Development of a Triage-Level Predictive Model for Hospitalization in the Emergency Department.

Area of Science:

• Emergency Medicine
• Health Services Research
• Clinical Informatics

Background:

• Emergency department (ED) overcrowding is a significant global health challenge.
• Effective patient transfer from the ED is crucial for mitigating congestion.
• Early prediction of hospitalizations using triage data can optimize patient flow.

Purpose of the Study:

• To develop and validate a predictive model for hospitalizations based on emergency department triage parameters.
• To assess the model's ability to forecast patient admissions and reduce ED crowding.

Main Methods:

• A historical cohort study involving 1436 patients from two tertiary medical centers.
• Logistic regression was employed to build a prediction model using variables like triage level, vital signs, and comorbidities.
• Model performance was evaluated using Area Under the ROC Curve (AUC) and Discrimination Slope (DS) in learning, testing, and validation groups.

Main Results:

• Key predictors for hospitalization included higher triage level, low O2 saturation (<95%), malignancy, cardiovascular disease, neurologic illness, weekend arrival, and fall season.
• The model demonstrated acceptable discrimination across all groups (AUCs ranging from 0.71 to 0.77).
• Specific findings: urgent triage (OR 1.45), low O2 saturation (OR 3.32), cardiovascular disease (OR 2.93).

Conclusions:

• The developed prediction model is easily implementable in hospital systems.
• It provides an expected number of ED patient hospitalizations, aiding management.
• The model can enhance patient flow and effectively reduce ED crowding.