CBT and SSRIs are the two treatments with the strongest evidence for anxiety disorders. CBT changes how the brain estimates threat through structured experience and skill-building; SSRIs reduce underlying reactivity by sustained increases in serotonin signaling. Either alone produces substantial improvement for most people. Combined treatment tends to work faster for moderate-to-severe symptoms and to produce equivalent or better long-term outcomes.
The two treatments with the strongest evidence for anxiety disorders are cognitive behavioral therapy and selective serotonin reuptake inhibitors, or SSRIs. The National Institute of Mental Health and the American Psychological Association both list the two as first-line. They work on different parts of the same problem. Knowing what each one is doing makes the choice between them, or the decision to use both, less mysterious.
What cognitive behavioral therapy is
Cognitive behavioral therapy is a structured talking treatment that focuses on the relationships between thoughts, feelings, and behavior. For anxiety specifically, it has three main moving parts.
The first is identification of automatic thoughts. A patient learns to notice the specific predictions and interpretations that show up in anxiety-provoking situations. "If I speak up in this meeting, I'll embarrass myself." "If my heart rate climbs, something is wrong with me." These are usually fast, half-conscious, and accepted as true without examination.
The second is testing those predictions. Not by arguing with them. By collecting data. The patient and therapist design small experiments in which the predicted bad outcome is exposed to reality. Speak up in the meeting once. Notice what actually happens. Compare it to the prediction. Do this repeatedly across enough situations that the prediction loses its grip.
The third is exposure. For panic disorder, social anxiety disorder, OCD, and post-traumatic stress disorder, the most active ingredient of CBT is exposure to the avoided thing. For generalized anxiety disorder, the work is more weighted toward worry-postponement, tolerance of uncertainty, and behavioral experiments around the predictions the worry makes. The avoidance is what keeps the fear alive. Approaching the avoided thing, in a graded and tolerable way, allows the nervous system to update its threat estimate. This is the part of CBT that does the most work, and it's also the part that requires the most care to do well.
CBT for anxiety typically runs 12 to 20 weeks. Most patients begin to notice change in the first four to six weeks. By the end of the protocol, the underlying patterns have usually shifted in ways that hold up after treatment ends. The single biggest factor in CBT outcomes is whether the patient actually does the work between sessions. The session itself is a small fraction of the active ingredient.
What SSRIs are doing
SSRIs are antidepressants that have, over decades of use, turned out to work even better for anxiety than for depression. The class includes sertraline, escitalopram, paroxetine, fluoxetine, and a few others. SSRIs are also first-line for major depressive disorder; for the depression-focused use of these medications, see DepressionResource.org.
The short mechanistic story is that they block the reuptake of serotonin from the synapse, which raises serotonin levels in the spaces between neurons. The longer and more accurate story is that this is the immediate pharmacology, and it doesn't, on its own, explain why anxiety improves. Serotonin levels change within hours. Anxiety improves over weeks. The current understanding is that the sustained increase in serotonergic signaling produces gradual changes in the way certain brain regions, particularly the amygdala and its prefrontal regulators, respond to potential threats. The reactivity comes down. The baseline becomes quieter. The same situations stop producing the same surge.
Patients on a working SSRI rarely describe it as feeling medicated. The more common description is that the things that used to bother them still happen, and they bother them less. The space between trigger and response widens. The grip of worry softens. Sleep often improves. The body feels less braced.
Most SSRIs take four to six weeks to show a clear effect at a therapeutic dose. The first one to two weeks can include side effects, the most common being nausea, jitteriness, mild headaches, sleep changes, and a temporary increase in anxiety. These usually fade. MedlinePlus maintains a plain-language summary of antidepressant effects and warnings. About 60 to 70 percent of patients respond well to the first SSRI tried. Those who don't usually respond to a different SSRI or to a different class.
A note on the FDA boxed warning
SSRIs and SNRIs in the United States carry an FDA boxed warning about an increased risk of suicidal thoughts and behaviors in children, adolescents, and young adults under 25 during the first few months of treatment or after a dose change. The warning is based on pooled analyses of clinical trials. It's worth understanding for what it is and what it isn't.
What the warning describes: a small but real increase in reported suicidal ideation and behavior, not in completed suicide, during the early weeks of treatment, most pronounced in patients under 25. The absolute risk increase in the trials was on the order of two percentage points (roughly 4 percent of patients on active medication versus 2 percent on placebo reported new or worsening suicidal thoughts).
What the warning doesn't mean. Most patients in this age group take SSRIs without issue. Untreated depression and untreated anxiety carry their own substantial risk of suicidal thoughts. The warning is a reason to monitor closely during the first weeks of treatment, not a reason to avoid the medication. The decision to start, switch, or stop an SSRI in a young adult is made with a clinician who knows the patient.
What follow-up should look like in the first weeks: close contact with the prescribing clinician, attention from family or a trusted person, and an explicit plan for what to do if new or worsening thoughts of self-harm appear. A typical schedule includes a follow-up visit at one to two weeks, then again at four to six weeks, with the option to reach the clinician between visits as needed.
How the two compare
The honest comparison is that they're roughly equivalent in efficacy for most anxiety disorders in the short term. CBT has an edge in the long term because the gains tend to persist after the treatment ends, while medication effects fade if the medication is stopped. Medication has an edge in the short term for people with severe symptoms who aren't yet able to do the active work of therapy.
Most clinical guidelines suggest starting with CBT when it's accessible, starting with an SSRI when symptoms are severe or when therapy isn't immediately available, and combining the two when severity is high or when one alone has been insufficient.
The choice in real life is shaped by access. Good CBT therapists aren't evenly distributed. A patient in a rural area, or with limited time, or with severe symptoms that make therapy hard to engage with, may need to start with medication. A patient who is hesitant about medication and has access to a skilled therapist may reasonably try CBT first. Neither sequence is wrong. The wrong move is treating the choice as ideological.
What doesn't work as well as people think
A short list, with no judgment about why people try them.
Pure talk therapy without a CBT structure has weaker evidence for anxiety disorders, though it can be useful for related life issues. Supportive therapy is helpful but tends not to move the symptoms themselves.
Benzodiazepines work fast and feel like a solution. Over time, they tend to maintain the underlying anxiety pattern because the brain learns to associate the medication with safety, which keeps the threat sensitivity intact. They have a narrower role than they're often used for.
Supplements marketed for anxiety, magnesium, ashwagandha, L-theanine, and others, have modest evidence at best. Some are reasonable. None replace the treatments above for clinically significant anxiety.
Cannabis is a special case. Low doses can transiently reduce anxiety. Regular use is associated with higher rates of anxiety disorders over time, and in some people produces the very panic and derealization symptoms it was used to treat. This pattern is common enough to be worth raising explicitly.
A note on benzodiazepines
Benzodiazepines (alprazolam, lorazepam, clonazepam, diazepam, and others) work quickly for anxiety and reliably reduce acute symptoms. They aren't first-line for most anxiety conditions, for reasons worth understanding rather than dismissing.
In 2020 the FDA expanded its boxed warning on benzodiazepines to include the risks of abuse, misuse, addiction, physical dependence, and withdrawal. Physical dependence can develop within several weeks of daily use even at prescribed doses. Withdrawal from benzodiazepines is medically serious, can include seizures, and requires a supervised taper.
Additional considerations: combining benzodiazepines with opioids or alcohol substantially increases the risk of fatal respiratory depression and is also a boxed warning. Older adults are at increased risk of falls, fractures, and cognitive effects on benzodiazepines.
Used in narrow, time-limited situations (a single procedure, or an acute crisis while another treatment ramps up), benzodiazepines have a role. Used daily for chronic anxiety, they tend to maintain the underlying pattern by training the brain to associate the medication with safety, and they carry meaningful risk over time. Most current guidelines for anxiety treatment recommend them sparingly, if at all, and with a clear exit plan.
The practical version
If anxiety is interfering with sleep, work, or relationships in a sustained way, the two interventions with the most evidence are CBT and SSRIs. They aren't mutually exclusive. They tend to work better together than apart for moderate-to-severe symptoms. Most patients improve substantially with one or both within three months of starting. A clinician can help match the starting point to the picture.
The starting conversation doesn't need to be elaborate. A primary care visit, a referral to a therapist who actually does structured CBT (worth asking about, since the term is used loosely), or a consultation with a psychiatrist are all reasonable first steps. The first appointment doesn't commit anyone to anything. It produces a plan, and a plan is usually the missing piece.