Substance and claimed effects

The substance is structured psychotherapy — repeated, hour-long conversations between a trained clinician and a patient, organized around a specific theoretical framework. The four modalities most commonly encountered in 2026 mental-health practice are cognitive behavioral therapy (CBT), eye-movement desensitization and reprocessing (EMDR), internal family systems (IFS), and psychodynamic therapy. Claimed effects: meaningful reduction of depression and anxiety symptoms Cuijpers et al. 2023; remission of post-traumatic stress symptoms de Jongh et al. 2024Mavranezouli et al. 2020; behavior change (avoidance, rumination, substance use, self-harm) Hofmann et al. 2012; improved emotion regulation and interpersonal functioning Shedler 2010; in some chronic-pain contexts, reduced pain and improved physical function Shadick et al. 2013. The entry covers all four modalities together because the dominant signal in the literature — across hundreds of head-to-head trials — is that the differences between them are smaller than the differences within them, and the shared mediator (the therapeutic alliance) carries more variance than the technique Wampold and Flückiger 2023. Scored consequences: mood (primary), short-term health, energy, sleep, focus, longevity, cumulative beauty (via stress / cortisol).

Evidence by addressing question

mechanism

CBT targets the cognition–behavior–emotion loop: identify distorted thoughts ("nobody likes me", "I'll fail this presentation"), test them against evidence, and pair the cognitive work with behavioral experiments (exposure, activity scheduling, response prevention). The mechanism is explicit and modular: each technique maps to a hypothesized maintaining factor of the disorder Hofmann et al. 2012. Behavioral activation alone (no cognitive work) produces effect sizes equivalent to full CBT for depression, suggesting the behavioral lever does much of the work Cuijpers et al. 2017.

EMDR couples imaginal exposure to a traumatic memory with a dual-attention task — typically bilateral eye movements following the therapist's fingers. The dominant mechanism account is the working-memory hypothesis: holding a vivid traumatic image in mind while performing a task that taxes the visuospatial sketchpad degrades the image's vividness and emotional charge, so the next time it's recalled it returns less distressing de Jongh et al. 2024. The bilateral character of the stimulation is not required for the effect; vertical eye movements, tapping, and other working-memory loads produce comparable results de Jongh et al. 2024. Alternative accounts (REM-like memory reconsolidation, parasympathetic activation) exist but remain less directly tested.

IFS models the mind as a collection of parts — exiled wounded child parts, protective manager parts, reactive firefighter parts — surrounding a core Self. Therapy is a guided internal dialogue in which the patient, from Self, gets to know each part, understands its protective function, and "unburdens" the wound the part was protecting. The mechanism is parts-mediated emotion regulation: rather than fighting an emotion or thought, the patient relates to it as a sub-personality with a history Hopwood et al. 2025. The mechanistic literature is thin; most of the evidence is theoretical and case-series rather than controlled mechanistic dismantling.

Psychodynamic therapy works through insight into unconscious patterns — particularly attachment-derived templates that play out in current relationships including, critically, the relationship with the therapist (the transference). Mechanism is interpretive and relational: the therapist names the pattern, the patient experiences it in the room, and the corrective emotional experience of being understood by an attuned other generalizes outward Shedler 2010. Long-term follow-up effect sizes grow after termination (0.59 at end-of-treatment, 0.98 at long-term follow-up), consistent with internalized capacity rather than skill memorization Shedler 2010.

The shared mediator. The therapeutic alliance — measured as the bond between patient and therapist plus their agreement on tasks and goals — predicts outcome across every modality studied, with a meta-analytic correlation of r = 0.28 with treatment outcome (k = 295 studies) Flückiger et al. 2018. The relationship is not just correlational: alliance measured early in treatment predicts later symptom change after controlling for early symptom change Wampold and Flückiger 2023. This is the layer underneath the technique that does much of the work.

evidence

CBT — depression and anxiety. The largest meta-analysis to date pooled 409 trials with 52,702 patients and found CBT effective versus waitlist (g = 0.81) and care-as-usual (g = 0.69), but found no significant superiority over other bona fide psychotherapies for depression Cuijpers et al. 2023. CBT versus pharmacotherapy: equivalent short-term, better long-term Cuijpers et al. 2023. CBT for anxiety disorders has the strongest support of any modality across panic, social anxiety, OCD, GAD Hofmann et al. 2012Carl et al. 2020. Recent placebo-controlled trials show smaller effects (g = 0.24 for anxiety, g = 0.15 for depression) than waitlist-controlled trials — the active control matters.

EMDR — PTSD. Over 30 RCTs support EMDR for PTSD de Jongh et al. 2024. Meta-analytic effect size versus waitlist is roughly g = -1.0 on PTSD symptoms de Jongh et al. 2024. In head-to-head trials versus trauma-focused CBT (TF-CBT) including prolonged exposure and cognitive processing therapy, EMDR is statistically equivalent on PTSD symptom reduction Mavranezouli et al. 2020. NICE, VA/DoD, ISTSS, WHO and APA all recommend EMDR for adult PTSD; APA gives it a conditional rather than strong recommendation, citing fewer high-quality trials than for PE and CPT APA 2017VA/DoD 2023NICE 2018.

Psychodynamic — depression, anxiety, personality disorders. Shedler's 2010 review synthesized eight meta-analyses (160 studies) and reported effect sizes for short-term psychodynamic psychotherapy of d = 0.97 at end-of-treatment, growing to d = 1.51 at long-term follow-up — comparable to or exceeding CBT for depression at long-term Shedler 2010. The Driessen 2015 meta-analysis confirmed efficacy for depression specifically (d = 0.49–0.69 versus controls) Driessen et al. 2015. Long-term psychodynamic therapy is one of the few modalities with controlled evidence for complex personality disorders, particularly borderline personality disorder Leichsenring and Rabung 2011.

IFS — early evidence. The strongest controlled study is Shadick 2013: a randomized trial of IFS versus psychoeducation for rheumatoid arthritis (n = 79) found significant reductions in pain, depression, and self-compassion at 1-year follow-up Shadick et al. 2013. A pilot effectiveness study for PTSD in adult survivors of multiple childhood trauma (n = 17) showed significant pre-post reductions in PTSD and depression with sustained gains at 1 month Hodgetts et al. 2021. A 2025 scoping review identified roughly 30 IFS outcome studies; depression is the only outcome with replicated statistical signal across multiple pilot trials Hopwood et al. 2025. SAMHSA's National Registry listed IFS as evidence-based in 2015, but that designation has been contested as outpacing the trial base.

Common-factor evidence. The "dodo bird verdict" — that bona fide therapies produce equivalent outcomes — has held up across decades of comparative trials Wampold and Flückiger 2023. Specific techniques account for 5–15% of outcome variance; common factors (alliance, expectancy, the therapist, the act of seeking help) account for 30–70% Wampold and Flückiger 2023. Therapist effects — the unexplained variance attributable to who the therapist is, independent of modality — explain another 5–10% of variance, larger than the differences between most pairs of modalities Lambert 2013.

protocol

Standard delivery across modalities: 50-minute sessions, once weekly, individually. A full course of CBT for depression or anxiety runs 12–20 sessions (~3–5 months) Hofmann et al. 2012. EMDR for single-incident adult PTSD frequently completes in 6–12 sessions; complex PTSD takes substantially longer de Jongh et al. 2024. Short-term psychodynamic ranges 16–40 sessions; long-term/intensive psychoanalytic work can run multiple years at multiple sessions per week Shedler 2010. IFS has no standardized session count; courses run from a few months to over a year depending on complexity. The APA's clinical guidance for PTSD specifies that most patients reach symptomatic remission within 8–15 sessions of a trauma-focused therapy when treatment is delivered with fidelity APA 2017.

Between-session work is part of the protocol for CBT (homework: thought records, behavioral experiments, exposure tasks) and is associated with better outcomes; psychodynamic and IFS do not formally assign homework, though processing between sessions is expected. Discontinuation is typically tapered (sessions move to fortnightly, then monthly) rather than abrupt.

contraindications

No absolute medical contraindications to talk therapy itself. Modality-specific cautions: prolonged exposure and EMDR for active psychosis, severe dissociation, or current suicidal crisis without stabilization are flagged by most clinical guidelines; trauma-focused work in those conditions begins with stabilization phases rather than trauma processing APA 2017VA/DoD 2023. Therapy is contraindicated as sole treatment for severe mania, acute psychosis, severe anorexia at medical risk, and active substance dependence requiring detox — psychiatric stabilization or medical management comes first. Deterioration risk on therapy is real: 3–10% of patients end treatment worse than they started, with rates closer to 7–15% in substance-use populations Cuijpers et al. 2019.

misconceptions

The dominant public misconception is that modality choice is the lever — that the question to optimize is "should I do CBT or EMDR?" The literature does not support this framing for most presenting concerns: across hundreds of head-to-head comparisons of bona fide therapies, between-modality effect-size differences are small and inconsistent, while therapist effects and alliance effects are larger and consistent Wampold and Flückiger 2023Lambert 2013. A second misconception: that therapy is for "crisis" or "severe" mental illness only — the largest evidence base is for mild-to-moderate depression and anxiety, with effect sizes comparable to first-line antidepressants Cuijpers et al. 2023. Third: that talking about a problem inevitably helps — supportive listening alone (without structure) performs worse than structured therapies in head-to-head trials for PTSD Mavranezouli et al. 2020. Fourth: that "trauma is in the body and talk therapy can't reach it" — popularized by van der Kolk 2014 — overstates the case; trauma-focused CBT and EMDR (both verbally mediated treatments with imaginal components) consistently show large effects on PTSD biomarkers including cortisol reactivity and amygdala activity.

alternatives

Pharmacotherapy is the main alternative. For moderate-to-severe depression, antidepressants and psychotherapy are roughly equivalent at end of treatment, with psychotherapy holding gains better at follow-up Cuijpers et al. 2018. Combined treatment outperforms either alone Cuijpers et al. 2018. For anxiety disorders, the same pattern holds Carl et al. 2020. For PTSD, trauma-focused therapy outperforms SSRIs and is preferred where access permits APA 2017. Other psychotherapies not the focus of this entry but with strong evidence in specific niches: dialectical behavior therapy (DBT) for borderline personality and chronic self-harm; acceptance and commitment therapy (ACT) for chronic pain and depression; interpersonal therapy (IPT) for depression; motivational interviewing for substance use. Self-guided iCBT shows real but smaller effects (~half the effect size of therapist-delivered CBT) and is reasonable for mild presentations or as a stepped-care first contact. Exercise produces antidepressant effects in the same ballpark as therapy for mild-to-moderate depression.

failure-modes

Common reasons therapy fails despite faithful attendance: (1) poor alliance / poor fit with the specific therapist — the single largest preventable cause of non-response; literature recommends reassessing fit by session 3–5 and changing therapist if the alliance hasn't formed Wampold and Flückiger 2023. (2) Premature dropout — meta-analytic dropout rate is 19.7%, varies little by modality (CBT 18.4%, psychodynamic 20.0%) Swift and Greenberg 2012. (3) Therapist inexperience — variance in outcome attributable to individual therapists is larger than variance between modalities Lambert 2013; trainee-delivered therapy underperforms experienced-clinician therapy. (4) Insufficient dose — bailing after 4–6 sessions before remission shows in clinical guideline expectations. (5) Wrong target — treating depression when the upstream driver is unmanaged sleep apnea, hypothyroidism, alcohol use, or domestic violence. (6) Active medical or social destabilizers — therapy does not work well in parallel with ongoing acute trauma exposure, food insecurity, or housing instability.

practicalities

U.S. private-practice fees in 2024 averaged $174 per 50-minute session for in-person therapy paid out of pocket, with state averages ranging from $122 in low-cost states to $227 in major metros Lipton 2024. With commercial insurance, in-network copays commonly fall in the $15–50 per session range; out-of-network reimbursement typically covers 50–80% of an "allowable amount" below the actual fee. Federally Qualified Health Centers and community mental health centers offer sliding-scale therapy down to $0 for qualifying income brackets. Online platforms (BetterHelp, Talkspace, Cerebral) compress costs to $60–100 per week but trade therapist choice and continuity. Finding an in-network therapist with availability is the friction; the typical waitlist for an in-network psychologist in urban U.S. markets in 2024 was 4–8 weeks. EMDR and IFS practitioners often charge a premium over generalist CBT clinicians due to specialized training (EMDRIA basic training ~ $1,500–2,500; IFS Institute Level 1 ~ $4,000).

stakes

Untreated major depression carries a 10-year mortality elevation of ~60% higher than population baseline after accounting for confounders, driven largely by suicide and cardiovascular causes. Untreated PTSD raises risk of substance dependence, chronic pain, and cardiovascular disease across decades. Untreated anxiety predicts later onset of depression and substance use. The felt-experience stakes: anhedonia, social withdrawal, irritability that costs friendships and jobs, the second drink that turns into nightly habit, the avoidance that shrinks a life by inches over years.

payoff

The trajectory for a successful course: weeks 1–4 build alliance and case formulation; weeks 4–8 see first measurable symptom reduction on standard scales (PHQ-9 dropping ~5 points, GAD-7 similar); weeks 8–16 consolidate; long-term follow-up at 6–12 months shows maintained or growing gains, particularly for psychodynamic and CBT Shedler 2010Cuijpers et al. 2023. Felt experience: the rumination loop quiets first, then the avoidance softens, then people in the patient's life start commenting that they seem different — partner, coworkers, family. PTSD remission for single-incident trauma is often achieved in 8–15 trauma-focused sessions with durable maintenance APA 2017. The "you in five years" projection: a patient who reaches remission and stays in remission has cardiovascular and mortality risk approaching never-depressed baseline over a decade.

The credibility range

Optimist case

Therapy is one of the most-replicated interventions in psychiatry. Pooling across modalities and conditions, the mean effect size versus no-treatment controls is large (d ≈ 0.7–1.0 for depression, anxiety, PTSD), maintained at follow-up, and comparable to or exceeding first-line pharmacotherapy on long-term outcomes Cuijpers et al. 2023Shedler 2010. Trauma-focused therapies produce remission — not just symptom reduction — in the majority of completers within months de Jongh et al. 2024. The mechanism is partially understood (alliance + dose + modality-specific technique), the dose-response is documented, and the most important variable — therapist–patient fit — is modifiable. For the catalogue reader trying to decide whether to start therapy at all, the evidence overwhelmingly favors trying.

Skeptic case

Recent placebo-controlled trials show meaningfully smaller effect sizes than older waitlist-controlled trials (CBT for anxiety: g = 0.24 versus active control). Much of the historical "effect of therapy" is the effect of receiving structured attention from a credentialed expert versus receiving nothing — not an effect of the technique. Publication bias in psychotherapy trials is substantial, allegiance effects (researchers' preferred modality outperforming in their own trials) are large and well-documented Cuijpers et al. 2023. Between-modality differences are small enough that the IFS evidence base — mostly uncontrolled pilots and one RCT in a non-psychiatric population — does not meet the bar for evidence-based status some of its proponents claim Hopwood et al. 2025. Therapy can harm: 5–10% of patients deteriorate during treatment Cuijpers et al. 2019. Dropout is high (~20%) and concentrated in the lowest-resourced patients, meaning the effect-size estimates from completer analyses overstate population benefit Swift and Greenberg 2012.

Author's call

Therapy works — meaningfully, replicably, for the major presenting conditions — and the choice of modality matters less than the choice of therapist. The evidence is strongest for CBT (across nearly every disorder), trauma-focused therapies (CBT and EMDR for PTSD), and psychodynamic for depression and personality presentations. IFS is in a "promising but undercooked" zone; the depression signal replicates but the trial base is too thin to recommend it over established modalities for that indication. The dodo-bird finding survives the placebo critique: even with smaller absolute effects, the consistency of equivalence across modalities is itself the finding. The actionable advice for a reader: start therapy if depression, anxiety, or PTSD is affecting your life; pick a modality that fits how you think (concrete-skills people often land in CBT; trauma-narrative-averse people often land in EMDR; introspective people often land in psychodynamic or IFS); reassess fit by session 5; switch therapists rather than dropping out if alliance isn't forming. Evidence rating: 4. Controversy rating: 3 — modality wars remain active among practitioners despite consistent equivalence findings; the IFS designation and the alliance-vs-technique debate are both live.

Stakeholder + incentive map

• Practitioner guilds (CBT centers, EMDRIA, IFS Institute, psychoanalytic institutes) — each has commercial interest in modality differentiation; certification fees and trainings depend on the perceived distinctiveness of the brand.
• Insurance payers — favor short-course manualized therapies (CBT) with measurable outcomes; reluctant to fund open-ended psychodynamic work; this skews the literature toward CBT-style trials with brief courses.
• Pharmaceutical industry — competes with therapy for the same patients; well-funded SSRI head-to-head trials versus shoestring-budget therapy trials creates asymmetric evidence.
• Government / clinical guideline bodies (APA, NICE, VA/DoD, ISTSS, AHRQ) — convergent in recommending CBT, EMDR, and trauma-focused therapies as first-line for their target conditions; conservative on novel modalities.
• Trauma-recovery community / lay public — heavily influenced by van der Kolk 2014 and the somatic-therapy ecosystem; tends to overweight body-based and parts-based modalities over traditional CBT.
• Skeptic / debunker community — points to dodo-bird, allegiance effects, and shrinking placebo-controlled effect sizes to argue that the therapy industry oversells the technique.

Population variability

Severity. Mild presentations respond well to self-guided iCBT, bibliotherapy, and brief interventions; severe presentations benefit more from in-person modality with experienced clinician and often combined pharmacotherapy Cuijpers et al. 2018. Complex PTSD (sustained childhood abuse, attachment trauma) responds less robustly to manualized short-course trauma-focused therapy; phase-based approaches (stabilization → trauma processing → integration) and longer courses are favored; IFS and EMDR have a strong clinical following here even though RCT evidence is thinner than for adult-onset single-incident PTSD. Personality disorders respond better to specialty long-term therapies (mentalization-based, transference-focused, DBT, schema therapy) than to short-course CBT. Age: CBT efficacy holds in youth and older adults; long-term psychodynamic work is overrepresented in young- and middle-adulthood. Cultural fit: most RCT samples are WEIRD (Western, educated); equity-of-access and equity-of-outcome data are weaker for racial and ethnic minorities in U.S. trials. Treatment-resistant cases: 30–40% of patients do not respond to a first adequately-delivered course of evidence-based therapy; switching modality or adding pharmacotherapy is the standard next step rather than continuing the same approach indefinitely.

Knowledge gaps

• Mechanism dismantling — most modalities have not been mechanistically decomposed. We know therapy works; we don't reliably know which component drives the effect in any individual case.
• Matching — there's no validated algorithm to match patient to optimal modality or therapist; the field's best advice remains "try one, change if it isn't working." Patient-preference and shared decision-making models are emerging but lack predictive validity.
• IFS evidence base — the modality's popularity is far ahead of its trial base; a multi-site RCT for IFS in depression or PTSD would substantially clarify the picture Hopwood et al. 2025.
• Telehealth equivalence — early evidence suggests video-delivered CBT and EMDR are non-inferior to in-person for most conditions, but long-term comparative data are still accumulating.
• Negative-effects measurement — most trials do not routinely measure adverse events; field-wide adoption of a standardized adverse-event protocol is overdue Cuijpers et al. 2019.
• Common-factor optimization — if alliance and therapist effects dominate outcome variance, the highest-leverage research is on therapist-training and feedback systems (e.g., routine outcome monitoring, deliberate practice). This is underfunded relative to modality-specific trials.

Scope vs. brief. The topic brief named CBT, EMDR, IFS, and psychodynamic plus four consequence dimensions (mood, anxiety, trauma, behavior) and the therapeutic alliance as shared mediator. The article covers all four modalities and all four consequences, with alliance positioned as the shared mediator in mechanism and again in the dodo-bird callout in evidence. No silent narrowing.

One-entry-vs-many call. Considered splitting into per-modality entries (CBT, EMDR, IFS, psychodynamic each as its own entry). Rejected because the dominant empirical finding — modality equivalence and alliance-as-mediator — only lands if the modalities are presented side by side. A reader who arrives at a CBT-only entry will leave thinking modality is the lever, which is the misconception this entry's job is to correct. Per-modality deep-dives may still warrant separate entries later (see "Separate-entry candidates" below).

Rating difficulties.

• beauty_cumulative. Dropped from 1 to 0. Long-term stress / cortisol reduction has a plausible chain to appearance but the article doesn't address appearance and the magnitude doesn't honestly clear the "real but slow contribution" bar. Cleaner zero than performative one.
• cost_burden 3 vs. 4. Out-of-pocket weekly at ~$174 × 50 weeks is ~$8.7k/year (tier 4). With in-network insurance copays it's ~$750–2,500/year (tier 2–3). Settled on 3 as the modal experience for an insured U.S. patient using in-network; an unflagged 4 would mislead the population that has access. Considered scoping audience to "no insurance" but the article applies broadly enough that audience scoping shrinks reach more than it clarifies.
• evidence 4 vs. 5. The trial base for CBT (and to a lesser extent EMDR, psychodynamic) easily clears the 5 anchor (multiple large RCTs, guideline-backed, consistent). What kept it at 4: recent placebo-controlled meta-analyses show meaningfully smaller absolute effects than older waitlist-controlled trials, and the IFS leg of the entry is genuinely pilot-tier. Rating reflects the floor across the four modalities, not just CBT's ceiling.
• controversy 3. The dodo-bird verdict, alliance-vs-technique, IFS designation, and EMDR's APA conditional rating are all live debates among reasonable specialists. Not 4 — there isn't a foundational fight about whether therapy works; the fight is about which mechanism story explains the effect.

Excluded by design.

• DBT, ACT, IPT, motivational interviewing — mentioned in alternatives only. Each warrants its own entry; DBT especially given the chronic self-harm / borderline population it serves.
• Couples and family therapy, group therapy formats — adjacent modalities with distinct mechanism profiles; flagged in out-of-scope, not covered.
• Ketamine-assisted and psilocybin-assisted therapy — separate intervention class with regulatory and dosing concerns that don't fit the talk-therapy frame.
• CBT-I (insomnia-specific) — flagged in out-of-scope; deserves its own entry under sleep.
• Complex PTSD treatment in depth — touched in population variability in research dossier but not in article body; merits a dedicated entry once we have a trauma category structure.

Separate-entry candidates for the backlog.

• CBT (standalone) — the evidence base supports a deep-dive entry covering disorder-by-disorder protocols, behavioral activation, exposure work, homework mechanics.
• EMDR (standalone) — the working-memory mechanism, the 8-phase protocol, and the bilateral-stimulation controversy each warrant more depth than this overview allows.
• The therapeutic alliance — given Wampold's argument that alliance is the highest-leverage variable, this could be a meta-entry on what a patient should look for / how to assess fit / how to repair alliance ruptures.
• DBT — strong evidence base, specific population, and the only therapy with reasonable RCT evidence for chronic suicidality.
• Antidepressants (SSRIs / SNRIs) — referenced repeatedly in alternatives as the comparison; needs its own entry.
• Finding a therapist (practical guide) — the fit / search / first-3-sessions material could carry a standalone practical entry.

Future links to wire in. Once the above entries exist, this one should cross-link in alternatives (antidepressants, DBT, ACT, IPT, exercise-for-depression), out-of-scope (CBT-I, ketamine, group therapy), and protocol (finding a therapist).

Hard call on tone. The dek and highlights both lead with the alliance-and-fit point rather than with a modality-specific pitch. This is the editorial bet: if a reader takes away only one thing, it should be "shop for the therapist, not the brand." If a reviewer wants the dek to open instead with the mood effect-size headline, that's a defensible alternative and worth flagging.