Digital CBT

What digital CBT is, and is not

Digital CBT (also called internet-delivered CBT, or iCBT) is not a new kind of therapy. It is CBT content delivered through a structured online or computer program: the same ideas about thoughts, behavior, avoidance, and exposure, worked through in guided modules rather than only in a room with a therapist.

That word "structured" matters. The evidence here belongs to specific programs that were built and tested for a diagnosed anxiety disorder. It does not automatically apply to a general wellness app, a mood tracker, a breathing exercise, a meditation library, or a chatbot, even if that product uses the phrase "CBT."

What the evidence shows

Pooling many trials, digital interventions produced a large average effect on anxiety, measured mostly against waiting lists or usual care. That is enough to reject the idea that digital delivery is a gimmick. Structured CBT content clearly survives being delivered online.

The signal also holds across the three core disorders, with estimates supporting effectiveness for generalized anxiety, panic disorder, and social anxiety. Panic programs showed particularly large effects. These are encouraging, but they belong to the specific programs tested in each diagnosis, and they should not be read as a promise of a set result for any product.

Why the large number is inflated

Most of these trials compared a program against an inactive control, such as a waiting list. Inactive controls make anxiety-therapy effects look bigger, because they offer no attention, no expectation of benefit, and no active treatment. So the large pooled number is best read as a signal that the format works, not as the amount of benefit you would get compared with another real treatment.

Guided or unguided?

Some programs include human support (guidance); others are fully self-guided. Guided programs scored somewhat higher on average, but the difference was not statistically significant. The safe reading is twofold: unguided programs can work, and support is not proven to be pointless. The practical question is not simply "guided or unguided." It is whether the level of support available is enough for you to actually start, keep going, and finish the program.

Is a program as good as a therapist?

This is the question most people want answered, and it has the thinnest evidence. When digital was compared head-to-head with face-to-face CBT, the difference was small and not statistically significant. But that rested on only a handful of comparisons and a few hundred participants. "Not significantly different" in a small, underpowered set of studies does not mean "equivalent." The comparison is promising enough to keep watching, but too thin to support swapping a therapist for a program.

What makes it actually work for a person

The label does not deliver the treatment. Benefit depends on real participation:

• Sticking with it: you have to return to the program, complete modules, and practice the exercises, including when symptoms or avoidance make that hard.
• Support: guidance can help you begin, troubleshoot, and keep going.
• The right program: a program built and tested for one disorder is not automatically right for another. A generalized-anxiety program is not automatically a panic or social-anxiety program.
• Access and setup: digital delivery can be a useful route when in-person therapy is delayed, unavailable, or declined. Access creates an opportunity; it does not guarantee engagement or that the program fits you.

What the evidence does not settle

• How long benefits last beyond the study window.
• Real-world completion rates outside of trials.
• How much guidance is enough.
• Which specific commercial programs carry the trial evidence.
• Whether larger, better comparisons would show digital truly matching face-to-face therapy.