Sign Up
Login
Try for Free
Login
The CAPS-5 (Clinician-Administered PTSD Scale for DSM-5) is the structured-interview reference standard for diagnosing PTSD: a trained clinician rates 20 symptoms against an index trauma, applies the DSM-5 algorithm, and sums a 0-80 severity total. VA, research, and trauma clinics rely on it. The write-up must name the form, show the algorithm, and keep PCL-5 numbers separate. This page covers CAPS-5 documentation, with a fictional sample.
Psychologists, psychiatrists, and trained clinicians or clinical researchers; distributed to credentialed professionals by the National Center for PTSD
Treating and referring clinicians, VA and veterans' claims examiners, disability and forensic reviewers, research teams, payers
250 to 600 words for the results summary · the interview itself runs about 45 to 60 minutes, plus scoring and integration
Clinician-administered structured diagnostic interview (30 items: 20 DSM-5 PTSD symptoms rated 0 to 4, plus onset, impairment, validity, and subtype items)
PTSD diagnostic evaluations, VA and veterans' assessments, treatment-outcome measurement and research trials, whenever a self-report screen needs diagnostic confirmation
Developed at the U.S. National Center for PTSD (request-gated, no listed fee); aligned with DSM-5 and DSM-5-TR adult criteria; no items or prompts reproduced
The CAPS-5 (Clinician-Administered PTSD Scale for DSM-5; Weathers, Blake, Schnurr, Kaloupek, Marx, and Keane; U.S. National Center for PTSD) is the structured interview the field treats as the reference standard for PTSD diagnosis. It is a 30-item interview: the 20 DSM-5 PTSD symptoms, each rated by the interviewer on a single 0 to 4 severity scale that combines frequency and intensity under prescribed rules (not multiplication), plus additional items covering onset and duration, distress and impairment, response validity, global severity, and the dissociative subtype. The 20 symptom ratings sum to a 0 to 80 total; a symptom counts toward diagnosis when rated 2 or higher. Three time-frame forms exist, and the difference is diagnostic, not cosmetic: Past Month supports a current diagnosis, Worst Month supports a lifetime determination, and Past Week measures short-interval change and cannot by itself establish a diagnosis, because DSM-5 requires symptoms lasting more than one month. All symptom inquiry is anchored to a defined index trauma, usually selected with the LEC-5. DSM-5-TR changed no adult PTSD criteria, so the CAPS-5 remains current; a revised CAPS-5-R with 0 to 10 item ratings and a 0 to 200 total has been published but is not yet distributed, and its totals must never be mixed with 0 to 80 CAPS-5 scores.
Documentation is where the instrument's precision is most often thrown away. The diagnosis comes from an algorithm, never from the total: at least one intrusion and one avoidance symptom, two negative-cognition-and-mood symptoms, and two arousal symptoms at threshold, plus duration over one month, clinically significant distress or impairment, and the exclusion judgment that the presentation is not better attributed to substances or another medical condition. The 0 to 80 total is a dimensional severity index riding alongside that determination, and the self-report PCL-5, which happens to share the 0 to 80 range, is a different instrument on a different metric whose 31 to 33 cut-points do not transfer. A note reading "CAPS/PCL score 38, positive for PTSD" fails on every one of these points at once. The interview feeds, and never replaces, the broader workup covered on the psychiatric diagnostic evaluation page.
Trauma-focused clinicians and diagnosing psychologists and psychiatrists use the CAPS-5 when a definitive PTSD determination matters: confirming what a screen suggested, resolving a complicated differential, or establishing a defensible baseline before treatment. The VA system is its natural habitat, in clinical care and in compensation and pension examinations, where the interview may be used but is not required by the disability questionnaire, and where the 2023 VA/DoD practice guideline suggests a validated structured interview to confirm the diagnosis. Researchers make it the dominant primary endpoint in PTSD treatment trials, administered by trained, calibrated, often blinded raters. Forensic, disability, and personal-injury evaluators use it because its structure makes findings reproducible, while keeping diagnosis, causation, and work capacity as separate questions. The note's readers mirror those uses: a claims examiner checking that a DSM-based diagnosis is actually supported, a reviewer distinguishing interview findings from the self-report screen, an attorney or independent examiner re-running the logic, and the next clinician re-administering against the same index event. Broader exposure history and Criterion A pathways belong on the trauma assessment page; the full diagnostic visit structure, including the mental status examination, lives on the MSE and biopsychosocial assessment pages.
No authority prescribes a CAPS-5 note format, but the instrument's own architecture dictates what a defensible summary must contain: which form and under what conditions, the index event at a structural level, the cluster arithmetic, the full algorithm, the total kept in its dimensional lane, the subtype and validity judgments, and change measured only against a true prior. Each section below carries the pitfall that most often undermines it.
Instrument, form, and administration conditions. Name the exact form and time frame (CAPS-5 Past Month, Worst Month, or Past Week) with the version date printed on the form, the administration date, modality (in person, video, or telephone), interpreter use, approximate duration, and any breaks or split sessions. Official materials carry inconsistent citation years (the webpage says 2013, the PDFs 2015, the forms April 2018), so cite the form actually administered. Pitfall: "CAPS administered." Which form? A Past Week total presented as a current diagnosis is the version error reviewers catch first, because the one-month duration criterion cannot come from a one-week window.
Index event, structurally. Record the event category and approximate period, the Criterion A pathway, and whether the index is one incident or a coherent set of closely related incidents, with the LEC-5 noted when used. Give enough to identify the same event at the next administration and no more: no graphic narrative, no scene-by-scene detail. The completed booklet can live in a restricted test-material section while the note carries this summary. Pitfall: Either extreme: "trauma history positive" with no identifiable index, or a graphic reconstruction that re-exposes the patient, inflates the record, and adds nothing to diagnosis, safety, or the referral question.
Cluster results, not item content. Report the number of symptoms rated 2 or higher in each cluster, for example B=3 of 5, C=1 of 2, D=4 of 7, E=3 of 6, with cluster severity sums when useful. That one line makes the algorithm auditable. Individual symptoms earn discussion only when they matter to treatment, differential, or risk, and then conceptually, in your own words. Pitfall: Twenty miniature trauma narratives, or interview prompts and rating anchors paraphrased into the note. The instrument is request-gated; its content does not belong in an open chart.
The full diagnostic determination. Make every element visible: Criterion A met, the B/C/D/E counts against their required minimums, duration over one month, clinically significant distress or impairment, and the exclusion judgment (no standalone score exists for it; the clinician documents that substances or another medical condition do not better explain the presentation), with symptoms attributed to the index event. Then state the determination: current PTSD met or not met. Pitfall: "CAPS-5 positive" or "met cutoff." The standard algorithm has no total-score cutoff, and a determination that does not show its elements cannot be audited or re-run.
Total severity, kept dimensional. Report the 0 to 80 total as a severity index alongside, never instead of, the determination. Twelve is a mathematical floor (six threshold symptoms at 2), not a validated cutoff; the circulating severity bands come from trial registrations, not the developers; characterize severity clinically and say so. Pitfall: "Total 38, therefore PTSD," or "moderate per the 23-34 band." One VA trial registration states cut scores "have not yet been established," and no developer-published band exists.
Dissociative subtype and validity judgments. State the subtype determination (present, absent, or not determinable): depersonalization and derealization are assessed by additional interview items that sit outside the 20 symptoms and contribute nothing to the 0 to 80 total. Record the overall response-validity judgment and any limits: fatigue, interruptions, interpreter constraints, inconsistent time framing. Pitfall: A "dissociation score" built by summing two of the 20 symptom items. Secondary pages, and even the National Center's own summary page in conflict with its booklet and training, misnumber these items; the booklet controls.
Change, monitoring, and the PCL-5 boundary. Compare only true priors: same form, same time frame, same index event, and say so. Name any benchmark with its source: the suggested 10-point response indicator, the 12-to-13-point reliable-change estimates, the post-treatment 8 and below and remission below 12 conventions, all sample-derived. Document the PCL-5 separately as self-report with its own metric, and never combine, convert, or average across CAPS-5, PCL-5, CAPS-IV, or CAPS-5-R. Pitfall: "CAPS/PCL score 38," a CAPS-IV total treated as continuous with a CAPS-5, or "recovered" declared from a 10-point drop while the algorithm is still met.
CAPS-5 RESULTS SUMMARY SKELETON Client: [initials] Age: [ ] Date(s): [ ] Interviewer: [ ] Form + time frame: [CAPS-5 Past Month / Worst Month / Past Week] (version date on form: [ ]) Modality: [in person / video / telephone] Interpreter: [ ] Duration: [ ] Breaks/split: [ ] Index event (structural): [category + approximate period; Criterion A pathway; single incident or coherent related series; LEC-5 administered: yes/no] No graphic narrative required. Cluster results (symptoms rated 2+): B (intrusion): [ ] of 5 C (avoidance): [ ] of 2 D (cognitions/mood): [ ] of 7 E (arousal): [ ] of 6 Cluster sums (optional): [ ] Diagnostic determination (all elements visible): A: [met/not] B >= 1: [ ] C >= 1: [ ] D >= 2: [ ] E >= 2: [ ] F duration > 1 month: [ ] G distress/impairment: [ ] H exclusion (substances / medical condition) addressed: [ ] Attribution to index event: [ ] Determination: [met / not met] Total severity: [ ]/80 (dimensional index; not the diagnosis; no invented cutoffs or bands) Dissociative subtype: [present / absent / not determinable] (additional items, outside the 20 and the total) Response validity + limitations: [ ] Change (only vs same form, time frame, index): [prior total, date, point change, named benchmark + source] Complementary measures: [PCL-5 self-report total, documented separately on its own metric] Plan: [treatment, monitoring instrument + interval, referrals] Clinician signature / credentials: Date:
Free to use and share, no signup. The PDF includes a one-page cheat sheet with section-by-section pitfalls and a pre-sign checklist; the DOCX is the blank results-summary skeleton, ready to adapt. Neither reproduces interview items, prompts, rating anchors, or scoring materials.
Scenario: an initial diagnostic evaluation in which the algorithm is met, written so every element is visible, the index event stays structural, and the PCL-5 stays in its own lane. All details are fictional.
Patient: R.T., 34 years · Setting: Outpatient psychiatric diagnostic evaluation · Clinician: S. Delgado, PhD, licensed clinical psychologist · Note date: 08/19/2026
Instrument and administration: CAPS-5 Past Month, adult form (version date April 13, 2018), administered in person on 08/19/2026 as part of a psychiatric diagnostic evaluation, by a psychologist trained in the National Center for PTSD administration curriculum. The interview lasted approximately 55 minutes with one brief break. The LEC-5 was completed first to survey lifetime exposure and select the index event.
Index event: The index trauma was a workplace vehicle collision in 2023 involving direct exposure to threatened death and serious injury, meeting Criterion A as a single incident. This structural description is sufficient to anchor today's ratings and to identify the same event at any future administration; a graphic narrative was not required for the referral question and is not recorded here. All symptom inquiry was anchored to this event.
Findings: Symptoms rated at severity 2 or higher: intrusion B=3 of 5, avoidance C=1 of 2, negative alterations in cognitions and mood D=4 of 7, and arousal and reactivity E=3 of 6, satisfying the required minimums in every cluster. Onset followed the collision, with duration well beyond one month (approximately eight months of active symptoms). Symptoms are associated with clinically significant occupational impairment (missed shifts, a route change to avoid driving demands) and interpersonal strain. The presentation is not better attributed to substances or another medical condition; alcohol use is minimal by history and screening, and medical records show no competing explanation. Depersonalization and derealization, assessed by the interview's additional subtype items, did not reach the threshold: dissociative subtype absent. Overall response validity was judged good: engagement was consistent, and reported symptoms matched observed affect and collateral occupational records.
Severity and complementary self-report: Total CAPS-5 severity was 32 of 80, the sum of the twenty 0 to 4 clinician ratings, recorded as a dimensional index of current symptom burden rather than as the diagnostic rule. Clinically, the burden is characterized as moderate; no published severity band is implied, because none is developer-endorsed. A PCL-5 completed the same day totaled 44 of 80 and is documented separately as self-reported symptom burden on its own metric: the two totals share a range, not a meaning, and neither replaces the other.
Determination and plan: Current PTSD criteria are met under the standard CAPS-5 scoring rule: Criterion A met, cluster minimums satisfied, duration exceeded, impairment present, exclusions addressed. Plan: trauma-focused evidence-based psychotherapy was reviewed and scheduled; the PCL-5 will track symptoms between sessions; response will be re-evaluated with a CAPS-5 Past Month re-administration anchored to the same index event, read against the suggested 10-point response indicator and the published 12-to-13-point reliable-change estimates, both sample-derived benchmarks rather than universal rules. The completed interview booklet is retained in the restricted test-material section of the record; this summary serves the open chart.
This sample is fictional and for educational purposes. It does not describe a real patient or record; the details are invented to show documentation structure and are not clinical guidance. No interview items, prompts, anchors, or scoring materials are reproduced, and the index event is described at a deliberately structural, non-graphic level.
Writing these after every session? BastionGPT drafts complete notes from bullets, dictation, or a transcript.
Generate a note from bulletsUnited States: the highest-stakes reader is often a VA adjudicator, and the law is specific about what the interview does and does not decide. Service connection for PTSD requires a DSM-based diagnosis conforming to 38 CFR 4.125(a) (which still incorporates DSM-5), medical evidence linking current symptoms to an in-service stressor, and credible supporting evidence that the stressor occurred (38 CFR 3.304(f)); the disability percentage then flows from occupational and social impairment under the 38 CFR 4.130 general rating formula, not from any instrument score, so a CAPS-5 total maps to no percentage and establishes neither stressor, nexus, credibility, nor employability (LAW). The PTSD disability benefits questionnaire requires DSM-5 diagnostic information and impairment evidence but no named instrument; initial examinations are restricted to qualified examiners (psychiatrists meeting VA board requirements, doctorate-level psychologists, specified supervised trainees), and 2025 exam-ordering guidance advises that the treating clinician should not, where possible, perform the examination (VA POLICY). The 2023 VA/DoD practice guideline suggests a validated structured interview such as the CAPS-5 or PSS-I-5 to confirm the diagnosis, with validated measures for monitoring, and labels itself decision support rather than a legal standard (CONVENTION). In civilian billing there is no CAPS-specific code: the interview ordinarily lives inside the psychiatric diagnostic evaluation codes 90791 or 90792, with the 96130 to 96139 testing family reserved for genuine standardized-testing work, and Medicare guidance adds that testing is not separately covered when the clinical interview could reasonably yield the information (PAYER POLICY). In research the CAPS-5 is the dominant trial endpoint; the 2024 FDA complete response letter in the MDMA program concerned trial conduct, not the instrument (CONVENTION).
Canada and Australia require the diagnosis, not the instrument. Veterans Affairs Canada's PTSD entitlement guideline (reviewed January 2025) works from DSM-5-TR, requires a clinical diagnosis from a qualified physician, nurse practitioner, or psychologist, and asks for supporting documentation as comprehensive as possible without naming the CAPS-5; the Table of Disabilities rates psychiatric impairment on functional domains, so the interview supports the diagnosis and severity picture while entitlement, service relationship, and assessment remain adjudicative questions (VAC POLICY). In Australia, veterans' claims run under legislative Statements of Principles built on DSM-5 definitions, with psychiatric compensation guidelines (amended July 2026) expecting a high-quality psychiatric report, usually from a psychiatrist, and departmental clinical resources name the CAPS-5 as the follow-up assessment after a positive PC-PTSD-5 screen, which is guidance rather than a statutory requirement (LAW and DVA POLICY). Phoenix Australia's national guideline prefers structured clinical interviewing over questionnaires for diagnosis; its NHMRC approval windows are time-limited (the updated recommendations reach their five-year mark in December 2026), so currency is worth checking at the time of writing (CONVENTION). Across all three countries the bottom line is identical and worth writing into reports: no authority mandates the CAPS-5 by name; what the law requires is a defensible DSM-based diagnosis with adequate supporting evidence, which is precisely what disciplined CAPS-5 documentation provides.
Instrument facts, psychometrics, and rights complete the picture. Access is free but request-gated: non-VA clinicians email the National Center for PTSD and certify advanced psychodiagnostic training, a master's degree in a clinical discipline, or research credentials, and specific CAPS-5 training (a four-part curriculum with simulated practice) is strongly encouraged; there are no official National Center translations of the CAPS-5, so translated research versions must be documented by language, source, and validation. Psychometrics from the initial veteran validation: diagnostic interrater kappa .78 to 1.00 depending on scoring rule, test-retest kappa .83, total-score alpha .88, interrater ICC .91, test-retest ICC .78, correlation .83 with the CAPS-IV and .66 with the PCL-5; a 2024 meta-analysis across 15 studies put global internal consistency at .92 with the two-item avoidance cluster predictably weakest, and a VA centralized telephone model with trained, calibrated raters reached a total-score interrater ICC near .98, which supports remote administration under controlled conditions without proving every remote administration equivalent. Change anchors are sample-derived: the suggested 10-point response indicator, reliable change of 12 to 13 points, post-treatment totals of 8 and below, and remission below 12 all trace to specific veteran samples and study conventions. On rights: as U.S. government work the instrument likely sits outside federal copyright, but the qualification gate, professional test security, and confidentiality still control, so completed forms belong in restricted records, blank stock stays secured, and public reproduction, native EHR item banks, and public web scorers are unsupported without written clarification from the Center. The CAPS-5 was developed at the U.S. National Center for PTSD, a division of the Department of Veterans Affairs. BastionGPT is not affiliated with, or endorsed by, the National Center for PTSD or the VA. This page reproduces no interview items, prompts, rating anchors, or scoring materials.
The numbers behind these errors are specific. The National Center's own materials carry a 2013 webpage citation, 2015 PDF citations, and April 2018 form dates; the initial validation put the CAPS-5's correlation with the self-report PCL-5 at only .66; and the published reliable-change estimates are 12 to 13 points, derived from male veteran samples. The BastionGPT Clinical Advisory Board sees the same errors most often in CAPS-5 documentation reviews:
BastionGPT is specifically trained, tuned, and clinically tested on psychiatric diagnostic evaluations and trauma assessments.
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.
Three time-frame forms of the same adult interview: Past Month, which supports a current PTSD diagnosis; Worst Month, which assesses the lifetime period when symptoms were worst and supports a lifetime determination; and Past Week, a short-interval symptom measure for treatment and research follow-up that cannot by itself establish a diagnosis, because DSM-5 requires duration beyond one month. The note names the exact form with the version date printed on it (the current adult forms are dated April 2018, while official citations vary between 2013 and 2015, so the form in hand controls). Two separate instruments complete the family and must not be blended into "the CAPS": the CAPS-CA-5 for children and adolescents 7 and older, and the published but not-yet-distributed CAPS-5-R, which rates items 0 to 10 and totals 0 to 200, a metric that can never be entered or compared as a 0 to 80 score.
The interviewer gathers frequency and intensity information for each of the 20 DSM-5 PTSD symptoms, anchored to the index trauma, and combines them under prescribed rules into a single 0 to 4 severity rating per symptom; the rating reflects typical severity during the interval, not the worst single episode, and the method is a structured combination, not frequency multiplied by intensity, whatever some explainer pages teach. A symptom counts toward diagnosis at severity 2 or higher. The 20 ratings sum to the 0 to 80 total, and cluster sums can be reported for intrusion (up to 20), avoidance (up to 8), negative cognitions and mood (up to 28), and arousal (up to 24). The additional interview items, covering onset and duration, distress and impairment, response validity, global severity, and the dissociative subtype, are separate clinical judgments that never add into the total. The scoring detail ships with the gated instrument; notes summarize results without reproducing anchors.
None, and any note built on a total cutoff is wrong on the instrument's own terms. The diagnosis comes from the algorithm: a qualifying Criterion A event, at least one intrusion symptom, one avoidance symptom, two negative-cognition-and-mood symptoms, and two arousal symptoms rated 2 or higher, duration beyond one month, clinically significant distress or impairment, and the exclusion judgment that substances or another medical condition do not better explain the presentation. Three numbers get misread into cutoffs: 12 is only the arithmetic floor six threshold symptoms create, 26 was a calibration rule inside the initial validation, and the tidy severity bands circulating online come from trial registrations rather than the developers, one of which states that cut scores "have not yet been established." A person can carry a fairly high total and fail the algorithm, or meet the diagnosis near the floor; the note reports the elements, the determination, and the total as a dimensional severity index beside it.
Everything except the 0 to 80 range. The CAPS-5 is a clinician-administered structured interview that yields a diagnostic determination: the interviewer clarifies, probes, attributes symptoms to the index event, and weighs clinical significance. The PCL-5 is a 20-item self-report that takes 5 to 10 minutes and serves screening, provisional identification, and symptom monitoring; its widely cited 31 to 33 cut-points index probable PTSD on the PCL-5 only. The totals correlate moderately (.66 in the initial validation), PCL-5 totals commonly run higher than CAPS-5 totals for the same person, and longitudinal change tracks in parallel without being identical, which is why the developers say the scores are not interchangeable. Chart them as separate labeled entries ("CAPS-5 Past Month total 32/80, current PTSD criteria met; PCL-5 self-report 44/80 same date") and never as a blended "CAPS/PCL" number; each instrument also keeps its own change benchmarks.
Structurally: the event category, the approximate date or period, the Criterion A pathway, and whether the index is a single incident or a coherent set of closely related incidents, for example "a workplace vehicle collision in 2023 involving direct exposure to threatened death and serious injury." That level lets the next administration re-anchor to the same event, supports the Criterion A conclusion, and explains the symptom relationship, which is everything the record needs. Scene-by-scene reconstruction, sensory detail, and intimate specifics that change nothing about diagnosis, safety, treatment, or the referral question add re-exposure risk and record bloat instead of clinical value, and trauma-informed record-keeping guidance runs minimum-necessary for exactly that reason. When several events qualify, name the one selected and why, keep it constant across items, and note the LEC-5 exposure survey; if the index changes at a later administration, say so, because the totals no longer form a clean same-event trajectory. The completed booklet can sit in a restricted test-material section while this summary serves the open, portal-visible chart.
As a determination from the interview's dedicated subtype items, stated as present, absent, or not determinable. Depersonalization and derealization are assessed by additional items that sit outside the 20 PTSD symptoms and contribute nothing to the 0 to 80 total; the subtype is specified when full PTSD criteria are met and clinically significant depersonalization or derealization is also established. The trap is numbering: because the 20 symptoms run intrusion 1 to 5, avoidance 6 to 7, cognitions and mood 8 to 14, and arousal 15 to 20, the last two symptom items are arousal symptoms, yet secondary pages, and even the National Center's own summary page in conflict with its booklet and training materials, describe a "dissociation score" summed from symptom items. A chart that follows that shortcut has added two Criterion E ratings and called them dissociation. Write the determination in words, from the subtype items, and skip item arithmetic entirely.
Benchmarks exist, and every one of them is sample-derived, so the note names its source. The National Center suggests roughly a 10-point drop as a response indicator while calling the evidence limited; published reliable-change estimates from two male veteran samples run 12 to 13 points; a post-treatment total of 8 or below marked clinically significant change in those samples; and totals below 12 have served as a remission convention, 12 being the arithmetic floor for a diagnosable configuration. Response, reliable change, loss of diagnosis, remission, and recovery are different claims: a 14-point drop can coexist with criteria still met, and losing the diagnosis is a statement about the algorithm, not the total. Valid comparison requires the same form, the same time frame, and the same index event, stated explicitly; CAPS-IV totals (a 0 to 136 metric), CAPS-5-R totals (0 to 200), and PCL-5 change all live on other scales and never mix. Pair the interview trajectory with function: work, sleep, avoidance behavior, and the patient's own goals.
No. It can establish the clinical diagnosis and make the severity picture transparent, which is real value in a claim, but the legal elements are separate. U.S. service connection for PTSD requires a DSM-conforming diagnosis, medical evidence linking current symptoms to an in-service stressor, and credible supporting evidence that the stressor occurred; the disability percentage then follows occupational and social impairment under the rating formula, so no CAPS-5 total maps to a percentage. The disability questionnaire requires DSM-5 diagnostic information but no named instrument (the CAPS-5 may be administered), initial examinations are restricted to qualified examiners, and current guidance advises against the treating clinician performing the exam where possible. Canada and Australia run parallel: Veterans Affairs Canada requires a qualified clinician's DSM-5-TR diagnosis with comprehensive documentation, and Australia's DVA works from legislative Statements of Principles with psychiatric reports, its clinical guidance naming the CAPS-5 only as recommended follow-up after screening. In every system the interview strengthens the evidence; the entitlement decisions belong to the adjudicator.
Yes. Give it the facts (form and time frame, administration conditions, the structural index event, cluster counts at threshold, duration, impairment, the exclusion judgment, subtype and validity determinations, and totals with any true prior) and it drafts the results summary: every algorithm element visible, the total kept dimensional with no invented cutoffs, change read against named sample-derived benchmarks, and the PCL-5 documented separately on its own metric, ready for your review. It can also cross-check a finished note for an unlabeled form, a Past Week total doing diagnostic work, conflated CAPS and PCL numbers, a missing exclusion judgment, or trauma detail beyond the minimum necessary, and it can draft the claims-ready and patient-facing variants of the same findings. 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.