What is CBT?

Treatment principles

The National Institute of Mental Health describes CBT as helping people identify and question unhelpful automatic thinking and change behaviour patterns. It places these methods within a wider clinical process that considers functioning, medical context, and the person’s goals (National Institute of Mental Health [NIMH], 2024). CBT involves what people practise and learn, rather than simply replacing negative thoughts with positive ones. A balanced appraisal can acknowledge a real difficulty while examining what follows from it.

Sources: NIMH, 2024

Components and formulation

Cognitive restructuring examines an interpretation and its supporting and conflicting evidence. Behavioural methods can include approaching previously avoided situations, testing a prediction, or rebuilding meaningful activity. A structured thought record organizes an experience into an event, interpretation, emotion, and response. Problem-solving works on actions that can change a practical situation. These activities serve different purposes. The review modules explain why exposure and response prevention for OCD, trauma-focused treatment for PTSD, and CBT-I for insomnia cannot be replaced by an undifferentiated reflection prompt. The component descriptions are an explanatory synthesis, not a claim that an isolated exercise has the same evidence as its parent treatment.

Sources: NIMH, 2024; NICE, 2005; U.S. Department of Veterans Affairs & U.S. Department of Defense, 2023; Edinger et al., 2021

What a comparison can establish

A randomized comparison estimates an average difference under the study’s conditions. A waitlist asks a different question from a credible active treatment. A change from baseline alone cannot separate a treatment effect from natural recovery, expectations, concurrent care, or regression to the mean. Standardized effects put different scales on a common statistical metric; they are not percentages of people cured. A confidence interval describes statistical uncertainty and should be read alongside the size of benefit and the patient-important outcome. A result with an interval crossing no difference does not establish equivalence. These interpretation principles follow the Cochrane Handbook (Schünemann, Vist, et al., 2024).

Sources: Schünemann, Vist, et al., 2024

Confidence and applicability

Evidence certainty concerns a defined outcome, comparison, and population. GRADE considers bias, inconsistent findings, indirectness, imprecision, and publication bias. A therapy’s familiar name is not a certainty rating. This review reproduces authors’ published ratings where accessible and otherwise gives a qualified narrative appraisal; it does not create a new formal GRADE assessment. Readers should ask whether the intervention, population, clinician support, and follow-up resemble their question. Separate adult and youth findings, and distinguish symptom change from remission, functioning, quality of life, and recurrence.

Sources: Schünemann, Higgins, et al., 2024

Implications for the apps

Umbrella Journal offers guided journaling and CBT-informed reflection tools; Ella offers AI-guided conversations and check-ins; Nuggets offers journaling and habit support. These are descriptions of functions, not measured treatment outcomes. Research on a complete therapeutic programme can inform a feature’s rationale. Demonstrating that the feature benefits users requires direct evaluation of the implemented product, its intended users, and its actual setting. A clinical topic label or a link to a paper does not close that evidence gap.

Sources: Umbrella Journal, 2026c; Umbrella Journal, 2026a; Umbrella Journal, 2026b

Everyday experiences and applicability

Everyday stress, relationship difficulties, low self-worth, and habit formation are distinct from diagnosed disorders. Choosing a topic or completing a reflection exercise does not establish a diagnosis. This library prioritizes adult evidence and explicitly labels mixed-age findings. Older adults, people with comorbid conditions, and people receiving concurrent treatment are not assumed to have the same outcomes as a general adult sample. Dedicated subgroup reviews and accessibility evaluation remain gaps in this review.