Authors: Merijn C. F. Mulders, Sevilay Vural, Lisanne Boekhoud, Tycho J. Olgers, Jan C. Ter Maaten, Hjalmar R. Bouma
Journal: American Journal of Emergency Medicine, 2024 October 11
Conclusion:
The prediction model developed in this study can help identify patients with infections at the ED who are safe for early discharge. This tool could help reduce unnecessary hospitalizations and optimize healthcare resource utilization.
Methods:
This prospective cohort study included adult, non-trauma patients presenting to the ED with a suspected infection and at least two SIRS criteria. Safe early discharge (SED) was defined as discharge within 24 hours without readmission or death within seven days. Multivariate logistic regression was used to develop a prediction model for SED, which was validated with k-fold cross-validation.
Results:
The study included 1381 patients, of which 354 (25.6%) met the criteria for SED. Predictors of SED included younger age, absence of comorbidities, living independently, low triage urgency (yellow or green), no ambulance transport or general practitioner referral, normal clinical impression scores, normal risk scores (qSOFA, PIRO, MEDS, NEWS, SIRS), stable vital signs, and absence of kidney or respiratory failure. The model had an area under the curve (AUC) of 0.824, indicating good predictive performance. Validation showed minimal performance decline, suggesting a robust model.