Sign Up
Login
Try for Free
Login
The clock drawing test is a brief constructional screen in which the patient draws a clock face and sets a requested time. It has no single version: at least a dozen named scoring systems use different ranges, conditions, and anchors. Primary care, geriatrics, and neuropsychology teams use it in cognitive screening. A score means nothing without its system. This page covers how to document clock drawing results, with a fictional sample note.
Any clinician after brief familiarization (physicians, nurses, occupational therapists, psychologists); embedded versions follow their parent instrument's rules; detailed systems reward trained raters
Primary care and geriatrics teams, memory clinics and neuropsychologists, hospital teams, driving and capacity evaluators, families receiving results, payers and quality reviewers
3 to 8 chart lines · administration about 2 to 5 minutes, plus disclosure and planning
Brief constructional cognitive screen (command or copy drawing, scored by a named system; ranges run from 0 to 3 up to 0 to 20)
Medicare wellness visits and memory concerns, screens alongside recall tasks, bedside cognitive checks, older-driver screens, serial monitoring of documented decline
A family of scoring systems, not one test; embedded MoCA and Mini-Cog clocks governed by their rights holders; no authority mandates any system; no test content reproduced
The clock drawing test asks a patient to draw a clock face, place the numbers, and set the hands to a requested time, and it is best understood as a family of tasks rather than one instrument. Recorded clinical use goes back to at least 1915 with head-injured soldiers, and the modern screening role grew out of the Boston tradition in the 1980s, but standardization never converged: the major named scoring systems (Shulman, 1986; Sunderland, 1989; Mendez, 1992; Rouleau, 1992; Freedman, 1994; the CLOX executive task, 1998; Freund's driving-oriented system, 2005) use different ranges, score directions, conditions, and requested times, and the clocks embedded in the MoCA (0 to 3) and Mini-Cog (0 or 2, deliberately binary) are components of those instruments, not standalone scores. Even "Shulman score" is not self-defining: current implementations number the same six categories 0 to 5 or 1 to 6, so a charted "clock 4" cannot be read without its anchors.
The load-bearing rule for documentation follows directly: a clock score is uninterpretable without five facts, the scoring system with its range and direction, the condition (command from instruction, or copy of a model), the drawing substrate (blank page or pre-drawn circle), the exact time requested, and the qualitative errors in words. Those variables change the task itself: command drawing taxes generation and planning while copy supplies the model, a pre-drawn circle removes contour planning, and the conventional "ten past eleven" is deliberate, forcing both halves of the clock face into use and requiring the spoken "ten" to be recoded to the numeral 2, which is exactly where stimulus-bound errors show themselves. This page covers the standalone clock; the MMSE page covers the screen that, contrary to common belief, contains no clock at all, and the SLUMS page covers the free instrument whose embedded clock is one of its items.
Primary care and geriatrics teams use a clock, alone or inside the Mini-Cog, for the Medicare annual wellness visit's required cognitive detection and whenever a patient or family raises a memory concern; the regulation requires detection by observation and collateral report and names no tool. Memory clinics and neuropsychologists use standalone administrations, often command plus copy, because the command-copy contrast is the clock's most clinically specific signal, and they retain the drawing for later re-scoring. Hospital teams reach for it at the bedside, where an abnormal clock during acute illness documents impairment at that moment but cannot separate delirium from baseline without collateral and a delirium instrument. Driving evaluators use the clock inside the AGS and NHTSA office battery (scored with Freund's system) as a referral signal, and Canada's 2026 national driver-fitness standard names an abnormal clock as a trigger for deeper assessment while expressly rejecting any cognitive test as the sole determinant. Capacity questions stay decision-specific in all three countries, with the clock as collateral evidence only, feeding a capacity evaluation rather than deciding it; fuller workups land in a neuropsychological report.
No US, Canadian, or Australian authority prescribes a clock documentation format or mandates a scoring system. What survives review is a record that fixes the five facts that give a clock score meaning, describes the errors in words, keeps the interpretation to a screen, and preserves the drawing. Each element below carries the pitfall that most often undermines it.
Context and indication. Record why the clock was drawn: the wellness-visit cognitive component, a patient or family concern, a bedside check during admission, a driving screen, or part of a larger instrument or battery. If the clock sits inside a MoCA, Mini-Cog, or SLUMS, say so and chart it as that instrument's component. Pitfall: An embedded clock detached from its parent: "MoCA clock 2/3" charted as a standalone CDT, or a Mini-Cog clock re-scored with a standalone rubric its conditions were never validated for.
Scoring system, range, and direction. Name the system and its anchors: "Shulman six-category system, implemented 0 to 5, higher better," "CLOX1 and CLOX2, each 0 to 15, lower worse," "Mini-Cog clock, 0 or 2." Implementations genuinely differ (Shulman circulates as both 0 to 5 and 1 to 6), so the name alone is not enough. Pitfall: "Clock 4." Without the system, range, and direction, the number cannot be interpreted, compared, or re-derived, and systems' totals (4/5, 4/7, 4/10, 4/15, 4/20) describe entirely different performances.
Condition, substrate, and requested time. Chart command versus copy (and both, in order, if both were given), blank page versus pre-drawn circle, and the exact time requested. These are test specifications, not trivia: a pre-drawn circle removes contour planning, copy removes generative demands, and 11:10 exists to test both hemifields and the ten-to-2 recoding that exposes stimulus-bound errors. Pitfall: Scores compared across different conditions or times as if serial data. A copy clock is a different observation from a command clock, and norms travel with the administration that produced them.
Administration integrity. Record repetitions, clarifications, cueing, restarts and which attempt was scored, spontaneous versus prompted correction, refusal, and any time limit. Keep the first attempt: abandoned contours and marginal work carry process information the final drawing hides. Pitfall: A silently discarded first clock. Protocols differ on restarts and self-correction, and a re-drawn clock scored as the only attempt misstates what happened.
Score plus the errors in words. Report the numeric score, then describe the drawing: number omissions, duplications, crowding or one-sided spacing, sequence, hand placement and length, stimulus-bound responses (hands drawn to the 10 and 11), perseveration, spatial disorganization, graphic difficulty. The qualitative account is what a later clinician can actually use. Pitfall: A score with no description. The number discards the pattern, and reviewers cannot tell a planning failure from a motor one, or re-score the performance, from "3/5" alone.
Confounds and validity. Document vision and hearing correction, dominant-hand motor status and tremor, language and literacy, education, analog-clock familiarity (now a demonstrated confound in younger, digital-native adults), medications, acute illness, and delirium status for inpatients. Conclude: interpretable, interpretable with caution, or not interpretable. Pitfall: An abnormal inpatient clock read as dementia during delirium, or a younger patient's unfamiliarity with analog clocks scored as cognitive impairment.
Interpretation, retention, and the next step. Keep the conclusion to a screen: what the result raises, what it does not establish (no clock diagnoses dementia, decides capacity, or clears driving), and the follow-up chain with an owner: collateral history, functional review, medication reconciliation, fuller assessment or referral. Scan the drawing into the record; the image outlives the score and permits later re-scoring. Pitfall: A screen charted as a diagnosis, a normal clock used to close a question the family has kept open, or the drawing binned so only an unreviewable number remains.
CLOCK DRAWING TEST DOCUMENTATION BLOCK Date: [ ] Setting: [ ] Administered by: [ ] Interpreted by: [ ] Indication: [wellness visit / memory concern / bedside check / driving screen / component of instrument or battery] Standalone or embedded: [standalone / MoCA clock x/3 / Mini-Cog clock 0 or 2 / SLUMS item; embedded clocks stay with their parent instrument] Scoring system + range + direction: [e.g. Shulman six-category, 0-5, higher better / CLOX1 + CLOX2, each 0-15 / other named system] Condition: [command / copy / both, in order] Substrate: [blank page / pre-drawn circle] Time requested: [e.g. ten past eleven, 11:10] Administration integrity: [repetitions, cueing, restarts, which attempt scored, spontaneous vs prompted correction, refusal] Score: [ ] Qualitative description: [omissions, duplications, spacing, sequence, hand placement and length, stimulus-bound response, perseveration, spatial planning, graphic difficulty] Copy condition result, if given: [score + description; command-copy contrast noted] Confounds: [vision, hearing, motor and tremor, language, literacy, education, analog-clock familiarity, medications, acute illness, delirium screen result if inpatient] Validity: [interpretable / with caution / not interpretable: reason] Interpretation: [screen result only; does not establish dementia, MCI, capacity, or driving fitness; considered with history and collateral] Drawing retained: [scanned to record: Y/N; first attempt kept] Plan: [collateral, functional review, medication reconciliation, fuller assessment or referral, repeat after stabilization] Owner + date: [ ] 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 any instrument's form, rubric text, or scoring tables.
Scenario: an abnormal command clock with a markedly better copy at a memory-concern visit, documented with the system, condition, and time named, the errors in words, and the follow-up chain set. All details are fictional.
Patient: R.D., 81 · Setting: Geriatric primary care, memory concern visit · Clinician: T. Nguyen, MD · Note date: 08/12/2026
Cognitive screening: Standalone clock drawing test administered after R.D.'s daughter reported six months of missed bills and repeated questions. Command condition first: blank page, instructed to draw a clock, put in all the numbers, and set the time to ten past eleven (11:10). Scored with the Shulman six-category system, implemented as 0 to 5 with higher scores better: 2/5. Reading glasses worn, hearing aids in, dominant right hand steady, no tremor; instructions repeated once at her request, single attempt, no restart. She has used analog clocks all her life.
Qualitative description: Contour drawn adequately. Numbers 1 through 8 present with 9 through 12 omitted, remaining numbers crowded into the right half of the face. Hands drawn pointing to the 10 and to the 11, a stimulus-bound response to the spoken instruction rather than the requested time. One number (4) written twice. No perseverative looping; pencil control intact.
Copy condition: A copy condition followed (examiner's model clock visible, same 11:10 setting): 5/5 by the same implementation, organized and complete, with accurate hand placement. The command-copy contrast, impaired generation with intact copying, suggests the burden falls on planning and executive control more than on visuoconstruction, and it is a screening observation, not a localization or a diagnosis.
Interpretation and disclosure: Abnormal cognitive screen at today's visit, consistent with the family's report; it does not by itself establish dementia, mild cognitive impairment, a cause, decisional capacity, or driving fitness. Result reviewed in plain terms with the patient and her daughter with the patient's permission: the drawing suggests thinking and planning should be looked at more closely, and it is not a diagnosis. Both clocks scanned to the diagnostic-media section of the record; the first attempt retained.
Plan: Collateral history and instrumental-function review completed today (finances, medications, cooking, driving: drives short daytime trips only). Medication reconciliation done; no new sedating or anticholinergic agents. Laboratory workup for reversible contributors ordered. Dedicated cognitive assessment visit scheduled 08/28/2026 with a longer instrument and functional inventory, neuropsychology referral to be decided there; driving discussed, with an occupational therapy driving evaluation to be arranged if the fuller assessment sustains concern. No dementia code entered from a screen. Follow-up owner: Dr. Nguyen; nurse call in one week to confirm the appointment.
This sample is fictional and for educational purposes. It does not describe a real patient or record; the scores, drawings, dates, and details are invented to show documentation structure and are not clinical guidance. No instrument's rubric text or scoring tables are reproduced.
Writing these after every session? BastionGPT drafts complete notes from bullets, dictation, or a transcript.
Generate a note from bulletsUnited States: the duty is to detect, never to use one tool. The Medicare annual wellness visit regulation (42 CFR 410.15) requires detection of cognitive impairment by direct observation with due consideration of patient, family, and caregiver report, and CMS deliberately names no instrument (LAW); a clock, alone or inside the Mini-Cog, is one conventional way to meet it. The cognitive assessment and care plan service (99483) is PAYER POLICY with documented elements far beyond any screen (independent historian, function and safety, neuropsychiatric symptoms, medication reconciliation, caregiver assessment, care plan), and a regional Medicare LCD illustrates the documentation principle for testing generally: name the instrument, summarize findings, and keep raw results available on request. An abnormal clock can initiate that pathway; it is not by itself 99483 documentation, and the CMS GUIDE dementia model standardizes care without mandating an instrument. Hospital pathways screen delirium with dedicated instruments (CAM, 4AT), and a clock drawn during acute illness documents impairment at that moment only (CONVENTION). The USPSTF's standing position on screening asymptomatic older adults is an insufficient-evidence statement, which makes the indication line (concern versus routine) worth charting. In driving evaluations, the AGS and NHTSA clinician's guide includes a clock, scored with Freund's system, in its CADReS office battery as a referral signal, and the underlying validation treated low scores as a reason for formal driving evaluation, not a licensing decision (CONVENTION).
Canada and Australia frame the clock explicitly as a trigger, not a verdict. Canada's National Safety Code Standard 6 for driver fitness (June 2026) names abnormalities on clock drawing, among other screens, as findings that should prompt deeper assessment, and in the same document states that cognitive tests lack established safe-or-unsafe cut points and must not be the sole determinant of driving competence; provincial licensing authorities make the decision, and physician reporting duties are provincial law (LAW for reporting, CONVENTION for the tool). Capacity is decision-specific in all three countries: Ontario's consent framework ties capacity to the particular decision, US federal and professional guidance analyzes understanding, appreciation, reasoning, and choice, and Australian frameworks presume capacity and require decision-specific, supported assessment, so a clock score is collateral cognitive evidence in a capacity evaluation, never the determination (LAW, varying by jurisdiction). Australia's Austroads driving standard does not name the clock at all, directing individualized medical, functional, and on-road evidence (REGULATORY STANDARD), and the national delirium care standard requires validated, culturally appropriate screening without naming a clock (CONVENTION). The evidence base supports the same restraint everywhere: pooled dementia accuracy is moderate and system-dependent, sensitivity for mild cognitive impairment is poor (about 0.63 for paper scoring in the largest meta-analysis, whether scoring was brief or detailed), and cultural, educational, and literacy effects are pervasive, so cutoffs do not travel across systems or populations (CONVENTION).
Rights and versions differ by system, and the clock now has a cohort-validity problem worth charting. Under the general US principle that copyright protects published expression rather than a method, the standalone scoring approaches can be applied clinically and described in original words, but the published rubric tables and figures belong to their journals and publishers: the rights holder of the Shulman system has stated the test is not copyrighted, the Freedman book is an all-rights-reserved Oxford University Press publication, CLOX reuse runs through the journal's permissions channel, and no blanket license was located for the Sunderland, Mendez, Rouleau, or Freund rubrics, so paraphrase, do not transcribe. The embedded clocks carry their instruments' terms: the MoCA permits qualifying clinical reproduction but prohibits web posting and unauthorized electronic versions, and its current policy still points to certification when an individual task like the clock is interpreted beyond the total score; the Mini-Cog is copyrighted with free unmodified clinical use and discrete-field EHR documentation recommended by its rights holder; the SLUMS is freely available. Digital clock products are proprietary: the leading device is FDA-listed under a premarket-exempt product code, which is not the same as FDA clearance or approval, and its several-hundred-feature process scores are not interchangeable with any paper system's score. Finally, analog-clock familiarity is now a documented confound: cognitively normal younger adults underperform on the clock (a 2024 study found Generation Z adults averaging 8.1 of 10 despite otherwise normal testing, and 2026 studies found clock errors in about 30 percent of younger adults against 7 percent of older adults), with no corrected norms yet, so familiarity belongs in the validity line. The MoCA, Mini-Cog, and DCTclock are trademarks or copyrighted instruments of their respective rights holders (MoCA Test Inc., Soo Borson, and Linus Health). BastionGPT is not affiliated with, or endorsed by, any of these rights holders or publishers. This page reproduces no test items, stimuli, norms, or scoring materials.
The numbers behind these errors are specific. In the largest meta-analysis (90 studies, 22,567 participants), paper clock scoring pooled a sensitivity of about 0.63 for mild cognitive impairment whether scoring was brief or detailed; naive and trained raters agreed on only about 60 percent of mildly impaired clocks (against more than 98 percent for clearly normal or impaired ones); pooled dementia accuracy differs by system (about 82 percent sensitivity for Shulman versus 73 for Sunderland); and cognitively normal younger adults now underperform on the clock for familiarity reasons, with no corrected norms yet. The BastionGPT Clinical Advisory Board sees the same errors most often in clock drawing 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.
By whichever named system the clinician chose, and the systems genuinely differ. Simple ordinal systems (Shulman's six categories; Sunderland's 1 to 10) rate the whole clock; detailed point systems (Mendez's 20 points, Freedman's 15, Rouleau's 10 plus a qualitative error taxonomy, the two 15-point CLOX conditions, Freund's 7-point driving-oriented scale) award points for discrete features; and the embedded clocks in the MoCA (0 to 3) and Mini-Cog (0 or 2, deliberately binary so untrained raters can apply a clean pass or fail) belong to their parent instruments. Ranges, directions, conditions, and requested times all vary, and even Shulman implementations number the same categories 0 to 5 or 1 to 6. That is why the documentation rule is to name the system with its range and direction, the condition, the substrate, and the exact time requested alongside the score.
Only within a named system, in a population it was studied in, and never as a diagnosis. Pooled accuracy differs by system (one meta-analysis reported about 82 percent sensitivity and 76 percent specificity for Shulman scoring against about 73 and 88 for Sunderland), comparative studies find no system consistently superior, and the field's own review concluded that for screening, simpler is about as good as complex. The weak point is mild impairment: across 90 studies, paper scoring pooled a sensitivity near 0.63 for mild cognitive impairment whether brief or detailed, so a normal clock cannot rule out early decline, and naive raters agree with experts on only about 60 percent of mildly impaired clocks. Report the named system's result, describe the errors, and let history, collateral, and a fuller assessment carry any decision.
Because 11:10 works the whole task. Setting it requires placing hands in both halves of the clock face, which can expose a neglected hemifield, and it requires recoding the spoken "ten" into the minute hand at the numeral 2 rather than pointing at the 10, which is exactly where a stimulus-bound response shows itself: hands drawn to the 10 and the 11 because those are the numbers that were said. Systems vary (CLOX uses 1:45, some protocols use 2:45 or 8:20, and Freedman's group noted that the popular "3 o'clock" tests neither hemifield balance nor recoding), which is one more reason the requested time is a test specification to chart, not a stylistic choice. A literal hand placement is an observation about how the drawing failed; it does not by itself identify a disease or a lesion.
It is the clock's most informative pattern, and it should be charted as a contrast, not a conclusion. Command drawing requires generating the concept, plan, and layout from instruction alone; copying supplies the model and drops those generative and executive demands while keeping the visuospatial and constructional ones. A markedly better copy therefore shifts suspicion toward planning and executive control rather than visuoconstruction, which is exactly what the CLOX task was designed to separate with its two 0-to-15 conditions. The defensible note reports both conditions with the same named system, describes each drawing, states the discrepancy as suggesting disproportionate executive burden, and routes it into history, collateral, medication review, and fuller assessment rather than into a diagnosis or an anatomical claim.
No. The embedded clocks were validated only inside their instruments, under their own conditions: the MoCA clock is drawn with no pre-drawn circle and scored 0 to 3 on contour, numbers, and hands, and current MoCA policy still points to certification when an individual task is interpreted beyond the total; the Mini-Cog clock uses a permitted pre-printed circle and a deliberate 0-or-2 binary. Re-scoring either with Shulman, Rouleau, or another standalone rubric applies cutoffs that were never validated on those administrations. Chart embedded clocks as components ("MoCA clock 2/3," "Mini-Cog clock 0/2") with a descriptive note if useful, and administer a standalone clock, under a named system and condition, when a standalone score is actually wanted. The SLUMS clock item follows the same rule, and the MMSE contains no clock at all.
Keep evaluating; the screen cannot close that question. Paper clock scoring has poor sensitivity for mild impairment (about 0.63 pooled, brief or detailed), reviews have concluded it should not be used to screen for mild cognitive impairment, and high-reserve patients can produce a normal drawing while declining. The Medicare wellness-visit regulation itself requires due consideration of concerns raised by family, friends, and caregivers, not just performance, so the collateral report is a required input, not an anecdote. Chart the specific examples (missed bills, repeated questions, medication slips), state that the normal screen does not exclude impairment, and document the next step: a longer instrument, an informant tool, functional review, and follow-up with an owner and date.
Sometimes, and it is now a documented validity issue rather than a hunch. A 2024 study found cognitively normal Generation Z adults averaging 8.1 of 10 on a standard clock system despite otherwise normal testing, generational work has shown the effect independent of measured cognition (some younger examinees write a digital "11:10" inside the circle), and 2026 studies reported clock errors in about 30 percent of younger adults against about 7 percent of older adults, with analog-clock literacy predicting clock performance but not a non-clock visuospatial task. No corrected norms or accepted replacement task exist yet, so the practical rule is documentary: ask about and record analog-clock familiarity, describe atypical responses rather than just scoring them, interpret cautiously in younger or digital-native patients, and never chart unfamiliarity as cognitive impairment.
It depends on whose material you touch. A clinic-designed generic sheet (its own instruction wording, blank page or its own circle) is the low-risk route, and the standalone scoring methods can be operationalized as EHR fields (system name, range, condition, time, score, errors) because a method is not copyrightable, though published rubric tables and figures belong to their journals and publishers, so paraphrase rather than transcribe, and the Freedman book is expressly all-rights-reserved. Branded instruments carry their own terms: the MoCA prohibits web posting and unauthorized electronic versions, and the Mini-Cog permits free unmodified clinical use, with its rights holder recommending discrete EHR fields for the component scores. Digital products are proprietary: the leading digital clock device is FDA-listed under a premarket-exempt product code, which is not FDA clearance or approval, and its machine-scored process metrics do not interchange with any paper system's score, so name the product and report its output as its own scale.
Yes. Give it the facts (system with range and direction, condition, substrate, time requested, score, the errors you observed, confounds, disclosure, and plan) and it drafts the full entry: the five specification facts fixed, the qualitative description in words, the screen-not-diagnosis line, and the follow-up chain with owner and date, ready for your review. It can also cross-check a finished note for a score without its system, an embedded clock charted as standalone, a missing validity line, or diagnostic weight a screen cannot carry, and it can draft the serial-comparison paragraph against a prior clock under the same system and condition. 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.