Can We Simplify Triage by Using Just GCS

Also posted over at Rogue Medic now at EMS Blogs. Go check out both sites.

-

A trauma system must provide the full gamut of available resources to seriously injured patients, but frequent overtriage of patients has financial and administrative ramifications that can ultimately jeopardize trauma system viability.^2,9[1]

This is an interesting study from 2002, that looks at the possibility of simplifying trauma triage criteria and improving the accuracy of trauma triage criteria, by making it just one trauma triage criterion.

-

Imagine a system where we intentionally ignore the usual horribly inaccurate trauma triage criteria, such as MOI (Mechanism Of Injury). And these are not even the usual criteria, because this is in a system with criteria already modifieded to avoid a lot of overtriage.

Field criteria for automatic trauma team activation were as follows:

gunshot wound to the torso or neck; gunshot wound to an extremity with loss of distal pulse; identification of a femur, tibial, or spinal fracture; any long bone open fracture; prehospital identification of airway compromise or the need for intubation in the field; systolic blood pressure < 90 mm Hg; heart rate < 60 beats/min or > 130 beats/min; spontaneous respiratory rate < 10 breaths/min or > 30 breaths/min; or paralysis in one or more extremities.

Other than the gunshot wound parameters, mechanism of injury criteria were not routinely used to determine trauma team activation for victims of falls and motor vehicle crashes.^[1]

I like the idea, but the way to assess this is not by hospital admissions.

-

The principal outcome evaluated in this study was the need for patient hospitalization to treat documented injuries.^[1]

This is the big problem with the study. The authors do admit that this is a limitation, patients may be admitted just for observation. However, the authors do not seem to see the way that GCS can be much more likely to correlate positively with admission. Nowhere else in the study does it mention that they limited this to admissions to treat documented injuries.

If the patient has a GCS of 14, how many doctors will admit the patient overnight, just for observation? How many just for observation patients does it take to affect the data?

-

-

-

-

The same is true for anything else that might be expected to result in a lot of admission for observation. The possibility of a head injury is so potentially catastrophic, that we have had exponential growth of CT scans of the head, because you can’t be too safe. Head CTs have become so common that we can expect it to increase some cancer rates over the next couple of decades. Just another example of being too safe.

Multivariate stepwise logistic regression analysis was performed to determine which variables within the data set were significant predictors of admission. The dependent variable in the analysis was admission versus nonadmission.^[1]

This is the wrong endpoint. A patient could have been transported to the local not trauma designated hospital and been admitted, as well. If the patient could have been adequately treated at a not trauma designated hospital, should that patient be considered to have appropriately bypassed the not trauma designated hospital just because the patient was admitted to a trauma designated hospital?

Any cause for admission is acceptable. I am being driven to the hospital for a respiratory problem (not yet diagnosed as pneumonia); there is a collision; everybody is uninjured, but the vehicles is not capable of being driven, so I continue to the hospital by ambulance. If I am transported as a trauma, I may expect to be admitted for the pneumonia, not for trauma – but I should expect that I will not be admitted for the lack of injuries due to trauma.

-

The present study was designed to evaluate traditional triage parameters for predicting hospitalization after MVC. The GCS score was the only prehospital physiologic parameter that provided a clinically relevant and measurable difference between patients who were hospitalized versus those who were discharged from the ED, excluding patients who were in extremis and died shortly after arrival. This reliability was not changed by the presence of positive prehospital anatomic criteria or prehospital airway issues. Three tiers of probability for admission were evident: a very high probability (96%) for GCS score 12, a high probability (73%) for GCS score of 13 to 14, and a relatively low probability (32%) for GCS score of 15. Neither the suspected presence or absence of drugs, alcohol, or a reported loss of consciousness compromised the accuracy of the GCS as a predictor for imminent hospitalization. This is not surprising, because disorientation, belligerence, mild confusion, or history of loss of consciousness does not significantly affect the GCS score. Retrospectively, had a GCS score of 14 been used to trigger trauma team activation, the overall undertriage rate of 19.1% would have improved to 4.4%.^[1]

There is certainly something to consider about this, but the validation needs to be something more objective than admission.

-

Footnotes:

-

^[1] A prehospital glasgow coma scale score < or = 14 accurately predicts the need for full trauma team activation and patient hospitalization after motor vehicle collisions.
Norwood SH, McAuley CE, Berne JD, Vallina VL, Creath RG, McLarty J.
J Trauma. 2002 Sep;53(3):503-7.
PMID: 12352488 [PubMed - indexed for MEDLINE]

.