Keep it simple: acute GCS score as a binary decision

Level of consciousness (LOC) is a measurement of arousal and response to external stimuli. An altered (from individual baseline) LOC can be caused by insufficient oxygenation, traumatic injury, or changes to the chemical environment of the brain. A standard scale to assess LOC is important for three reasons: (1) communication among healthcare providers; (2) guidance for diagnostic workup and therapeutic intervention; and (3) potentially guiding prognosis. In 1974, Dr.’s Graham Teasdale and Bryan J. Jennett published Assessment of coma and impaired consciousness: a practical scale, also known as the Glasgow Coma Scale (GCS). The GCS quantifies LOC and is composed of three objective tests: eye, verbal, and motor responses. The lowest possible total GCS is 3, while the highest is 15.

GCS was initially developed for “repeated bedside assessment” in a neurosurgical unit to detect “changing states” of consciousness and to measure “duration of coma.” It has since become a widely used assessment tool for mental status, incorporated into acute care medicine and taught as a core component of trauma and life support courses. The goal of the GCS is to predict clinically significant outcomes in altered LOC, significant brain injury, and as a way to guide medical decision-making. However, developers of the GCS stated in 1978: “We have never recommended using the GCS alone, either as a means of monitoring coma, or to assess the severity of brain damage or predict outcome” (Teasdale & Jennett).

Since then, several studies have been published that have cast doubt on the utility of the GCS, along with its reliability in clinical practice. For example, Dr. Steven M. Green wrote an editorial for the Annals of Emergency Medicine in 2011 strongly advocating against using the GCS, stating in comparison to the total GCS score, simple unstructured clinical judgment alone can be just as accurate, and that GCS itself has poor reliability. For example, Gill et al. (2005) examined the inter-rater reliability of GCS in the emergency department (ED). Between 19 attending ED physicians, they found an agreement percentage for an exact GCS of 32% and GCS-motor (GCS-m) of 72% (n=116). Thirteen possible GCS values were capped at 120 combinations of its components – a GCS score of 4 predicted a mortality rate of 48% if calculated at E1V1M2, 27% if at E1V2M1, and 19% at E2V1M1 (Healey et al.). Riechers II et al. assessed physician knowledge of the GCS through an anonymous, voluntary survey and found that poor knowledge of GCS components among physicians surveyed was a result of a lack of routine use of the scale and the complexity of the scale itself.

So how do we fix the GCS?  Reichers et al. proposed either an improvement in training strategy or implementation of a simpler scale. Gill et al. (2006) observed that just 3 of the 6 points of the GCS-m score significantly defined total GCS score performance. By collapsing the motor scale to just these items, she formed the Simplified Motor Scale (Obeys Commands, Localizes Pain, Withdrawal or Less Response). This simplified test was independently validated in the ED setting.

A recent study by Kupas et al. also suggests simplification of the total GCS score to a binary decision point of GCS-m score less than 6 (i.e., patient “does not follow commands”) for trauma-related decisions, such as when to intubate. The team performed a retrospective analysis of the Pennsylvania Trauma System Foundation’s registry which includes trauma patients admitted to the state’s Level I, II, III, and IV trauma centers (n=393,877). Results showed that differences between total GCS scores less than or equal to 13 and GCS-m scores less than 6 (patient “does not follow commands”) were below a prespecified 5% threshold for clinical importance – [Sensitivity: 2.5 to 4.9%, Specificity: -1.2 to -2.0%]. The study had two noteworthy limitations: (1) the population sample was from a single state and may not be representative elsewhere, although PA includes large urban, suburban, and rural areas, and (2) approximately half of the first reported GCS scores were determined by ED staff, rather than pre-hospital providers – making it difficult to determine if outcomes would have differed if one or the other had routinely provided GCS scores. Barazian et al. showed that GCS assessment performed by pre-hospital personnel was generally two points lower than in-hospital assessment, but followed a strong correlation between the two, independent of the time between score determinations.

Does the patient follow commands: Yes/No

As assessment of LOC is a vital practice in patient care, standardized tools are necessary that account for scope of practice and environmental pressures (i.e., cognitive ergonomics). Total GCS may have its place in long-term care (e.g., a neurosurgical unit) to assess change in LOC. In acute care, not error in the design of the Glasgow Coma Scale, but how it is used is the main contributor to its lack of utility. The answer is a simpler and easier to use scale. The simple, evidence-based binary assessment by Dr. Kupas and his team is very practical and appealing when considering its ease of use for providers at all levels of education. Looking forward, we should focus on making all assessment and information sharing more efficient by removing complexity when possible. 

Ameer Khalek is a MPH Candidate of the GWU Milken Institute School of Public Health