We conclude that the risk of mortality among ED patients could be accurately predicted by using common clinical signs and biochemical tests.
Medical patients admitted to Emergency Department (ED) are highly heterogeneous in terms of disease spectrum and severity.
Model I included 10 factors and had a posterior probability of 5.5%; model II included 8 factors with 4% posterior probability; and model III included 11 factors with 3.8% posterior probability.
Several risk factors were present in different models.
The average age of participants was 65.8 years (range: 16–105 years).
The Bayesian Model Averaging method within the Cox’s regression model was used to identify independent risk factors for mortality.
In the development cohort, the incidence of 30-day mortality was 9.8%, and the following factors were associated with a greater risk of mortality: male gender, increased respiratory rate and serum urea, decreased peripheral oxygen saturation and serum albumin, lower Glasgow Coma Score, and admission to intensive care unit.
The Bayesian Model Average (BMA) algorithm identified 3 most parsimonious models for predicting the risk of 30-day mortality.
The three models included the following factors: gender, respiratory rate, peripheral oxygen saturation, duration of illness, GCS, ICU admission, serum urea, glycaemia, serum albumin, alanine aminotransferase (ALT), and high-sensitivity C-reactive protein (hs CRP).