How to Document a Patient’s Level of Consciousness

Documenting a patient’s level of consciousness means translating a complex neurological state into standardized, reproducible language that any clinician can interpret and act on. The Glasgow Coma Scale remains the most widely used tool for this purpose, but it is far from the only option, and over fifty years of clinical experience have revealed significant gaps in how it captures consciousness. Choosing the right scale, recording its components correctly, noting confounders, and reassessing at consistent intervals all determine whether a bedside note is genuinely useful or just a number that obscures more than it reveals.

Why Structured Scales Beat Narrative Descriptions

Before the Glasgow Coma Scale appeared in 1974, clinicians described consciousness in loose terms like “stuporous,” “obtunded,” or “semi-comatose.” These words meant different things to different people, and the confusion sometimes led to delayed treatment. Graham Teasdale and Bryan Jennett developed the GCS specifically to replace that ambiguity with a numerical system that could be communicated quickly and tracked over time.¹ The GCS was designed for three purposes: objectively measuring consciousness, assessing trends, and enabling accurate communication between clinicians.²

That said, a structured scale is only as good as the person using it. A vague chart note reading “GCS 8” tells less than most clinicians think it does, because different combinations of subscale scores can produce the same total. The real documentation value comes from recording the individual components, noting what was actually tested, and flagging anything that might have skewed the result.

The Glasgow Coma Scale and How to Record It Properly

The GCS evaluates three domains: eye opening (scored 1–4), verbal response (scored 1–5), and motor response (scored 1–6). The total ranges from 3 (deep coma) to 15 (fully alert). In clinical shorthand, you will often see something like “E3V4M5 = 12,” which tells the reader exactly which behaviors were observed in each domain rather than just giving the sum.

Recording subscale scores individually matters because two patients with the same total can be in very different neurological states. A large multicenter study examining over 69,000 GCS assessments across 2,455 patients found that every total score between 4 and 14 mapped to more than one disorder-of-consciousness diagnosis, with the greatest diagnostic ambiguity falling in the 7–11 range.³ A patient scoring E2V3M6 (total 11) and a patient scoring E3V2M6 (total 11) have the same number but different clinical pictures. If you only document the total, that distinction vanishes from the medical record.

A related problem is inconsistency in how clinicians actually perform the exam. A systematic review found that more than 90% of publications using GCS scoring referenced the older 14-point version rather than the current 15-point version, that the timing of the initial assessment varies widely, and that confounders like intubation or sedation are often unreported.⁴ Good documentation, then, should include not just the score but also the time it was taken, whether the patient was intubated (which makes the verbal score untestable), and any medications that might depress consciousness.

When the Motor Score Tells You the Most

Among the three GCS components, the motor response carries the most prognostic weight. This is partly practical: you can always test motor response, even in intubated or facially injured patients where eye and verbal responses are unreliable. But it also reflects the neuroanatomy of consciousness. The motor score captures the integrity of pathways from the cortex through the brainstem, so a declining motor response often signals worsening brain function before other signs appear.

Research on severe traumatic brain injury patients has shown a strong link between the best motor response at admission and long-term recovery. In one study, patients whose best motor response was localizing pain or obeying commands had substantially better outcomes than those showing only abnormal posturing or no response at all.⁵ When the full GCS is hard to obtain, some trauma protocols now accept the motor subscale alone as a reasonable proxy for consciousness level.

AVPU for Rapid Triage

Not every clinical situation calls for the full GCS. In prehospital settings and rapid triage, clinicians frequently use the AVPU scale: Alert, responds to Voice, responds to Pain, Unresponsive. It takes seconds, needs no equipment, and gives a quick snapshot of where the patient falls on the consciousness spectrum.

AVPU and GCS are not interchangeable, but they do map onto each other reasonably well. One study of prehospital data found optimal cutpoints mapping Alert to a GCS of 14–15, Verbal to 10–13, Pain to 7–9, and Unresponsive to 3–6, with substantial overall agreement.⁶ Another study found that AVPU and GCS had comparable ability to predict 48-hour mortality, with the discriminative power of AVPU for death within 48 hours being nearly equivalent to the GCS.⁷ The 48-hour mortality rates in that study climbed steeply across AVPU categories: about 1% for Alert, 4% for Verbal, 18% for Pain, and over 53% for Unresponsive.

The tradeoff is granularity. AVPU’s middle categories, Verbal and Pain, are coarse. That same prehospital mapping study showed that classification accuracy was high for Alert and Unresponsive patients but much weaker for those in the Verbal and Pain categories. If your patient falls somewhere in the middle, AVPU alone may not capture meaningful changes, and a full GCS becomes necessary.

In children under five, the relationship between AVPU and GCS shifts. A UK study of over 2,300 children with head injuries found that the correlation between AVPU and median GCS scores differed by age group and that for infants, AVPU categories of V, P, or U had a higher diagnostic yield for intracranial injury than a GCS below 15.⁸ This suggests AVPU may actually outperform GCS for very young children in certain triage contexts.

The FOUR Score as a More Comprehensive Alternative

The Full Outline of UnResponsiveness (FOUR) score was developed to address several GCS shortcomings, particularly in patients who are intubated. It evaluates four components: eye response, motor response, brainstem reflexes, and respiration, each scored from 0 to 4. Because it drops the verbal component entirely and adds brainstem and respiratory assessments, it works for patients on ventilators without needing the awkward “VT” notation that GCS requires when someone has a tube in their throat.

A systematic review comparing the two scales in intensive care units and emergency departments found that the FOUR score demonstrated slightly higher overall reliability and validity than the GCS, making it a promising alternative assessment tool.⁹ The FOUR score can also detect locked-in syndrome and distinguish between different types of brainstem dysfunction, neither of which the GCS does well. For ICU documentation, it fills meaningful gaps.

Pupil Assessment and Brainstem Reflexes

A consciousness assessment is incomplete without checking the pupils and at least a few brainstem reflexes. Pupil size, symmetry, and reactivity to light are among the most important bedside indicators of whether the brainstem pathways that support consciousness are intact. The corneal reflex, which tests the trigeminal and facial nerve pathways, is another fundamental component, particularly for neuroprognostication after cardiac arrest and in brain death determination.¹⁰

Document pupil findings as specifically as possible: size in millimeters, whether they are equal or asymmetric, whether they react briskly or sluggishly to light, and whether there is a consensual response. “PERRL” (pupils equal, round, reactive to light) is a common shorthand, but it loses value when the pupils are not equal or reactive. Any deviation from normal deserves its own description.

Automated pupillometry is increasingly common in neurocritical care units and adds a layer of objectivity. These devices measure the Neurological Pupil index (NPi), a standardized score from 0 to 5. A retrospective study found that patients who developed significant neurological deterioration had lower mean NPi values than those who remained stable, and that an NPi below 3.0 or a declining trend warranted closer attention.¹¹ A large multicenter study has further suggested that repeated automated pupillometry could improve continuous monitoring of disease progression at the bedside.¹²

An interesting recent finding involves the pupillary “light-off” reflex, which measures how the pupil rebounds after light is removed rather than how it constricts when light hits it. This reflex correlated strongly with consciousness levels and was absent in a much higher proportion of unresponsive patients than the standard light reflex. Of 25 unresponsive patients, 84% still had a preserved light reflex, but only 28% had a fully preserved light-off reflex, suggesting that the light-off reflex may be a more sensitive marker of impaired consciousness.¹³

Documenting Consciousness in Sedated Patients

In the ICU, patients are often sedated, and the GCS becomes unreliable for measuring intrinsic consciousness when drugs are suppressing it. The Richmond Agitation-Sedation Scale (RASS) is the standard tool in this context. It runs from +4 (combative) through 0 (alert and calm) down to −5 (unarousable), and international guidelines recommend hourly RASS documentation for mechanically ventilated patients.¹⁴

A clinical trial found that using a formal sedation protocol guided by RASS scores kept patients more consistently in the target sedation range (scores of −1, 0, and +1) throughout their ICU stay.¹⁵ This matters because oversedation can mask neurological deterioration, and undersedation can lead to agitation that mimics a worsening state. The documentation should clearly separate the sedation level from the underlying consciousness level whenever possible, noting both the RASS score and any GCS assessment performed during a sedation holiday or lightening trial.

The interaction between sedation and other monitoring tools deserves attention. The pupillometry study mentioned earlier found that sedation depth accounted for roughly 20% of the relationship between lower pupillary reactivity and neurological deterioration.¹¹ In other words, deep sedation doesn’t just confuse the clinical exam; it also affects the objective measurements that are supposed to help when the exam is unreliable. Recording the sedation regimen alongside any consciousness assessment is not optional; it is essential for interpreting the data later.

Serial Assessment and Catching Changes

A single consciousness assessment is a snapshot. Its real value emerges when compared to previous assessments. Patients with altered consciousness are among the most difficult to evaluate, and tracking trends over time is what enables faster recognition of life-threatening neurological events.¹⁶ A GCS of 10 means something very different when the patient was at 14 two hours ago versus when they have been stable at 10 for the past three days.

Practical documentation tips for serial assessment: record each evaluation at consistent intervals (typically hourly in acute settings, though this varies by protocol), always note the time, and flag any change from the previous assessment explicitly. A note reading “GCS E3V4M5 = 12, decreased from E4V5M6 = 15 at 0800, new finding of sluggish left pupil” tells the next clinician far more than just “GCS 12.” The direction of change, the speed of change, and any new asymmetries are the signals that drive clinical decisions.

Common Documentation Pitfalls

Several factors undermine the reliability of consciousness documentation. A systematic review examining GCS reliability found conflicting results regarding the effect of the observer’s experience level, the underlying pathology causing reduced consciousness, and the presence of intubation or sedation.¹⁷ Inconsistency in how clinicians perform the GCS exam is a persistent problem: studies have documented substantial variation in the sequence of testing, the type of stimulus applied for pain responses, and the resulting accuracy of scores.¹⁸

Inter-rater reliability, meaning how consistently different clinicians score the same patient, is often lower than you would hope. A study comparing palliative care nurses and trained researchers found only moderate agreement on GCS total scores, with the nurses and first researcher agreeing about 49% of the time.¹⁹ Even the two researchers agreed only about 79% of the time. This gap matters because a chart documenting a “decrease” from GCS 11 to 9 could reflect genuine neurological change or simply two clinicians applying the scale differently.

Locked-in syndrome is perhaps the most dangerous documentation pitfall. A patient with locked-in syndrome is fully conscious but unable to move anything except, in the classic form, their eyes. Unless the examiner specifically looks for voluntary vertical eye movements or blinking to command, these patients can be mistaken for comatose or vegetative.²⁰ Documenting that eye-tracking and blinking-to-command were tested, even when the result is negative, protects against this error.

Pediatric Consciousness Assessment

The standard GCS cannot be applied to infants and young children because the expected normal verbal and motor responses change with developmental stage. A child under five will not follow a command like “show me two fingers,” and a baby’s best verbal response is crying, not oriented speech. Pediatric versions of the GCS were developed specifically to address this, adapting the verbal and motor subscales to age-appropriate responses while preserving the same scoring structure.²¹

When documenting a child’s consciousness level, you should note the version of the scale used and the child’s age, because what counts as a “normal” score shifts with development. A one-year-old who cries in response to pain and withdraws a limb is showing age-appropriate behavior, not an impaired response. Failing to account for this leads to artificially low scores and potentially unnecessary interventions.

Differentiating Disorders of Consciousness

Once you move beyond acute triage, documenting consciousness also means distinguishing between specific diagnostic categories: coma, unresponsive wakefulness syndrome (formerly called vegetative state), minimally conscious state, and post-traumatic confusional state. These categories carry different prognostic and treatment implications, but telling them apart at the bedside is genuinely difficult. The misdiagnosis rate between unresponsive wakefulness syndrome and minimally conscious state has been estimated at 37–43%.²²

The key clinical difference is whether the patient shows any reproducible evidence of awareness. A patient in an unresponsive wakefulness syndrome may open their eyes and have sleep-wake cycles but shows no purposeful behavior. A minimally conscious patient shows intermittent but definite signs like visual tracking, reaching for objects, or following simple commands. Accurately differentiating these states requires multiple assessments across different times of day, because patients in a minimally conscious state can fluctuate and may look unresponsive during one exam and clearly aware during another.²³

Delirium adds another layer of complexity, particularly in older adults. The hallmark features of delirium, including inattention, fluctuating awareness, and altered consciousness, overlap with those of dementia, making it easy to mistake one for the other. Specialized screening tools such as the Confusion Assessment Method (CAM) and the 4AT have been developed to detect delirium even in the presence of moderate or severe dementia.²⁴ When documenting a patient whose consciousness fluctuates, noting the specific screening tool used and the time of assessment helps distinguish delirium from a stable baseline cognitive impairment.

EEG-Based Monitoring in the Operating Room

In anesthesia, documenting consciousness level means something different than in the emergency department. Here, the goal is to ensure the patient stays unconscious throughout a procedure while avoiding excessively deep anesthesia. The Bispectral Index (BIS) monitor, derived from processed EEG data, generates a score from 0 (no cortical activity) to 100 (fully awake), with values below 60 generally indicating a low probability of awareness.²⁵

The BIS has real limitations. It performs reasonably well at confirming that someone is awake but has low sensitivity for confirming that someone is truly unconscious. Some patients may show BIS values above 60 despite being clinically asleep, while others may be conscious at values below 50. One study found that both BIS and a competing index (the Patient State Index) showed wide enough variation that some patients were erroneously classified as unconscious when they were still responsive.²⁶ After electroconvulsive therapy, over a quarter of patients woke up at BIS values below 50, which would normally be interpreted as deep unconsciousness, likely because the post-seizure EEG pattern confuses the algorithm.²⁷

For documentation purposes, the takeaway is that BIS values should always be interpreted alongside clinical signs, not in isolation. Recording the BIS trend along with the anesthetic agents used, the clinical responses observed, and any confounding factors (seizures, hypothermia, pre-existing EEG abnormalities) gives future readers a complete picture rather than a single number that could mislead.

The Neuroanatomy That Explains Why Consciousness Is Hard to Measure

Consciousness depends on the reticular activating system, a network of nuclei in the brainstem that sends projections upward through the thalamus to the cerebral cortex. This system was first described in 1949 through experiments showing that stimulating it produced wakefulness and that damaging it produced coma.²⁸ Loss of consciousness, whether from anesthesia, trauma, or disease, involves disruption of this ascending pathway at various levels.²⁹

This anatomy explains why no single bedside test perfectly captures consciousness. A lesion in the brainstem may abolish arousal entirely, producing coma. Damage to the cortex may leave the brainstem intact, so the patient opens their eyes and has sleep-wake cycles but shows no awareness, the pattern seen in unresponsive wakefulness syndrome. Patchy cortical damage may leave some islands of function, producing the intermittent awareness of a minimally conscious state. Each of these patterns requires different documentation strategies, different assessment tools, and different clinical responses, which is why understanding even the basic architecture helps clinicians choose the right documentation approach for the patient in front of them.