CBT vs Medication for Anxiety: What Head-to-Head Trials Actually Show

“Should I try therapy or medication?” is one of the most common questions people with anxiety ask, and the answer they usually get is some version of “both work.” That’s true as far as it goes. But the evidence behind that answer is less tidy than it sounds, because the studies that actually compare CBT and medication head to head are fewer, harder to design and more limited than the studies that test each one separately.

Most evidence is indirect

The bulk of the evidence comes from trials that compare CBT with a control group (often a waiting list or treatment as usual) and trials that compare a medication with a placebo pill. Both kinds of trial show that their respective treatments work better than the control. But the size of the advantage depends heavily on what the control is, as we discuss in understanding effect sizes.

Comparing the results of two separate trials — one of CBT, one of medication — is tempting but unreliable. The participants may differ, the measures may differ, the severity may differ, and the comparison groups don’t mean the same thing. A waiting-list control and a pill placebo are fundamentally different: one involves receiving nothing, the other involves receiving a convincing inert treatment with regular clinical contact.

What head-to-head trials find

A systematic review of direct comparisons between psychotherapy and pharmacotherapy for anxiety disorders concluded that the two were broadly similar in acute effects. Differences were small and inconsistent across conditions [1]. A network meta-analysis of 234 randomized trials across anxiety disorders reached a similar conclusion: both CBT and medications showed efficacy, with overlapping confidence intervals in most direct and indirect comparisons [2].

The broad finding is one of similarity, not clear superiority of either approach. But that similarity comes with important caveats.

The blinding problem

This is the fundamental design issue and it doesn’t have a clean solution. In a drug trial, both groups can swallow identical-looking pills, so neither participants nor (ideally) clinicians know who has the active drug. This controls for expectation fairly well, though not perfectly.

In a therapy trial, there is no equivalent. Participants know whether they are receiving therapy, and most comparison conditions don’t credibly look like real therapy to participants. Boot and colleagues argued that even “active” psychological control conditions fail to equalize expectations, and that almost no psychological intervention studies check whether the groups expected the same benefit [6].

This means that when CBT is compared with a pill (whether active or placebo), the CBT group’s advantage may include an expectation component that the pill comparison group doesn’t share. Direct comparisons are not comparing like with like.

The follow-up question

One frequently cited advantage of CBT is that it lasts longer. After treatment ends, patients who received CBT tend to maintain gains, while patients who stop medication may relapse.

This is supported by some data, but the comparison is uneven. In typical follow-up designs, the CBT group has completed treatment (the skills, in theory, persist), while the medication group has discontinued their medication. In practice, many patients stay on medication long-term. Comparing CBT-after-treatment with medication-after-discontinuation doesn’t tell you how CBT compares with ongoing medication.

Hofmann and colleagues, reviewing meta-analyses of CBT efficacy, noted that CBT’s durability is one of its most cited advantages but acknowledged that the evidence base for long-term comparisons specifically against medication is limited [4].

What guidelines recommend

NICE’s guideline for generalized anxiety and panic disorder recommends offering CBT first for most anxiety disorders, with medication (typically an SSRI) as an alternative when CBT is not available or not preferred, or as an addition when CBT alone isn’t sufficient [5]. This reflects both the evidence of similar acute efficacy and the theoretical advantage of CBT’s durability, as well as practical considerations like side effects and patient preference.

Guidelines are summaries of committee judgment, not simple readouts of evidence. Different guideline bodies weigh the same evidence differently.

What this means for reading comparisons

When you see a claim that CBT is “as effective as” or “more effective than” medication:

1. Check whether the comparison is direct or indirect. Were the same participants randomized to one or the other, or are results from separate trials being compared?
2. Check what the comparison group was. CBT vs waiting list and an SSRI vs pill placebo are not comparable designs.
3. Check the time point. Equal at end of treatment doesn’t mean equal at one year.
4. Check the blinding. Was expectation controlled in both arms?
5. Check whether medication was continued or stopped. Follow-up comparisons where the medication group discontinues are testing withdrawal, not long-term medication.

For why the choice of outcome measure matters in these comparisons, see what an anxiety study actually measures.

What remains uncertain

Whether CBT and medication differ in efficacy for specific anxiety disorders remains unclear. Whether combining them is better than either alone depends on the condition and possibly the individual. Most importantly, individual variation in response to both is large, and we don’t yet have reliable ways to predict who will benefit more from which approach. The question “which is better?” may be less useful than “which is better for this person, given their preferences, circumstances and access?”

Choosing between treatment options. The evidence supports both CBT and medication for anxiety disorders, and the choice depends on your circumstances, preferences, access and how you respond. Talk to a GP or mental health professional who can help you weigh the options. Don’t stop or change prescribed medication without speaking to your prescriber. If you are in crisis or thinking about harming yourself, contact emergency services or a crisis line now: in the US, call or text 988; in the UK and Ireland, Samaritans on 116 123.