Sign Up
Login
Try for Free
Login
The SLUMS (Saint Louis University Mental Status examination) is a 30-point cognitive screen developed by Tariq and colleagues at Saint Louis University with the St. Louis VA and published in 2006, interpreted through education-adjusted bands. Primary care, geriatrics, and VA teams use it as a no-fee alternative to the licensed MMSE. It screens; it never diagnoses. This page covers how to document SLUMS results, with a fictional sample note.
Health-care and social-service professionals after reviewing SLU's training video (annual review advised); interpreted by the treating clinician; official paper form used at no fee, copyright retained by SLU
Primary care and geriatrics teams, VA and post-acute clinicians, memory clinics and neuropsychologists, payers and case-management programs, families receiving results
4 to 8 chart lines · administration about 7 minutes (clinical team estimate)
Clinician-administered cognitive screen (11 items, total 0 to 30, education-adjusted three-band interpretation)
Cognitive screening on patient or family concern in adults 60 and older, geriatric and post-acute baselines, Medicare wellness-visit cognitive detection, medication-authorization pathways in Alberta
Free for professional clinical use on SLU's official paper form, copyright retained (not public domain); no US, Canadian, or Australian law mandates it; named by Alberta Blue Cross and one Missouri program
The Saint Louis University Mental Status examination is a clinician-administered, 11-item cognitive screen scored out of 30, published in 2006 by Syed Tariq, Nina Tumosa, John Chibnall, Mitchell Perry, and John Morley of Saint Louis University's Division of Geriatric Medicine, developed and first studied with the Geriatric Research, Education and Clinical Center at the St. Louis VA Medical Center, which is why the official form is headed VAMC SLUMS Examination. In aggregate the items sample orientation, registration and attention, delayed verbal recall, calculation, timed animal naming, working memory, clock construction, figure recognition and size judgment, and recall from a brief narrative; the developer describes the functions each item taps, but item groups are not validated stand-alone domain scores. SLU intends it for adults 60 and older, advises against use for cognitive dysfunction attributable to traumatic brain injury, developmental disability, or schizophrenia, recommends administration no more than annually in most cases, and lists 23 professional language and regional localizations (English-Canada and English-Australia among them). The current US-English form carries no version number or revision date, so a reproducible note records the form source and access date.
Two facts govern the documentation. First, the official interpretation is not a single cutoff but a three-band, education-adjusted system printed on the form: for a person with a high-school education, 27 to 30 is the normal band, 21 to 26 the mild neurocognitive disorder band, and 1 to 20 the dementia band; for less than a high-school education, 25 to 30, 20 to 24, and 1 to 19. A raw 26 is therefore normal under one table and in the mild band under the other, a chart that records only "SLUMS 26/30" cannot be reproduced, and the printed labels are screening bands, not diagnoses; SLU states plainly that a diagnosis requires fuller assessment by a qualified clinician. Second, the SLUMS is not public domain, whatever many pages say (the label traces to a 2011 New England Journal of Medicine essay contrasting it with the newly enforced MMSE copyright): SLU retains the copyright and charges no fee for health professionals to use the official paper form in clinical work and research, and the LOINC entry for the instrument carries a Saint Louis University and Department of Veterans Affairs copyright notice. The licensed MMSE is the instrument it is most often chosen to replace; the MoCA and the Mini-Cog are the other common comparators, and none of the four independently establishes a neurocognitive diagnosis, capacity, or fitness to drive.
Primary care and geriatrics teams use it when a patient, family member, or clinician raises a cognitive concern in an adult 60 or older, and as the cognitive-detection element of a Medicare annual wellness visit, where federal regulation requires detection of impairment by observation and report and names no instrument. VA clinicians use it heavily by convention and workflow (it was developed with the St. Louis VA GRECC and disseminated through VA training), though national VHA materials do not require it by name; post-acute and home-based primary care teams use it for baselines, while skilled-nursing assessment uses the CMS resident assessment instrument's own cognitive interview instead. Two payers and programs name it: Alberta Blue Cross accepts the SLUMS and RUDAS in place of the MMSE for dementia-medication authorization, and Missouri's home and community based services manual requires it when there is concern about a participant's ability to self-direct care, while forbidding its use as the sole basis for denying services. In Australia the 75-and-over health assessment must cover cognition without a named tool and dementia-drug subsidy remains keyed to the MMSE, and Canada's national task force recommends against instrument-based screening of asymptomatic older adults, so the indication leads the note. A positive screen routes to a neuropsychological report or a fuller memory workup, the bedside findings sit in the mental status exam, and any decision-specific question goes to a capacity evaluation, because the screen decides none of those things.
SLU publishes administration instructions and printed bands, not a note template, and no US, Canadian, or Australian authority prescribes one. What survives review is a record that names the form and access date, records education and the interpretation table applied, reports the total out of 30 within the printed band as a screening band, documents the conditions that could invalidate the result, keeps the interpretation to a screen with function and collateral beside it, and closes the referral loop. Each element below carries the pitfall that most often undermines it.
Form, localization, and access date. Chart the official SLU form and its localization (US-English, English-Canada, another language) with the access date, since the current form carries no version number, and that it was administered in person on paper by a professional who has reviewed SLU's training video. SLU says the examination was not designed for telephone or virtual administration, that fully virtual accommodations remain unvalidated, and that item wording is not to be changed. Pitfall: "SLUMS 24" with no form, language, or mode, a video or telephone administration charted as standard, or a translation improvised at the bedside.
Education recorded, table named. Record highest completed grade or years of schooling, highest credential, and country and language of schooling, then name the table applied (high-school education or less than high-school education). SLU publishes no rule for a GED, interrupted schooling, or foreign credentials; when the choice would change the band, chart the raw total with both band implications and say the interpretation rests on history and function. Pitfall: "26/30, normal" with no education line, when the same 26 sits in the mild band under the other table; a GED silently recoded as high school.
Total out of 30, in the printed band, as a screen. Write x/30 and the printed screening band for the named table ("23/30; high-school table; falls in the printed 21 to 26 mild neurocognitive disorder screening band"). Do not report a percentage or a pass or fail; the form's lowest bands begin at 1, so a total of zero has no printed band and is described as profoundly abnormal performance requiring interpretation, not assigned to one. Pitfall: "Patient has mild neurocognitive disorder per SLUMS," "failed the SLUMS," or a zero forced into the dementia band the form never printed.
Administration conditions and validity. Document the setting, alertness and engagement, primary language and any interpreter, corrective lenses and hearing aids in place and adequate (SLU says not to administer when the patient cannot see or hear adequately despite correction), pain, fatigue, mood, acute illness, sedation, interruptions, deviations, whether every item was completed (no partial or prorated total), and whether the result is valid and interpretable. Pitfall: A precise total charted beside inadequate hearing or an unavailable language localization; a partial score generated because the patient stopped.
Pattern in prose, no domain subscores. Describe relative performance broadly (greater difficulty with delayed retrieval and clock construction, stronger orientation and attention) without reproducing item content, and do not chart item groups as validated stand-alone domain measures; factor studies disagree about the instrument's structure. Pitfall: "Memory 9/13, visuospatial 3/5" reported as domain scores, or the answers and stimuli written into the note.
Screen, not diagnosis, with function and collateral beside it. State that the SLUMS is a screening measure and does not independently establish a neurocognitive diagnosis, decisional incapacity, or driving unfitness; then record patient and informant concerns, change from baseline, medication and medical review, delirium and mood assessment, ADL and IADL status, and safety (medications, finances, driving, wandering, firearms). Pitfall: A normal band used to close the question when the spouse reports new medication errors, or a low band written up as dementia.
Referral chain, owner, and the record. Document collateral history, functional assessment, reversible contributors, indicated examination and laboratory workup, referral for comprehensive cognitive, geriatric, neurologic, or neuropsychological assessment where concern persists, a decision-specific capacity or driving evaluation where those questions arise, the owner and date, and the completed form scanned to the record; a declined screen is charted as offered and declined with the alternative assessment used, never as a zero or a band. Pitfall: A positive screen that ends at a number, or a refusal entered as 0/30.
SLUMS DOCUMENTATION BLOCK Date: [ ] Setting: [ ] Administered by: [ ] Interpreted by: [ ] Form: [official SLU form; localization: US-English / English-Canada / other language] Accessed: [date] Mode: [in person, paper] Trained administrator: [training video reviewed; annual review date] Education: [highest grade / years; credential; country + language of schooling; GED or foreign credential noted] Table applied: [high-school education / less than high-school education] Conditions: [quiet setting; alert and engaged; primary language; interpreter; lenses and hearing aids adequate; pain, fatigue, mood, illness, sedation; interruptions or deviations; all items completed] Validity: [valid and interpretable / not interpretable: reason; no partial total] Total: [ ]/30 Printed band for the named table: [normal 27-30 or 25-30 / mild neurocognitive disorder 21-26 or 20-24 / dementia 1-20 or 1-19]; zero: no printed band, described not assigned Band under the other table (if education uncertain): [ ] Pattern (prose, no subscores): [ ] Interpretation: [screening band only; does not establish a neurocognitive diagnosis, incapacity, or driving unfitness] Concerns + function: [patient / informant concerns; change from baseline; ADLs, IADLs; safety: medications, finances, driving] Reversible contributors: [delirium, depression, medications, sensory, metabolic] Plan: [collateral / functional assessment / workup / referral / capacity or driving evaluation if indicated] Owner + date: [ ] Record: [completed form scanned] Declined: [offered + declined + alternative assessment; never a score] 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 examination's items, stimuli, or scoring instructions.
Scenario: a cognitive-concern visit in geriatric primary care where the raw score lands in the printed normal band under the correct education table but would sit in the mild band under the other, and a spouse's report of medication errors keeps the evaluation open. All details are fictional.
Patient: W.H., 79 · Setting: Geriatric primary care clinic, cognitive concern visit · Clinician: T. Nakamura, DO · Note date: 08/12/2026
Screening: Saint Louis University Mental Status examination, official SLU US-English paper form (accessed 08/2026), administered in person in a quiet private room by the clinic's nurse (training video reviewed 02/2026), scored and interpreted by me. Indication: spouse's concern about medication management raised at the last visit. Patient alert and engaged; corrective lenses and hearing aids in place and adequate; primary language English, no interpreter; no acute illness, no sedating medication changes; all items completed. Education: 10 completed years, no high-school credential, US schooling in English; less-than-high-school table applied.
Result: Total 26/30, which falls within the printed normal screening band of 25 to 30 for the less-than-high-school table. For transparency: the same total would fall in the printed 21 to 26 mild neurocognitive disorder band under the high-school table. Relative pattern: greater difficulty with delayed verbal retrieval, stronger performance across the other broadly sampled functions; no domain subscores reported. Prior formal screen: none on record. Result is a screening finding and does not establish or exclude a neurocognitive diagnosis, decisional incapacity, or driving unfitness.
Concerns and function: With the patient's permission the spouse was interviewed: about six months of a pill organizer found double-filled or skipped, two missed appointments, and repeating questions; patient acknowledges occasional forgetting, denies getting lost, manages meals and self-care, drives locally, and handles cash but the spouse now pays the bills. PHQ-2 negative; no features of delirium; sleep and hearing reviewed. Because the collateral history describes functional change despite a normal-band screen, the evaluation stays open.
Plan: Medication reconciliation completed and pharmacy synchronization arranged; laboratory evaluation for reversible contributors ordered per clinic protocol; structured instrumental-activities inventory and a fuller cognitive assessment scheduled 08/26/2026 with the spouse present, memory-clinic referral to follow if concern persists; interim: spouse to supervise the pill organizer with the patient's agreement, driving discussed with a plan to assess formally at the follow-up visit; return precautions for acute change. Follow-up owner: Dr. Nakamura; nursing call within one week to confirm the appointment.
Administrative: Completed official form scanned to the chart; total, table applied, band, and interpretation entered in the cognitive-screen field with the form access date. Cognitive detection for the wellness-visit component documented from observation, patient report, and spouse report; no diagnosis coded from the screen; SLU's guidance of no more than annual administration noted for the next routine screen.
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: the requirement to look is federal, the tool is a choice, and payer acceptance is not universal. The Medicare annual wellness visit regulation (42 CFR 410.15, LAW) requires detection of cognitive impairment by direct observation with due consideration of patient report and concerns raised by family, friends, or caregivers, and names no instrument; the SLUMS may serve that component when clinically appropriate, but a chart should never say Medicare required it. The cognitive assessment and care plan service (99483, PAYER POLICY, CMS page modified May 2026) requires a far broader evaluation with an independent historian, function, safety, medications, behavioral symptoms, supports, and a written care plan, and a SLUMS result alone does not meet it; a Palmetto GBA article effective January 2026 names the Mini-Cog, GPCOG, or short MoCA for the documented cognitive assessment under that service, does not name the SLUMS, and requires the full raw scoring to be available for review, in its jurisdictions only (PAYER POLICY), so payer acceptance must be verified per policy and region. Missouri's home and community based services manual is a rare authority that names the SLUMS (PAYER POLICY, one state program): it shall be used when there is concern about a participant's ability to self-direct care, cannot be the sole basis for denying covered services, requires the last completed grade or degree to be recorded and the completed form scanned into the electronic case record, and requires further capacity evaluation. National VHA materials do not require the SLUMS by name despite its VA origins (CONVENTION), skilled-nursing assessment uses the resident assessment instrument's own cognitive interview (PAYER POLICY, with FY2027 MDS expansion effective October 2026 carrying no SLUMS element), and capacity and driving remain decision-specific evaluations that a screen can prompt but never decide (CONVENTION).
Canada, Australia, and the evidence. Canada's national task force reaffirmed in 2024 its recommendation against instrument-based screening of asymptomatic community-dwelling adults 65 and older, excluding anyone with symptoms or concern (CONVENTION); Alberta Health Services recommends the SLUMS among several tools, published an April 2026 administration guide that treats the clock's two criteria as all-or-none, obtained provincial copyright permissions before implementing cognitive tools in Connect Care, and reports that Alberta Blue Cross accepts the SLUMS and RUDAS in place of the MMSE for medication authorization (CONVENTION and PAYER POLICY, one province); the June 2026 CCMTA driver-fitness model standard says no cognitive test can be the sole determinant of driving ability (CONVENTION applied through provincial LAW). Australia's MBS health assessments (items 224, 225, 226, 227, 701, 703, 705, 707) cover cognition and mood without a named tool, a March 2026 change removed outdated screening-test requirements from affected assessments, dementia-drug subsidy remains keyed to the MMSE, and Austroads uses individualized functional assessment (PAYER POLICY and CONVENTION); an English-Australia localization exists and establishes no Australian norms or endorsement. On the evidence: the 2006 study of 702 VA patients (mean age 75) reported sensitivity and specificity of 95% and 76% for mild neurocognitive disorder and 98% and 100% for dementia in the high-school stratum, and 92% and 81% and 100% and 98% in the less-than-high-school stratum, comparable to the MMSE for dementia and better for mild impairment; a 2014 comparison in 136 educated veterans found statistically equivalent discrimination for the SLUMS, MoCA, and Short Test of Mental Status (AUC 0.74, 0.77, and 0.77 for MCI; 0.98, 0.96, and 0.97 for dementia); the validation program was concentrated in male veterans (all 357 surviving participants in a 7.5-year follow-up were men); a 2023 study of 602 Black and White veterans found Black veterans 1.99 times more likely to be classified in the dementia band; a Chinese-version study of 367 participants found normal-versus-MCI AUCs of only 0.32 to 0.54; a 2024 study of 263 neuropsychology referrals found the official cutoffs classified 55.1% correctly and proposed research thresholds (24 or below for MCI, 17 or below for dementia) that have not replaced the form; reliability data are thin (alpha .709 and one-year stability .723 in 108 veterans; an earlier report near .57; no item-level inter-rater coefficient for the clock or figure items), and a 2022 review of 20 psychometric studies advised cautious use. Report the original figures with their strata and population, and never age-adjust or transport the bands to a translation without local validation.
Rights are precise: free is not public domain. Saint Louis University makes the official SLUMS form available at no fee for health-care and social-service professionals to administer in clinical work and research after reviewing its training video, and states that it retains the copyright; the LOINC entry for the instrument carries a Saint Louis University and Department of Veterans Affairs copyright and used-with-permission notice; SLU's copyright policy notes that absence of a notice does not place a work in the public domain and that reproduction, distribution, public display, and derivatives are owner-controlled; and the widely repeated "public domain" label traces to a 2011 New England Journal of Medicine essay written to contrast the SLUMS with the MMSE's enforced copyright. Three uses differ: entering the total, education, band, and interpretation in a record and scanning the completed official form are ordinary clinical documentation; reproducing the items, stimuli, wording, or scoring logic inside an EHR, or building a derivative electronic form, has no public authorization on SLU's current page (Alberta obtained permissions before its EHR build); and a commercial public web test that reproduces the examination has no published licence, while a score-only calculator that accepts a total and an education category without displaying items is a lower-risk question that SLU has not endorsed. Wording is not to be changed, translations come from SLU's 23 localizations rather than the bedside, and telephone or virtual administration is outside the design. The Saint Louis University Mental Status examination is copyright of Saint Louis University, developed with the St. Louis VA Geriatric Research, Education and Clinical Center, and the official form is available from SLU. BastionGPT is not affiliated with, or endorsed by, Saint Louis University or the Department of Veterans Affairs. This page reproduces no items, stimuli, or scoring instructions and describes the instrument in original prose.
The numbers behind these errors are specific. A raw score of 26 is in the printed normal band under the less-than-high-school table and in the mild neurocognitive disorder band under the high-school table; the form's lowest bands begin at 1, leaving zero unassigned; the original 2006 study of 702 VA patients reported 95% sensitivity for mild impairment in the high-school stratum while a 2024 study of 263 neuropsychology referrals found the official cutoffs classified 55.1% correctly and Black veterans were 1.99 times more likely to land in the dementia band; and SLU retains the copyright it does not charge for. The BastionGPT Clinical Advisory Board sees the same errors most often in SLUMS 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.
Eleven items are scored to a total out of 30, sampling orientation, registration and attention, delayed verbal recall, calculation, timed animal naming, working memory, clock construction, figure recognition and size judgment, and recall from a brief narrative. Interpretation is education-adjusted, with two tables printed on the official form: with a high-school education, 27 to 30 is the normal band, 21 to 26 the mild neurocognitive disorder band, and 1 to 20 the dementia band; with less than a high-school education, 25 to 30, 20 to 24, and 1 to 19. The labels are screening bands derived from the original 2006 study's clinical classifications, not diagnoses, and SLU states that a diagnosis requires fuller assessment by a qualified clinician. Two details most pages miss: the lowest bands begin at 1, so a total of zero has no printed band, and a raw 26 changes bands solely with the education table, which is why the table is always named. Do not report a percentage, a pass or fail, or item-group subscores as validated domain measures.
The official form offers two categories, high-school education and less than high-school education, and SLU asks administrators to record years of schooling or highest degree; it publishes no rule for a GED obtained later, interrupted schooling, foreign credentials, learning disability, or language of instruction. So document what you know ("completed grade 10; GED at 32; English-language US schooling; reads independently"), name the table you applied, and when the choice would change the band, chart the raw total with both band implications and say the interpretation rests on history and function ("26/30: normal band under the less-than-high-school table, mild band under the high-school table; interpretation qualified"). Do not convert a foreign credential to "high school" from years alone unless a recognized equivalency or local policy supports it, and do not silently recode a GED. That dual statement is a documentation solution, not an SLU rule, and it is far better than hiding the uncertainty.
No, and not without asking. Saint Louis University makes the official paper form available at no fee for health-care and social-service professionals to administer in clinical work and research after reviewing its training video, and states that it retains the copyright; the LOINC entry for the instrument carries a Saint Louis University and Department of Veterans Affairs copyright and used-with-permission notice, SLU's copyright policy says the absence of a notice does not make a work public domain, and the popular "public domain" label traces to a 2011 New England Journal of Medicine essay contrasting the SLUMS with the MMSE's enforced copyright. Three uses differ: charting the total, education, band, and interpretation and scanning the completed form are ordinary documentation; reproducing the items, stimuli, wording, or scoring logic inside an EHR or a derivative electronic form has no public authorization on SLU's current page (Alberta Health Services obtained provincial permissions before its Connect Care build); and a commercial public web test that reproduces the examination has no published licence, while a score-only calculator that takes a total and an education category is lower-risk but not SLU-endorsed. Wording may not be changed, and translations come from SLU's 23 localizations. Contact SLU at the address on its page before any item-level digital build.
It depends on the exact policy, so verify rather than assume. The Medicare annual wellness visit regulation requires detection of cognitive impairment by observation and report and names no instrument, so the SLUMS may serve that component when clinically appropriate and no test is federally mandated. The cognitive assessment and care plan service (99483) requires a much broader evaluation (independent historian, function, safety, medications, behavioral symptoms, supports, a written care plan), which a SLUMS alone cannot satisfy, and a Palmetto GBA article effective January 2026 names the Mini-Cog, GPCOG, or short MoCA for the documented cognitive assessment under that service, does not name the SLUMS, and requires the full raw scoring to be available, in its jurisdictions only. Missouri's home and community based services manual requires the SLUMS for a specific self-direction concern and forbids using it alone to deny services; Alberta Blue Cross accepts the SLUMS and RUDAS in place of the MMSE for medication authorization; Australia's health-assessment items require cognition to be assessed without a named tool, and dementia-drug subsidy there is keyed to the MMSE. "A validated cognitive screen" does not mean every payer accepts every instrument for every service.
Choose the SLUMS when a no-fee official paper screen with an education-adjusted three-band output is wanted, mild impairment detection is an explicit goal, and the patient fits the design (60 or older, adequate sensory function, in-person administration in an available localization). Keep or choose the MMSE when continuity with prior serial scores or a payer or pathway that keys to it (Australian dementia-drug subsidy, some Canadian plans) matters and the organization has licensed access; the 2006 study found the two comparable for dementia and the SLUMS better for mild neurocognitive disorder. Choose the MoCA when a local referral, research, or payer pathway expects it or longitudinal MoCA data exist, mindful of its permission and certification rules; a 2014 comparison in educated veterans found statistically equivalent discrimination for the SLUMS, MoCA, and Short Test of Mental Status. Choose the Mini-Cog when a three-minute first-pass screen by any trained staff member is the need, accepting less multidomain narrative and no education adjustment; and consider the RUDAS when cultural, language, or educational fairness is central. None establishes a diagnosis, capacity, or driving fitness.
Never as a score. For a refusal: record that the SLUMS was offered and its purpose explained, that the patient declined, the reason only if volunteered, that no items were administered and no total or band was assigned, how cognition was assessed instead (observation, history, collateral, functional review), any communication, sensory, distress, or language factors, and the follow-up plan; do not enter 0/30 or imply that refusal is evidence of impairment. For a partial administration: no prorated total, per Alberta's guide, and per SLU no wording changes to get through an item. For an invalid attempt (hearing or vision inadequate despite correction, no suitable language localization, delirium or acute illness, a telephone or video attempt): chart "attempted but not completed; no total or interpretation assigned; administration not valid because..." and the accommodations or alternative evaluation arranged. Approved enlarged visual material with lenses in place is a documented accommodation, not an invalidation.
Not by phone or video as a standard administration: SLU states the examination was not designed for telephone or virtual administration and that a valid study of fully virtual accommodations has not been done, so a remote attempt is documented as nonstandard and not compared with the printed bands. For low vision, SLU allows corrective devices in place and enlarged visual material in specified circumstances, and Alberta's guide likewise permits enlarged versions while prohibiting partial scoring; if the patient still cannot adequately see or hear despite correction, do not administer, and chart the attempt as not completed with no score. Document the accommodation exactly ("approved enlarged visual material used; lenses in place; material adequately perceived"). Item wording is not to be changed, and there is no published SLUMS-specific inter-rater reliability for the clock or figure items; Alberta's all-or-none rule for each clock criterion standardizes scoring without being a reliability study.
Good in its original setting, less certain elsewhere. The 2006 study of 702 VA patients reported sensitivity and specificity of 95% and 76% for mild neurocognitive disorder and 98% and 100% for dementia in the high-school stratum, and 92% and 81% and 100% and 98% in the less-than-high-school stratum, comparable to the MMSE for dementia and better for mild impairment; a 2014 comparison in 136 educated veterans found statistically equivalent discrimination for the SLUMS, MoCA, and Short Test of Mental Status. The limits: the validation program was concentrated in male veterans (all 357 survivors in a 7.5-year follow-up were men); a 2023 study of 602 Black and White veterans found Black veterans 1.99 times more likely to be classified in the dementia band; a Chinese-version study of 367 participants found normal-versus-MCI AUCs of only 0.32 to 0.54 despite dementia AUCs near 0.8; a 2024 study of 263 neuropsychology referrals found the official cutoffs classified 55.1% correctly and proposed research thresholds that have not replaced the form; reliability data are thin (alpha about .71 in one veteran sample, an earlier estimate near .57); and a 2022 review of 20 psychometric studies advised cautious use pending better norms and reliability evidence. Report the numbers with their strata and population, and never transport the bands to a translation without local validation.
Yes. Give it the facts (form and localization, education and the table applied, conditions, total, pattern, informant report, function, prior results, and the plan) and it drafts the full entry: the printed band for the named table with the other table's band where the education choice is uncertain, the validity conclusion, the screen-not-diagnosis line, function and collateral, and the referral chain with owner and date, ready for your review. It can also check a finished note for an unnamed education table, a band charted as a diagnosis, a zero forced into a band, an invalidated attempt scored, a normal band used to dismiss reported decline, and a positive screen with no plan, and it never needs the examination's items 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.