Skip to content
MyFreud

Therapy for Anxiety: What Works and What to Ask For

Therapy is not one thing, and the difference between the kinds matters. What the evidence supports for anxiety, and the words that get you the right treatment.

5 min read

Pop-art illustration of two people sitting across a small table in conversation, hands resting on the tabletop, with an orange city skyline through the window behind them.

Key takeaways

  • Therapy is not one thing, and the kind you are offered changes what happens. Asking for a named approach rather than for "therapy" is the single most useful thing you can do at a first appointment.
  • Cognitive behavioural therapy has the strongest evidence base for anxiety disorders. A meta-analysis of 41 placebo-controlled trials found a moderate effect on the target disorder, which is a real effect of an ordinary size rather than a transformation.
  • The active ingredient is approaching what you avoid, not talking about your childhood. Any anxiety therapy that never asks you to do something uncomfortable between sessions is unlikely to be the treatment with the evidence.
  • Course length is shorter than most people expect. Anxiety protocols typically run 8 to 20 sessions rather than years, and open-ended talking without a plan is a different activity.
  • Medication and therapy are not rivals. Both work, the combination is reasonable, and access is often what decides the order rather than what is theoretically best.

Therapy for anxiety is not one thing, and which kind you get changes what actually happens in the room. Cognitive behavioural therapy has the strongest evidence, it is structured, it is shorter than most people expect, and it will ask you to do things between sessions that you would rather not do.

That last part is not a side effect of the treatment. It is the treatment, and knowing that in advance is the difference between finishing a course and concluding that therapy does not work for you.

What kind of therapy works best for anxiety

Cognitive behavioural therapy has the strongest evidence base across the anxiety disorders, and the effect is moderate rather than dramatic. A meta-analysis of 41 randomised placebo-controlled trials, covering 2,843 people with generalised anxiety, panic disorder, social anxiety, obsessive compulsive disorder or post-traumatic stress disorder, found a moderate effect of CBT on the symptoms of the target disorder and smaller effects on other anxiety symptoms, depression and quality of life. [carpenter-2018-cbt-anxiety]

Two things about that number are worth holding on to. It is measured against a placebo rather than a waiting list, which is a much harder comparison and the reason the effect looks smaller than the figures quoted in most articles. And a moderate average means a meaningful number of people did not improve, which is information you deserve before you start rather than after.

Guidelines follow this. For generalised anxiety and panic disorder, a structured psychological intervention based on CBT principles is the recommended psychological treatment. [nice-cg113-anxiety-therapy]

Why the uncomfortable part is the part that works

Anxiety is maintained by avoidance, so a treatment that never asks you to approach anything cannot correct it. The prediction driving an anxiety disorder is that some situation will be unbearable or dangerous. Every time you avoid it, or get through it using a precaution, the prediction survives untested, and it quietly gets stronger.

What CBT does is arrange for that prediction to be tested under conditions where it can actually be disproved. This is why the homework matters more than the conversation, and why a course that consists entirely of discussing how the week went is unlikely to move anything.

Where the change comes from in a course of anxiety therapy Illustrative
0 25 50 75 100 Contribution to symptom change 88 Testing a prediction in real life 74 Dropping a precaution 30 Understanding where it came from 42 Feeling heard in the session
0 25 50 75 100 Contribution to symptom change 18 Testing a prediction in real life 16 Dropping a precaution 64 Understanding where it came from 78 Feeling heard in the session

A schematic of the mechanism described in this section and in the sources cited. Not measured data.

Feeling heard is not worthless, and a therapist you cannot talk to is a bad therapist regardless of method. The point is that it is not sufficient on its own for an anxiety disorder, and a course built entirely around it will usually leave the avoidance exactly where it was.

What the different therapies actually involve

The names are not interchangeable, and knowing roughly what each one does makes the choice yours rather than the waiting list’s.

  • CBT identifies the specific predictions you make, then tests them through planned tasks between sessions. Structured, time-limited, homework-based. This is the default to ask for.
  • Exposure-based work is often a component of CBT rather than a separate thing, and it dominates where the anxiety attaches to a specific situation.
  • Acceptance and commitment therapy works on your relationship to anxious thoughts rather than on their content, and asks what you would be doing if anxiety were not deciding. Reasonable where CBT has been tried, or where the anxiety is diffuse.
  • Counselling is supportive rather than protocol-driven. Genuinely useful for a difficult period in your life, and not the same as treatment for an anxiety disorder.
  • Psychodynamic therapy works on patterns and their origins over a longer course. Its evidence base for anxiety specifically is thinner than CBT’s.

Our guide to how to deal with anxiety covers what to do while you wait, and the anxiety hub sets out what the condition is and how it differs from ordinary worry.

Therapy or medication

Both work, and this is one of the few places in mental health where a meta-analysis has compared them directly. Pooled across anxiety disorders, psychological therapy and medication both produce meaningful improvement, and the combination is a reasonable option where symptoms are severe. [bandelow-2015-anxiety-efficacy]

The practical differences matter more than the average effect sizes. Medication generally acts sooner and asks less of you week to week. Therapy asks considerably more and its gains tend to hold up better once it stops. Availability is often what actually decides: where a course of therapy is a year away, starting medication is not a compromise, and it does not prevent therapy later.

What to take to a first appointment

Tick anything you can already answer. This is a preparation prompt rather than a test, and it produces no diagnosis.

0 of 6 ticked

When to seek help

Speak to a doctor or a therapist if anxiety has been getting in the way of work, sleep or relationships for more than a couple of weeks and your own efforts are not shifting it. That threshold is deliberately set at interference rather than at intensity, because how bad anxiety feels is a poor guide to whether it needs treating.

Go sooner if you are avoiding things you need to do, if you are drinking or using something to manage the feeling, or if panic attacks have started. Ask for cognitive behavioural therapy by name, say what you have stopped doing, and ask how many sessions the course runs to.

The practical side of that, which is where most people stall, is covered separately: our guide to finding a therapist explains what the licence letters actually mean, what a session costs, and why contacting eight clinicians rather than two is the change that most reliably gets somebody seen.

Go urgently if you have thoughts of harming yourself.

How MyFreud can help

Whether anxiety tracks events or arrives on its own is the question a first appointment turns on, and it is very hard to reconstruct from memory. MyFreud gives you daily mood tracking that shows the pattern across weeks, so you arrive with a record rather than an impression.

Download MyFreud and start today: App Store or Google Play.

Frequently asked questions

What type of therapy is best for anxiety?

Cognitive behavioural therapy has the strongest evidence base across anxiety disorders, and the version used is usually tailored to the specific problem rather than generic. A meta-analysis of 41 randomised placebo-controlled trials covering generalised anxiety, panic, social anxiety, OCD and post-traumatic stress found a moderate effect of CBT on target symptoms. Other approaches can help, particularly where CBT has been tried without benefit, but if you are choosing without other information, CBT is the reasonable default to ask for.

How long does therapy for anxiety take?

Most structured anxiety protocols run somewhere between 8 and 20 sessions, and the number depends more on the specific problem than on how severe it feels. Panic disorder protocols are often at the shorter end, social anxiety and OCD at the longer. If you are offered six sessions of unstructured talking, or an open-ended arrangement with no plan and no endpoint, it is fair to ask what the treatment is and how you will both know whether it is working.

Does therapy actually work for anxiety, or does it just help you cope?

It changes symptoms rather than only teaching you to tolerate them, which is what the placebo-controlled trials are measuring. The mechanism is that anxiety is maintained by avoidance: every situation you escape confirms the prediction that it would have been unbearable. Therapy works by arranging for that prediction to be tested under conditions where it can be disproved, which is why the uncomfortable part is the part that does the work.

Should I do therapy or take medication for anxiety?

Both have good evidence, and the choice is usually practical rather than principled. Medication tends to act sooner and requires less of you week to week; therapy takes more effort and its gains tend to persist better after it stops. The combination is a reasonable option for moderate to severe anxiety. In practice, availability and waiting times often decide the order, and starting medication first does not close off therapy later.

What should I say at a first appointment about anxiety?

Say what you have stopped doing. Clinicians can work with "I no longer take the motorway" or "I have turned down two promotions because of presentations" far more usefully than with "I feel anxious all the time", because avoidance is what treatment targets and it is also what marks the difference between distress and a disorder. Bring how long it has been happening and one recent concrete example, and name the therapy you are asking for.

References

  1. 1.Carpenter JK, Andrews LA, Witcraft SM, Powers MB, Smits JAJ, Hofmann SG ( 2018). Cognitive behavioral therapy for anxiety and related disorders: a meta-analysis of randomized placebo-controlled trials. Depression and Anxiety. doi:10.1002/da.22728
  2. 2.Bandelow B, Reitt M, Rover C, Michaelis S, Gorlich Y, Wedekind D ( 2015). Efficacy of treatments for anxiety disorders: a meta-analysis. International Clinical Psychopharmacology. doi:10.1097/YIC.0000000000000078
  3. 3.National Institute for Health and Care Excellence ( 2011). Generalised anxiety disorder and panic disorder in adults: management (CG113). NICE. nice.org.uk .