Sign Up
Login
Try for Free
Login
The Richmond Agitation-Sedation Scale (RASS) is a ten-level observational scale, published by Sessler and colleagues in 2002, that places a patient on one continuum from deep sedation through calm alertness to severe agitation. ICU and step-down nurses, intensivists, respiratory therapists, and pharmacists use it to set and check sedation targets and to gate delirium screening. This page covers how to document RASS assessments, with a fictional sample.
Bedside ICU and step-down nurses at each assessment, with intensivists, advanced practice providers, respiratory therapists, and pharmacists reading and acting on it; no license, training certificate, or publisher qualification level applies, and the rights holder permits free clinical use
Relieving nurses at handoff, intensivists and the rounding team who set the sedation target, pharmacists reviewing sedative and analgesic orders, respiratory therapists timing awakening and breathing trials, delirium screeners, quality staff counting ICU Liberation bundle assessments, and later the reviewers who reconstruct a sedation course from the chart
1 to 3 chart lines per assessment (time, assessed value, ordered target, stimulus where relevant, sedation on board, action if off target) · assessment takes under a minute at the bedside
Clinician-observed ten-level ordinal scale on one continuum (negative values for increasing sedation depth, zero for calm and alert, positive values for increasing agitation; the stimulus needed to obtain a response is built into the negative values)
Continuous sedation of mechanically ventilated adults, titration of sedative and analgesic infusions, spontaneous awakening trials, the arousal check before every CAM-ICU, agitation episodes in the ICU, emergency department, and step-down unit, palliative sedation with the RASS-PAL, and every handoff and transfer that carries a sedation status
Sessler and colleagues (AJRCCM, 2002) with longitudinal validation by Ely and colleagues (JAMA, 2003); copyright registered by Virginia Commonwealth University in 2012 with free use for health systems, clinicians, and researchers; named by the SCCM PADIS guideline and the ICU Liberation bundle, mandated by no law; described here for documentation, no level descriptors reproduced
The Richmond Agitation-Sedation Scale is a bedside observational scale of sedation and agitation developed at Virginia Commonwealth University in Richmond, Virginia, by a team that was multidisciplinary by design: the eight authors of the November 2002 paper in the American Journal of Respiratory and Critical Care Medicine (Sessler, Gosnell, Grap, Brophy, O'Neal, Keane, Tesoro, and Elswick) included physicians, nurses, pharmacists, and a biostatistician, and the first reliability panel was two physicians, two nurses, and one pharmacist. The scale has ten levels on a single continuum: four positive values describe increasing agitation, from restlessness to behavior that endangers the patient or staff; zero describes a patient who is awake and calm without any prompting; and five negative values describe increasing depth of sedation, separated by how much stimulation it takes to get a response and by the quality of that response. The assessment runs as a sequence rather than a menu. The clinician first watches the undisturbed patient, and a patient who is alert, or who is restless or agitated, is rated from observation alone. If the patient is not spontaneously alert, the clinician speaks, saying the name and asking for eye opening and eye contact, and rates by whether eye contact is made and how long it is held. Only when there is no response to voice does the clinician use physical stimulation, and the two deepest values are separated by whether that produces any response at all. The value therefore carries the stimulus inside it: a negative number already tells the next reader whether voice or touch was needed. The wording that defines each level is the scale's copyrighted text and is not reproduced here; the published scale is the source of record. Ely and colleagues validated the scale over time at Vanderbilt (JAMA, June 2003): in 38 patients, two nurses agreed with a weighted kappa of 0.91 across 290 paired observations, against 0.64 for the Glasgow Coma Scale; in 275 mechanically ventilated patients, RASS values correlated with the GCS (r 0.91), with an attention screen (r 0.78), and with bispectral EEG (r 0.63), and tracked change across consecutive ICU days. That Vanderbilt group, now the Critical Illness, Brain Dysfunction, and Survivorship Center, made the RASS the arousal step that precedes the CAM-ICU, which is why the two are usually taught together.
The load-bearing fact for documentation is that a RASS value describes an observed state, and the record needs two more things before it describes sedation management: the target the team ordered and the sedation the patient had on board when the value was taken. The scale does not need a target to be valid; Ely and colleagues reported that their ICU had no sedation protocol and no routinely identified objective sedation targets during the validation, and the scale performed anyway. What the target and the drug context add is intent: whether an assessed value is what the team wanted, an overshoot after a bolus, an unintended drift, or a deliberate exception such as deep sedation under neuromuscular blockade. Three boundaries follow. The RASS is not a consciousness scale for brain injury: the Glasgow Coma Scale describes eye, verbal, and motor responses after injury and does not measure sedation depth, so in a sedated neurological patient the two are charted side by side rather than one standing in for the other. The RASS is not a delirium tool: it decides whether the patient is arousable enough for the CAM-ICU (at the two deepest values the screen is recorded as unable to assess), and the delirium features are a separate assessment, so a calm and alert patient can still be delirious. And the RASS is not a pain scale: PADIS asks for a behavioral pain instrument such as the CPOT or BPS in patients who cannot self-report, and pain is assessed and treated before sedation is escalated. Alternatives and variants have their own names and are charted under them: the Riker Sedation-Agitation Scale (1999) is the other sedation scale the SCCM guidelines treat as valid and reliable; the older Ramsay scale runs in the opposite direction and separates agitation poorly; the RASS-PAL (Bush and colleagues, 2014) is a palliative care modification with pilot evidence only; the modified RASS (Chester and colleagues, 2012) adds an attention item for ward delirium screening; and the adult scale has been validated in critically ill children (Kerson and colleagues, 2016; Tapia and colleagues, 2021) without adult targets automatically transferring.
Critical care nurses do most of the charting: on admission, at the interval the unit protocol sets, with each titration of a sedative or analgesic infusion, before and after a bolus, before every CAM-ICU, and before, during, and after a spontaneous awakening trial. Intensivists and advanced practice providers write the target as a value or a range in the sedation order and revisit it on rounds; pharmacists read the assessed values against the target when they reconcile infusion rates, cumulative doses, and benzodiazepine exposure; respiratory therapists time the spontaneous breathing trial to the awakening trial and need the RASS trajectory to do it; and quality staff count documented arousal assessments for the ICU Liberation bundle. Beyond the ICU, emergency and step-down nurses use it during agitation episodes and the continuous sedation of ventilated patients waiting for a bed, procedural teams use it beside their required monitoring, and palliative care teams use the RASS-PAL during palliative sedation and agitated delirium, naming the variant. The value travels with the patient: a transfer-of-care summary should carry the current target, the last assessed value with its time, active infusions, the last CAM-ICU result, and the awakening trial status, and an incident report after a self-extubation or a restraint-related injury should record the RASS at the moment of the event with its precipitant. Neighbors win in four places: the GCS when the question is consciousness after brain injury, the CAM-ICU for delirium, the CPOT or BPS for pain, and, on general wards where the RASS was never validated, a ward opioid sedation score or the ACVPU consciousness item inside an early warning score such as MEWS or NEWS2.
No law, payer, or guideline prescribes a RASS note format; what exists is a copyrighted scale, a validation literature that ties it to goal-directed sedation, guideline and bundle conventions that expect a documented target and regular arousal assessments, and unit protocols that set the interval. What survives review is an entry that carries a clock time and its place in the course of care, the ordered target as written, the assessed value with the stimulus where it is not self-evident, the sedation and analgesia on board at that moment, what was done when actual and target differed, the awakening trial trajectory and the CAM-ICU it leads to, and an honest entry rather than a number under neuromuscular blockade. Each element below carries the pitfall that most often undermines it.
Time, assessor, and the moment in the course of care. Give every assessment a clock time and place it: scheduled per protocol, after a titration, minutes after a bolus, during a procedure, before or during an awakening trial, on arrival from the emergency department. The interval is protocol, not regulation: Critical Care BC asks for depth of sedation every four hours or more often, Vanderbilt's trauma protocol monitors each shift, one UK cardiothoracic unit requires hourly entries, and the SCCM ICU Liberation data manual counts documented level-of-arousal assessments per 24 hours with a goal of at least six. Name the assessor, and when a value was taken within about half an hour of a bolus or a rate change, say so in the same line, because a post-bolus value describes a transient peak rather than the infusion. Pitfall: A column of values with no times, so nobody can tell which one followed the 13:10 propofol bolus and which one describes the running infusion.
The ordered target, as written. Chart the target exactly as the order states it, a single value or a range, beside every assessed value or in the header of the flowsheet the values sit in, and name the order it comes from and the date it was last reviewed. Do not translate the order into a phrase such as light sedation, and do not silently narrow a range to one number. If the ordered target is deeper than the guideline preference, record the indication (neuromuscular blockade, status epilepticus, refractory intracranial hypertension, severe ventilator dyssynchrony); if no target has been ordered, obtain one and chart the request, because a value with no target cannot be read as on or off course. Vanderbilt's trauma protocol states that "goal and actual RASS should be specified on ALL patients," and Critical Care BC asks for a documented daily RASS goal with sedatives titrated to it. Pitfall: Assessed values with no target anywhere in the record, or an order for -1 to 0 charted as a target of -2 because that is what the unit usually uses.
The assessed value, with the stimulus where it matters. Follow the sequence (observe, then voice, then physical stimulation) and record the value it produces, not the value the infusion rate would predict. A negative value already implies the stimulus that was needed, so a stimulus note is not required beside every number, but write it when the picture is ambiguous: eyelid swelling, hearing loss or a language barrier, a neurological injury that alters eye opening, a patient who looks calm with eyes open but does not attend to voice, or a value that disagrees with the last one. A patient asleep who wakes fully to voice is scored by that response, not as deeply sedated. Where the observed behavior would let two raters pick different values, add three or four words of behavior; a 2025 study found an ICU nurse and two trained raters chose the identical value in only 35 percent of encounters and flagged ambiguity in 16 percent. Pitfall: A zero charted because the eyes were open, or a deep value charted because they were closed, with nothing to show what the patient did when spoken to.
Sedation and analgesia on board. Record what the patient had at the time of the assessment: each sedative and analgesic infusion with its rate, any bolus with the clock time, the last rate change and when, and the non-drug measures in use. In an electronic record this is often a matter of aligning the RASS row with the medication administration record rather than retyping doses, but the note that explains an off-target value should name the drug and the time in words. When agitation is the finding, chart the pain assessment made with a behavioral tool before any sedative response, because PADIS puts routine pain assessment and treatment ahead of sedative escalation and the ICU Liberation bundle keeps pain, sedation choice, and delirium as separate elements. Pitfall: A value taken twenty minutes after a procedural bolus read as the effect of the infusion, or an agitated patient given more sedative with no pain score anywhere near the entry.
Off target: recognition, cause, action, recheck. When the assessed value differs from the target, the entry has four parts. Recognition: state the direction, deeper or lighter than target, rather than leaving the reader to compare two columns. Cause: for a deeper value, a recent bolus, drug accumulation, a new neurological change, or a deliberate exception; for a lighter or agitated value, pain, hypoxemia or dyssynchrony, a full bladder or other discomfort, withdrawal, delirium, a procedure, or an environmental trigger. Action: the titration made under the medication order or unit protocol, with the new rate and time, or the reason the team chose not to titrate, and the notification when the pattern or the indication needs a prescriber. Recheck: the follow-up value at the interval the order or protocol sets. The scale itself orders nothing; the titration authority is the order. Pitfall: A value two levels deeper than target followed by "continue current infusion," or a score-to-sedative reflex with no cause considered.
Awakening trial trajectory and the CAM-ICU link. For a spontaneous awakening trial, chart the pre-trial value with its time, the safety screen result, the time each sedative was stopped and which infusions continued, the values during the trial with their times, the criteria watched for failure (sustained agitation, respiratory distress or a high respiratory rate, desaturation, an arrhythmia, or other instability under your protocol) and whether any occurred, the coordination with the respiratory therapist for the breathing trial, and, if sedation was restarted, the reason, the time, and the rate. Once the patient is above the arousal gate and a delirium screen is due, record the CAM-ICU result after the trajectory, not before it; at the two deepest RASS values the screen is recorded as unable to assess, any lighter value allows it, and a calm and alert value does not by itself say anything about delirium. Pitfall: "Propofol held per SAT" with no values, times, or restart rate, or a CAM-ICU charted as negative on a patient whose RASS that hour made it unassessable.
Neuromuscular blockade, restraint, and handoff. Under effective neuromuscular blockade, do not manufacture a value: the scale depends on visible responses that paralysis abolishes, so chart that the RASS cannot be assessed because of pharmacologic paralysis, carry forward the last assessment before the blocker with its time, record the sedative and analgesic regimen and the blocker's start time, and name the unit's monitoring strategy for sedation depth, processed EEG if used; the 2016 SCCM guideline calls for sedation and analgesia before and during blockade, and the 2026 SCCM guideline for ARDS treats scale-based versus non-scale evaluation before blockade, and monitoring versus no monitoring during it, as open questions with very low certainty evidence. When agitation ends in restraint, the RASS describes the behavior and does not authorize the restraint: document the behavior and precipitant, the less restrictive measures tried, the order, the times applied and removed, and the response, as 42 CFR 482.13(e) and your policy require. At handoff and transfer, carry the target, the last value with its time, the infusions, the last CAM-ICU, and the awakening trial status. Pitfall: RASS -5 charted every hour on a paralyzed patient as if it were a sedation assessment, or "RASS +3, restraints applied" with nothing else.
RASS DOCUMENTATION BLOCK (assessment with target and context) Date: [ ] Time: [ ] Setting: [ICU / step-down / ED / palliative unit] Assessor: [ ] Scale version: [RASS / RASS-PAL / pediatric use: state it] Ordered target: [single value or range, exactly as written] Order date and prescriber: [ ] Indication if deeper than the unit default: [neuromuscular blockade / seizures / ICP / severe dyssynchrony / other] Assessed RASS: [ ] Stimulus if not self-evident: [observation / voice / physical] Behavior in a few words (when ambiguous): [ ] Confounders: [hearing, language, eyelid swelling, neurological injury, sleep] Moment in care: [scheduled / after titration / [ ] minutes after bolus / procedure / awakening trial / arrival] Sedation and analgesia on board: [drug, rate, last change and time / bolus and time / non-drug measures] Relation to target: [at target / deeper than target / lighter or agitated] Likely cause: [ ] Action under the order or protocol: [new rate, time / none, reason] Prescriber notified: [who, time, response] Recheck due: [time] Recheck value: [ ] Awakening trial (if today): pre-trial RASS and time [ ] safety screen [pass / fail: why] sedative stopped [drug, time] values during [ ] failure criteria observed [none / which] breathing trial [with RT, time, result] restarted [drug, rate, time, reason] CAM-ICU link: [performed at RASS [ ], time, result / unable to assess at -4 or -5 / due at ] Neuromuscular blockade: [RASS not assessable, paralyzed since [time] / last pre-blockade RASS and time / sedation and analgesia running / monitoring strategy] Agitation event: [behavior, precipitant, pain score and tool, measures tried, medication and time, restraint order and times if used, response] Handoff: [target, last RASS and time, infusions, last CAM-ICU, awakening trial status] 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 scale's level descriptors or scoring card.
Scenario: a US academic medical ICU, day three of mechanical ventilation for pneumonia with acute hypoxemic respiratory failure in a 64-year-old woman; the day nurse charts the ordered target beside each assessed value, a spontaneous awakening trial that leads into a CAM-ICU, an expected overshoot after a procedural bolus, an agitation episode with its cause and a brief restraint, and the handoff. All details are fictional.
Patient: L.M., 64 · Setting: Medical ICU, academic hospital; day 3 of mechanical ventilation · Clinician: J. Ortega, RN, CCRN · Note date: 09/14/2026
Target, overnight context, and pre-trial assessment (06:50): RASS as built in the unit flowsheet, assessed by J. Ortega, RN, at 06:50 on 09/14/2026. Ordered target: RASS -1 to 0, written by the intensivist in the sedation order of 09/12/2026 and reaffirmed on rounds 09/13/2026; the unit protocol (MICU-SED-04) sets assessment every two hours, with every titration, and before every CAM-ICU, which is policy rather than a regulatory interval. On board: propofol 30 mcg/kg/min, unchanged since 22:00 on 09/13/2026; fentanyl 75 mcg/h. Assessed 06:50: RASS -2; she opened her eyes to her name but did not hold eye contact; no physical stimulation needed. Deeper than target by one level with no bolus in the preceding six hours, consistent with overnight accumulation. CPOT 1. Night values of -1 at 02:00 and -2 at 04:00 were charted against the same target. Spontaneous awakening trial safety screen at 07:00, per protocol: no active seizures or alcohol withdrawal, no escalating sedative doses for agitation, no neuromuscular blockade, no active myocardial ischemia, no raised intracranial pressure concern; screen passed and the trial planned with the respiratory therapist (R. Chen, RRT).
Awakening trial, breathing trial, and the CAM-ICU (07:05 to 08:45): Propofol stopped 07:05; fentanyl continued at 75 mcg/h for analgesia. 07:35: RASS -1, opens eyes to her name and holds eye contact, squeezes hands on request; RR 20, SpO2 97 percent on FiO2 0.50 and PEEP 8, synchronous with the ventilator. 08:05: RASS 0, awake and calm without prompting, nods appropriately, no reaching for lines; RR 22, SpO2 96 percent, heart rate 92, no arrhythmia. No failure criteria (no sustained agitation, RR under 35, SpO2 above 88 percent, no distress or arrhythmia). Spontaneous breathing trial started 08:10 by R. Chen, RRT, on pressure support; stopped 08:32 for RR 38 with accessory muscle use and SpO2 90 percent, returned to prior settings, result and settings in the respiratory flowsheet. Sedation not restarted: at RASS 0 and comfortable, the target is met without an infusion; the propofol order was left active for use if needed. CAM-ICU performed 08:30 at RASS 0, above the arousal gate: negative, recorded in the delirium flowsheet; next screen due 20:00 under the once-per-shift policy. Intensivist informed at 08:45 of both trial results; plan from rounds: target stays -1 to 0 with no continuous sedative unless required, fentanyl decreased to 50 mcg/h at 09:00 for CPOT 0 to 1.
Procedural bolus and an expected off-target value (13:10 to 14:30): 13:10: right radial arterial line replaced by the intensivist; propofol 30 mg intravenous bolus and fentanyl 50 mcg given under the procedural order, both charted in the medication administration record with times. 13:30: RASS -3, deeper than the target of -1 to 0; she moved and briefly opened her eyes when her name was called but made no eye contact; no physical stimulation needed. Cause recorded: expected transient effect of the 13:10 bolus, no infusion running, no new neurological finding, pupils equal and reactive. No titration to make (no sedative infusion); recheck set for 14:00 per protocol. 14:00: RASS -1, eye contact to voice held; 14:30: RASS 0, target met. No prescriber notification required under the protocol for a bolus-related value that returned to target within the recheck interval.
Agitation episode: cause, intervention, response (16:20 to 17:05): 16:20: RASS +3; after repositioning she became restless, then repeatedly reached for the endotracheal tube and pulled at the new arterial line; SpO2 89 percent with high peak pressures and ventilator asynchrony. Cause search: suctioned for a thick secretion plug at 16:21, SpO2 95 percent by 16:24; CPOT 5 with grimacing and muscle rigidity, fentanyl 50 mcg bolus given 16:25 under the pain order; bladder catheter patent and draining; no new medication started today, no alcohol history; CAM-ICU negative this morning and repeated at 16:50 (below). Soft bilateral wrist restraints applied 16:23 under the intensivist's order for the immediate physical safety of the airway after verbal reassurance, hand-holding, and reorientation failed to stop her reaching for the tube; the order, the less restrictive measures tried, her behavior, and her response are documented in the restraint record as hospital policy and 42 CFR 482.13(e) require. No additional sedative given; the intensivist agreed at 16:30 to treat pain and the airway first and to reassess before any propofol. 16:50: RASS 0, calm, follows commands, ventilator synchronous; CAM-ICU repeated at RASS 0 because of the acute behavior change: negative. Restraints removed 17:05 at the earliest point her behavior allowed, RASS 0 sustained for 15 minutes, removal time charted. Daughter present from 17:15 and updated.
Handoff and record (18:45): Handoff to the night nurse at 18:45 with the sedation state in one line: target RASS -1 to 0 per the 09/12/2026 order, reaffirmed today; last assessed RASS 0 at 18:30 by observation alone; no sedative infusion running since 07:05, propofol order active; fentanyl 50 mcg/h, last bolus 16:25; awakening trial passed today and breathing trial stopped at 08:32 for tachypnea; CAM-ICU negative at 08:30 and 16:50, next due 20:00; restraint episode 16:23 to 17:05 documented in the restraint record; recheck RASS every two hours and with any titration per MICU-SED-04. Values entered in the sedation flowsheet beside the target and the medication administration record; this narrative filed as the nursing note for 09/14/2026. Every RASS value in this record is an observation at its clock time; none was estimated from the infusion rate.
This sample is fictional and for educational purposes. It does not describe a real patient or record; the values, times, doses, dates, and details are invented to show documentation structure and are not clinical guidance. No scale descriptors or scoring card are reproduced.
Writing these after every session? BastionGPT drafts complete notes from bullets, dictation, or a transcript.
Generate a note from bulletsUnited States: no LAW names the RASS, and the rules that bind sit one step from the bedside. Under LAW, the Medicare Conditions of Participation require a medical record that contains practitioner orders, nursing notes, medication records, vital signs, and the other information needed to monitor the patient's condition, with entries complete, dated, timed, and authenticated (42 CFR 482.24), and a nursing assessment and care plan for every inpatient (42 CFR 482.23); neither names a sedation scale, an interval, or a target field, so a RASS entry is defensible under them when it lets a reader reconstruct what was ordered, observed, given, and done. The restraint rules are LAW with specific documentation content: 42 CFR 482.13(e) allows restraint only to ensure the immediate physical safety of the patient, staff, or others, requires the least restrictive effective intervention and an order from a physician or other licensed practitioner, requires discontinuation at the earliest possible time regardless of the length of the order, and requires the record to show the patient's behavior and the intervention used, the alternatives or less restrictive interventions attempted, the condition or symptoms that warranted it, and the patient's response; a RASS value is none of those things, and it is not an order. CMS's anesthesia services condition (42 CFR 482.52) and its interpretive guidelines (S&C-10-09 of December 2009, revised by S&C-11-10 of January 14, 2011) place anesthesia, deep sedation, and monitored anesthesia care for procedures under pre- and post-anesthesia evaluation and rescue-capacity requirements along the ASA continuum; they name no sedation scale, and neither memorandum mentions the continuous sedation of mechanically ventilated ICU patients, so the ICU record's obligations come from the general record and nursing conditions and from professional standards rather than from the anesthesia condition. Under CONVENTION, the SCCM PADIS guideline (Devlin and colleagues, Critical Care Medicine, 2018) suggests light rather than deep sedation in mechanically ventilated adults (a conditional recommendation on low-quality evidence), achieved by daily sedative interruption or nurse-protocolized targeted sedation, and suggests propofol or dexmedetomidine over benzodiazepines; the 2025 focused update (Lewis and colleagues) adds a conditional suggestion for dexmedetomidine over propofol where light sedation or delirium reduction is the highest priority and otherwise leaves the 2018 framework in place. Neither document sets a documentation interval, a single universal target number, or a rule that the target, the drug time, and the CAM-ICU must share a row, and the definitions of light sedation used across the trials and guidelines differ, which is one more reason the patient's own ordered target is the reference point. The SCCM ICU Liberation bundle turns the convention into a metric: its data collection manual defines the C element by the number of documented level-of-arousal assessments per 24 hours using a standard scale such as the RASS or SAS, with a goal of at least six, and treats a RASS of -4 or -5 as exempt from the D element delirium assessment; in the Collaborative's 15,226 adults across 68 ICUs, complete bundle performance was associated with lower odds of death within seven days (adjusted hazard ratio 0.32), next-day mechanical ventilation (adjusted odds ratio 0.28), coma (0.35), and delirium (0.60) (Pun and colleagues, 2019), and a documented target sedation level was independently associated with higher odds of an awakening trial (adjusted odds ratio 1.68) and a breathing trial (1.46) being performed (Balas and colleagues, Chest, 2022). The Critical Care Societies Collaborative's Choosing Wisely list (2014) advises against deep sedation of ventilated patients without a specific indication and without daily attempts to lighten it. Joint Commission provision-of-care standards on sedation and anesthesia address pre-sedation assessment, monitoring, and recovery for procedures and name no scale, and its restraint standards parallel the CMS rules (CONVENTION with accreditation force). The CDC NHSN ventilator-associated event algorithm (January 2026 protocol) is built on ventilator settings, temperature, white cell count, antimicrobials, and microbiology, contains no sedation element, and states that it is for surveillance and is not a clinical definition algorithm, so a sedation practice or a missing RASS never triggers a VAE. Under PAYER POLICY, no CMS national coverage rule, NCCI edit, or Hospital IQR or MIPS measure keys payment or a score to a RASS value; ICU Liberation metrics are a voluntary quality program, and a hospital's own RASS documentation metric is local policy. Under neuromuscular blockade, the 2016 SCCM guideline (Murray and colleagues) states as good practice that analgesic and sedative drugs should be used before and during blockade, the 2026 SCCM guideline for adults with ARDS suggests either a scale-based or a non-scale evaluation of sedation depth before blockade and either monitoring or no monitoring during it, both on very low certainty evidence, and a 2022 meta-analysis of 24 studies and 1,235 patients found a correlation of 0.66 between the RASS and the bispectral index (Heavner and colleagues), so processed EEG is related to but not interchangeable with the scale; awareness with paralysis was reported in 2.6 percent of 383 ventilated emergency department patients (the ED-AWARENESS study, 2021) and in 3.4 percent of patients in the rigorously conducted studies pooled by the same group, which is the reason the record shows sedation and analgesia running rather than a manufactured value. Three habits deserve the same label: an hourly RASS by regulation, a target of -2 for every patient, and mandatory BIS under paralysis are unit conventions or assumptions, not rules any of these authorities state.
Canada and Australia name no scale and set no interval, and the reliability record explains why the stimulus and the drug time belong in the entry. In Canada, no federal or provincial LAW names the RASS. The clearest guidance is provincial: Critical Care BC's ICU Liberation guide (Provincial Health Services Authority, July 22, 2025; CONVENTION) states that depth of sedation should be assessed every four hours or more using a validated tool, naming the RASS, that each patient should have a daily RASS goal documented with sedatives titrated to meet it, that the goal and the effectiveness of analgesic and sedative choices are discussed on daily interdisciplinary rounds, and that RASS and pain scores are documented every four hours or more to support handover. Health Standards Organization's HSO 11001:2026 Critical Care Services, the standard relevant to Accreditation Canada programs, was in public review as a draft dated March 2026 as of September 2026; the draft requires that organizational leaders provide the team with validated assessment tools for clients receiving critical care and describes reassessment with changes in clinical status, with input from clients or care partners, or at intervals defined by the organization, and it names no sedation scale, so no Accreditation Canada interval or RASS requirement should be asserted and the draft should be cited as a draft until adoption is published. The Choosing Wisely deep-sedation recommendation also appears on the Canadian critical care list. In Australia, the ANZICS-led SPICE III trial (Shehabi and colleagues, New England Journal of Medicine, 2019; 3,904 patients) targeted a RASS of -2 to +1 in both arms and found 90-day mortality of 29.1 percent in each, which is why that range appears in Australasian order sets as a trial target rather than a mandate. The NSQHS Recognising and Responding to Acute Deterioration Standard (second edition; CONVENTION with accreditation force) requires observation charts, track-and-trigger systems, and escalation without naming a consciousness or sedation instrument; general ward charts record consciousness as AVPU or ACVPU and opioid-induced sedation on a short ward sedation score, neither of which is a RASS, so a RASS on an Australian ward is a local choice and the ward sedation score should never be charted as one. The ACSQHC Delirium Clinical Care Standard (2021) asks for a validated tool rather than naming one, and its supporting material lists the modified RASS among tools used to monitor for incident delirium; the joint college guideline for patients sedated for acute behavioural disturbance (CICM IC-22 and ANZCA PS63) describes response to voice or stimulation as the way to monitor depth of sedation, treats tools such as the RASS as potentially helpful, and asks that drug, dose, time, and response be recorded, within its own scope of acute behavioural disturbance rather than routine ICU sedation. Two boundaries hold everywhere. Version: the RASS-PAL (Bush and colleagues, 2014) rests on a pilot of 13 clinicians rating 10 patients at five time points with inter-rater intraclass correlations of 0.84 to 0.98, whose authors asked for further validation, particularly for delirium; the modified RASS (Chester and colleagues, 2012) adds an attention item and, in 95 medical inpatients, detected delirium with 64 percent sensitivity and 93 percent specificity as a single assessment and 74 and 92 percent when used serially; and the adult scale was validated in critically ill children across 100 encounters (weighted kappa 0.825 between nurse and researcher; Kerson and colleagues, 2016) and 139 episodes in 14 pediatric ICUs (weighted kappa 0.946; Tapia and colleagues, 2021), so the note names the version and does not import adult targets. Reliability: Sessler and colleagues (2002) reported agreement of r 0.956 and weighted kappa 0.73 among five investigators in 192 encounters and r 0.964 and kappa 0.80 between a nurse educator and 27 trained bedside nurses in 101 encounters, and Ely and colleagues (2003) a weighted kappa of 0.91 between nurses; in routine practice, Fuchita and colleagues (2025) found that an ICU nurse and two trained investigators chose the identical value in only 35 percent of 79 encounters, that nurses recorded a lighter value than the investigators in 52 percent, and that at least one assessor noted uncertainty in 16 percent, and a 2026 prospective audit of 76 nursing assessments in a UK cardiothoracic unit found 75 percent accurate against an expert rater but only 33 percent documented hourly as that unit's policy required, with poorer agreement at deeper sedation (Szentgyorgyi and colleagues). Completion and accuracy are different problems, and both are the reason to write the behavior and the drug time when the number is not self-explanatory.
The Richmond Agitation-Sedation Scale was published by Sessler and colleagues in the American Journal of Respiratory and Critical Care Medicine in November 2002; the article is copyright 2002 the American Thoracic Society, and Virginia Commonwealth University holds a registered copyright in the scale itself (2012; VCU Tech 12-063, inventors Sessler, Grap, Brophy, and Elswick), which the university describes as allowing free use by health systems, clinicians, and researchers while pharmaceutical companies must license it for clinical trials comparing sedative medications, and which its technology transfer office lists as available for licensing to industry. The Vanderbilt ICU Delirium and Cognitive Impairment Study Group, now the Critical Illness, Brain Dysfunction, and Survivorship Center, validated the scale and built it into the CAM-ICU workflow; icudelirium.org states that the CAM-ICU and its educational materials, which carry the RASS on the pocket card, are copyright 2002 E. Wesley Ely and Vanderbilt University, that the group deliberately made them unrestricted for implementation and clinical use with the copyright line requested on pocket cards and educational materials, and that other uses go to [email protected]. That permission is Vanderbilt's, for Vanderbilt's materials; it is not a general license to republish the scale's level descriptors, which this page describes and does not reproduce. BastionGPT is not affiliated with, or endorsed by, the authors or the rights holder. This page reproduces no test items, stimuli, norms, or scoring materials.
The numbers behind these errors are specific. In the original validation, five investigators (two physicians, two nurses, and one pharmacist) agreed at r 0.956 and weighted kappa 0.73 across 192 encounters, and a nurse educator and 27 trained bedside nurses at r 0.964 and kappa 0.80 (Sessler and colleagues, 2002); in contemporary practice an ICU nurse and two trained investigators chose the identical value in only 35 percent of 79 encounters, nurses recorded a lighter value than the investigators in 52 percent, and an assessor flagged uncertainty in 16 percent (Fuchita and colleagues, 2025), and a 2026 audit of 76 nursing assessments found 75 percent accurate but only 33 percent documented hourly as the unit required, with most discrepancies one level (Szentgyorgyi and colleagues). On the target side, a goal-directed sedation guideline in one medical ICU raised the share of orders carrying a patient-specific RASS goal from 21.3 to 85.4 percent and RASS assessments per 24 hours from 4.7 to 11.4, while the share of assessments at goal rose only from 31.4 to 44.1 percent (DeGrado and colleagues, 2011); across the ICU Liberation Collaborative, a documented target sedation level was associated with higher odds of an awakening trial (adjusted odds ratio 1.68) and a breathing trial (1.46) being performed (Balas and colleagues, 2022), and complete bundle performance with lower odds of coma (0.35) and delirium (0.60) in 15,226 adults (Pun and colleagues, 2019). No published audit counts how often a RASS entry lacks its target, its clock time, or the sedation on board; the bundle metrics and the compliance reviews described under compliance considerations are what turn those omissions into a quality finding. The BastionGPT Clinical Advisory Board sees the same errors most often in RASS documentation reviews:
BastionGPT is specifically trained, tuned, and clinically tested on critical care nursing and sedation documentation.
See how clinicians use it day to day on the AI medical 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.
The RASS has ten levels on one continuum. Positive values (+1 to +4) describe increasing agitation, from restlessness to behavior that endangers the patient or staff; zero is a patient who is awake and calm without prompting; negative values (-1 to -5) describe increasing depth of sedation, separated by how much stimulation it takes to get a response and by the quality of that response. The assessment is a sequence: observe the undisturbed patient first, and rate an alert or agitated patient from observation alone; if the patient is not alert, speak, say the name and ask for eye opening and eye contact, and rate by whether contact is made and how long it is held; only if voice produces nothing, use physical stimulation, and the two deepest values differ by whether anything happens at all. A negative value therefore already tells the next reader which stimulus was needed. The wording that defines each level is the scale's copyrighted text (Virginia Commonwealth University, 2012), so score from the published card your unit uses and describe the behavior in your own words when a value could be read two ways. It takes under a minute at the bedside, and it is neither a pain scale nor a delirium screen.
Both, every time the flowsheet allows it. The target belongs in the record exactly as the order states it, a single value or a range (-1 to 0, or -2 to +1), with the order date, never translated into a phrase such as light sedation and never narrowed to one number. The assessed value sits beside it with its time, so each entry reads as at target, deeper, or lighter; Vanderbilt's trauma protocol states that "goal and actual RASS should be specified on ALL patients," and Critical Care BC asks for a documented daily RASS goal with sedatives titrated to it. If the order is a range, chart the range and note which end the team is aiming for when rounds decided that. If no target has been ordered, obtain one and chart the request, because a value with no target cannot be read as on or off course and a goal-directed protocol has nothing to titrate to; in one medical ICU, a goal-directed guideline raised orders with a patient-specific RASS goal from 21.3 to 85.4 percent. If the ordered target is deeper than the guideline preference for light sedation, preserve it, record the indication (neuromuscular blockade, status epilepticus, refractory intracranial hypertension, severe dyssynchrony), and clarify an unexplained deep target with the team rather than substituting your own; a target of -2 for everyone is a habit, not a guideline.
No law or accreditor in the United States, Canada, or Australia sets a RASS interval, and PADIS does not either; the interval comes from the medication order and the unit protocol, plus the clinical moments that call for an unscheduled value (a titration, a bolus, a procedure, an awakening trial, a change in behavior, and the arousal check before every CAM-ICU). Real protocols differ: Critical Care BC (2025) asks for depth of sedation every four hours or more, Vanderbilt's trauma protocol monitors each shift, one UK cardiothoracic unit requires hourly entries, some units require a value with every titration of a continuous sedative, and the SCCM ICU Liberation data manual counts documented level-of-arousal assessments per 24 hours with a goal of at least six. Follow your protocol, cite it in the note as policy, and remember that frequency and accuracy are separate problems: in the 2026 UK audit, 75 percent of nursing values matched an expert rater but only 33 percent were documented hourly as required. A row of hourly numbers is not evidence of good sedation management if the target, the drug context, and the response to off-target values are missing, and a correctly assessed patient can still fail a local frequency metric.
The RASS is the arousal check that comes first. In the Vanderbilt workflow, a RASS of -4 or -5 means the patient cannot be assessed for delirium and the CAM-ICU is recorded as unable to assess for that attempt; any value from -3 through +4 clears the gate and the delirium features are assessed, so -3 is not automatically too sedated. The SCCM ICU Liberation data manual treats -4 and -5 the same way, as exempt from the delirium element. Chart the RASS with its time first and the CAM-ICU result after it, and repeat the screen when the patient becomes assessable again if one is due; the RASS and the CAM-ICU are linked but not synchronous, since sedation is assessed many times a day while delirium is usually screened once a shift, so the absence of a CAM-ICU beside every RASS is not a defect. A RASS of 0 says the patient is calm and alert; it says nothing about delirium, which is answered only by the CAM-ICU features, and a patient can be delirious at 0, at -2, or at +2. Delirium itself is a medical diagnosis documented by the provider; the screen is nursing documentation.
Do not manufacture one. The scale depends on visible responses, and a paralyzed patient will look unresponsive whatever their awareness, so a charted -5 under effective blockade is not a sedation assessment and can hide inadequate sedation. Chart that the RASS cannot be assessed because of pharmacologic paralysis, with the blocker and its start time; carry forward the last assessment taken before the blocker with its time; record the sedative and analgesic regimen with rates; and name the unit's strategy for judging sedation depth during blockade, whether processed EEG or clinical signs. The 2016 SCCM guideline on sustained blockade states as good practice that analgesic and sedative drugs should be used before and during blockade, and the 2026 SCCM guideline for adults with ARDS suggests either a scale-based or a non-scale evaluation before blockade and either a monitoring-based strategy or no monitoring during it, both on very low certainty evidence, so BIS is not mandatory; a 2022 meta-analysis found a correlation of only 0.66 between the RASS and the bispectral index, so the two are related and not interchangeable. The reason the entry matters is awareness with paralysis: 2.6 percent of 383 ventilated emergency department patients in the ED-AWARENESS study and 3.4 percent in the rigorously conducted studies pooled by the same group. When the blocker wears off, resume scoring and write the first assessable value with its time.
As a trajectory, not an event. Record the pre-trial value with its time and the target; the safety screen result under your protocol (the original ABC trial excluded patients on a sedative infusion for active seizures or alcohol withdrawal, on escalating sedation for agitation, on neuromuscular blockade, or with active myocardial ischemia or raised intracranial pressure); the time each sedative was stopped and which infusions, usually the analgesic, continued; the values during the trial with their times and the behavior that produced them; the failure criteria you watched for (sustained agitation, respiratory distress or a high respiratory rate, desaturation, arrhythmia, hemodynamic instability) and whether any occurred; the coordination with the respiratory therapist for the breathing trial and its result; and, if sedation was restarted, the reason, the time, and the rate (the ABC protocol restarted at half the prior dose, a protocol choice rather than a rule). Once the patient has cleared the arousal gate and a delirium screen is due, chart the CAM-ICU after the trajectory. Daily interruption traces to Kress and colleagues (2000) and the paired trial to Girard and colleagues (2008); PADIS names daily interruption and nurse-protocolized targeted sedation as the ways to achieve light sedation, and the sample on this page shows one full day.
Chart the one that answers the question, and name it. The RASS answers how sedated or agitated the patient is and whether a delirium screen can proceed. The Glasgow Coma Scale answers level of consciousness after brain injury through eye, verbal, and motor components; the two correlated at r 0.91 in the Vanderbilt validation but measure different things, and in a sedated neurological patient both are charted, with the sedative named as a confounder of the GCS. The Riker Sedation-Agitation Scale (1999, seven levels, weighted kappa 0.92 in its validation) is the other tool the SCCM guidelines treat as valid and reliable; a unit uses one or the other, and no conversion between them belongs in a chart. The Ramsay scale is older, runs in the opposite direction, and separates agitation poorly; if a legacy order uses it, chart under that name and ask for the order to be rewritten. Pain is a separate construct: PADIS asks for the CPOT or BPS in patients who cannot self-report, and pain is assessed and treated before sedation is escalated, because an agitated RASS is often a pain finding. On general wards and in Australian observation charts, consciousness is recorded as AVPU or ACVPU inside an early warning score (see the MEWS and NEWS2 page) and opioid-induced sedation on a short ward sedation score; neither is a RASS, and the RASS was validated in critical care populations rather than in every hospitalized adult.
For a health system's own clinical use, yes, on the rights holder's stated terms. Virginia Commonwealth University registered a copyright in the scale in 2012 (VCU Tech 12-063) and states that the protection allows free use by health systems, clinicians, and researchers while pharmaceutical companies must license it for clinical trials comparing sedative medications; its technology transfer office lists the scale as available for licensing to industry. So a hospital printing the scale on its own flowsheet or building it into its own EHR sits inside that permission, with attribution to Sessler and colleagues (2002) and Ely and colleagues (2003) as the convention. A software vendor redistributing the descriptor wording in a commercial product, a paid app, or a public website that republishes the complete scale is a different case: VCU's statement names health systems, clinicians, and researchers, not public republication, and the wide appearance of the table on calculator sites is not a license. The Vanderbilt CIBS Center's permission covers its own CAM-ICU materials, which carry the RASS on the pocket card; it asks for its copyright line and refers other uses to [email protected], and it is not a substitute for VCU's position on the scale itself. Free, non-commercial, or attributed does not by itself create a license, which is why this page describes the scale in original words and reproduces no level descriptors.
Yes. Give it the facts (the ordered target and its date, each assessed value with its clock time and the moment in care, the stimulus or behavior where it was ambiguous, the infusions and boluses with times, the pain score and tool, any awakening trial sequence and CAM-ICU result, any blockade or restraint event, and the recheck) and it drafts the full entry: target beside actual, the relation to target with its cause, the action taken under the order, the trial trajectory and the delirium screen in order, and the handoff line, ready for your review. It can also check a finished note for values with no target, a target rewritten, a value with no time or drug context, a sedative given before a pain assessment, a manufactured value under paralysis, a CAM-ICU at an unassessable level, or an awakening trial charted as a single word. 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 September 2026:
Educational content, not legal or billing advice. Sample notes are fictional. Follow your organization's policies and your board, payer, and jurisdiction requirements.