Does CBT work? Evidence, criticism and limits

In short

• CBT is the most researched psychotherapy, not the "proven best". For depression it barely differs from other bona fide psychotherapies: the difference is g = 0.06, where 0 means no difference.
• The same therapy gives different numbers depending on the opponent. For depression, CBT scores 0.98 against a waitlist, 0.60 against usual care and 0.55 against a placebo pill.
• After corrections the effect is more modest than the marketing. Remove weak control groups and poor-quality trials, account for unpublished studies, and in one reassessment the effect of psychotherapy for depression drops from 0.70 to 0.31. Small, but real.
• The "Dodo bird" is half right. For depression, bona fide therapies are roughly equal on average. For OCD and phobias, exposure-based methods produce very large effects.
• CBT has not "stopped working". The 2015 claim of a falling effect did not hold up in stricter analyses, although the debate is not fully closed.
• Therapy can do harm. Unwanted events are expected in roughly 5–20% of people in psychotherapy; in a UK survey about 1 in 20 reported lasting bad effects.
• Apps help a little; guided online CBT works about as well as face-to-face. For depression, apps show g = 0.28: around 12 people need to use one for one extra person to improve noticeably because of it.
• Skill does not come from a single course. Therapists who learned CBT online without supervision were no better on recorded sessions than therapists with no training at all.
• CBT-based coaching exists, but its evidence is thinner. For work stress, perfectionism and procrastination, CBT methods show small to moderate, sometimes large effects, but the studies are small. Coaching is for development goals; at signs of a disorder, a coach refers to a doctor or psychotherapist.

Timeline

The argument over whether psychotherapy works, and which kind works best, has been running for almost 90 years. Details and sources are in the chapters below.

1. 1936

   Saul Rosenzweig suggests different therapies work through common factors and borrows the Dodo's line from Alice in Wonderland

   Evidence

2. 1952

   Hans Eysenck: by his review, psychotherapy does no better than spontaneous recovery

   Evidence

3. 1975

   Luborsky and colleagues: comparative studies find almost no differences, the "Dodo bird verdict"

   Evidence

4. 1977

   First randomised trials of cognitive therapy for depression

   Cognitive line

5. 1997

   Wampold and colleagues: differences between bona fide therapies cluster around zero

   Evidence

6. 2012

   Hofmann and colleagues review 106 meta-analyses of CBT

   Evidence

7. 2014

   Furukawa and colleagues: a waitlist may act as a nocebo

   Evidence

8. 2015

   Johnsen and Friborg: "the effect of CBT for depression is falling"

   Evidence

9. 2016

   Cuijpers and colleagues: CBT's effects have been "considerably overestimated"

   Evidence

10. 2017

    Two re-analyses find no consistent decline

    Evidence

11. 2018

    The "Was Eysenck right after all?" reassessment: after corrections, psychotherapy for depression is around 0.3

    Evidence

12. 2022

    Umbrella review: psychotherapy and medication give comparable, mostly small effects

    Evidence

13. 2023

    Meta-analysis of 409 CBT trials for depression

    Evidence

14. 2024

    Meta-analysis of 176 trials of smartphone apps

    Evidence

15. 2025

    Metapsy synthesis: psychotherapy across 12 problems and world regions

    Evidence

How to read these notes: evidence labels

Almost every claim here is a claim about research itself, so each label tells you how firmly you can stand on it.

CBT stands for cognitive behavioural therapy. These are the closing notes of the series, and they put CBT itself to the test. Every claim carries a label and a source in brackets, for example proven by research; Cuijpers et al., 2023.

Where sources disagree, both sides are given. Myths are marked with a "Myth vs fact" callout and collected in one table at the end.

How do we know a therapy helped?

One person feeling better after therapy does not prove the therapy did it.

Picture someone with depression who has twelve sessions with a psychologist and then feels better. Did the therapy help? Not necessarily. Depression often lifts on its own. The person may have changed jobs, made up with someone close, or simply made it to spring. That is why research on psychotherapy is built almost entirely on comparison.

The randomised trial

A randomised controlled trial (RCT) works like this. Take, say, 200 people with the same problem and split them into groups by chance: some get the therapy, others get something else or nothing yet. The second group is the control group. Chance is what makes the groups differ only in treatment, not in age, severity or motivation, so any difference in outcome can be credited to the therapy.

Lists of evidence-based treatments, such as the one kept by the Society of Clinical Psychology (Division 12 of the American Psychological Association), rest on RCTs, because only they allow strong conclusions about cause and effect clinical consensus; APA Division 12, FAQ.

Meta-analysis and its relatives

One trial is not enough: its result could be a fluke. So the next step is a meta-analysis, which pools the results of many trials into a single average. It has a well-known weakness: garbage in, garbage out. If the original studies are biased, a meta-analysis neatly adds up their errors too.

There are several variants:

• A network meta-analysis compares several treatments at once, including pairs that were never tested head to head. If treatment A was compared with a waitlist and so was treatment B, A can be compared with B indirectly.
• An individual participant data meta-analysis pools each participant's data rather than study averages. That lets researchers ask who benefits more, for example people with mild or with severe depression.
• An umbrella review is a review of meta-analyses.
• Metapsy is an open, continuously updated database of psychotherapy trials run by researchers from more than 25 universities, led by Vrije Universiteit Amsterdam documented fact; VU Amsterdam, Metapsy. Anyone can use it at metapsy.org.

Sources for this chapter: APA Division 12, FAQ · VU Amsterdam, Metapsy · Furukawa et al., 2014

Effect sizes in plain language

The key number in any therapy study is not "significant" but "how much". It can be translated into shares of 100 people.

What g and d mean

An effect size shows how far the therapy group, on average, ends up from the control group. It is measured in standard deviations, that is, relative to how much people in a group differ from one another anyway. Papers write it as d (Cohen's d) or g (Hedges' g, the same thing with a correction for small samples). Meta-analyses of psychotherapy usually report g.

The psychologist Jacob Cohen proposed rough benchmarks: 0.2 small, 0.5 medium, 0.8 large. According to methodological reviews, Cohen himself saw these cut-offs as arbitrary and advised using them only when a field has no benchmarks of its own clinical consensus; Jané et al., Benchmarks; Cohen, 1988, not opened.

Into percentages

Below, values of g are converted into intuitive measures with standard formulas, assuming a normal distribution. This is our own calculation, not study data; the method was checked against Cuijpers et al. (2016): at g = 0.75 it gives the same NNT of 3.86 as the paper (interpretation; our calculation).

NNT (number needed to treat) is the number of people who need treatment for one more person to improve noticeably than would without it. The lower the NNT, the stronger the treatment. Here NNT assumes 19% of the control group improve, as in Cuijpers et al. (2016); a different assumption gives a different NNT.

The table's main lesson: even with a "large" effect, the two groups overlap by about two thirds. Plenty of people without therapy do better than plenty of people with it. An average shows how the whole group shifts, not what will happen to any one person.

Small is not useless

Cuijpers and colleagues (2014) showed there is no universal threshold for a "clinically relevant" effect. For depression they tentatively put the smallest difference patients themselves notice at around 0.24 proven by research, a tentative estimate; Cuijpers et al., 2014. So an effect of about 0.3 is small but may still be felt.

Response and remission

Besides effect sizes, studies count proportions of people:

• Response usually means symptoms fall by 50% or more.
• Remission means symptoms drop below the threshold for the disorder.
• Deterioration means a reliable worsening, larger than measurement error.

In Hofmann and colleagues' review (2012), response rates to CBT across reviews ranged from 38% for OCD to 82% for body dysmorphic disorder (preoccupation with an imagined flaw in appearance). On waitlists they ranged from 2% for bulimia to 14% for generalised anxiety disorder proven by research; Hofmann et al., 2012.

Sources for this chapter: Cuijpers et al., 2014 · Cuijpers et al., 2016 · Hofmann et al., 2012 · Jané et al., Benchmarks

What hundreds of studies show about CBT

The largest summaries of CBT research agree on the main point: it works, but how well depends on the problem a person brings.

A review of 106 meta-analyses (2012)

Stefan Hofmann and colleagues found 269 meta-analyses of CBT published since 2000 and examined a representative sample of 106 in detail proven by research; Hofmann et al., 2012. Their conclusions:

• The strongest support is for anxiety disorders, somatoform disorders (distressing bodily symptoms not explained by a medical illness), bulimia, anger problems and general stress.
• Among anxiety disorders, effects are large for OCD and specific phobias and at least medium for social anxiety, panic and PTSD. For PTSD, CBT and EMDR (eye movement desensitisation and reprocessing) came out equal.
• For depression: a medium effect against waitlists, mixed results against other active treatments, and possible bias from unpublished studies.
• Limits named by the authors themselves: few studies and small samples for many problems, little long-term data, and almost no meta-analyses on ethnic minorities or people on low incomes.

A meta-analysis of 409 trials for depression (2023)

The largest work to date on CBT for depression is the meta-analysis by Pim Cuijpers and colleagues: 409 trials and 52,702 patients. The authors describe it as the largest meta-analysis of a single type of psychotherapy for a single disorder proven by research; Cuijpers et al., 2023.

Trial quality improved over time: more low-risk-of-bias studies, fewer waitlists, larger samples proven by research; Cuijpers et al., 2023.

And one number that rarely makes the adverts. According to The Mental Elf's review of this meta-analysis, fewer than half the people who received CBT for depression were classed as responders, and around a third reached remission proven by research; Cuijpers et al., 2023, as summarised by Babbage & Loades, 2023.

The 2022 umbrella review

Falk Leichsenring, John Ioannidis and colleagues pooled 102 meta-analyses (3,782 trials, 650,514 patients) of psychotherapy and medication for major adult mental disorders. They included only meta-analyses that assessed study quality and excluded weak comparators such as waitlists proven by research; Leichsenring et al., 2022.

• Psychotherapy vs placebo or usual care: 0.34.
• Medication vs placebo or usual care: 0.36.
• Psychotherapy vs medication head to head: 0.11, not significant.
• Combined treatment vs either alone: a further 0.31.

The authors conclude that after more than 50 years of research, treatment effects remain limited. They speak of a "ceiling" in current research and call for a change of approach (the authors' own account; Leichsenring et al., 2022). Note that this is about psychotherapy and medication in general, not CBT specifically.

The Metapsy synthesis (2025)

In 2025 a team led by Mathias Harrer and Cuijpers pulled the Metapsy database together across 12 problems: 1,029 studies and 85,952 patients. These figures are for psychotherapy in general, not just CBT proven by research; Harrer et al., 2025.

These figures are not corrected for the weak control groups discussed next. They are most useful for comparing problems: psychotherapy has a much bigger effect on phobias and OCD than on psychosis.

The evidence and guidelines for specific disorders are covered in Episodes 7–10 (depression, anxiety, OCD and PTSD, insomnia, eating disorders, addictions, psychosis), and borderline personality disorder in Episodes 11–12.

Sources for this chapter: Hofmann et al., 2012 · Cuijpers et al., 2023 · Babbage & Loades, 2023 · Leichsenring et al., 2022 · Harrer et al., 2025

Who is CBT beating?

An effect size means nothing without "compared with what". The most common opponent in these studies is also the weakest.

One therapy, three different numbers

Control groups come in different strengths. From weakest to strictest:

1. Waitlist: people are told help is coming, in two months.
2. No treatment.
3. Usual care: whatever the person would have received anyway.
4. Psychological placebo: sessions without the "active" techniques.
5. Placebo pill.
6. Active comparison: another proper therapy or a medication.

In 2016 Cuijpers and colleagues sorted 144 CBT studies (11,030 patients) by what CBT was compared with proven by research; Cuijpers et al., 2016:

Against other psychotherapies for depression, as we saw, the figure is 0.06 proven by research; Cuijpers et al., 2023. Same method; only the opponent changes. Numbers resting on 3–5 studies are unreliable.

In anxiety disorders, about 80% of trials compared CBT with a waitlist proven by research; Cuijpers et al., 2016.

The waitlist as a nocebo

A nocebo is a placebo in reverse: getting worse because of expectations. In Furukawa and colleagues' network meta-analysis (2014), people promised nothing at all had better odds of improving than people on a waitlist: an odds ratio of 2.9 (95% confidence interval 1.3–5.7) proven by research; Furukawa et al., 2014.

One possible explanation: someone told "wait, help is coming" looks for help less on their own and changes less in their life (interpretation). But there is a serious caveat: after adjusting for unpublished studies the difference did not hold, and an analysis restricted to high-quality studies was not possible. The authors themselves call the finding tentative (the authors' own account; Furukawa et al., 2014).

Sources for this chapter: Cuijpers et al., 2016 · Cuijpers et al., 2023 · Furukawa et al., 2014

What inflates the impressive numbers

Unpublished studies, weak designs, researchers' loyalties and unblinded assessment all make results look better than they are. This applies to every school of therapy, not only CBT.

Unpublished studies

Publication bias arises when studies with negative results are less likely to be published. A meta-analysis sees only what was published, so the picture looks rosier. Statistical methods can estimate how many studies have gone missing, such as "trim and fill". In Cuijpers et al. (2016) this correction lowered CBT's effect for depression from 0.75 to 0.65, and for generalised anxiety from 0.80 to 0.59. For panic and social anxiety no missing studies were found proven by research; Cuijpers et al., 2016.

Study quality

Of the 144 studies in that paper, only 25, or 17.4%, met all four quality criteria proven by research; Cuijpers et al., 2016.

Today trial quality is judged with the Cochrane RoB 2 tool ("risk of bias"). It asks five questions documented fact; Sterne et al., 2019:

1. Were participants allocated to groups fairly?
2. Did people receive the treatment they were assigned?
3. Is too much data missing?
4. Did the person rating the outcome know who was in which group?
5. Did the authors pick the most favourable of several results to report?

Researcher allegiance

Researcher allegiance means that when a study is run by a supporter of one of the treatments being compared, that treatment tends to win. Thomas Munder and colleagues (2013) pooled 30 meta-analyses and found a moderate link between researchers' allegiance and results: r = 0.26 proven by research; Munder et al., 2013. The link held across treatment types and patient groups.

It cuts every way. Hofmann and colleagues give their own example: the only review in their sample where CBT had a lower response rate than psychodynamic therapy (for personality disorders) was carried out by psychodynamically oriented researchers documented fact; Hofmann et al., 2012. Studies run by CBT's supporters deserve the same caution.

Blind assessment

Assessment is blind when the rater does not know who is in which group. In a meta-analysis of CBT for schizophrenia (Jauhar et al., 2014), the overall effect on symptoms was small, 0.33. In studies where raters knew who had received therapy, it was 0.62; where they did not, 0.15, roughly a quarter of the size proven by research; Jauhar et al., 2014. The authors conclude that CBT has a small effect on the symptoms of schizophrenia, and that it shrinks once sources of bias are taken into account.

"Was Eysenck right after all?"

In 1952 the British psychologist Hans Eysenck published a review concluding that psychotherapy did no better than spontaneous recovery documented fact; Eysenck, 1952; more in Episode 1. Almost 70 years later Cuijpers and colleagues gave their reassessment of psychotherapy for depression a provocative title: "Was Eysenck right after all?" documented fact; Cuijpers et al., 2019.

• Uncorrected, psychotherapy against control groups came to g = 0.70.
• Only 23% of the studies had a low risk of bias.
• After removing waitlists, restricting to good-quality studies and accounting for unpublished ones, the effect fell to 0.31 proven by research; Cuijpers et al., 2019.
• According to The Mental Elf's review of the paper, after the strictest corrections a small significant effect remained for CBT and for supportive counselling; for problem-solving therapy, interpersonal therapy and behavioural activation there were too few high-quality studies proven by research; Cuijpers et al., 2019, as summarised by Gant, 2019.

The authors' answer to their own question is "no": the effect is smaller than it looked, but it exists and is above the threshold patients notice (the authors' own account; Cuijpers et al., 2019).

Sources for this chapter: Cuijpers et al., 2016 · Sterne et al., 2019 · Munder et al., 2013 · Hofmann et al., 2012 · Jauhar et al., 2014 · Eysenck, 1952 · Cuijpers et al., 2019 · Gant, 2019 · Munder et al., 2019

How to read a psychotherapy headline: 10 questions

Everything above adds up to a checklist. Here it is, applied to an invented headline.

An honest version of the same headline might read: "In a small study run by the app's own developers, students who used it reported less anxiety than students who were waiting. No comparison with medication, no long-term data."

If you remember only three questions:

1. Compared with what? People who just waited, or real help?
2. How big is the effect? Is there a number, not just the word "proven"?
3. Who checked? An independent team, blind assessment, a replication?

Sources for this chapter: APA Division 12, FAQ · Cuijpers et al., 2016 · Harrer et al., 2025 · Linardon et al., 2024 · Dash, 2017 · Jauhar et al., 2014 · Leichsenring et al., 2022 · Munder et al., 2013 · Sterne et al., 2019 · Linden & Schermuly-Haupt, 2014

The Dodo bird: are all therapies equal?

For almost 90 years people have argued over whether different psychotherapies work equally well. The answer seems to depend on the problem.

Where the name comes from

In Lewis Carroll's Alice in Wonderland, soaked animals run round in a circle to dry off, with no start and no finish. At the end the Dodo declares: "Everybody has won, and all must have prizes." In 1936 the American psychologist Saul Rosenzweig borrowed the line for a paper on the "implicit common factors" in different psychotherapies. He suggested that different methods work about equally well because they share so much: the relationship with the therapist, hope, a convincing explanation of what is happening to the person documented fact; Rosenzweig, 1936, cited via Wampold et al., 1997; original not opened.

In 1975 Lester Luborsky and colleagues reviewed comparative studies of psychotherapies and found almost no differences in outcome. From then on the "Dodo bird verdict" became a common phrase documented fact; Luborsky et al., 1975, as described in Wampold et al., 1997; original not opened.

Wampold's meta-analysis (1997)

Bruce Wampold and colleagues tested the Dodo hypothesis on head-to-head comparisons of bona fide therapies, meaning therapies intended to be therapeutic, with a theory behind them and a trained therapist. Placebos and "alternative" conditions were excluded. The results clustered around zero. Even under the most generous assumptions, the upper bound for the true difference between therapies was about 0.20. The differences did not grow with publication year, so better methods did not uncover them, and they did not depend on how dissimilar the therapies were. The authors concluded that bona fide therapies are roughly equally effective proven by research; Wampold et al., 1997.

For scale, Wampold cites benchmarks: psychotherapy against no treatment about 0.82, psychotherapy against placebo about 0.48 proven by research, 1990s data; Lambert and Bergin's benchmarks as cited in Wampold et al., 1997.

Objections

Those who think methods do differ replied:

• Wampold himself notes that his result does not mean all therapies work equally well for every disorder and every person the author's own account; Wampold et al., 1997.
• Wampold's "bona fide" therapies included methods that did not need proven efficacy, only an intention to be therapeutic (critics' interpretation, second-hand; Dodo bird verdict, Wikipedia; the originals, Hunsley & Di Giulio, 2002, and Chambless, 2002, not opened).
• For anxiety disorders, a re-analysis of Tolin's data (Baardseth et al., 2013) found a small-to-medium advantage for CBT on measures of the disorder's own symptoms, but no superiority on broader measures proven by research; Baardseth et al., 2013.
• For OCD and phobias, exposure-based methods, meaning deliberately facing what a person fears, produce effects above 1 proven by research; Harrer et al., 2025; Hofmann et al., 2012. Effects that size are hard to explain by common factors alone (interpretation).

Where the Dodo is right, and where it isn't

The upshot: for depression the Dodo seems to be right; for several anxiety disorders, probably not (interpretation; based on Cuijpers et al., 2023; Baardseth et al., 2013). There is no final answer.

The UK's NICE guideline on depression in adults (NG222, 2022) reflects the same picture: instead of one "best" treatment it offers a menu of options for the person to choose from together with a clinician. CBT is one of them clinical consensus; NICE, 2022.

Sources for this chapter: Wampold et al., 1997 · Rosenzweig, 1936 · Luborsky et al., 1975 · Dodo bird verdict, Wikipedia · Baardseth et al., 2013 · Harrer et al., 2025 · Hofmann et al., 2012 · Cuijpers et al., 2023 · NICE, 2022

Relationship or technique?

If therapies are roughly equal, maybe it all comes down to the relationship with the therapist? The honest answer: we don't know yet.

Specific factors are the techniques of a particular method: testing thoughts, exposure, behavioural activation. Common factors are present in every therapy: the relationship, empathy, expectations, a believable explanation of the problem. Common-factors advocates, starting with Rosenzweig, hold that these explain most of what therapy achieves. Wampold develops the idea in his "contextual model" the author's own account; Wampold et al., 1997.

The alliance

The therapeutic alliance is the agreement between client and therapist on goals and on what to do, plus the bond between them. The largest meta-analysis (Flückiger et al., 2018) pooled 295 studies and more than 30,000 patients. The alliance correlates with outcome at r = 0.28; in online therapy the figure is almost the same, r = 0.275 proven by research; Flückiger et al., 2018.

But this is a correlation, a link, not a proven cause. Perhaps the alliance is good because the person is already improving, not the other way round (interpretation).

What it takes to call something a mechanism

Cuijpers, Reijnders and Huibers (2019) reviewed what is known about mechanisms. All studies of mediators, the factors through which therapy affects outcome, are so far correlational. To call a factor a mechanism you need to show it changes before improvement, that more of it means more effect, rule out third causes, confirm it experimentally, and have a strong theory. No common factor and no specific factor meets these criteria yet proven by research, review; Cuijpers et al., 2019, Annual Review.

The same question, what actually changes in therapy, is explored further in Episode 13 on the third wave and process-based therapy.

Sources for this chapter: Wampold et al., 1997 · Flückiger et al., 2018 · Cuijpers, Reijnders & Huibers, 2019

Has CBT stopped working?

In 2015 a paper claimed CBT's effect on depression was falling. Two years later the claim was challenged, and stricter analyses found no decline.

What Johnsen and Friborg found

The Norwegian researchers Tom Johnsen and Oddgeir Friborg analysed 70 studies of CBT for depression conducted between 1977 and 2014 and reported that the effect had declined over the years (proven by research, conclusion disputed; Johnsen & Friborg, 2015, as summarised by Jarrett, 2015, and Dash, 2017).

Their hypothesis: CBT used to be presented as the "gold standard", so people arrived full of hope; as studies began showing it was no better than other methods, hope and faith in it may have faded, and the effect with them (interpretation; the authors' hypothesis as reported by Jarrett, 2015). The hypothesis is unproven.

The responses

• Method. The analysis included non-randomised studies; effects were calculated only "before and after" within groups, with no comparison against a control; only people who completed treatment were counted documented fact; Dash, 2017.
• Cristea et al., 2017. A team including Cuijpers and Hollon kept only randomised trials and added 30 more. Publication year mattered only in some before-and-after analyses and in none of the comparisons with a control group. The trend appeared only in US studies. The authors' conclusion: the linear "fall" is most likely a spurious finding proven by research; Cristea et al., 2017.
• Ljótsson et al., 2017. A Swedish group found that the decline levelled off from about 1995 onwards proven by research; Ljótsson et al., 2017, as described in Friborg & Johnsen, 2017.
• Friborg and Johnsen's reply (2017). The original authors disagreed with their critics and maintain there is a decline (the authors' own account; Friborg & Johnsen, 2017).
• The 2023 meta-analysis. CBT's effect on depression did not fall across publication periods: before 2001, 2001–2010 and after 2011 proven by research; Cuijpers et al., 2023.

There is also a conflict of interest in the debate: some authors of the re-analysis are trained CBT therapists, as The Mental Elf's review points out documented fact; Dash, 2017.

The upshot: depending on the analysis, the effect is falling, fell until the mid-1990s, or is not changing. The stricter analyses, using only randomised trials and comparing with a control, show no decline (interpretation; Dash, 2017; Cristea et al., 2017; Cuijpers et al., 2023).

Sources for this chapter: Johnsen & Friborg, 2015 · Jarrett, 2015 · Dash, 2017 · Cristea et al., 2017 · Ljótsson et al., 2017 · Friborg & Johnsen, 2017 · Cuijpers et al., 2023

Where CBT is weak, and can therapy do harm?

Every method has limits and unwanted effects. Knowing about them is the client's right, not a reason to avoid help.

Where CBT is not the first choice or helps little

• Psychosis. A small effect, smaller still with blind assessment proven by research; Jauhar et al., 2014; little impact on relapse and hospital admissions proven by research; Hofmann et al., 2012. Here CBT is an add-on to treatment led by a doctor.
• Bipolar disorder. Small-to-medium effects; there is little evidence for CBT as a stand-alone treatment, so it is used as an add-on to medication proven by research; Hofmann et al., 2012.
• Opioid and alcohol dependence. CBT does less well than medication-based substitution treatment proven by research; Hofmann et al., 2012.
• Personality disorders. The data are mixed; for borderline personality disorder there are specialised methods, DBT and schema therapy (Episodes 11–12).
• Anxiety in later life. In one review CBT added nothing beyond relaxation training proven by research; Hofmann et al., 2012.
• Depression overall. Even here, where the evidence is richest, fewer than half of people respond proven by research; Cuijpers et al., 2023, as summarised by Babbage & Loades, 2023.

In psychosis, bipolar disorder, severe depression with suicide risk and addictions, CBT can only be one part of treatment led by a doctor (interpretation; based on Hofmann et al., 2012).

Side effects of psychotherapy

The German researchers Michael Linden and Marie-Luise Schermuly-Haupt (2014) proposed these distinctions clinical consensus; Linden & Schermuly-Haupt, 2014:

Examples of unwanted events from their review: new or worsening symptoms, strain in relationships up to divorce, problems at work, stigma, dependence on therapy and the therapist. They estimate that unwanted events, including failure and deterioration, should be expected in roughly 5–20% of people in psychotherapy. Of 132 trials of psychological treatments, only 21% monitored harm, and just 3% described how they collected the data and what the events were proven by research, review; Linden & Schermuly-Haupt, 2014.

The UK survey: 1 in 20

Mike Crawford and colleagues (2016) surveyed 14,587 people who had received psychological treatment for depression and anxiety in 184 services across England and Wales. 763 of them, 5.2% or roughly 1 in 20, reported lasting bad effects from therapy proven by research, survey; Crawford et al., 2016.

Lasting bad effects were reported more often by people from sexual and ethnic minorities and by those unsure which therapy they had received. They were reported less often by people who had been given enough information before therapy started, and by those over 65 proven by research; Crawford et al., 2016. This was a one-off survey: it shows associations, not causes.

Sources for this chapter: Jauhar et al., 2014 · Hofmann et al., 2012 · Cuijpers et al., 2023 · Babbage & Loades, 2023 · Linden & Schermuly-Haupt, 2014 · Crawford et al., 2016

Other cultures, online therapy and apps

Most of the data come from rich Western countries. What do we know about other cultures and about therapy through a screen?

Does CBT work outside the West?

According to the 2025 Metapsy synthesis, 83–86% of psychotherapy data come from rich Western countries. The authors found no sign that psychotherapy works less well in low- and middle-income countries or in non-Western cultures proven by research; Harrer et al., 2025.

Cultural adaptation means adjusting a method to a person's language, values, the role of family, and the way they themselves describe their symptoms. What is known:

• Hall et al., 2016: 78 studies, 13,998 participants. Adapted treatments against a range of other conditions: g = 0.67; against the unadapted version of the same treatment: 0.52, but only 9 such direct comparisons exist. Language, country and ethnic match between therapist and client made no difference to outcome proven by research; Hall et al., 2016.
• Rathod et al., 2018: a review of 12 reviews. Only one thing can be said with confidence: adapted treatments beat usual care; they have rarely been compared with unadapted versions. Adaptation worked better when it took account of the person's values and beliefs and how they describe symptoms; in psychosis, treatments worked better when relatives attended sessions proven by research, review; Rathod et al., 2018.

Guided online CBT

Guided online CBT is an internet programme based on CBT that a person works through with regular support from a clinician. Andersson and colleagues' review (2019) reports an updated meta-analysis of 20 trials comparing it with face-to-face CBT: a difference of g = 0.05, so the effects are equivalent. For anxiety disorders the quality of this evidence is low. Deterioration was 3.4% in treatment groups and 7.6% in control groups proven by research; Andersson et al., 2019.

Karyotaki and colleagues (2021) pooled individual data from 36 studies (8,107 people) to compare online CBT with and without support. Response: 48% with support versus 37% without. With mild symptoms there was little difference; the more severe the depression, the more support helped. By 6–12 months the gap had narrowed proven by research; Karyotaki et al., 2021.

In a survey of 175 organisations in eight European countries, only 1.9% would recommend online CBT alone for severe depression, and 27% a blended format combining face-to-face and online work clinical consensus, survey; Andersson et al., 2019.

Smartphone apps

Linardon and colleagues' meta-analysis (2024) pooled 176 trials of apps proven by research; Linardon et al., 2024:

• Depression: g = 0.28, NNT = 11.5 (33,567 participants).
• Generalised anxiety: g = 0.26, NNT = 12.4 (22,394 participants).
• Effects were larger for apps with CBT features and, for depression, apps with a chatbot.

These are small effects: about 12 people need to use an app for one more person to improve noticeably because of it. According to The Mental Elf's review, apps look better against groups that did nothing than against groups getting some kind of help, so part of the effect may be a "digital placebo". The same review notes that only 6% of the studies met all five quality criteria (interpretation; Valentine, 2024).

Sources for this chapter: Harrer et al., 2025 · Hall et al., 2016 · Rathod et al., 2018 · Andersson et al., 2019 · Karyotaki et al., 2021 · Linardon et al., 2024 · Valentine, 2024

Who can deliver CBT, and how it is learned

The method proven in trials is the method in the hands of trained, supervised therapists. Without that, its effect is not guaranteed.

Why supervision matters

In trials, therapists are trained, follow a manual and receive supervision, regular review of their work with a more experienced clinician. How much this matters was shown in a trial by Rakovshik and colleagues (2016). 61 therapists in routine practice were randomly split into three groups: an online CBT course with a worksheet for reviewing their own work; the same course plus supervision over Skype; or training later, after the study. Competence was rated from recordings of real sessions. The supervised group gained more competence than the other two. The unsupervised group did no better than those not yet trained proven by research; Rakovshik et al., 2016.

What usually fits a psychologist's or coach's training, and what needs clinical qualifications

This is a general framework, not legal advice: requirements for psychologists, coaches and psychotherapists vary by country, and the law is not covered here (interpretation).

Two frameworks apply to any practitioner, whether psychologist, psychotherapist or coach. Confidentiality and its limits. What a client says is not passed on without their consent, but the rule has limits: for example, a serious risk to the life of the client or others, and legal requirements. These are explained honestly at the first meeting, and the specific rules depend on the country's law (interpretation). Dual relationships. When a practitioner is also the client's friend, relative, boss or business partner, it gets in the way of the work and creates a risk of exploitation, so such situations are avoided (interpretation). BABCP's code of ethics, for instance, says that the safety of service users and others must come before other loyalties, that personal and practice-based relationships must be kept clearly distinct, and that the relationship with a service user must not be abused sexually, emotionally, financially or in other ways documented fact; BABCP Standards of Conduct, Performance and Ethics, 2009, revised 2022.

How certification works

Examples of requirements from established organisations documented fact; organisations' websites, November 2026:

• Academy of Cognitive and Behavioral Therapies (A-CBT, USA). Founded in 1996 as a non-profit to certify CBT clinicians. Certification requires a professional degree and a licence for independent practice (where a country has no licensing, documents are reviewed case by case), at least 40 hours of training, at least a year of practice and 10 clients, plus a recording of a real session and a written case conceptualisation, both rated by the academy's reviewers documented fact; A-CBT.
• The Beck Institute (USA) runs tiered certification: clinician, master clinician, supervisor. A theory stage comes first, then an assessment of clinical work documented fact; Beck Institute.
• BABCP (UK). Accreditation as a CBT psychotherapist requires a recognised mental health "core profession", or a portfolio showing equivalent training documented fact; BABCP.
• EABCT, the European Association for Behavioural and Cognitive Therapies, publishes a list of national associations it accredits; in November 2026 it held 32 associations from 25 countries, including Russia, Ukraine, Georgia, Estonia and Turkey. No associations from Azerbaijan or Kazakhstan appeared on its member or accreditation lists at that time documented fact; EABCT.

How to choose a clinician and a training provider

The clinician: questions for the first meeting. It is normal to ask them, and a good clinician answers without irritation (interpretation; compiled from BABCP and A-CBT requirements; Crawford et al., 2016 — people given enough information about therapy beforehand were less likely to report lasting harm):

1. What is your core training, and where did you learn CBT?
2. Do you have supervision at the moment?
3. Do you belong to a professional association, which code of ethics do you follow, and where could I take a complaint?
4. How do you understand my problem, and what method do you suggest? Why that one?
5. Roughly how many sessions, how often, and how long is each?
6. How will we know therapy is helping? Will we measure progress?
7. What might be unpleasant or hard, such as exposure, and what happens if I get worse?
8. Will there be tasks between sessions?
9. What stays between us, who sees the notes, and when would you have to break confidentiality?
10. When would you refer me to a psychiatrist or another specialist?

Warning signs (interpretation): guaranteed results; "a method for everything"; the clinician can't explain what they are doing and why; won't name their training or supervision; puts on pressure — to buy a big block of sessions, to avoid doctors, to stop medication; blurs boundaries — friendship, business offers, romantic involvement with a client; talks about other clients in a recognisable way; goes months without checking whether things are improving, and blames the client when they aren't.

Registers. In some countries you can check a qualification in a public register. One example is the CBT Register of the UK's BABCP: searching by name or membership number shows whether someone holds accredited or registered status. The register is accredited by the UK's Professional Standards Authority, and those on it follow BABCP's code of ethics documented fact; BABCP, CBT Register. Where no such register exists — Azerbaijan currently has no EABCT-accredited association — you are left asking directly about training, hours and supervision (interpretation).

The training provider: a short checklist (interpretation; compiled from the requirements of A-CBT, BABCP, EABCT and Rakovshik et al., 2016):

1. Is the school, or its association, on EABCT's accredited list or a similar body's?
2. Are standards published: hours of theory, hours of supervision, number of supervised clients?
3. Is real work assessed, such as a session recording rated for competence, rather than only a test?
4. Who teaches, and what are their qualifications?
5. Is supervision part of the training?
6. Are there promises like "become a certified CBT therapist in a weekend"?

For free reading on the evidence, try metapsy.org (open trial databases), The Mental Elf reviews, BPS Research Digest, and the APA Division 12 site (div12.org), which lists treatments with their level of evidence.

Sources for this chapter: Rakovshik et al., 2016 · A-CBT · Beck Institute · BABCP · BABCP, CBT Register · BABCP Standards of Conduct, Performance and Ethics · EABCT · Crawford et al., 2016

CBT beyond diagnoses, and cognitive behavioural coaching

What people most often bring to a psychologist or coach, what we know about it, and where the line with therapy runs.

Many people come not with a disorder but with a difficulty: stress at work, perfectionism, low confidence, putting things off, outbursts of anger. CBT models are used here too. But there is less evidence than for treating disorders: studies are often small, frequently with students, and outcomes are rated by participants themselves (interpretation).

Five common requests

• Work stress and burnout. In a meta-analysis of 36 studies (55 programmes, 2,847 people), workplace stress management programmes had a moderate average effect (d = 0.53), and cognitive behavioural programmes the largest: d = 1.16, against 0.50 for relaxation proven by research; Richardson & Rothstein, 2008. The WHO describes burnout as resulting from chronic workplace stress that has not been successfully managed: exhaustion, mental distance or cynicism towards one's job, and reduced professional efficacy. In ICD-11 it is an "occupational phenomenon", not a medical condition documented fact; WHO, 2019. Similar symptoms occur in depression, so persistent exhaustion is worth taking to a doctor (interpretation).
• Perfectionism. In a meta-analysis of 15 randomised trials (912 people, mean age 23), CBT reduced perfectionism (g = 0.57–0.89 across scales) and symptoms of depression (g = 0.60), anxiety (g = 0.42) and eating disorders (g = 0.61) proven by research; Galloway et al., 2022. How perfectionism differs from OCD is covered in Episode 9.
• Self-esteem. In Melanie Fennell's model, low self-esteem rests on a negative overall conclusion about oneself, the "bottom line" ("I'm not good enough"). It is covered over by "rules for living" such as "everything must be perfect"; when a rule is under threat, anxiety and self-criticism rise, and safety behaviours seem to confirm the conclusion (the model author's account; Fennell, as described by Kolubinski et al., 2018). In a review of 7 studies (821 people, all in the UK), self-esteem rose clearly after weekly sessions (1.12) and less after a one-day workshop (0.34). These are before-and-after changes without a control group, so the effect is probably overstated proven by research; Kolubinski et al., 2018.
• Procrastination. In a meta-analysis of 12 randomised trials (718 people), psychological treatments reduced procrastination a little (g = 0.34), and the CBT subgroup moderately (g = 0.55). But that subgroup had only 3 studies and 236 people, with a high risk of bias proven by research; Rozental et al., 2018. How behavioural activation tackles a postponed task is covered in Episode 6.
• Anger. In a meta-analysis of 96 studies, psychological treatments for anger (cognitive, CBT, relaxation, skills training and others) had a moderate average effect (0.76); about 8 sessions appeared adequate, with little added benefit beyond that proven by research; Saini, 2009.

What cognitive behavioural coaching is

Cognitive behavioural coaching uses the same models — the links between thoughts, emotions and actions, Ellis's ABC model (Episode 3), problem solving — but for development goals: a career decision, a negotiation, a presentation, rather than treating symptoms (interpretation). The core books are Michael Neenan and Stephen Palmer's edited volume Cognitive Behavioural Coaching in Practice (2012) and Neenan and Windy Dryden's Cognitive Behavioural Coaching (3rd edition, 2020; originally titled Life Coaching: A Cognitive Behavioural Approach) documented fact; Routledge.

One working framework is Stephen Palmer's PRACTICE model, an adaptation of B. Wasik's seven-step problem-solving sequence: Problem identification; Realistic, relevant goals; Alternative solutions; Consideration of consequences; Target the most feasible solution; Implementation of Chosen solution; Evaluation documented fact; Palmer, 2008.

What the coaching evidence shows

• In a randomised study, group cognitive behavioural, solution-focused life coaching increased goal striving, well-being and hope compared with a waitlist proven by research; Green, Oades & Grant, 2006; sample size and details not verified.
• In a 2022 meta-analysis (20 studies, 957 people), psychologically informed workplace coaching had a large effect on goal attainment (g = 1.29) and a moderate one on self-efficacy (0.59). No significant differences were found between popular approaches, cognitive behavioural coaching included proven by research; Wang et al., 2022.

To be honest: this evidence is weaker than the evidence for CBT as therapy. There are fewer studies, samples are small (the review includes studies of 10–14 people), outcomes are more often self-rated, and comparisons with active controls are rare (interpretation; based on Wang et al., 2022). CBT's effects for disorders don't automatically carry over to coaching.

Where coaching ends and therapy begins

Coaching is for people without a mental disorder who want to reach a goal; therapy is treatment of a disorder (interpretation). The line matters: coaches without mental health training may miss subtle signs of a disorder, and pushing someone with depression towards goals beyond their current capacity can do harm (interpretation; Cavanagh, 2005, as cited in Campone, 2014).

A coach refers to a doctor or psychotherapist when they see (interpretation; based on the table above and Episodes 7–10):

• low mood or loss of interest nearly every day for more than two weeks;
• thoughts of death or self-harm — in that case go straight to If you need help now;
• panic attacks, avoidance that is shrinking someone's life, compulsive rituals;
• alcohol, drugs or food as the main way of coping;
• after-effects of trauma: flashbacks, nightmares;
• a goal that hasn't moved for months because a symptom, not the plan, is in the way.

Coaching and therapy can run in parallel if the client wants this and the professionals have agreed their roles. The coach still does not treat the disorder, and only a doctor prescribes medication (interpretation).

Sources for this chapter: Richardson & Rothstein, 2008 · WHO, 2019 · Galloway et al., 2022 · Kolubinski et al., 2018 · Rozental et al., 2018 · Saini, 2009 · Neenan & Palmer, 2012 · Neenan & Dryden, 2020 · Palmer, 2008 · Green, Oades & Grant, 2006 · Wang et al., 2022 · Campone, 2014

Talking about CBT honestly: the series in one conclusion

CBT passed the test — not as a miracle, but as good treatment with its own limits.

The series began with history (Episode 1): Aaron Beck taught patients not to take a thought at its word but to test it like a hypothesis. Science has done the same to his method. Does CBT work? Yes: it is one of the most researched psychotherapies, and in hundreds of trials it has helped more than no help at all. For anxiety disorders, OCD and phobias it does well. For depression it works moderately, about as well as other bona fide therapies, and as well as antidepressants straight after treatment.

CBT's supporters Daniel David, Ioana Cristea and Stefan Hofmann call it the "gold standard" of psychotherapy, but explain that by this they mean "the best we currently have", not an ideal. They acknowledge that effects against usual care are small to medium, against other active treatments small or zero, and that there is clearly room for improvement (the authors' own account; David, Cristea & Hofmann, 2018). The authors disclose their ties to CBT organisations.

Choosing your words

This matters for more than honesty. Johnsen and Friborg speculated that selling CBT as the "gold standard" may have raised expectations, and that disappointment may then have lowered its effect (interpretation; hypothesis as reported by Jarrett, 2015). The hypothesis is unproven, but it is one more argument for sober framing.

Sources for this chapter: David, Cristea & Hofmann, 2018 · Cuijpers et al., 2016, 2023 · Jarrett, 2015

Myths about CBT: a quick check

Every myth from these notes in one table.

Test yourself

1. Why is CBT's effect bigger against a waitlist than against usual care? A waitlist is a weak opponent, possibly even a nocebo. For depression: 0.98 versus 0.60.
2. What does g = 0.5 mean in plain language? The average person after therapy feels better than 69 of 100 people in the control group, yet the two groups still overlap by about 80%.
3. How does CBT compare with other bona fide psychotherapies for depression? Barely different: g = 0.06, non-significant in most sensitivity analyses.
4. What is the "Dodo bird verdict", and where does it seem not to hold? The claim that bona fide therapies are roughly equal; for OCD and phobias, exposure-based methods produce very large effects.
5. Why was the claim of a "falling" CBT effect challenged? It included non-randomised studies, before-and-after effects without a control and completers only; in analyses of randomised trials compared with a control, there was no decline.
6. What should you ask a clinician before starting therapy? What the method is, what to expect, what might be unpleasant, and how you will both know it is helping.
7. How does cognitive behavioural coaching differ from CBT as therapy, and when does a coach refer on? Same models, but for development goals rather than treatment; the evidence is weaker. At signs of a disorder — persistent low mood, thoughts of self-harm, panic, rituals, dependence — a coach refers to a doctor or psychotherapist.

Key takeaways

• A number without an opponent means nothing. Always ask: compared with what?
• CBT works, moderately. Clearly better than no help, more modestly better than usual care, and about level with other good therapies for depression.
• The evidence is strongest for anxiety disorders, OCD and phobias; weaker for psychosis, bipolar disorder and addictions.
• Therapy can do harm, and that is worth discussing up front.
• The method in the trials is the method in the hands of a trained, supervised clinician.
• CBT-based coaching is for goals, not treatment; its evidence is weaker, and the line with therapy should be clear to both sides.
• The honest formula: "one of the methods with the strongest evidence base", not "the proven best".

Sources

1. Andersson, G., Titov, N., Dear, B. F., Rozental, A., & Carlbring, P. (2019). Internet-delivered psychological treatments: From innovation to implementation. World Psychiatry, 18(1), 20–28. https://mhpsshub.org/wp-content/uploads/2024/09/Andersson-et-al.-2019-Internet-delivered-psychological-treatments-from-.pdf
2. APA Division 12 (Society of Clinical Psychology). Frequently Asked Questions — Research-supported psychological treatments. https://div12.org/?p=3006
3. A-CBT (Academy of Cognitive and Behavioral Therapies). Become a Certified CBT Therapist; A-CBT History. https://www.academyofcbt.org/certification · https://www.academyofcbt.org/a-cbt-history
4. Baardseth, T. P., et al. (2013). Cognitive-behavioral therapy versus other therapies: Redux. Clinical Psychology Review, 33. https://iro.uiowa.edu/esploro/outputs/journalArticle/Cognitive-behavioral-therapy-versus-other-therapies-Redux/9984359947402771
5. Babbage, C., & Loades, M. (2023). Is cognitive behavioural therapy the best we've got for depression? The Mental Elf. https://nationalelfservice.net/?p=163263
6. BABCP. Core Professions. https://babcp.com/Core-Professions/
7. Beck Institute for Cognitive Behavior Therapy. CBT Certification. https://beckinstitute.org/certification/
8. Crawford, M. J., Thana, L., Farquharson, L., Palmer, L., Hancock, E., Bassett, P., Clarke, J., & Parry, G. D. (2016). Patient experience of negative effects of psychological treatment: Results of a national survey. British Journal of Psychiatry, 208(3), 260–265. https://pure.royalholloway.ac.uk/en/publications/patient-experience-of-negative-effects-of-psychological-treatment
9. Cristea, I. A., Stefan, S., Karyotaki, E., David, D., Hollon, S. D., & Cuijpers, P. (2017). The effects of cognitive behavioral therapy are not systematically falling: A revision of Johnsen and Friborg (2015). Psychological Bulletin, 143(3), 326–340. https://scholars.mssm.edu/en/publications/the-effects-of-cognitive-behavioral-therapy-are-not-systematicall/
10. Cuijpers, P., Turner, E. H., Koole, S. L., van Dijke, A., & Smit, F. (2014). What is the threshold for a clinically relevant effect? The case of major depressive disorders. Depression and Anxiety, 31(5), 374–378. https://research.vu.nl/en/publications/what-is-the-threshold-for-a-clinically-relevant-effect-the-case-o/
11. Cuijpers, P., Cristea, I. A., Karyotaki, E., Reijnders, M., & Huibers, M. J. H. (2016). How effective are cognitive behavior therapies for major depression and anxiety disorders? A meta-analytic update of the evidence. World Psychiatry, 15(3), 245–258. https://pmc.ncbi.nlm.nih.gov/articles/PMC5032489
12. Cuijpers, P., Karyotaki, E., Reijnders, M., & Ebert, D. D. (2019). Was Eysenck right after all? A reassessment of the effects of psychotherapy for adult depression. Epidemiology and Psychiatric Sciences, 28(1), 21–30 (online 2018). https://doi.org/10.1017/S2045796018000057
13. Cuijpers, P., Reijnders, M., & Huibers, M. J. H. (2019). The role of common factors in psychotherapy outcomes. Annual Review of Clinical Psychology, 15, 207–231. https://research-portal.uu.nl/en/publications/the-role-of-common-factors-in-psychotherapy-outcomes/
14. Cuijpers, P., Miguel, C., Harrer, M., Plessen, C. Y., Ciharova, M., Ebert, D., & Karyotaki, E. (2023). Cognitive behavior therapy vs. control conditions, other psychotherapies, pharmacotherapies and combined treatment for depression: A comprehensive meta-analysis including 409 trials with 52,702 patients. World Psychiatry, 22(1), 105–115. https://research.vu.nl/en/publications/cognitive-behavior-therapy-vs-control-conditions-other-psychother/
15. Dash, S. (2017). Battle of the meta-analyses: Is CBT becoming less effective over time? The Mental Elf. https://www.nationalelfservice.net/treatment/cbt/battle-of-the-meta-analyses-is-cbt-becoming-less-effective-over-time/
16. David, D., Cristea, I., & Hofmann, S. G. (2018). Why cognitive behavioral therapy is the current gold standard of psychotherapy. Frontiers in Psychiatry, 9, 4. https://www.frontiersin.org/articles/10.3389/fpsyt.2018.00004/full
17. Dodo bird verdict. Wikipedia (secondary source, used to navigate the debate). https://en.wikipedia.org/wiki/Dodo_bird_verdict
18. EABCT. Find a member association; Associations accredited by EABCT. https://eabct.eu/?p=163 · https://eabct.eu/training-and-accreditation/accredited-association
19. Eysenck, H. J. (1952). The effects of psychotherapy: An evaluation. Journal of Consulting Psychology, 16, 319–324. https://psychclassics.yorku.ca/Eysenck/psychotherapy.htm
20. Flückiger, C., Del Re, A. C., Wampold, B. E., & Horvath, A. O. (2018). The alliance in adult psychotherapy: A meta-analytic synthesis. Psychotherapy, 55(4), 316–340. https://www.societyforpsychotherapy.org/the-alliance-in-adult-psychotherapy/
21. Friborg, O., & Johnsen, T. J. (2017). The effect of cognitive–behavioral therapy as an antidepressive treatment is falling: Reply to Ljótsson et al. (2017) and Cristea et al. (2017). Psychological Bulletin, 143(3), 341–345. https://munin.uit.no/handle/10037/12727
22. Furukawa, T. A., Noma, H., Caldwell, D. M., et al. (2014). Waiting list may be a nocebo condition in psychotherapy trials: A contribution from network meta-analysis. Acta Psychiatrica Scandinavica, 130(3), 181–192. https://pesquisa.bvsalud.org/ses/resource/pt/mdl-24697518
23. Gant, E. (2019). Psychotherapy for adult depression: Is it as good as it's cracked up to be? The Mental Elf. https://www.nationalelfservice.net/treatment/psychotherapy/psychotherapy-adult-depression/
24. Hall, G. C. N., Ibaraki, A. Y., Huang, E. R., Marti, C. N., & Stice, E. (2016). A meta-analysis of cultural adaptations of psychological interventions. Behavior Therapy, 47(6), 993–1014. https://wou.edu/wp/ibarakia/files/2018/08/Hall-et-al.-2016-BT.pdf
25. Harrer, M., Miguel, C., van Ballegooijen, W., … Cuijpers, P. (2025). Effectiveness of psychotherapy: Synthesis of a "meta-analytic research domain" across world regions and 12 mental health problems. Psychological Bulletin, 151(5), 600–667. https://pure.amsterdamumc.nl/en/publications/effectiveness-of-psychotherapy-synthesis-of-a-meta-analytic-resea/
26. Hofmann, S. G., Asnaani, A., Vonk, I. J. J., Sawyer, A. T., & Fang, A. (2012). The efficacy of cognitive behavioral therapy: A review of meta-analyses. Cognitive Therapy and Research, 36(5), 427–440. https://pmc.ncbi.nlm.nih.gov/articles/PMC3584580
27. Jané, M. B., et al. Benchmarks (online guide to effect sizes). https://matthewbjane.quarto.pub/Benchmarks.html
28. Jarrett, C. (2015). Cognitive Behavioural Therapy for depression is losing its efficacy. BPS Research Digest. https://bps.org.uk/research-digest/cbt-depression-losing-its-efficacy
29. Jauhar, S., McKenna, P. J., Radua, J., Fung, E., Salvador, R., & Laws, K. R. (2014). Cognitive-behavioural therapy for the symptoms of schizophrenia: Systematic review and meta-analysis with examination of potential bias. British Journal of Psychiatry, 204(1), 20–29. https://researchprofiles.herts.ac.uk/en/publications/cognitive-behavioural-therapy-for-the-symptoms-of-schizophrenia-s/
30. Johnsen, T. J., & Friborg, O. (2015). The effects of cognitive behavioral therapy as an anti-depressive treatment is falling: A meta-analysis. Psychological Bulletin, 141(4), 747–768. https://pubmed.ncbi.nlm.nih.gov/25961373/
31. Karyotaki, E., Efthimiou, O., Miguel, C., et al. (2021). Internet-based cognitive behavioral therapy for depression: A systematic review and individual patient data network meta-analysis. JAMA Psychiatry, 78(4), 361–371. https://mhpsshub.org/wp-content/uploads/2024/09/KAryotaki_Depression.pdf
32. Leichsenring, F., Steinert, C., Rabung, S., & Ioannidis, J. P. A. (2022). The efficacy of psychotherapies and pharmacotherapies for mental disorders in adults: An umbrella review and meta-analytic evaluation of recent meta-analyses. World Psychiatry, 21(1), 133–145. https://doi.org/10.1002/wps.20941
33. Linardon, J., Torous, J., Firth, J., Cuijpers, P., Messer, M., & Fuller-Tyszkiewicz, M. (2024). Current evidence on the efficacy of mental health smartphone apps for symptoms of depression and anxiety: A meta-analysis of 176 randomized controlled trials. World Psychiatry, 23(1), 139–149. https://research.vu.nl/en/publications/current-evidence-on-the-efficacy-of-mental-health-smartphoneapps-/
34. Linden, M., & Schermuly-Haupt, M.-L. (2014). Definition, assessment and rate of psychotherapy side effects. World Psychiatry, 13(3), 306–309. https://pmc.ncbi.nlm.nih.gov/articles/PMC4219072/
35. Ljótsson, B., Hedman, E., Mattsson, S., & Andersson, E. (2017). The effects of cognitive–behavioral therapy for depression are not falling: A re-analysis of Johnsen and Friborg (2015). Psychological Bulletin, 143(3), 321–325.
36. Luborsky, L., Singer, B., & Luborsky, L. (1975). Comparative studies of psychotherapies: Is it true that "everyone has won and all must have prizes"? Archives of General Psychiatry, 32(8), 995–1008.
37. Munder, T., Brütsch, O., Leonhart, R., Gerger, H., & Barth, J. (2013). Researcher allegiance in psychotherapy outcome research: An overview of reviews. Clinical Psychology Review, 33(4), 501–511. https://www.ethicalpsychology.com/2016/01/researcher-allegiance-in-psychotherapy.html
38. Munder, T., Flückiger, C., Leichsenring, F., Abbass, A. A., Hilsenroth, M. J., Luyten, P., Rabung, S., Steinert, C., & Wampold, B. E. (2019). Is psychotherapy effective? A re-analysis of treatments for depression. Epidemiology and Psychiatric Sciences, 28(3), 268–274. https://busqueda.bvsalud.org/portal/resource/es/mdl-30058524
39. NICE (2022). NICE creates new menu of treatment options for those suffering from depression (guideline NG222). https://www.nice.org.uk/news/nice-creates-new-menu-of-treatment-options-for-those-suffering-from-depression
40. Rakovshik, S. G., McManus, F., Vazquez-Montes, M., Muse, K., & Ougrin, D. (2016). Is supervision necessary? Examining the effects of Internet-based CBT training with and without supervision. Journal of Consulting and Clinical Psychology, 84(3), 191–199. https://eprints.glos.ac.uk/10999/
41. Rathod, S., Gega, L., Degnan, A., Pikard, J., Khan, T., Husain, N., Munshi, T., & Naeem, F. (2018). The current status of culturally adapted mental health interventions: A practice-focused review of meta-analyses. Neuropsychiatric Disease and Treatment, 14, 165–178. https://www.dovepress.com/the-current-status-of-culturally-adapted-mental-health-interventions-a-peer-reviewed-fulltext-article-NDT
42. Rosenzweig, S. (1936). Some implicit common factors in diverse methods of psychotherapy. American Journal of Orthopsychiatry, 6(3), 412–415.
43. Simons, P. (2022). Major review finds limited effectiveness for medication and therapy. Mad in America (secondary summary of Leichsenring et al., 2022). https://www.madinamerica.com/?p=229411
44. Sterne, J. A. C., Savović, J., Page, M. J., et al. (2019). RoB 2: A revised tool for assessing risk of bias in randomised trials. BMJ, 366, l4898. https://eprints.whiterose.ac.uk/150579/1/bmj.l4898.full.pdf
45. Valentine, L. (2024). Apps for depression and anxiety: Big new meta-analysis supports effectiveness. The Mental Elf. https://www.nationalelfservice.net/mental-health/depression/apps-for-depression-and-anxiety-big-new-meta-analysis-supports-effectiveness/
46. VU Amsterdam. Metapsy. https://vu.nl/en/research/more-about/metapsy
47. Wampold, B. E., Mondin, G. W., Moody, M., Stich, F., Benson, K., & Ahn, H. (1997). A meta-analysis of outcome studies comparing bona fide psychotherapies: Empirically, "all must have prizes." Psychological Bulletin, 122(3), 203–215. https://doi.org/10.1037/0033-2909.122.3.203
48. BABCP. CBT Register. https://portal.babcp.com/therapist/register
49. BABCP. Standards of Conduct, Performance and Ethics (adopted 16 July 2009; revised 5 December 2022; an updated version takes effect on 1 May 2027). https://babcp.com/about/who-are-babcp/our-policies/standards-of-conduct-performance-ethics/
50. Campone, F. (2014). At the border: Coaching a client with dissociative identity disorder. International Journal of Evidence Based Coaching and Mentoring, 12(1), 1–13 (cites Cavanagh, 2005). https://radar.brookes.ac.uk/radar/file/f0eb1fe8-6d5a-4afd-8c6d-2182e2c52b6d/1/vol12issue1-paper-01.pdf
51. Galloway, R., Watson, H., Greene, D., Shafran, R., & Egan, S. J. (2022). The efficacy of randomised controlled trials of cognitive behaviour therapy for perfectionism: A systematic review and meta-analysis. Cognitive Behaviour Therapy, 51(2), 170–184. https://doi.org/10.1080/16506073.2021.1952302
52. Green, L. S., Oades, L. G., & Grant, A. M. (2006). Cognitive-behavioral, solution-focused life coaching: Enhancing goal striving, well-being, and hope. The Journal of Positive Psychology, 1, 142–149.
53. Kolubinski, D. C., Frings, D., Nikčević, A. V., Lawrence, J. A., & Spada, M. M. (2018). A systematic review and meta-analysis of CBT interventions based on the Fennell model of low self-esteem. Psychiatry Research, 267, 296–305. https://eprints.kingston.ac.uk/41225
54. Neenan, M., & Dryden, W. (2020). Cognitive Behavioural Coaching: A Guide to Problem Solving and Personal Development (3rd ed.). Routledge. https://www.routledge.com/Cognitive-Behavioural-Coaching-A-Guide-to-Problem-Solving-and-Personal-Development/Neenan-Dryden/p/book/9780367461621
55. Neenan, M., & Palmer, S. (Eds.) (2012). Cognitive Behavioural Coaching in Practice: An Evidence Based Approach. Routledge. https://www.karnacbooks.com/product/cognitive-behavioural-coaching-in-practice-an-evidence-based-approach/32563
56. Palmer, S. (2008). The PRACTICE model of coaching: Towards a solution-focused approach. Coaching Psychology International, 1(1), 4. https://www.ukcentreforrebt.com/wp-content/uploads/2019/04/Practice_CPI_Article_1_1_2008.pdf
57. Richardson, K. M., & Rothstein, H. R. (2008). Effects of occupational stress management intervention programs: A meta-analysis. Journal of Occupational Health Psychology, 13(1), 69–93. https://activityinsight.pace.edu/krichardson/intellcont/Richardson%20JOHP%202008-1.pdf
58. Rozental, A., Bennett, S., Forsström, D., Ebert, D. D., Shafran, R., Andersson, G., & Carlbring, P. (2018). Targeting procrastination using psychological treatments: A systematic review and meta-analysis. Frontiers in Psychology, 9, 1588. https://discovery-pp.ucl.ac.uk/10054305/6/Rozental_fpsyg-09-01588.pdf
59. Saini, M. (2009). A meta-analysis of the psychological treatment of anger: Developing guidelines for evidence-based practice. Journal of the American Academy of Psychiatry and the Law, 37(4), 473–488. https://jaapl.org/content/37/4/473
60. Wang, Q., Lai, Y.-L., Xu, X., & McDowall, A. (2022). The effectiveness of workplace coaching: A meta-analysis of contemporary psychologically informed coaching approaches. Journal of Work-Applied Management, 14(1), 77–101. https://www.emerald.com/insight/content/doi/10.1108/JWAM-04-2021-0030/full/html
61. World Health Organization (2019, 28 May). Burn-out an "occupational phenomenon": International Classification of Diseases. https://www.who.int/news-room/detail/28-05-2019-burn-out-an-occupational-phenomenon-international-classification-of-diseases

If you need help now

If your life is in danger, call your local emergency number: 112 in most of Europe, 911 in the US and Canada, 999 in the UK, 000 in Australia. For support lines by country, see Help now.

Image credits

This page uses no third-party images.