Do I Need Medication for Anxiety, or Can Therapy Alone Work?
By John Nicholas Jung Shumate, MD, JD | The Jeong Center for Psychiatry and Psychotherapy
Last updated: May 17, 2026
Whether you need medication for anxiety depends on the severity, duration, and functional impact of your symptoms. Therapy alone is often effective for mild to moderate anxiety; medication becomes more important when anxiety is severe, physiologically intense, or hasn't responded to therapy. For many people, the combination of medication and therapy is more effective than either one alone.
The word anxiety comes from the Latin angere: to choke, to constrict. It is one of the oldest words in medicine, and one of the most physically honest. Twenty centuries of clinical progress and we have not improved on it. The person sitting in front of you does not say "I have generalized anxiety disorder." They say "I feel tight." They say "I can't breathe." They say "something is squeezing me and I don't know what it is." The diagnosis gives the constriction a name. It does not, by itself, loosen the grip.
Nora Chen* did not use the word anxiety when she first came to see me. She said she felt "tight" (tight in her chest, tight in her thinking, tight in the way she held herself through the workday, which she described with the precision of someone accustomed to billing her time in six-minute increments).
She was a litigator in her late thirties, the kind of person who remembered colleagues' birthdays and ran meetings with an efficiency her partners admired and her friends found slightly alarming. She had been managing her anxiety for fifteen years, she told me, though "managing" was not quite the right word.
Managing implies control.
What she described was closer to negotiation: a daily, exhausting bargain with a part of herself that would not be quieted, could not be outworked, and had the disconcerting habit of intensifying precisely when everything in her life was going well.
She had one question. Should she take medication, or could therapy handle this on its own?
It is the most common question I hear (particularly from high-achieving professionals), and the one with the most unsatisfying answers. Most of what is written online falls into one of two camps: pro-medication content from health systems that treats an SSRI like a solution, or anti-medication content from therapy advocates who treats it like a concession. The honest answer lives in neither camp, and it requires more than a paragraph to get right.
Does therapy work for anxiety without medication?
Yes, often.
For mild to moderate generalized anxiety, cognitive behavioral therapy (CBT) produces meaningful improvement in roughly 50–60% of patients, according to APA treatment guidelines. Exposure and response prevention (ERP) is even more effective for panic disorder and specific phobias. Psychodynamic therapy, which is central to how I practice, works differently: rather than targeting symptoms directly, it addresses the underlying patterns that generate anxiety in the first place.
What therapy does that medication cannot is build understanding.
A good therapist helps you recognize what your anxiety is actually about, which is rarely what it appears to be about on the surface. The person who cannot sleep before a presentation may not have a presentation problem. They may have a relationship to failure, or to visibility, or to the expectations of people they have never quite managed to disappoint, that has been running quietly in the background for decades. Therapy is the process of making that architecture visible. And visibility (as any litigator could tell you) is where the real leverage lives.
The catch is that therapy takes time.
It requires showing up, tolerating discomfort, and engaging in the work of self-examination, which is not passive and not always comfortable and not something that yields to the kind of disciplined effort that high-achieving people are used to applying to everything else in their lives. Insight does not respond to harder work. It responds to patience, which is a different skill entirely.
What does medication actually do for anxiety?
Medication reduces the physiological intensity. If your nervous system is running hot (racing heart, constricted chest, the low-grade alarm that never fully switches off) an SSRI or SNRI can lower the baseline enough that you can function, sleep, and engage in the kind of self-reflection that therapy requires. The Latin had it right: anxiety constricts. Medication loosens the grip.
What medication does not do is teach you anything.
It does not help you understand why you are anxious, what patterns in your life sustain the anxiety, or why the alarm keeps sounding long after the danger (if there ever was a specific danger) has passed. For most anxiety presentations, medication is not a permanent solution. It is a tool (sometimes an essential one) that creates the conditions for deeper work.
I should be direct about my own philosophy here. I believe in evidence-based prescribing, and I have access to the full range of psychiatric medications. But I do not believe in fruitlessly medicating away symptoms that are really tied to other things in a patient's life. An SSRI will not repair a marriage. Buspirone will not resolve a career that has quietly become unbearable.
When I prescribe, I want to know what I am treating and why, and I want the patient to understand that too.
I was not always this clear about it. Early in my residency at Harvard, I reached for the prescription pad more quickly than I should have (not out of carelessness, but out of a particular kind of clinical impatience, the understandable desire to relieve suffering fast when someone is sitting across from you in obvious distress). It took several patients who taught me more than I taught them to understand that the fastest intervention is not always the deepest one. And the deepest one is not always fast enough.
When is the combination more effective than either one alone?
For moderate to severe anxiety disorders, the evidence is fairly clear: combined treatment outperforms either approach in isolation. A major meta-analysis published in World Psychiatry found that for generalized anxiety disorder, the combination of pharmacotherapy and psychotherapy produced significantly better outcomes than either alone (not just in symptom reduction, but in sustained recovery and daily functioning over time).
Here is what this looks like in practice.
A patient arrives with anxiety that has been worsening for months. They are not sleeping. Their work has slipped. They have started avoiding situations that used to be routine (social gatherings, difficult conversations, even driving). The anxiety has become so physiological that sitting still feels difficult, let alone sitting with the kind of emotional material that good therapy surfaces.
Starting medication first (or simultaneously) lowers the floor.
It does not solve the problem; it makes the problem approachable. And then therapy (real therapy, not a checklist of coping strategies handed out like pamphlets) can do what it does best: help the person understand what is driving the anxiety, not merely manage its symptoms.
The integrated model is especially helpful here. When one clinician handles both medication and therapy, the prescriber and the therapist share the same understanding of who you are, because they are the same person. I can adjust a medication based on what I am hearing in our sessions, and I can use what I observe about a medication's effects to sharpen the therapeutic work.
That is precision medicine, and it is the kind of care that disappears when a fifteen-minute med check is the only contact a patient has with their psychiatrist.
How do I know which approach is right for me?
There is no formula, but there are real considerations.
Therapy alone is often a strong starting point when your anxiety is mild to moderate, when it is connected to identifiable life circumstances or patterns, and when you are able to function (even if functioning feels harder than it should). Many of my patients in this category do very well without medication, and some are surprised by how much shifts when they begin to understand what their anxiety has been trying to communicate.
Medication becomes more important when the anxiety is severe, when it is disrupting sleep or work or relationships in ways that are accelerating, when panic symptoms are prominent, or when a reasonable course of therapy has not produced sufficient improvement.
There is nothing weak or deficient about needing medication. The brain is an organ, anxiety involves real neurobiology, and sometimes the biology needs to be addressed directly before the psychology can do its work.
What I tell every patient: you should be able to ask your psychiatrist why they are recommending one approach over another, and you should get an answer that makes sense to you. Not a vague reassurance, not a dismissal, but a real explanation grounded in your specific situation.
If your prescriber does not ask about your life, your history, and what the anxiety actually feels like from the inside, that is worth noticing. If a therapist dismisses medication categorically, without understanding your clinical picture, that is also worth noticing. The question is not medication or therapy. The question is what does your anxiety need, and how do we find out.
This is exactly the kind of conversation I have with patients in a first session. If you are weighing your options and want to think it through with someone who takes the question as seriously as you do, I'd welcome that conversation.
Nora and I worked together for several sessions before the question of medication came up again.
She raised it, not me. By then she had begun to see something she had not expected: that her anxiety was not a malfunction but a signal, one she had been overriding with competence and willpower for so long that she had forgotten it carried any information at all. We started a low-dose SSRI (not because therapy had failed, but because it became clear that the constriction was loud enough that it was interfering with the work we were trying to do together).
The medication did not change who she was. It gave her enough room to hear herself think.
She still bills in six-minute increments. She still remembers the birthdays. But the tightness she described in that first session has loosened, and what she found underneath it was not the emptiness she had feared but something quieter, less finished, that she is still learning to name. My office, late in the afternoon after she leaves, holds the particular stillness of a conversation that went somewhere neither of us expected.
It is the best part of this work, and the part no prescription pad can capture.
Frequently Asked Questions
Can anxiety be treated without medication?
Yes, in many cases. Mild to moderate anxiety is frequently and effectively treated with therapy alone (particularly CBT, ERP, and psychodynamic psychotherapy). The key factors are severity, duration, and how much the anxiety is interfering with daily life. For some patients, therapy produces lasting change that endures well beyond the end of treatment.
What type of therapy is best for anxiety?
The most evidence-supported therapies for anxiety disorders are cognitive behavioral therapy (CBT) for generalized anxiety, exposure and response prevention (ERP) for panic disorder and phobias, and psychodynamic psychotherapy for anxiety rooted in deeper relational and identity patterns. In my practice, I draw from multiple modalities depending on what a particular patient needs (which is one of the advantages of working with a psychiatrist who also does therapy).
How long does anxiety medication take to work?
Most antidepressants typically take two to four weeks to begin showing effects, with full therapeutic benefit often reached at six to eight weeks. Benzodiazepines work faster (within thirty minutes to an hour) but carry risks of dependence and are not recommended as a long-term solution in most cases. The timeline is one reason I have detailed conversations with patients about expectations before prescribing.
*The patients and individuals described are composites. They are drawn from real clinical encounters, real lives, and real systemic failures, but their names, biographical details, and identifying circumstances have been changed, combined, and reimagined to protect the privacy of the people whose experiences inspired them. The emotional and medical truths are preserved; the particulars are not. This is a form of fidelity, not of deception: the goal is to honor what these stories reveal about the human experience of illness and care, while ensuring that no one's private life becomes public without their consent. Prior results do not guarantee future results in any particular case.
J. Nicholas Jung Shumate, MD, JD is a Harvard-trained psychiatrist and sees patients throughout the Boston, Brookline, Cambridge, and Newton, MA region and supervises trainees at Harvard Medical School. He completed his residency training at the Harvard Psychiatry Training Program at Beth Israel Deaconess Medical Center.
