Predictive Models for Dengue Severity, Mortality and Hospitalization: A Systematic Review and Meta-Analysis Protocol

Résumé fourni par la source

Background: Dengue fever affects millions of people worldwide, with clinical outcomes ranging from mild illness to severe complications. Accurate prediction of adverse outcomes remains challenging for clinicians, particularly in resource-limited settings where dengue is endemic. Clinical prediction models have emerged as promising tools to support decision-making, but their performance and applicability across different populations and settings remain unclear. This systematic review aims to identify, evaluate, and synthesize evidence on the predictive performance of clinical models for severity, mortality, and hospitalization in dengue patients.Methods: We will search PubMed, LILACS, Web of Science, Scopus, and Embase for studies reporting clinical predictive models for dengue mortality and/or hospitalization outcomes, with no restrictions on year or language. Two reviewers will independently screen articles, extract data, and assess risk of bias using the TRIPOD checklist and PROBAST tool. We will extract information on model characteristics, predictor variables, performance metrics, including area under the curve, sensitivity, and specificity, validation approaches and study limitations. Meta-analysis will be conducted when sufficient homogeneous data is available, reporting the pooled Area Under the ROC Curve (AUC) of the models.Discussion: This review will provide a comprehensive assessment of existing predictive models for dengue outcomes, identifying high-performing models suitable for clinical implementation and highlighting gaps that require further research.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.