Risk prediction models for imported infectious diseases: A systematic review

Area of Science:

• Epidemiology
• Public Health
• Mathematical Modeling

Background:

• Accurate prediction models for imported infectious diseases are vital for global health security, including early warning and resource allocation.
• Existing models for imported infectious diseases exhibit substantial heterogeneity in structure, reporting, and performance, hindering practical application.

Purpose of the Study:

• To systematically review and analyze the landscape of predictive models for imported infectious diseases.
• To identify commonalities, variations, and limitations in current modeling approaches, focusing on predictors, performance metrics, and validation strategies.

Main Methods:

• Conducted a systematic review following PRISMA guidelines, searching PubMed, Scopus, and Embase.
• Extracted data from 273 studies (308 models) using the CHARMS framework, focusing on model characteristics and validation.
• Categorized models by transmission route (respiratory, vector-borne, fecal-oral) and modeling approach (dynamic, statistical, machine learning).

Main Results:

• Most models focused on respiratory or vector-borne diseases, primarily in airport settings, estimating disease burden.
• Limited reporting of predictive performance (54.6%) and validation (38.5% internal, 9.9% external) was observed.
• Predictors varied by transmission route, with environmental factors for vector-borne and mobility for respiratory/fecal-oral diseases.

Conclusions:

• Imported infectious disease models are constrained by limited scenario diversity and inadequate reporting of performance and validation.
• Future models require improved transparency, broader cross-border context consideration, and robust external validation.
• Enhanced modeling is essential to improve utility in global public health decision-making.