The Role of Glasgow Coma Scores in Modeling APACHE®IV Scores in Neuro ICU Patients (2008)

Undergraduates: Kristina Riemen, Hakimi, R., Olson, B. Halley, N. Thoyre, S. Olson, D., Riemen, K.

Faculty Advisor: Suzanne Thoyre
Department: Nursing

---

Purpose: To explore how different Glasgow Coma Scores (GCS) obtained during the first 24-hours of admission to ICU will impact the computed values of APACHE®IV where the APACHE®IV is a predictor of mortality and length of stay of Neurological Critical Care patients. Background/Significance: APACHE®IV scores are calculated from the lowest GCS in the first 24-hours. However, the GCS in neurologically injured patients may vary significantly during the first 24-hours and we were interested in exploring whether, with the neuro-ICU patients, the highest GCS in the first 24-hours (when used to compute APACHE®IV scores) is superior in predicting mortality and length of stay (LOS). Methods: We created three models for predicting mortality and LOS for 50 patients in the Neurological ICU that were in an ongoing IRB approved study. Each model used the APACHE®IV calculator changing only the GCS. In the first model we used the lowest GCS in the first-twenty four hours. In the second model we used the option of not including a GCS. In the third and final model we used the highest GCS in the first twenty-four hours. Results: Regression analyses were performed using SAS 9.1 for Windows (Cary, NC). The standard model using the lowest GCS was a significant predictor for mortality (r2=0.078, p=0.049), but not for LOS (r2=0.044, p=0.15). The model omitting the GCS did not significantly predict mortality (r2=0.056, p=0.098) or LOS (r2=0.073, p=0.057). The model using highest GCS was a significant