Risk factors present (max 15) — a checklist tally, not a validated prediction score. For a validated ICU model use E-PRE-DELIRIC (at admission) or PRE-DELIRIC (after 24 h).
Bands: Few 0–3 · Several 4–6 · Many 7–10 · Very many 11+ (heuristic)
Evidence of acute change from mental status baseline, OR fluctuation in mental status in the past 24 hours (e.g., on RASS/GCS or a previous delirium assessment)? Anchor the baseline early — ask family/collateral or check the H&P on admission; a normal baseline may be presumed in a younger patient admitted from home with no known cognitive disease. No informant? Documented fluctuation over the past 24 hours (RASS/GCS or a prior delirium assessment) still satisfies Feature 1.
Letters task: say S-A-V-E-A-H-A-A-R-T (10 letters, ~1 every 3 sec); patient squeezes on each "A".
Can't do letters (deaf, language barrier, non-verbal)? Use the Pictures ASE — show 5 pictures to memorize, then 10 recognition pictures (yes/no); >2 errors out of 10 = positive (same 0–10 scale as the letters task).
Feature 3 is read from the RASS you documented above — present whenever the RASS is anything other than alert and calm (zero), so there is nothing to enter here.
Ask the 4 yes/no questions, then the command. Disorganized thinking present if combined errors >1.
- Will a stone float on water?
- Are there fish in the sea?
- Does one pound weigh more than two pounds?
- Can you use a hammer to pound a nail?
Command: "Hold up this many fingers" (examiner holds up 2), then "Now do the same thing with the other hand" — do not repeat the number of fingers. If the patient cannot move both arms, for the second part ask them to "Add one more finger". Any wrong answer or failure to complete the entire command counts as one error.
Limitations: score against the patient's baseline; reduced accuracy in primary neurologic injury/TBI, aphasia, or deep sedation. In documented dementia specificity falls, so a positive result is less certain and is confirmed clinically (shown in older ED patients). Missed cases are predominantly hypoactive — screen every shift.
A positive CAM-ICU is a screen, not a diagnosis — confirm clinically and exclude mimics.
Delirium vs mimics
Depression — can look hypoactive; mood-driven, attention less affected.
Receptive aphasia / focal stroke — focal deficit, not global inattention → neuro exam + imaging.
Non-convulsive seizures / status — unexplained ↓LOC → EEG.
Catatonia — posturing, mutism, waxy flexibility.
Intoxication / withdrawal — substance history; treat the cause.
These can coexist with delirium — known dementia does not exclude superimposed delirium.
Logged this session — scratchpad, clears on reload, not saved to the record
No assessments logged this session.
SAT / SBT safety screen & failure criteria
SAT — do NOT interrupt sedation if: active seizures · alcohol withdrawal on sedation · escalating sedation for agitation · neuromuscular blockade · MI in prior 24 h · raised ICP.
SAT failure → stop SAT, restart sedation at HALF the prior dose if: sustained anxiety/agitation/pain · RR >35 for >5 min · SpO₂ <88% for >5 min · acute dysrhythmia · ≥2 signs of respiratory distress.
SBT — proceed only if: SpO₂ ≥88% on FiO₂ ≤50% & PEEP ≤7.5 · no agitation · no ischemia in 24 h · hemodynamically stable · no raised ICP.
SBT failure → abort, resume support if: RR >35 or <8 · SpO₂ <88% · abrupt mental-status change · dysrhythmia · ≥2 distress signs.
Loop: pass SAT → proceed to SBT; fail SAT → restart sedation at half the prior dose; fail SBT → resume full ventilatory support (restart sedation at half the prior dose if needed); re-screen in 24 h.
eCASH: minimal sedation is the default — analgesia first, lightest RASS that meets the goal, reassess each shift.
Safety screen first — reddest parameter wins; a single red = hold & discuss with the team.
Progression: passive ROM → active ROM / bed exercises → sit / edge of bed → stand / transfer → ambulate.
Glasses & hearing aids in? Other sensory deficits addressed? (Environment — whiteboard, daytime lights — lives under Dr. DRE on the Treatment tab.)
Check SpO₂, Hgb; MI, stroke, pulmonary embolism.
Fever, leukocytosis, cultures pending? Occult sepsis?
Urinary retention or constipation? Bladder scan; disimpact or catheterize if indicated.
Consider non-convulsive seizures / status — esp. with unexplained ↓LOC. Consider EEG.
Volume status, intake, electrolytes. Parenteral thiamine promptly in at-risk patients — do not delay glucose for hypoglycemia (ASAM 2020: in either order or together; EFNS still advises thiamine before carbohydrate). At-risk (malnutrition / alcohol use): 100–300 mg IV daily. Suspected Wernicke — guidelines diverge (low-certainty evidence): EFNS 200 mg IV TID or RCP 500 mg IV TID × 2–3 days then 250 mg taper.
Na, Mg, Ca, glucose, BUN/Cr, acid-base, liver function. Pain/retention/constipation?
Subdural hematoma (recent fall or anticoagulation)? Sleep deprivation?
Algorithm overview (visual) — the Step cards below are the working checklist
Quick dosing reference
Quetiapine 12.5–25 mg q12h PO
Dexmedetomidine 0.2–0.7 mcg/kg/hr (preferred when agitation prevents weaning)
Lorazepam — withdrawal only (titrate to RASS)
Full doses, cautions & deliriogenic-med review → Medications
Restraints are the last resort after de-escalation, environment, device removal, and 1:1 observation — use the least-restrictive option, re-evaluate per institutional policy, and discontinue at the earliest possible time.
These are conventional / expert starting references for short-term agitation control — not RCT- or guideline-calibrated delirium doses. Dexmedetomidine is a sedative infusion for ventilated patients, not a PRN antipsychotic-equivalent.
Facility name is set in the header field at the top of the page — it prints on every generated PDF.
This label prints on the protocol documents. The CAM-ICU tab itself implements the CAM-ICU only — units using the ICDSC (8 items over the shift; total ≥ 4 = positive) score it on their own worksheet and use this tool for the prevention, review, and treatment content.
Deeper sedation (RASS ≤ −3) is associated with longer ventilation, more delirium, and higher mortality (observational evidence plus one small RCT; PADIS is conditional, low-certainty). Use it only for a specific indication — it is not a delirium-prevention default.
Settings also auto-save in this browser. Save settings writes a
settings.json; Load settings reads one back — for
sharing a configured protocol or moving it between machines.