Spatio-Temporal COVID-19 Modeling: A Global Systematic Review of Data Integration, Equity, and Lessons for Pandemic Preparedness

The COVID-19 pandemic generated an unprecedented volume of spatially and temporally resolved data, enabling rapid development of spatio-temporal models for surveillance, forecasting, and policy support. However, the evolution, geographic distribution, and equity implications of these models remain insufficiently synthesized. This study presents a global systematic review of 363 peer-reviewed studies published between January 2020 and August 2025 using publicly available data. Following PRISMA 2020 guidelines, studies were classified by geographic scale, modeling approach, data streams, and analytical purpose. The results indicate that Bayesian and compartmental models remained dominant throughout the pandemic, although methodological diversity increased over time with the growing use of machine learning and hybrid frameworks integrating mobility, environmental, and socio-demographic data. Data integration was more common than previously reported. Approximately 30% of studies relied on a single data stream, while 70% incorporated multiple sources, although most multi-source approaches combined only two data types and relatively few studies integrated three or more. Geographic coverage was uneven, with a strong concentration of studies in high-income regions and persistent underrepresentation of low- and middle-income contexts. Models incorporating finer spatial scales and socio-demographic variables more frequently supported geographically targeted interpretation of risk, vulnerability, testing access, and intervention needs. Overall, the findings highlight the importance of multi-source data integration, improved geographic representativeness, and transparent uncertainty communication, alongside the need for FAIR-aligned and equity-aware data infrastructures to strengthen future pandemic preparedness.