How well does talking therapy really work for anxiety? A team of researchers went looking for the most demanding evidence they could find: randomised trials where CBT was tested against a placebo, published since 2017.
They found ten such trials, covering 1,250 people. Seven of the ten focused on PTSD.
The results were more modest than you might expect. CBT did outperform placebo, but the advantage was small, and smaller than earlier reviews had suggested. Its effect on accompanying depression wasn't statistically clear at all.
Some context matters here. Comparing therapy against a placebo is a deliberately hard test, because a psychological placebo usually still means real time with a supportive professional. So this measures the extra benefit of CBT's specific techniques, not whether therapy beats doing nothing. The authors also note that most of these trials studied PTSD, and very few placebo-controlled trials of other anxiety disorders have appeared recently.
What does that mean for you? CBT remains among the best-supported treatments for anxiety, and honest appraisals like this one are how the field keeps its claims proportionate and shows where more evidence is still needed.
If anxiety is weighing on you, a psychologist can talk through the options and what the evidence genuinely supports.
Setting up a fair test
Cognitive behavioural therapy is a structured talking therapy that helps people identify and change the unhelpful thought patterns and behaviours maintaining anxiety. To establish how well it works, researchers look for randomised controlled trials, in which participants are assigned at random to receive either the treatment or a comparison condition, reducing the chance that other factors explain the results.
The strictest version of this design uses a psychological placebo. Participants still receive regular attention from a professional and a plausible-sounding intervention, but without CBT's specific techniques. That isolates the added value of CBT itself from the general benefit of being cared for by someone.
What this meta-analysis did
A meta-analysis statistically combines results across multiple studies to produce a more precise pooled estimate than any single trial can offer. This one gathered randomised, placebo-controlled trials of CBT for anxiety-related disorders in adults published since 2017. Ten trials involving 1,250 participants met the criteria, and seven of the ten concerned post-traumatic stress disorder.
What it found
Results are reported using Hedges' g, an effect size describing how large a difference is in standardised form, independent of the particular scales used across studies. By convention, roughly 0.2 is a small effect, 0.5 medium and 0.8 large.
Across all ten trials, CBT outperformed placebo on the main anxiety-related symptoms with g = 0.24, a small effect that was statistically significant at p < 0.05, meaning it is unlikely to be due to chance. Its effect on co-occurring depression symptoms was smaller at g = 0.15 and did not reach statistical significance. Narrowing the analysis to the seven PTSD trials reduced the effect on target symptoms further to g = 0.14, though it remained statistically significant.
Two technical checks strengthen confidence in these figures. Heterogeneity, meaning how much results varied between studies, was very low, so the individual trials largely agreed rather than pulling in different directions. There was also no evidence of publication bias, the tendency for striking positive results to be published more often than unremarkable ones, which can otherwise inflate pooled estimates.
Why it matters, and its limits
The headline finding is that these recent, rigorously placebo-controlled effects are smaller than those reported in earlier meta-analyses of CBT for anxiety. That does not mean CBT has stopped working. It means this particular demanding style of evidence, drawn from the last five years, tells a more modest story than older reviews did.
The concentration of the evidence is an important limitation. Most of these trials, and therefore most of the pooled result, come from PTSD research. Very few placebo-controlled trials of other anxiety-related disorders have been published in this period, so the same pattern cannot yet be asserted with equal confidence across the board.
None of this suggests therapy is not worth trying. It reflects a field being appropriately careful about how large a benefit to promise. If anxiety is affecting your life, a psychologist can help you understand your options and what the evidence supports for your situation.