A common question for anyone facing depression is whether medication or talking therapy works better. This study examined young people with depression and anxiety, comparing those who took an antidepressant, specifically an SSRI, on its own against those who combined it with cognitive behavioural therapy, a form of talk therapy that helps a person notice and change unhelpful thinking patterns.
The combination worked faster overall, with real improvement evident by four weeks. The more interesting result concerns timing. The additional benefit of CBT over medication alone did not clearly emerge until week twelve. The therapy component takes longer to show its effect, and the wait is worthwhile.
Younger people with milder symptoms, and those whose main difficulty was depression rather than anxiety, tended to improve fastest. Boys, older teenagers and those with more severe symptoms took longer to respond.
The practical implication is worth holding onto. Starting therapy and medication together and feeling no different straight away is entirely normal, because the benefit of therapy often takes three months to appear properly. That is a reason to stay with treatment rather than abandoning it early.
If you are struggling, reaching out to a psychologist or GP is a solid first step.
What the study actually did
This analysis pooled data from three large trials funded by the US National Institutes of Health: two in children and adolescents with major depression, and one in those with anxiety disorders. In each trial, some participants received an SSRI (selective serotonin reuptake inhibitor, a type of antidepressant that increases available serotonin in the brain) on its own, while others received an SSRI combined with CBT (cognitive behavioural therapy, a structured talk therapy that targets unhelpful thought and behaviour patterns).
Rather than just comparing average outcomes, the researchers used Bayesian Hierarchical Models (BHMs) — a statistical approach that allows data from different studies and individuals to be combined while still estimating how each person's own trajectory of symptoms changes over time. This let them look not just at whether combined treatment worked, but at *when* it started working, and for *whom* it worked fastest or slowest.
What it found
Across both depression and anxiety, SSRI+CBT produced significantly greater symptom reduction than SSRI alone by week 4 (p<0.001 — meaning this difference was very unlikely to be due to chance). This advantage held at weeks 8 and 12. But when the researchers specifically tested when CBT started adding measurable benefit *on top of* the SSRI, that additive effect didn't reach statistical significance until week 12.
The study also looked at moderators — patient characteristics that predict a faster or slower response. Younger participants, those with milder symptoms at the start, and those with depression rather than anxiety tended to respond fastest to combined treatment. Boys, older adolescents, minoritized children, those with more severe baseline symptoms, and those with externalizing disorders (conditions marked by outward behaviours like conduct problems, rather than inward distress) tended to respond more slowly.
Why it matters
This adds weight to existing evidence that combining medication and therapy is generally more effective than medication alone for young people with depression or anxiety. It also offers a nuance: the medication's effects and the added value of CBT don't necessarily arrive on the same timeline. CBT involves learning and practising new skills, so it may simply take longer for those skills to translate into measurable symptom change — even though the trial shows they eventually do. The researchers frame this as an early step toward tailoring treatment: understanding who tends to respond quickly versus slowly could eventually help clinicians set realistic expectations or adjust treatment plans for individual patients.
Limits worth knowing
The authors note several caveats. The moderators (like severity or demographic group) weren't measured consistently across all three trials, the number of observations per participant varied over time, and the studies didn't include genetic or pharmacokinetic data (information about how individuals metabolise the SSRI), which could otherwise help explain some of the differences in response speed.
If any of this resonates with your own experience, please know that treatment can take time to show its full benefit — and reaching out to a psychologist or doctor is a genuinely good next step.