Tests of cognitive function, including delirium
Tests screen, they do not diagnose
- Cognitive tests screen and quantify; they do not diagnose
- The diagnosis comes from the history, the collateral history and the functional impact. A score is a number to track, not a label
- Three questions, three different tools
1. Is this delirium? -> 4AT, CAM
2. Is there cognitive impairment, and how much? -> MMSE, MoCA, ACE-III, RUDAS
3. Which syndrome is it? -> domain-specific testing, imaging, neuropsychology
Domains to cover
- Attention (the delirium domain), orientation, episodic memory, language, visuospatial, executive function, praxis, social cognition
Epidemiology
- Cognitive impairment is present in ~30-40% of older medical inpatients and is undetected in most without routine screening
- Delirium is missed in up to 2/3 of cases when no structured tool is used
- Screening instruments have far lower specificity in acute illness, low education, non-English speakers and sensory impairment - false positives are common in hospital
What each test measures - and its blind spot
What each test is actually measuring - and its blind spot
| Test | Score | Strength | *Blind spot* |
|---|---|---|---|
| MMSE | /30 | Familiar, tracks progression, PBS criterion for cholinesterase inhibitors | *Does NOT test executive function or social cognition - often normal in bvFTD and early vascular dementia*; heavily education- and language-dependent; ceiling effect |
| MoCA | /30, <26 abnormal (add 1 point if <=12 yr education) | More sensitive for MCI; tests executive function, abstraction, delayed recall | Ceiling in advanced disease; needs literacy and vision/hearing |
| ACE-III | /100 | Best for subtyping - separate attention, memory, fluency, language, visuospatial subscores | Takes ~20 min |
| RUDAS | /30 | Designed for cultural and linguistic diversity and low literacy; interpreter-friendly | Less sensitive to mild deficits |
| KICA-Cog | Validated for Aboriginal and Torres Strait Islander people | Specific to that population | |
| AMT / AMT4 | /10, /4 | Fast bedside | Crude; orientation-heavy |
| Clock drawing | Executive + visuospatial in 60 seconds | Not a stand-alone test | |
| 4AT | /12, >=4 = possible delirium | No training needed, works in the drowsy or uncooperative patient, the ward standard | Does not separate delirium from dementia by itself |
| CAM / CAM-ICU | Algorithm | High specificity for delirium | Needs training; CAM-ICU for the ventilated |
| IQCODE / informant scale | Asks the informant about change over 10 years - the only tool that establishes a baseline | Requires a reliable informant | |
| GDS-15 | /15 | Screens depression, which mimics dementia | Not valid in moderate-severe dementia |
Attention tests - the delirium bedside battery
- Months of the year backwards (the most useful single test), days of the week backwards, serial sevens, digit span, "WORLD" backwards
Ancillary investigations in the cognitive workup
- Structural imaging (CT/MRI) - excludes tumour, subdural, hydrocephalus, stroke burden; shows hippocampal or lobar atrophy pattern
- FDG-PET - metabolic pattern separates AD (temporoparietal + posterior cingulate) from FTD (frontal/anterior temporal) from DLB (occipital, with the cingulate island sign)
- Tc-99m HMPAO SPECT - measures regional cerebral perfusion; characteristic hypoperfusion patterns support a subtype where FDG-PET is unavailable
- DaT SPECT - striatal dopamine transporter uptake; abnormal in DLB, normal in AD
- CSF and plasma biomarkers (Abeta42/40, p-tau181, p-tau217) - specialist settings
- EEG - non-convulsive status, Creutzfeldt-Jakob disease, encephalitis
- Formal neuropsychological assessment - the definitive test: for young onset, high-premorbid-function patients, atypical presentations, medico-legal and capacity questions, and to separate depression from dementia
Using the tools together
1. Screen for delirium first - an abnormal score in a delirious patient is uninterpretable as evidence of dementia
2. Get the collateral history and the functional impact - IADLs (finances, medications, driving, cooking) fail before ADLs
3. Choose the instrument to fit the patient - language, education, hearing, vision, literacy, cultural background
4. Re-test after recovery - cognitive assessment 6-8 weeks after an acute illness or delirium resolves, not during it
Conditions for a valid test
- Hearing aids in, glasses on, well lit, quiet, patient not in pain, not sedated, not hypoglycaemic
- An uncorrected hearing loss will produce a false-positive score in any of these tests
- Interpreter if needed - a family member is not an adequate interpreter for cognitive testing
- Document education, first language, occupation and premorbid function alongside the score
Interpreting a score
- A single score is a snapshot; serial scores are the information
- A normal MMSE does not exclude dementia - high premorbid function, frontal or visuospatial presentations, or early disease
- An abnormal score does not diagnose dementia - delirium, depression, deafness, low education, anxiety and non-English-speaking background all lower it
- *Serial testing is only comparable if the same test is used* - MMSE and MoCA are not interchangeable
After the test
- If delirium: treat as delirium, find the cause, defer any diagnostic labelling of dementia, and arrange follow-up cognitive assessment after recovery
- If cognitive impairment without delirium: complete the dementia workup (bloods, imaging, depression screen, medication review), determine the syndrome, and disclose the diagnosis properly
- If normal but the informant is concerned: believe the informant - proceed to more sensitive testing (ACE-III or neuropsychology) rather than reassuring
- If the patient is distressed by testing: stop; the relationship matters more than the score
Practical consequences
Practical consequences that follow from the assessment
- PBS access to cholinesterase inhibitors requires a documented baseline score (MMSE >=10 or an equivalent validated instrument) and demonstrated benefit at review
- Driving: cognitive test scores alone do not determine fitness to drive - *assess per Austroads Assessing Fitness to Drive, with on-road assessment where needed*; know your jurisdiction's reporting obligations
- Capacity is decision-specific and is NOT determined by a test score - a person with an MMSE of 18 may retain capacity for some decisions and not others; assess understanding, retention, weighing and communication for the decision in question
- Advance care planning, enduring power of attorney and guardianship should be addressed while capacity remains
- Aged care assessment (ACAT), carer support, safety review (driving, cooking, finances, wandering, firearms)
Confounders
- Delirium, dementia, depression ("the three Ds") - frequently coexist
- Sensory impairment - hearing and vision; the commonest correctable cause of a falsely low score
- Low education, non-English-speaking background, cultural difference - major sources of test bias
- Anticholinergic and sedative burden
- Obstructive sleep apnoea, hypothyroidism, B12 deficiency, alcohol
- Capacity, guardianship and elder abuse assessments
- Driving fitness and work capacity
Expected trajectories
- Expected trajectory in untreated Alzheimer disease: MMSE falls ~2-4 points per year
- An abrupt fall is delirium, depression, a drug or a new stroke - not progression
- MCI: ~10-15% per year converts to dementia; a proportion revert to normal, especially where depression, sleep apnoea or drugs were the cause
- Cognitive testing after delirium: many patients do not return to baseline for weeks to months - re-test at 6-8 weeks, not at discharge
- Practice effects occur with repeated administration at short intervals - space serial testing at least 6 months apart where possible
🔒
12 more sections, plus exam facts
Premium unlocks every note across every specialty, and the full exam fact library behind it.
Get premium access