Development of a bedside score to predict dengue severity

Area of Science:

• Epidemiology
• Infectious Diseases
• Public Health

Background:

• New Caledonia faced a severe dengue outbreak in 2017 with a high burden on healthcare services.
• A predictive model for dengue severity was needed to optimize patient management and hospital resource allocation.

Purpose of the Study:

• To develop a local operational model for predicting dengue severity.
• To create a comprehensive patient score based on clinical and biological parameters.

Main Methods:

• Retrospective analysis of hospitalized dengue patients from January to July 2017.
• Univariate and multivariate analyses to identify risk factors for severe dengue.
• Development of predictive models using a descending step-wise method, validated on 2018 data.

Main Results:

• Out of 383 patients, 34% developed severe dengue; 3.4% died.
• Key predictors of severity included age, comorbidities, alert signs, low platelets, prolonged prothrombin time, elevated AST/ALT, and prior dengue infection.
• Gender-specific models showed high predictive accuracy (AUC 0.80-0.88).

Conclusions:

• Robust and efficient bedside scores were developed to predict dengue severity.
• The proposed spreadsheet tool can aid health practitioners in managing dengue outbreaks and improving patient care.
• The models aim to enhance patient management and optimize hospitalization flow during severe dengue outbreaks.