin receptor TrkB and engage  BDNF-dependent signaling that supports synaptic plasticity.⁵ Because the affinity of these drugs for TrkB is in the micromolar range, brain concentrations sufficient to engage the receptor accumulate over weeks of treatment, which the authors propose as one contributor to the gradual onset of clinical effect.⁵ Seen this way, GAD is less a simple chemical imbalance than a disorder of maladaptive synaptic and circuit plasticity. That perspective links its altered prefrontal control, sensitized threat circuitry, slow treatment response, and relapsing course into a single picture, and it points toward restoring the regulatory connections that may be disrupted by chronic stress.

The Persistence Problem In GAD

The burden in generalized anxiety disorder (GAD) is easy to picture as untreated illness. A large part of it is different: patients who improve with treatment and then, over months and years, do not stay well. GAD is a chronic and relapsing condition, and its course over time, not only the response to a first prescription, accounts for much of its burden.

A course measured in years

Over the first five years, prospective data show GAD following a chronic course of remission and relapse rather than resolving, and retrospective studies suggest this pattern can persist for as long as two decades.¹ In a 12-year prospective study, the probability of recovering from GAD was 58%, and among patients who did recover, the probability of recurrence was 45%.² GAD also rarely runs its course alone. It usually co-occurs with another mood, anxiety, or substance use disorder,³ and in the same cohort, comorbid major depression reduced the probability of recovering from GAD by roughly half.² For many patients, the realistic expectation is not a single course of treatment followed by cure, but long-term management of an illness that remits and returns, all the more so when another psychiatric condition is present.

Relapse when treatment stops

The chronic course also shapes how treatment is used. Clinical guidance advises continuing medication for about 12 months after improvement before tapering, specifically to lower the risk of relapse.³ That caution is supported by relapse-prevention data. A meta-analysis of relapse-prevention trials, including six in GAD, examined antidepressant discontinuation across 28 trials (5,233 patients with anxiety disorders, obsessive-compulsive disorder, or PTSD). Stopping the antidepressant roughly tripled the odds of relapse compared with continuing (odds ratio 3.11), and over follow-up ranging from 8 to 52 weeks, relapse occurred in 36.4% of patients who discontinued versus 16.4% of those who continued.⁴ These data suggest that improvement does not necessarily persist after treatment is withdrawn.

The burden of an illness that persists

Because the course is long and relapsing, its burden accumulates. GAD is associated with substantial impairment in work and social functioning, reduced quality of life, and high direct and indirect costs.⁵ When symptoms recur or never fully clear, that impairment is not a single episode but a recurring feature of a patient’s life.

What it adds up to

Measured over years rather than weeks, GAD care is defined less by whether a patient responds than by whether that response persists. Patients may improve and then relapse, comorbidity can make recovery more difficult, and stopping treatment raises the risk of relapse. These data describe a chronic, relapsing condition that calls for long-term management rather than a single course of treatment.

Measuring Symptoms In GAD

Generalized anxiety disorder (GAD) is treated over long periods, and treatment decisions depend on whether symptoms are improving. Measurement-based care uses a validated outcome to standardize symptom-change measurement alongside clinical judgment.

A validated instrument for GAD

The GAD-7 is a seven-item self-report scale developed and validated in primary care.¹ Among 2,739 participants, 965 completed a structured psychiatric interview; a cutoff of 10 or higher gave a sensitivity of 89% and a specificity of 82%. Internal consistency was 0.92 (Cronbach alpha) and test-retest reliability 0.83 (intraclass correlation). Scores of 5, 10, and 15 may be interpreted as mild, moderate, and severe. The researchers identified severity assessment and monitoring of change as intended uses, while noting that responsiveness to change remained untested.

Screening recommendations

In 2023, the US Preventive Services Task Force (USPSTF) recommended screening for anxiety disorders in adults aged 19 to 64 years, including pregnant and postpartum persons, a grade B recommendation for people without a diagnosed mental health disorder or recognized symptoms.² For adults 65 and older, it issued an I statement, finding the evidence was insufficient to assess screening because data on screening-tool accuracy and on the benefits and harms of screening in older adults are lacking. The grade B recommendation relies on indirect evidence. Two small trials directly evaluating screening programs (n=918) showed no symptom difference at 13 to 22 weeks, so the recommendation instead links the accuracy of screening tools with the effectiveness of treatment for detected anxiety; thus, a positive screen requires confirmatory assessment. Potential harms include unnecessary referrals, overdiagnosis and treatment, labeling, and stigma, though no direct evidence on screening harms was found.

Guidelines recommendations for measurement frequency

Guidelines recommend repeating assessments without prescribing an instrument. The UK-based National Institute for Health and Care Excellence (NICE) issued “Generalised anxiety disorder and panic disorder in adults: management” (CG113) in 2011 and updated in 2019 with the recommendation that GAD care include identification, assessment, education, and active monitoring.³ It places drug treatment alongside high-intensity psychological therapy and recommends reviewing the effectiveness and side effects of drug treatment every two to four weeks for the first three months, but does not specify a rating scale. The Canadian clinical practice guidelines for anxiety disorders (2014) list eight first-line agents for GAD, and notes that the Clinical Global Impression scale is brief enough for use at each appointment and recommends therapy for at least 12 to 24 months for most patients.⁴

What the outcome evidence shows

Controlled evidence for measurement-based care varies across mental health disorders. A meta-analysis of seven randomized trials found higher remission with measurement-based care than usual care in five trials (1,518 patients; 52.8% vs 43.0%; odds ratio [OR] 1.83, 95% confidence interval [CI] 1.12 to 2.97) and higher adherence (OR 1.68, 95% CI 1.22 to 2.30); response did not differ, but five of the seven trials carried a high risk of bias.⁵ In anxiety, CALM randomized 1,004 primary care patients, 75% of whom had GAD, to collaborative care built around web-based symptom tracking or usual care.⁶ On the primary outcome, the 12-item Brief Symptom Inventory, the intervention group had significantly lower anxiety symptoms than usual care at 6, 12, and 18 months (mean differences were -2.49, -2.63, and -1.63, respectively). Response and remission were also higher; at 12 months, response was 63.7% vs. 44.7% and remission was 51.5% vs. 33.3%. Importantly, the investigators stated that the tracking component cannot be isolated from the rest of the intervention.

Application to GAD

The GAD-7 is a validated, disorder-specific instrument for tracking GAD symptoms, and guidelines suggest specific re-assessment timepoints. The USPSTF recommendation addresses case findings in adults without a diagnosed disorder, not monitoring during treatment. What is still missing is a trial in GAD that isolates the effects of measurement itself, so the evidence for measurement-based care remains indirect rather than definitive.