Richmond Agitation-Sedation Scale — a reproducible 10-point scale for titrating ICU sedation from unarousable (−5) to combative (+4), used to steer infusions toward an individualized target
Before validated sedation scales, ICU charts described patients as "somewhat sedated," "a bit agitated," or "resting comfortably" — language that meant different things to different nurses, on different shifts, in different units. The Richmond Agitation-Sedation Scale (RASS), developed at Virginia Commonwealth University and published by Sessler et al. in 2002, replaced that ambiguity with a 10-point ordinal scale, from −5 (unarousable) to +4 (combative), that any trained bedside clinician can score identically.
Each level of RASS describes both the direction (agitation vs. sedation) and the intensity of the patient's state, anchored to observable behavior rather than interpretation:
+4 Combative — overtly violent, immediate danger to staff +3 Very agitated — pulls at tubes/catheters, aggressive +2 Agitated — frequent non-purposeful movement, fights ventilator +1 Restless — anxious, movements not aggressive or vigorous 0 Alert and calm — the neutral, desired resting state −1 Drowsy — not fully alert, sustained (>10s) awakening, eye contact, to voice −2 Light sedation — briefly (<10s) awakens with eye contact to voice −3 Moderate sedation — any movement (but no eye contact) to voice −4 Deep sedation — no response to voice, but movement/eye opening to physical stimulation −5 Unarousable — no response to voice or physical stimulation
The scale is deliberately symmetric around zero: agitation and sedation are treated as two ends of the same continuum, not separate problems, which is exactly how sedative and analgesic titration behaves in practice.
Before RASS (and its close relative, the Sedation-Agitation Scale, SAS), sedation depth was rarely documented as a discrete, trended value. That made it nearly impossible to answer basic clinical questions: Is this patient more sedated than yesterday? Has the propofol infusion been reduced appropriately after last night's dose increase? Is the whole unit drifting toward deeper sedation than intended?
Validation studies scored RASS against blinded observers and found excellent agreement (weighted kappa 0.91 overall, and 0.94 between critical care nurses specifically), even across different patient populations — medical, surgical, and cardiac ICU. That reliability is what allows RASS to function as a shared clinical language: a "−2" charted by the night nurse means the same thing to the day team, the intensivist, and the pharmacist adjusting the drip.
RASS is now embedded in nearly every major ICU sedation guideline (SCCM PADIS 2018), ventilator liberation protocol, and delirium-screening workflow (the CAM-ICU is only interpretable once RASS confirms the patient is arousable enough to test). It is the common unit of measurement that all downstream sedation decisions are built on.
RASS is not scored by impression — it is scored by following a fixed, stepwise stimulus escalation. The assessor observes first, escalates to verbal stimulation only if needed, and escalates to physical stimulation only if voice fails. Wherever the patient responds is where the scale stops, which is what makes the score reproducible between different clinicians and different shifts.
Step 1 — Observe (scores +4 to +1 or 0): Watch the patient without any stimulus. Is there overt aggression (+4)? Pulling at lines (+3)? Frequent non-purposeful movement (+2)? Anxious but non-aggressive movement (+1)? Or is the patient simply alert and calm (0)? If the patient is not spontaneously alert or agitated, move to step 2.
Step 2 — Verbal stimulation (scores −1 to −3): Say the patient's name in a normal, then progressively louder voice, and ask them to look at the assessor. Sustained (>10 s) eye contact or response scores −1 (drowsy). Brief (<10 s) eye opening/contact scores −2 (light sedation). Any movement at all in response to voice, without eye contact, scores −3 (moderate sedation). If there is no response to voice, move to step 3.
Step 3 — Physical stimulation (scores −4 or −5): Shake the shoulder, then rub the sternum if needed. Any movement to physical stimulation scores −4 (deep sedation). No response of any kind to voice or touch scores −5 (unarousable).
The rule is strict: never skip a step, and stop at the first level that produces a response — do not continue escalating "just to check."
The stepwise design does two things simultaneously: it minimizes unnecessary stimulation of an already-comfortable patient (you never touch someone who responds to your voice), and it forces a consistent stopping rule so two different nurses land on the same number.
Because the test is graded to stop at the mildest stimulus that produces a response, RASS assessment is also, by design, gentle — a light-sedation patient is never sternal-rubbed just to confirm a score, since the voice step already answered the question. This matters for patient comfort and for avoiding the confound of a painful stimulus artificially raising an otherwise stable agitation score.
RASS is typically paired with the CAM-ICU (Confusion Assessment Method for the ICU) for delirium screening — but CAM-ICU is only valid at RASS ≥ −3, since a patient who is too sedated to engage cannot be meaningfully assessed for delirium. The assessment sequence therefore gates the entire downstream delirium-monitoring workflow.
A structured 3-step test sounds simple, but before its adoption many "agitation" scores were really scores of noxious over-stimulation — assessors jumping straight to a sternal rub. The fixed escalation order is what makes RASS a measurement instrument rather than an impression.
For decades, standard ICU practice was to sedate patients deeply and continuously, largely to ensure ventilator synchrony and prevent self-extubation. Landmark trials in the 2000s–2010s reversed that default: current best-practice guidance calls for setting an explicit, individualized target RASS — usually light sedation, roughly −2 to 0 — and treating deep sedation as the exception that requires justification, not the starting point.
Trials such as the Awakening and Breathing Controlled (ABC) trial, the MENDS trial, and the SLEAP (SLEep in the intensive care unit and Awakening trial with Propofol) study each compared protocolized light sedation against usual, deeper sedation practice. Across this evidence base, lighter targets were associated with shorter time on the ventilator, shorter ICU stay, and no increase in adverse psychological outcomes such as PTSD — overturning the earlier assumption that heavier sedation was simply "safer" or "kinder."
The 2018 SCCM Pain, Agitation/sedation, Delirium, Immobility, and Sleep disruption (PADIS) guideline formalizes this: it recommends using light sedation (arousable, spontaneously or with mild stimulation) rather than deep sedation in critically ill, mechanically ventilated adults, unless clinically indicated otherwise (e.g., therapeutic paralysis, severe ARDS with dyssynchrony, status epilepticus, elevated intracranial pressure).
A single target is not applied to every patient uniformly. The multidisciplinary team — intensivist, bedside nurse, and pharmacist — sets and documents a target RASS range for each patient at each point in their course, and that number becomes the reference the whole team titrates against for that shift.
Common reasons to deliberately target deeper sedation than −2 to 0 include: refractory intracranial hypertension, severe ventilator dyssynchrony despite optimization, active seizures, or use of neuromuscular blockade (which mandates deep sedation, since a paralyzed, undersedated patient cannot show distress). Outside of those indications, the guideline default is light sedation, reassessed and re-targeted as the clinical picture evolves.
Setting the target before adjusting the dose reframes titration from "how much drug is this patient getting" to "how far is this patient from where we intend them to be" — which is exactly the comparison the titration workflow in the next stage is built around.
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| General ventilated ICU patient | −2 to 0 | Arousable, cooperative, able to participate in spontaneous breathing trials | Shorter vent time, less delirium, earlier mobilization |
| Severe ARDS / ventilator dyssynchrony | −3 to −4 | Deeper sedation to achieve lung-protective ventilation when light sedation fails | Reduces dyssynchrony-driven lung injury |
| Neuromuscular blockade in use | −5 (mandatory) | Paralyzed patients cannot signal distress; deep sedation/analgesia is required | Prevents awareness during paralysis |
| Elevated intracranial pressure | −4 to −5 | Deeper sedation blunts ICP surges from coughing, straining, stimulation | Protects cerebral perfusion pressure |
Once a target RASS is set, sedation becomes a closed feedback loop rather than a fixed order: assess the patient, compare the score to the target, adjust the infusion rate accordingly, and reassess again shortly after. Too agitated relative to target means the dose is increased; oversedated relative to target means it is decreased; at target means the current rate is held.
At each reassessment, the bedside nurse compares the current RASS to the documented target and follows a simple decision rule:
• Current RASS above target (more agitated/awake than intended) → increase the sedative infusion rate, typically in 10–25% increments, and reassess sooner (often within 30 minutes)
• Current RASS at target → maintain the current infusion rate; continue routine reassessment interval
• Current RASS below target (more sedated than intended) → decrease the infusion rate, again reassessing to confirm the patient moves back toward target without rebound agitation
This loop is deliberately incremental rather than aiming for an instant correction — because sedatives have context-sensitive half-lives that vary with infusion duration, overcorrecting in either direction risks overshooting the target on the next assessment.
Trials comparing nurse-implemented, protocolized sedation titration (where bedside nurses adjust infusion rates within a pre-approved range according to the RASS-based algorithm) against traditional physician-ordered, as-needed adjustments consistently found the protocolized approach achieved target RASS faster and maintained it more consistently — because the person doing hourly reassessment is also the person empowered to act on it immediately, without waiting for the next physician order.
Daily sedation interruption ("sedation vacation") is often layered on top of continuous titration: the infusion is paused once daily to reassess the patient's neurological baseline and confirm the ongoing need for sedation at all, then restarted at a lower rate if appropriate — directly testing whether the current target itself should be revised.
Titration is only meaningful if reassessment is frequent enough to catch drift in either direction — a target that is set once and never re-checked against actual RASS scores provides no real protection against oversedation.
The clinical case for light, target-directed sedation rests on a consistent finding across observational and randomized studies: sedation deeper than clinically necessary is not a neutral, "safe" default — it is associated with measurably worse outcomes, including longer time on the ventilator, longer ICU stay, and substantially higher odds of delirium.
Deeper-than-necessary sedation drives downstream harm through several linked pathways:
• Prolonged immobility — deeply sedated patients cannot participate in early mobilization or physical therapy, contributing to ICU-acquired weakness • Delayed spontaneous breathing trials — a patient who cannot be lightened enough to test readiness for extubation stays ventilated longer by definition, which itself carries risk (ventilator-associated pneumonia, further sedative exposure) • Disrupted sleep-wake architecture — many sedatives (particularly benzodiazepines) suppress normal sleep architecture rather than replacing it, compounding delirium risk • Reduced ability to detect neurological change — a heavily sedated patient masks evolving problems (stroke, worsening encephalopathy) that would be apparent in a lighter, arousable patient
Each of these mechanisms independently lengthens the ICU course, and they compound: more ventilator days increases delirium risk, and delirium itself independently predicts longer stay and worse long-term cognitive outcomes.
Delirium is consistently the most sensitive marker of oversedation harm in the literature. Benzodiazepine exposure and deeper sedation depth are among the most reproducible modifiable risk factors for ICU delirium identified in large cohort studies, and delirium itself is independently associated with longer mechanical ventilation, longer hospital stay, higher mortality, and long-term cognitive impairment that can persist for a year or more after discharge (post-intensive care syndrome).
This is the core rationale behind bundled ICU care approaches (such as the ABCDEF bundle: Assess/prevent/manage pain, Both spontaneous awakening and breathing trials, Choice of sedation, Delirium assessment, Early mobility, Family engagement) — RASS-guided light sedation is not an isolated intervention but one link in a chain explicitly designed to reduce delirium and its downstream consequences.
Higher compliance with ABCDEF-bundle elements — including maintaining light, target-directed sedation — has been associated in large multicenter cohorts with a dose-dependent reduction in hospital mortality, delirium, and coma, underscoring that the RASS target is not a documentation exercise but a measurable lever on patient outcomes.