Predicting Outcomes in Emergency Medical Admissions Using a Laboratory Only Nomogram

Area of Science:

• Emergency Medicine
• Biostatistics
• Medical Informatics

Background:

• Emergency department triage and laboratory data are crucial for predicting patient outcomes.
• The Acute Illness Severity (AIS) model previously showed utility in predicting mortality.
• A need exists to better explain the prognostic factors within the AIS model.

Purpose of the Study:

• To develop and describe a nomogram to visually represent and explain the AIS model.
• To predict 30-day in-hospital survival for emergency medical admissions.

Main Methods:

• Logistic regression analysis of 96,305 emergency medical admissions (2002-2016) to link mortality with laboratory data.
• Transposition of the validated AIS model into a Kattan-style nomogram using Stata.
• Inclusion of Manchester triage category and specific biochemical markers (albumin, sodium, potassium, urea, RDW, troponin).

Main Results:

• The AIS model, incorporating laboratory data, demonstrated strong predictive performance (AUROC 0.85).
• High sensitivity (94.4%) and negative predictive value (99.1%) were observed.
• The nomogram effectively maps predictor variables to a probability axis for bedside and educational use.

Conclusions:

• A Kattan-style nomogram is presented to illustrate the prognostic factors of the AIS score.
• The nomogram serves as an accessible tool for clinicians to understand and explain patient prognosis.
• This visual aid enhances the interpretability of the AIS model for predicting survival.