Corrections

These tools are built to be accurate. If something is wrong, please flag it.

Errors in clinical and philosophical content matter. If you notice a factual inaccuracy, a misattribution, outdated information, or a modality or thinker that's been misrepresented, use the form below to let us know. You don't need to have the correct answer — flagging the problem is enough.

You don't need to have the correct answer — flagging the problem is enough.

You can also contact Matthew Sorg directly at matt@matthewsorg.com. Email and name are kept private and only used to follow up on substantive corrections.


Policy

Clinical reference tools carry an obligation to accuracy. When errors are identified — factual, attributional, evidential, or structural — they are corrected promptly and documented here.

Epoché Clinical follows these principles for corrections: errors are never silently revised; substantive corrections are logged with the date, what changed, and which pages were affected; corrections are distinguished from editorial updates (rewording, formatting, added context).


Correction Log

Substantive corrections are documented below. Editorial improvements (rewording, formatting, added context) are not logged unless they change the meaning of clinical information.

8 September 2026 — depression and anxiety continuation
Additional depression and anxiety evidence assessments. A further bounded review checked 22 approach–condition associations and added 21 population-specific assessments: 14 for depression and 7 for anxiety. One supportive-psychotherapy association remains pending because the checked guideline names a different form of treatment. The new records preserve null primary outcomes, active-comparator limits and distinctions between acute treatment and relapse prevention. All 82 earlier assessments are unchanged. The catalogue contains 103 scoped assessments, with 410 approach–condition associations still pending; the wider audit and clinical sign-off remain incomplete.
8 September 2026
Further scoped evidence assessments and study-status corrections. A bounded source review added 43 assessments covering specific populations and outcomes. Flash Technique’s statement that no randomized study existed was corrected; Open Dialogue’s trial-status notes now reflect the published ODDESSI results and their transdiagnostic crisis-care population. MBT’s guideline attribution was corrected, and NARM’s research notes distinguish therapist perceptions from measured patient outcomes. Source links accompany corrected statements. Single-case reports, null trial results and disagreements between guidelines remain explicit. The catalogue now contains 82 scoped assessments, with 431 approach–condition associations still pending. This is partial source coverage, not completion of the wider audit or expert clinical sign-off.
September 2026
Evidence labels now require a sourced assessment. The previous classification rules could infer trial support from a guideline flag or descriptive text, and could transfer an overall label to an unrelated condition. The generated labels on modality and condition pages, comparisons, Evidence Lookup, and the catalogue summary for AI tools now require an explicit assessment. Review progress now distinguishes source checks, condition-specific assessments and expert review. Entries without an assessment or source check display “Assessment not yet completed”; frameworks are identified separately. Incomplete assessment does not mean ineffective. Unreviewed numerical outcome cards and 130 automatically generated evidence sentences in entry introductions have been removed; 130 generated treatment-association sentences now describe catalogue links. Comparative summaries are withheld while their sources are reviewed. Original guidelines and official sources are linked with their recommendation scope. These changes do not complete the review of authored introductions, lineage teaching material, safety guidance or historical claims. The February 2025 change below did not fully resolve the classification problem.
August 2026
Duplicate modality entries merged: Complicated Grief Treatment and Prolonged Grief Disorder Treatment. The two entries were one protocol recorded twice. Both named M. Katherine Shear as founder, both dated 2005, both described the same 16-session manualized treatment, and both linked to the same Columbia training program under its old and new names: Shear's Center for Complicated Grief was renamed the Center for Prolonged Grief when DSM-5-TR adopted Prolonged Grief Disorder in 2022. Because they sat as separate entries, each page carried a superiority claim that implicitly ranked the protocol against itself. The records are now merged at /modalities/prolonged-grief-disorder-treatment/, which keeps both names, both citations, and the outcome data that only the Complicated Grief Treatment record carried; the former URL 301-redirects there. Modality count 152 to 151.
March 2026
Three new interactive tools. Evidence Lookup: select a condition and see every modality ranked by evidence tier with sortable effect sizes and guideline status. Blind Spot Checker: select the modalities you use in practice and discover gaps in condition coverage, epistemological range, tradition representation, and equity/cultural adaptations — with suggested additions. Reading List Generator: filter 464 sources by tradition, condition, or keyword to build curated reading lists with key texts highlighted. 357 total pages, 9 interactive tools.
March 2026
Three new features: global search, knowledge graph, PDF export. Global search (⌘K / Ctrl+K) added across all 297 pages — modalities, thinkers, conditions, and vignettes. Knowledge graph page added at /modalities/graph/ — D3 force-directed visualization of all 151 modalities with 429 connection links, colored by tradition, filterable, zoomable, and clickable. PDF export added to Compare tool — print-optimized stylesheet hides UI chrome and adds attribution. Quick-action entry points added to homepage. Stale modality counts corrected on compare, client-fit, and quiz pages (141→151, 17→16 traditions). 354 total pages.
March 2026
Performance overhaul, taxonomy cleanup, 3 new lineage pages, source URL coverage. Modalities index page reduced from 374KB to 116KB (search data externalized to async load). Compare page reduced from 302KB to 30KB. Google Fonts moved from render-blocking CSS @import to non-blocking preloaded link tags. Google Analytics made async. Three rogue tradition tags fixed (Couples → Family Systems, Eastern-Influenced → Contemplative, Existential-Humanistic → Humanistic) — clean 16 traditions. Three new lineage pages added: Attachment, Contemplative, and Crisis Intervention — all 16 traditions now have lineage coverage (14 pages). Source URLs added to 460 of 464 sources (99% coverage). Stale modality counts on compare, client-fit, and quiz pages corrected from 141 to 151. Sources promoted to main navigation. Source citations added to condition and vignette page templates. Homepage quick-action entry points added. Total: 151 modalities, 22 conditions, 14 lineage pages, 102 thinkers, 464 sources, 353 pages.
February 2025
Guideline-recommended evidence tier now condition-specific. Previously, a modality marked as guideline-recommended for one condition (e.g., EMDR for PTSD) would display as "guideline-recommended" on every condition page where it appeared. The evidence tier logic on condition pages now checks whether the guideline recommendation applies to the specific condition being viewed. A modality is only labeled "guideline-recommended" on a condition page if it has a condition-specific guideline flag (ob[conditionId].gl === true). This prevents misleading evidence claims — for example, EMDR is guideline-recommended for PTSD but not for depression, and the condition pages now reflect this distinction.
February 2025
Content expansion: 7 new modalities, 4 new conditions, 3 new lineage pages. New modalities added: Process Group Therapy, Pluralistic Therapy, Compassionate Mind Training, Rumination-Focused CBT, EFT Tapping, Naikan Therapy, Feldenkrais Method. New conditions: Insomnia, Perinatal Mood Disorders, Adult ADHD, Sexual Dysfunction. New lineage pages: Expressive Arts, Grief & Loss, Integrative & Eclectic. All 148 modalities now have lineage tags and related-modality pairs. Total: 148 modalities, 22 conditions, 11 lineage pages, 87 thinkers.
February 2025
Vignette disagreement callouts added. Each of the 22 clinical vignettes now includes a "Where Approaches Genuinely Disagree" section identifying real clinical tensions between modalities for that case — not just different emphasis, but substantive disagreements about what to do. Examples: whether to expose or stabilize first in panic, whether to explore the meaning of obsessions in OCD, whether to work with the child or the parents in child trauma.
February 2025
Site migration to static pages. All modality pages, condition pages, vignette pages, and philosophy pages migrated from single-page JavaScript application to individually-indexed static HTML pages. Content preserved; no substantive clinical information changed during migration.
February 2025
Condition code fixes. Two modalities (AF-EMDR, Trauma-Sensitive Yoga) had an orphaned condition code `dis` corrected to `diss` (Dissociative Disorders). Neurofeedback and Biofeedback had `adhd` and `tbi` condition codes removed (no corresponding condition entries exist in the dataset).
February 2025
Vignette modality name fixes. Five vignette formulation keys corrected: MI → Motivational Interviewing, Psychodynamic → Short-Term Psychodynamic, Family-Based Treatment → FBT / Maudsley, DBT (Adolescent) → DBT, Complicated Grief Treatment → Prolonged Grief Disorder Treatment.
February 2025
Tradition consolidation. 25 traditions consolidated to 16 for improved navigability. Merges: Jungian → Psychoanalytic, Adlerian → Psychoanalytic, Existential-Humanistic → Humanistic, Transpersonal → Existential, Trauma (AF-EMDR) → Trauma-Focused, Systemic → Family Systems, Feminist → Social Justice, Experiential → Expressive, Eclectic/NLP → Integrative. Original tradition preserved in data as `t_original`.