Scholarlink

Background <p>Predicting where and when epidemics will spread is essential for guiding early public health responses, yet remains challenging when surveillance data are sparse. Mobility-based approaches offer a promising alternative by leveraging human movement patterns to predict spatial spread without requiring detailed epidemiological data.</p> Methods <p>We evaluated the ability of mobility-informed effective distance to predict the relative timing of epidemic onset for three respiratory pathogens: severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), seasonal influenza virus, and respiratory syncytial virus (RSV). We compared effective distances derived from county-level commuting data, airline traffic flows, and a theoretical radiation model within a metapopulation framework. Analyses were conducted at the county level across South Carolina for all three pathogens and at the state level across the United States for influenza and SARS-CoV-2.</p> Results <p>Effective distance reliably predicted the relative order of epidemic onset across pathogens and spatial scales. Across all analyses, epidemics arrived earlier in larger, more connected populations, and effective distance outperformed geographic distance as a predictor of onset timing. In metapopulation simulations of a hypothetical respiratory pathogen, radiation-model effective distance was highly correlated with simulated arrival times (Spearman’s ρ = 0.95, 95% CI [0.90 to 0.98]), substantially outperforming commuting-based effective distance (ρ = 0.54, 95% CI [0.24 to 0.75]). In empirical analyses at the county level in South Carolina, radiation-model effective distance consistently outperformed commuting-based measures for the coronavirus disease 2019 (COVID-19) first wave (ρ = 0.62 vs. 0.39), the 2022–2023 influenza season (ρ = 0.57 vs. 0.44), and the 2022–2023 RSV season (ρ = 0.70 vs. 0.47). At the state level, airline-based effective distance predicted arrival timing for COVID-19 (ρ = 0.46, 95% CI [0.19 to 0.66]) and onset timing for the 2024/2025 influenza season (ρ = 0.45, 95% CI [0.14 to 0.66]).</p> Conclusions <p>Mobility-informed models capture consistent spatiotemporal patterns of epidemic spread across pathogens and scales. These findings support their use as scalable, data-efficient tools for early outbreak preparedness.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Mobility-informed metapopulation models predict the spatio-temporal spread of respiratory epidemics across scales

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月18日

查看全文
  • Aakash Pandey,
  • Lu Zhong,
  • Lior Rennert

摘要

Background

Predicting where and when epidemics will spread is essential for guiding early public health responses, yet remains challenging when surveillance data are sparse. Mobility-based approaches offer a promising alternative by leveraging human movement patterns to predict spatial spread without requiring detailed epidemiological data.

Methods

We evaluated the ability of mobility-informed effective distance to predict the relative timing of epidemic onset for three respiratory pathogens: severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), seasonal influenza virus, and respiratory syncytial virus (RSV). We compared effective distances derived from county-level commuting data, airline traffic flows, and a theoretical radiation model within a metapopulation framework. Analyses were conducted at the county level across South Carolina for all three pathogens and at the state level across the United States for influenza and SARS-CoV-2.

Results

Effective distance reliably predicted the relative order of epidemic onset across pathogens and spatial scales. Across all analyses, epidemics arrived earlier in larger, more connected populations, and effective distance outperformed geographic distance as a predictor of onset timing. In metapopulation simulations of a hypothetical respiratory pathogen, radiation-model effective distance was highly correlated with simulated arrival times (Spearman’s ρ = 0.95, 95% CI [0.90 to 0.98]), substantially outperforming commuting-based effective distance (ρ = 0.54, 95% CI [0.24 to 0.75]). In empirical analyses at the county level in South Carolina, radiation-model effective distance consistently outperformed commuting-based measures for the coronavirus disease 2019 (COVID-19) first wave (ρ = 0.62 vs. 0.39), the 2022–2023 influenza season (ρ = 0.57 vs. 0.44), and the 2022–2023 RSV season (ρ = 0.70 vs. 0.47). At the state level, airline-based effective distance predicted arrival timing for COVID-19 (ρ = 0.46, 95% CI [0.19 to 0.66]) and onset timing for the 2024/2025 influenza season (ρ = 0.45, 95% CI [0.14 to 0.66]).

Conclusions

Mobility-informed models capture consistent spatiotemporal patterns of epidemic spread across pathogens and scales. These findings support their use as scalable, data-efficient tools for early outbreak preparedness.