Sign Up
Login
Try for Free
Login
The Y-BOCS (Yale-Brown Obsessive Compulsive Scale) is a clinician-rated measure of OCD severity published by Goodman and colleagues in 1989, scored 0 to 40 across obsession and compulsion subtotals, with a 0 to 50 second edition and child versions. OCD specialists, therapists, and prescribers use it for baselines and treatment tracking. It rates severity; it never diagnoses. This page covers how to document Y-BOCS results, with a fictional sample note.
Clinician-administered semi-structured interview (psychologists, psychiatrists, therapists trained on the scale); self-report and parent-report adaptations exist and are charted as such; use requires the rights holder's permission
OCD specialty and ERP programs, prescribers, utilization reviewers and payers, neurosurgical and DBS programs, clinical trials
3 to 8 chart lines per administration · severity interview about 15 to 30 minutes, longer with the checklist at intake (clinical team estimate)
Clinician-rated OCD severity scale (10 items scored 0 to 4; obsession and compulsion subtotals 0 to 20; total 0 to 40; second edition 0 to 5 per item, total 0 to 50)
OCD baseline severity, ERP and medication response tracking, level-of-care and medical-necessity documentation, neurosurgical referral, trial endpoints
Licensed intellectual property of OCD Scales, LLC: paper use needs permission, EHR, app, and web use need a paid license; response and remission definitions are expert consensus, not law
The Yale-Brown Obsessive Compulsive Scale is a clinician-administered, semi-structured measure of current OCD symptom severity, published in two 1989 Archives of General Psychiatry papers by Wayne Goodman, Lawrence Price, Steven Rasmussen, and colleagues. It has three working parts: a symptom checklist that establishes which obsession and compulsion content is present, a target list that names the symptoms the interview will rate, and a severity scale of ten items scored 0 to 4, five for obsessions and five for compulsions, giving two 0 to 20 subtotals and a 0 to 40 total for the past week. The checklist is never added into the total; the scale was built so that people with different symptom content can be compared on aggregate burden. The Y-BOCS-II (Storch, Rasmussen, Price, and colleagues, 2010) widened each item to 0 to 5 for a 0 to 50 total to relieve the ceiling, replaced the resistance-to-obsessions item (confusing for patients taught in ERP not to resist) with an obsession-free interval item, and folded avoidance into severity; the CY-BOCS (1997) and CY-BOCS-II (2019) adapt the interview for children and adolescents with caregiver input.
Two facts govern the documentation. First, the number is a severity rating, not a diagnosis, a level-of-care rule, or a remission declaration: no band establishes OCD, a low score after treatment does not undo an established diagnosis, and a high score on a self-report adaptation does not create one. Second, the scale is licensed intellectual property. As stated at ocdscales.com in August 2026, any use requires explicit permission, noncommercial paper use for a clinician's own patients requires prior permission from Dr. Goodman, and a paid licensing agreement is required for commercial use, EHR incorporation, apps, and other digitalization, which the terms extend to ambient-AI capture of a paper administration; the many forms and calculators circulating online are not evidence of authorization. Charts therefore record scores, subtotals, targets, and interpretation without pasting item wording. Serial results follow the outcome measure note conventions, the score sits inside a psychotherapy progress note or medication management note, and the goals it anchors live in the treatment plan.
OCD specialty clinics and ERP therapists use it as the severity anchor at intake and at treatment reviews, prescribers use it to document medication trials against a baseline, and intensive outpatient, partial hospital, and residential programs use it to communicate severity and trajectory to utilization reviewers, who apply generic level-of-care criteria that name no instrument. Neurosurgical and deep brain stimulation programs use it in eligibility workups, where three different authorities are routinely conflated (the FDA indication has no number, FDA labeling says benefit below 30 was not established, and the International OCD Foundation's clinical criteria use 28 or higher). Clinical trials use it as the standard primary outcome with protocol-defined responder rules. In Canada, Ontario Health's OCD quality standard names it as the recommended severity and monitoring instrument without setting any admission threshold, and Ontario's out-of-country funding checklist asks for standardized severity results with the CY-BOCS as an example; in Australia, Better Access requires an outcome measure of the clinician's choosing and OCD programs describe suitability for ERP, readiness, and risk rather than a cutoff. When a score is charted, it belongs in a psychotherapy progress note with the ERP work it informs, and the level-of-care rationale it supports belongs in the prior authorization request and the treatment plan review.
No US, Canadian, or Australian law prescribes a Y-BOCS note format, and no scale authority makes a band a diagnosis or a care level. What survives review is a record that names the version and administration mode, keeps the checklist and targets as the referent for the severity ratings, reports the total with both subtotals and denominators, names the band framework if one is used, states change with the definition it invokes, and pairs the number with function, avoidance, and insight. Each element below carries the pitfall that most often undermines it.
Version and mode. Chart original Y-BOCS (0 to 40), Y-BOCS-II (0 to 50), CY-BOCS, or CY-BOCS-II, and whether it was a clinician interview, a self-report adaptation, or a parent report, with the date, rater, informants, and the past-week rating window. Pitfall: A Y-BOCS-II charted as 0 to 40 (or original bands pasted onto a 0 to 50 result), a self-report total spliced into a clinician-rated series, or a raw score trended across editions as if the metric had not changed.
Checklist, targets, then severity. At intake, establish content with the symptom checklist and name the target symptoms before rating severity; at follow-up, review and update the targets so the ratings have a documented referent. Record the current content areas (for example contamination, responsibility and checking, covert reviewing, reassurance seeking) in your own words. Pitfall: A severity score with no identified symptoms behind it, or the checklist treated as optional at initial administration.
Total with both subtotals. Report x/40 with the obsession and compulsion subtotals and their denominators (obsessions 11/20, compulsions 9/20, total 20/40); for the second edition, x/50. Keep the arithmetic to the ten severity items; the checklist never enters the total. Pitfall: "Y-BOCS 20" with no denominator, no subtotals, and no edition, which cannot be compared with anything later.
Bands named by framework. If you use a severity band, name it: the legacy original-scale bands printed by the rights holder (0 to 7 subclinical, 8 to 15 mild, 16 to 23 moderate, 24 to 31 severe, 32 to 40 extreme); the 2022 empirical original-scale bands from 5,140 people (0 to 13, 14 to 21, 22 to 29, 30 to 40); or the 2025 empirical second-edition bands (0 to 14, 15 to 21, 22 to 34, 35 to 50). A 28/40 is "severe" under one framework and "moderate" under another, so report the raw score and the framework. Pitfall: "Y-BOCS 28, severe OCD" with no framework, "0 to 7 means no OCD," or 16 treated as a diagnostic cutoff (the 2022 study called that trial-entry threshold arbitrary).
Change with its definition. State the baseline date and score, the current score, absolute change, and percentage reduction (baseline minus current, divided by baseline), on the same edition, mode, and interval. Then say which convention you invoke: the 2016 expert consensus defines full response as at least 35% reduction plus CGI-I of 1 or 2 sustained at least a week, partial response as 25% to under 35% with CGI-I no worse than 3, and remission as no longer meeting criteria or a total of 12 or less with CGI-S 1 or 2; a 2024 individual-participant meta-analysis found 30% and 15 statistically optimal and still endorsed the consensus. Pitfall: "36% reduction, responder" with no CGI anchor, remission inferred from a percentage alone, or 25% described as more conservative than 35% (it is less stringent).
Function, avoidance, insight, accommodation. Document work, school, relationships, self-care, and community function; avoided settings and delegated tasks; covert rituals and reassurance seeking; family accommodation; and insight, separately from the total, because the original scale under-weights avoidance and the core ten items do not directly capture insight. Pitfall: A falling total read as restored function while the patient still cannot use the kitchen or attend work, or insight inferred from the number.
Decision link, without item text. Tie the result to the ERP target or hierarchy change, the response-prevention focus, medication continuation or optimization, the level-of-care rationale, or a diagnostic reassessment when score and function diverge; and keep item wording, anchors, and forms out of the note, which the license does not permit you to reproduce. Pitfall: A number filed with no consequence, or a chart template that pastes the interview prompts into the record.
Y-BOCS DOCUMENTATION BLOCK Date: [ ] Setting: [ ] Rater: [ ] Informants: [ ] Version: [original Y-BOCS (0-40) / Y-BOCS-II (0-50) / CY-BOCS / CY-BOCS-II] Mode: [clinician interview / self-report adaptation / parent report] Rating window: [past week] Assessment interval: [ ] Checklist / targets: [reviewed at intake / updated today; current content areas in own words: ...] Score: obsessions [ ]/20 compulsions [ ]/20 total [ ]/40 (or /50) Band (framework named): [legacy original bands / 2022 empirical original bands / 2025 second-edition bands]: [ ]; not a diagnosis Baseline: [date] [ ]/[ ] Change: [ ] points, [ ]% reduction Response classification (definition named): [2016 consensus: full (35%+ with CGI-I 1-2, sustained 1 week) / partial (25% to under 35%, CGI-I 3 or better) / remission (no longer meets criteria, or 12 or less with CGI-S 1-2) / protocol-specific / numerical change only] CGI-I / CGI-S: [ ] Duration sustained: [ ] Function, avoidance, insight, accommodation: [work, school, relationships, self-care; avoided settings; covert rituals; reassurance; family accommodation; insight] Decision link: [ERP targets / response prevention / medication / level of care / diagnostic reassessment] Next administration: [date + same version and mode] Clinician signature / credentials: Date:
Free to use and share, no signup. The PDF includes a one-page cheat sheet with element-by-element pitfalls and a pre-sign checklist; the DOCX is the blank documentation block, ready to adapt. Neither reproduces the scale items, checklist, anchors, or forms.
Scenario: a scheduled review during outpatient ERP, with a clinician-administered original Y-BOCS compared to the intake baseline, the change classified under the 2016 consensus with its CGI anchor, persistent avoidance documented separately from the falling score, and the treatment plan adjusted. All details are fictional.
Patient: K.P., 31 · Setting: Outpatient OCD specialty clinic, ERP session 7 review · Clinician: N. Adeyemi, PhD · Note date: 08/12/2026
Measure: Y-BOCS, original 1989 version, clinician-administered by the treating psychologist, past-week window; symptom targets reviewed and updated before rating. Today: obsessions 11/20, compulsions 9/20, total 20/40. Baseline at intake 07/01/2026 (same version and mode, checklist completed): obsessions 15/20, compulsions 13/20, total 28/40. Change: 8 points, a 28.6% reduction over six weekly ERP sessions. Under the legacy original-scale bands today's total sits in the moderate range (baseline severe); under the 2022 empirical bands both fall in the mild-to-moderate boundary region. Raw scores govern; the band is descriptive only.
Symptom targets and function: Current content areas: contamination-related threat with hand-washing, responsibility and checking of appliances, covert mental reviewing, and reassurance seeking from her partner. Ritual frequency and duration reduced by her report and partner report; covert reviewing persists in the evenings. Shared-kitchen avoidance remains marked (still preparing no meals at home, eating out or relying on her partner), and she has not yet returned to the two client sites she stopped visiting in the spring. Insight fair: recognizes the feared outcomes as unlikely most days. Family accommodation: partner still confirms appliance status on request several times a day.
Response classification: CGI-I rated 3 (minimally improved), sustained for two weeks. Consistent with the 2016 expert-consensus definition of partial response (25% to under 35% reduction with CGI-I no worse than 3); does not meet full response (35% or more with CGI-I 1 or 2) or remission. Score change reflects reduced overt ritualizing more than restored function; avoidance and accommodation are carrying part of the improvement.
Plan: Continue weekly ERP with the hierarchy re-weighted toward avoided settings (home kitchen meal preparation, the two client sites) and covert reviewing, and a response-prevention agreement with the partner to taper appliance reassurance starting this week. Between-session practice logged daily. Medication unchanged pending the prescriber review scheduled 08/26/2026; this score and the functional summary sent to the prescriber today. Next Y-BOCS at session 13, same version and mode. Outpatient level of care remains appropriate: engaged, completing practice, no safety concerns.
Administrative: Score, subtotals, version, mode, and CGI entered in the outcome-measure record; no item wording or form content stored in the note. Instrument administered under the practice's permission from the rights holder.
This sample is fictional and for educational purposes. It does not describe a real patient or record; the scores, dates, and details are invented to show documentation structure and are not clinical guidance.
Writing these after every session? BastionGPT drafts complete notes from bullets, dictation, or a transcript.
Generate a note from bulletsUnited States: no law requires the Y-BOCS, and the pressure comes from payers, programs, and one device authorization. Medicare's outpatient psychotherapy documentation expectations (PAYER POLICY) call for an individualized plan, diagnosis and goals, type and frequency of services, periodic evaluation, progress, and a reasonable expectation of improvement, and neither CMS nor a current contractor article names the Y-BOCS; a 2026 UnitedHealthcare residential policy requires a need for 24-hour structure, an individualized evidence-based plan with weekly updates, active treatment, a reasonable expectation of further improvement, and evidence of likely deterioration at a lower level, again with no score threshold, so specialty programs that use a Y-BOCS cutoff should attribute it to their own criteria, not to the instrument or the payer (CONVENTION). Do not assume 96127 applies to a clinician-administered severity interview or that its payment rules are uniform; record the service actually performed, the instrument and mode, the result and interpretation, and verify the payer (PAYER POLICY). Deep brain stimulation is where three authorities get collapsed into one number: the FDA humanitarian device exemption for bilateral stimulation of the anterior limb of the internal capsule (H050003, February 2009, adjunct to medication and alternative to anterior capsulotomy for chronic, severe, treatment-resistant OCD in adults who have failed at least three SSRIs) contains no Y-BOCS number in its indication (LAW); FDA labeling states that safety and probable benefit were not established below a Y-BOCS of 30, reflecting the 26-patient experience that selected above 30 (LAW, an evidence boundary); and the International OCD Foundation's clinical criteria use a Y-BOCS of 28 or higher with at least five years of primary OCD, three serotonin-reuptake inhibitor trials including clomipramine, an antipsychotic augmentation attempt, and at least 20 ERP sessions (CONVENTION). Two 2026 HDE supplements concerned manufacturing and process matters, not the indication. A neurosurgical referral note keeps those three apart and adds the insurer's own requirements. Trials use change from baseline as the continuous endpoint with protocol-defined responder rules; the historic entry threshold of 16 on the original scale was described as arbitrary by the 2022 multinational analysis, and the protocol, not a scoring website, governs the endpoint (CONVENTION).
Canada and Australia name no threshold either, and the evidence rewards modesty about bands and change. Ontario Health's 2020 OCD quality standard names the Y-BOCS as the recommended severity and monitoring instrument (CONVENTION), Ontario's February 2024 out-of-country funding checklist requires specialist confirmation of medical necessity with diagnosis, comorbidity, standardized severity results (CY-BOCS given as an example), function, prior medication and psychotherapy, and outcomes but sets no cutoff (PAYER POLICY), Canada's leading intensive program screens admission severity with the Florida Obsessive Compulsive Inventory rather than a Y-BOCS number, and CADTH characterized DBS for OCD as emerging in 2023, with a Canadian research protocol using a minimum of 24. Australia's Better Access guidance (March 2026) requires an outcome tool of the clinician's choosing and does not name the Y-BOCS (PAYER POLICY), and public program descriptions turn on diagnosis, suitability for ERP, readiness, and risk (CONVENTION). On the evidence: the 1989 rater study (four raters, 40 patients) reported excellent inter-rater reliability, a 2015 synthesis of 144 studies found mean alpha .866, test-retest .848, and intraclass correlation .922 with real variation across samples, and the CY-BOCS validation reported alpha .87 with a compulsion-subtotal intraclass correlation of .68 in youth; the second edition's initial study (130 adults) reported severity alpha .89 and reliability coefficients above .85, and a 2025 Spanish Y-BOCS-II study of 1,805 adults found alpha .92 with only marginal fit for the usual two-factor model. The 2022 empirical bands (5,140 people aged 5 to 82) achieved only about 55% to 57% exact category agreement in validation samples and low negative predictive value at the diagnostic boundary, so bands are descriptors, not partitions, and a low score cannot rule OCD out. Self-report adaptations agree only moderately with clinician totals (correlations around .45 to .65 in one carefully diagnosed sample, with clinicians rating compulsions higher), the rights holder frames its self-rated versions as decision support, and no scale-author source prescribes an administration interval; charts state the local convention and keep version and mode constant.
Rights are the fact most pages omit. As stated at ocdscales.com in August 2026, any use of the scales requires explicit permission: noncommercial paper use for a clinician's own patients requires prior permission from Dr. Goodman, permission granted to one clinician does not automatically extend to colleagues, and a paid licensing agreement is required for commercial use, EHR incorporation, apps, software, and other digitalization, which the terms extend to ambient-AI capture used to create an EHR record. No free-use exception covers a public web tool that scores responses; publication in any medium, website posting, and distribution are not permitted; adaptations and translations require prior authorization; permitted copies keep the copyright and proprietary markings and published work cites the primary papers. The one instrument in the suite the rights holder allows to be reproduced on paper without permission is the DSM-5 Level 2 repetitive thoughts and behaviors measure adapted from the FOCI. The site names the licensing entity inconsistently (OCD Scales, LLC and OCD Scales, Inc.), and OCD Scales, LLC filed US trademark applications for Y-BOCS and CY-BOCS in November 2024, so a license request should confirm the exact counterparty. Freely posted forms and calculators are not evidence of authorization; this page links to none and hosts nothing. Y-BOCS, CY-BOCS, and their second editions are the intellectual property of OCD Scales, LLC and the instrument's authors. BastionGPT is not affiliated with, or endorsed by, OCD Scales, LLC or the authors. This page reproduces no items, checklist content, anchors, or scoring forms and describes the scoring architecture in original prose.
The numbers behind these errors are specific. The 2022 empirical bands from 5,140 people reached only about 55% to 57% exact category agreement and place a 28/40 in "moderate" where the legacy bands say "severe"; the 2016 consensus makes 35% one component of response, not the definition, and a 2024 individual-participant meta-analysis of 1,235 trial participants found 30% and a post-treatment total of 15 statistically optimal; the FDA DBS indication carries no Y-BOCS number while its labeling says below 30 was not established and IOCDF criteria say 28 or higher; and the rights holder's terms require permission for paper use and a paid license for any digital use. The BastionGPT Clinical Advisory Board sees the same errors most often in Y-BOCS documentation reviews:
BastionGPT is specifically trained, tuned, and clinically tested on behavioral health progress notes and screening documentation.
See how clinicians use it day to day on the AI therapy notes page.
Many BastionGPT users report saving more than 90 minutes per day on documentation.
HIPAA-compliant with a signed BAA on every plan. Your data is never used to train models. BastionGPT drafts, you review and sign.
Ten clinician-rated items scored 0 to 4, five for obsessions and five for compulsions, give two 0 to 20 subtotals and a 0 to 40 total for the past week; the symptom checklist identifies content first and never enters the arithmetic. The Y-BOCS-II scores each item 0 to 5 for a 0 to 50 total. Bands are conventions, and there are three non-equivalent frameworks: the legacy original-scale bands printed by the rights holder (0 to 7 subclinical, 8 to 15 mild, 16 to 23 moderate, 24 to 31 severe, 32 to 40 extreme), the 2022 empirical original-scale bands from 5,140 people (0 to 13 subclinical, 14 to 21 mild, 22 to 29 moderate, 30 to 40 severe), and the 2025 empirical second-edition bands (0 to 14, 15 to 21, 22 to 34, 35 to 50). None is a diagnostic threshold or a level-of-care rule, exact agreement with clinician global severity in the 2022 validation samples was only about 55% to 57%, and a low score cannot rule OCD out. Chart the raw score with subtotals and name the framework if you use a band.
The original 1989 scale rates ten items 0 to 4 for a 0 to 40 total and has more than three decades of trial, response-threshold, and longitudinal literature behind it. The 2010 second edition rates each item 0 to 5 for a 0 to 50 total to relieve the ceiling among the most severe presentations, replaces the resistance-to-obsessions item (which had the weakest psychometrics and confuses patients taught in ERP not to resist) with an obsession-free interval item, integrates avoidance into severity, and revises the checklist; a 25/40 and a 25/50 do not sit in the same part of their ranges, original bands must not be pasted onto second-edition results, and conversion back to the original requires additionally rating the replaced item. The CY-BOCS (1997) and CY-BOCS-II (2019) adapt the interview for children and adolescents and require developmentally informed reconciliation of child and caregiver report. Charts name the edition and never trend a raw score across editions.
Whatever definition you name, and the best-known one has several parts. Percentage change is baseline minus current, divided by baseline. The 2016 international expert consensus defines full response as at least a 35% reduction plus a CGI-I of 1 or 2 sustained at least one week; partial response as 25% to under 35% with CGI-I no worse than 3; remission as no longer meeting diagnostic criteria or, without a structured interview, a total of 12 or less with CGI-S of 1 or 2, sustained a week; and recovery as remission held for at least a year. Earlier signal-detection work found around 30% (2005) and at least 35% (2013) best matched global improvement, and a 2024 individual-participant meta-analysis of 1,235 adult trial participants found 30% and a post-treatment total of 15 statistically optimal while endorsing continued use of the consensus. So distinguish numerical change ("decreased 36%"), study-specific responder status, consensus response with its CGI and duration components, and remission, which is never inferred from a percentage alone.
At initial administration, yes: the designed sequence is checklist and target list first, severity ratings second, so the ten ratings have identified symptoms as their referent. At follow-up you do not have to re-run the full checklist, but you should review and update the targets and record the current content areas in your own words before rating, and note any new content. What reviewers should not see is a severity number with no symptoms behind it, or the checklist treated as an optional extra. Two related administration questions: purely mental compulsions (reviewing, neutralizing, internal checking, repeating) are compulsions and are rated in the compulsion half, described behaviorally without assuming that no visible ritual means no compulsion; and the interview stays neutral about feared content, clarifying what the patient does with uncertainty rather than adjudicating whether the feared event is likely, which the note can state in a sentence.
Not without a license. As stated at ocdscales.com in August 2026, any use of the scales requires explicit permission; noncommercial paper use for your own patients requires prior permission from Dr. Goodman, and permission to one clinician does not automatically extend to colleagues; a paid licensing agreement is required for commercial use, EHR incorporation, apps, software, and other digitalization, and the terms expressly treat ambient-AI capture used to create an EHR record as digitalization. No free-use exception covers a public web tool that scores responses; publication in any medium, website posting, distribution, adaptation, and translation all require authorization. Recording a total, subtotals, and interpretation is different from reproducing prompts or a form, but item-level or automated workflows should be confirmed in the license. The site names its entity both as OCD Scales, LLC and OCD Scales, Inc., and trademark applications for Y-BOCS and CY-BOCS were filed in November 2024, so confirm the exact counterparty. The DSM-5 Level 2 repetitive thoughts and behaviors measure adapted from the FOCI is the one suite instrument the rights holder allows on paper without permission.
No authority reviewed requires a specific score for intensive outpatient, partial hospital, or residential care: payers apply generic level-of-care criteria (24-hour structure, an individualized evidence-based plan, active treatment, expected improvement, likely deterioration at a lower level) that name no instrument, and programs that use a cutoff should attribute it to their own criteria. The score is most useful beside concrete impairment, ERP participation and targets, trajectory, accommodation, and why a lower level cannot yet work. For deep brain stimulation, keep three authorities apart: the FDA humanitarian device exemption (H050003, 2009) for adults with chronic, severe, treatment-resistant OCD who have failed at least three SSRIs contains no Y-BOCS number in its indication; FDA labeling says safety and probable benefit were not established below a Y-BOCS of 30; and the International OCD Foundation's clinical criteria use 28 or higher with at least five years of primary OCD, three serotonin-reuptake inhibitor trials including clomipramine, an antipsychotic augmentation attempt, and at least 20 ERP sessions. "FDA requires 28" and "FDA requires 30" are both wrong. Canadian and Australian programs and funders ask for standardized severity results, function, risk, and prior treatment, with no national cutoff.
Explicitly, and in two parts. State the numerical change ("total decreased from 30 to 18, a 40% reduction") and then the functional reality ("occupational and community functioning remain substantially restricted; the patient continues to avoid the kitchen, has not returned to work, and the family still performs the checking"). Consider whether the drop reflects real improvement, reduced contact with triggers, or a shift from overt rituals to avoidance and accommodation, and adjust the ERP targets toward avoided settings and covert rituals. Do not classify this as consensus full response or remission when the CGI-I sits outside the response range and impairment persists; write "40% numerical reduction with persistent major functional impairment." The original scale under-weights avoidance by design (the second edition integrated it), and insight is not directly captured by the core items, so both are charted separately every time.
Self-report adaptations are acceptable as clearly named self-report measures used through an authorized workflow and interpreted as adjunctive: record who completed it, whether it was reviewed with the clinician, and that the mode differs from an interview. They are not interchangeable with the clinician version (correlations around .45 to .65 in one carefully diagnosed clinical sample, with the worst agreement on resistance items and clinicians rating compulsions higher), the rights holder frames its self-rated versions as decision support that cannot generate a diagnosis, and a self-report total should never be spliced into a clinician-rated series without noting the mode change. For brief patient-completed screening or dimension-level tracking, the 18-item OCI-R and the 20-item DOCS are validated self-report options that do not replace a clinician severity interview for level-of-care or trial-comparable decisions; verify each instrument's own permission terms, since a license for one authorizes nothing for another.
Yes. Give it the facts (version and mode, targets, subtotals and total, baseline and current with dates, CGI, function, avoidance, insight, accommodation, medication status, and the plan) and it drafts the full entry: denominators in place, the band with its framework, change with the definition invoked, function and insight documented separately, and the decision link with dates and owners, ready for your review. It can also check a finished note for a band charted as a diagnosis or care level, an unnamed edition or mode, a percentage called response with no CGI anchor, a score with no targets, and a falling total standing in for restored function, and it never needs the instrument's item text to do so. BastionGPT is HIPAA-compliant with a signed BAA on every plan, and your data is never used to train models.
The instrument facts and compliance claims on this page trace to these sources, last verified August 2026:
Educational content, not legal or billing advice. Sample notes are fictional. Follow your organization's policies and your board, payer, and jurisdiction requirements.