How Do We Know When Psychotherapy Helps—or Harms?

Judging meaningful change, fair comparisons, uncertainty and unwanted effects

Level: Introductory
Reading load: 10 core readings in 5 units
Prerequisites: No clinical training is required. How Does Psychotherapy Produce Change? or How to Read Psychological Science provides useful preparation. Some syntheses use technical statistics; the guides explain the main distinctions.
Last reviewed: September 2026

Introduction

To say that therapy helps is to begin a necessary account. What improved, and did the improvement matter to the person? What would have happened with other care? Was anyone harmed, and did the treatment cause that harm? The passage of time adds a further question: what remains after treatment has ended?

This path develops a way to evaluate those questions. It begins with the outcomes people value and a method for distinguishing change large enough to separate from measurement error from reaching an outcome that matters in a person's life. It then examines control conditions, missing studies and researchers' commitments to particular therapy approaches before turning to unwanted effects and evidence of deterioration. The final pair asks what follow-up trials and routine services can tell us.

Conceptual papers help formulate the questions; empirical analyses show how particular choices affect the answers. The examples include couples, adults with depression, panic disorder and English psychological therapy services. They offer tools for judgment while leaving the limits of each population and setting visible.

By the end, you should be able to state the outcome and comparison behind a treatment claim, distinguish individual change from a treatment effect, identify a plausible bias, and explain what evidence about harms and duration is still missing.

Unit 1 of 5

What counts as a worthwhile benefit?

Cuijpers asks whose goals and which outcomes matter. Jacobson and Truax then make two different judgments explicit: whether a change exceeds measurement noise and whether the resulting status meets a chosen standard. Statistical rules can inform a judgment about benefit without supplying all of its values.

Reading
1 of 10
Targets and Outcomes of Psychotherapy
Cuijpers, P. (2019). Targets and outcomes of psychotherapies for mental disorders: An overview. World Psychiatry, 18(3), 276–285. https://doi.org/10.1002/wps.20661
Foundational argument · Major synthesis · Accessible · Full work: 10 pages
Compare symptoms, functioning, personal goals, unwanted effects and costs as competing parts of an evaluation.
Why you’re reading it. Specify what a favorable treatment claim would need to capture before judging its evidence.
A question to consider. For a person seeking help with panic, propose a symptom outcome and a valued life outcome that could move in different directions.
Reading
2 of 10
What Counts as Meaningful Individual Change?
Jacobson, N. S., & Truax, P. (1991). Clinical significance: A statistical approach to defining meaningful change in psychotherapy research. Journal of Consulting and Clinical Psychology, 59(1), 12–19. https://doi.org/10.1037/0022-006x.59.1.12
Foundational argument · Methodological contribution · Moderate · Full work: 8 pages
Use the couples example to separate reliable change from the cutoff used to classify recovery.
Why you’re reading it. Turn the broad question of benefit into explicit, inspectable measurement decisions.
A question to consider. If a client improves substantially but remains outside the selected reference range, how should that result be described without erasing either progress or remaining difficulty?

Unit 2 of 5

What should treatment be compared with?

A comparator is part of the question. Mohr’s framework introduces the choices; the later meta-analysis examines how those choices and their implementation relate to reported effects. Keep randomization within a trial separate from observational comparisons between differently designed trials.

Reading
3 of 10
Choosing a Control Condition
Mohr, D. C., Spring, B., Freedland, K. E., Beckner, V., Arean, P., Hollon, S. D., Ockene, J., & Kaplan, R. (2009). The selection and design of control conditions for randomized controlled trials of psychological interventions. Psychotherapy and Psychosomatics, 78(5), 275–284. https://doi.org/10.1159/000228248
Foundational argument · Methodological contribution · Moderate · Full work: 10 pages
Connect each control condition to the hypothesis and to the care participants actually receive.
Why you’re reading it. Choose a comparison that can answer the intended question about benefit.
A question to consider. Would adding therapy to usual care answer the same question as comparing it with an equally credible alternative therapy? Specify the claim each design could support.
Reading
4 of 10
How Control Design Changes Treatment Comparisons
Mohr, D. C., Ho, J., Hart, T. L., Baron, K. G., Berendsen, M., Beckner, V., Cai, X., Cuijpers, P., Spring, B., Kinsinger, S. W., Schroder, K. E., & Duffecy, J. (2014). Control condition design and implementation features in controlled trials: A meta-analysis of trials evaluating psychotherapy for depression. Translational Behavioral Medicine, 4(4), 407–423. https://doi.org/10.1007/s13142-014-0262-3
Major synthesis · Methodological contribution · Challenging · Full work: 17 pages
Examine comparator definitions and support for control therapists alongside the effect estimates.
Why you’re reading it. Apply the framework to an evidence synthesis while retaining uncertainty about causal explanations.
A question to consider. A trial reports a large advantage, but only the experimental therapists received supervision: what competing interpretations should a reviewer investigate?

Unit 3 of 5

How dependable is the body of evidence?

Some evidence never reaches journals; other evidence may be shaped by the commitments of its investigators. Driessen traces funded trials to uncover missing results. Munder examines allegiance across reviews. These are different threats and require different evidence and safeguards.

Reading
5 of 10
What Happens When Unpublished Trials Are Included?
Driessen, E., Hollon, S. D., Bockting, C. L. H., Cuijpers, P., & Turner, E. H. (2015). Does publication bias inflate the apparent efficacy of psychological treatment for major depressive disorder? A systematic review and meta-analysis of US National Institutes of Health-funded trials. PLOS ONE, 10(9), Article e0137864. https://doi.org/10.1371/journal.pone.0137864
Methodological contribution · Major synthesis · Moderate · Full work: 23 pages
Track the denominator of funded, initiated, published and recoverable studies before comparing effect estimates.
Why you’re reading it. Understand how assembling a different evidence base changes a treatment claim.
A question to consider. If recovering unpublished trials reduces an effect estimate, what can you now conclude about benefit, and what remains unknown about selectively reported outcomes?
Reading
6 of 10
Researcher Allegiance and Treatment Outcomes
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://doi.org/10.1016/j.cpr.2013.02.002
Major synthesis · Critical reappraisal · Challenging · Full work: 11 pages
Separate the observed allegiance–outcome association from explanations involving bias, expertise or genuine treatment differences.
Why you’re reading it. Evaluate a threat to credibility without replacing evidence appraisal with suspicion about researchers.
A question to consider. Propose one design safeguard and one additional observation that would help distinguish allegiance bias from informed preference for a more effective treatment.

Unit 4 of 5

How should unwanted effects be detected and attributed?

Linden and Schermuly-Haupt broaden the set of outcomes worth noticing and distinguish observation from attribution. Ebert and colleagues then investigate one defined outcome: reliable worsening of depression symptoms. A useful measure can still leave important harms unmeasured.

Reading
7 of 10
Recognizing Unwanted Effects of Psychotherapy
Linden, M., & Schermuly-Haupt, M.-L. (2014). Definition, assessment and rate of psychotherapy side effects. World Psychiatry, 13(3), 306–309. https://doi.org/10.1002/wps.20153
Foundational argument · Methodological contribution · Accessible · Full work: 4 pages
Use the proposed definitions to distinguish an unwanted event, its severity and its possible relationship to treatment.
Why you’re reading it. Build an evaluation that makes room for difficulties outside the primary symptom measure.
A question to consider. If someone reports relationship conflict during therapy while their symptoms improve, what should be recorded before deciding whether treatment caused harm?
Reading
8 of 10
Measuring Deterioration in Guided Self-Help
Ebert, D. D., Donkin, L., Andersson, G., Andrews, G., Berger, T., Carlbring, P., Rozenthal, A., Choi, I., Laferton, J. A. C., Johansson, R., Kleiboer, A., Lange, A., Lehr, D., Reins, J. A., Funk, B., Newby, J., Perini, S., Riper, H., Ruwaard, J., … Cuijpers, P. (2016). Does Internet-based guided-self-help for depression cause harm? An individual participant data meta-analysis on deterioration rates and its moderators in randomized controlled trials. Psychological Medicine, 46(13), 2679–2693. https://doi.org/10.1017/s0033291716001562
Methodological contribution · Major synthesis · Challenging · Full work: 15 pages
Compare deterioration risks, measurement rules and follow-up uncertainty; treat the subgroup analyses cautiously.
Why you’re reading it. Apply reliable-change reasoning to a specific adverse outcome without mistaking it for a complete safety assessment.
A question to consider. How would you revise the claim “this treatment does not cause harm” if its evidence measured only short-term worsening on depression questionnaires?

Unit 5 of 5

Does benefit last and matter in ordinary care?

The panic-disorder trial follows patients through treatment and discontinuation, with changing analysis populations. The IAPT study examines outcomes across services, with a different denominator and a different causal limitation. Together they show why a favorable endpoint is only part of an evaluation.

Reading
9 of 10
Treatment Response and Durability in Panic Disorder
Barlow, D. H., Gorman, J. M., Shear, M. K., & Woods, S. W. (2000). Cognitive-behavioral therapy, imipramine, or their combination for panic disorder: A randomized controlled trial. JAMA, 283(19), 2529–2536. https://doi.org/10.1001/jama.283.19.2529
Landmark empirical study · Methodological contribution · Moderate · Full work: 8 pages
Follow the changing denominators across acute treatment, maintenance and follow-up, using the corrections for affected details.
Why you’re reading it. Judge durability without equating responder-only comparisons with the original randomized comparison.
A question to consider. If relapse is less frequent among one group’s earlier responders, what information about selection, discontinuation and missing data is needed before claiming a lasting advantage?
Reading
10 of 10
What Routine Services Can Tell Us About Outcomes
Clark, D. M., Canvin, L., Green, J., Layard, R., Pilling, S., & Janecka, M. (2018). Transparency about the outcomes of mental health services (IAPT approach): An analysis of public data. The Lancet, 391(10121), 679–686. https://doi.org/10.1016/s0140-6736(17)32133-5
Landmark empirical study · Methodological contribution · Moderate · Full work: 8 pages
Identify who enters the reported outcome rates and distinguish service-level prediction from a tested intervention effect.
Why you’re reading it. Bring measurement and causal scrutiny into the practical question of how ordinary care performs.
A question to consider. Design a short service report that includes improvement, deterioration, access and follow-up, and explain which causal claim it still could not establish.

Scope and omissions

This path teaches evaluation of psychotherapy research, not how to choose or deliver treatment for an individual. It develops measurement, comparisons, evidence selection, attribution of unwanted effects and the interpretation of follow-up and routine-care data. It is not a complete course in clinical-trial statistics.

Several syntheses concern adult depression, where the selected methodological questions are unusually visible. Their conclusions cannot be generalized automatically to other conditions, children, unguided self-help or different health systems. Severe and rare adverse events, patient narratives, costs, equity and long-term functioning need evidence beyond these core works. Optional readings extend harm measurement, allegiance, trial design and follow-up in Indian primary care. Later paths in this area develop adaptation and implementation more fully.

Synthesis

A worthwhile treatment effect joins a meaningful outcome to a credible comparison and an adequate period of observation. Improvement within people is not yet evidence of what treatment caused. Missing studies, reporting choices, unwanted effects, and loss to follow-up can alter the account. The readings make benefit and harm questions to investigate explicitly, including whose experience an average result may conceal.

Evaluate the claim that a psychological treatment produces a worthwhile, lasting benefit. Identify the population, an outcome the person would value, the comparison care and the period of follow-up. Distinguish the observed change within people from the effect attributed to treatment.

Specify a rule for meaningful improvement and one unwanted outcome that the main symptom measure would miss. Explain whose data enter the estimate, what happens to missing observations, and how people who leave early are represented. Identify a plausible design or reporting bias and the evidence needed to investigate it.

Write a conclusion that states what is supported, what remains uncertain and what finding would change your judgment. Then describe the additional evidence needed before transferring that conclusion to another service or population.

Return to How Does Psychotherapy Produce Change? to distinguish evidence of benefit from an explanation of its mechanism. For a deeper examination of measurement, follow What Do Psychological Tests Measure?.

To examine collaboration, empathy, rupture repair and the evidence for relational change, continue with How Does the Therapeutic Relationship Contribute to Change?.