Kim Byul Nim, Jo Junwoo, Oh Chunyoung, Moon Sanghyeok, Abdulali Arsen, Lee Sunmi
Department of Applied Mathematics, Kyung Hee University, Yongin, Republic of Korea.
Department of Mathematics Education, Chonnam National University, Gwangju, Republic of Korea.
Front Public Health. 2025 Jun 18;13:1586786. doi: 10.3389/fpubh.2025.1586786. eCollection 2025.
The effective reproduction number ( ) is a key indicator for monitoring and controlling infectious diseases such as COVID-19, where transmission patterns can differ substantially across demographics, regions, and phases of the pandemic. In this study, we propose a novel, network-based approach to empirically estimate using detailed transmission data from South Korea. By reconstructing infector-infectee pairs, our method incorporates local factors like mobility and social distancing, offering a more precise perspective than traditional methods.
We acquired infector-infectee pair data from the Korea Disease Control and Prevention Agency (KDCA) for 2020-2021 and built infection networks to derive empirical . This framework allows us to examine regional differences and the effects of social distancing measures. We also compared our results with Cori's , which employs incidence data and serial interval distributions, to highlight the advantages of an infection network-based strategy.
Our empirical uncovered three distinct patterns. Early in the outbreak, when case numbers were low, remained near 1, indicating limited transmission. During superspreading events, our estimates showed sharper peaks than Cori's method, demonstrating higher sensitivity to sudden changes. As the Delta variant emerged, our values converged with Cori's, underscoring the utility of network-based methods for capturing nuanced shifts during high-variability phases.
Incorporating infection networks into estimation thus provides decision-makers with timely insights for targeted interventions. Empirically reconstructing infection networks and directly estimating reveal real-time transmission dynamics often overlooked by aggregated approaches. This method can significantly improve outbreak forecasts, inform more precise public health policies, and strengthen pandemic preparedness.
有效繁殖数（ ）是监测和控制COVID-19等传染病的关键指标，在大流行期间，不同人群、地区和阶段的传播模式可能存在显著差异。在本研究中，我们提出了一种新颖的基于网络的方法，利用韩国的详细传播数据对 进行实证估计。通过重建感染者与被感染者的配对，我们的方法纳入了流动性和社交距离等局部因素，提供了比传统方法更精确的视角。
我们从韩国疾病控制与预防机构（KDCA）获取了2020 - 2021年感染者与被感染者的配对数据，并构建感染网络以得出实证 。该框架使我们能够研究地区差异以及社交距离措施的影响。我们还将结果与采用发病率数据和序列间隔分布的科里（Cori）的 进行比较，以突出基于感染网络策略的优势。
我们的实证 揭示了三种不同模式。在疫情初期，病例数较少时， 接近1，表明传播有限。在超级传播事件期间，我们的估计显示出比科里方法更尖锐的峰值，表明对突然变化具有更高的敏感性。随着德尔塔变异株出现，我们的 值与科里的 值趋同，强调了基于网络的方法在高变异阶段捕捉细微变化的效用。
将感染网络纳入 估计为决策者提供了及时的见解，以便进行有针对性的干预。通过实证重建感染网络并直接估计 ，揭示了汇总方法常常忽视的实时传播动态。这种方法可以显著改善疫情预测，为更精确的公共卫生政策提供依据，并加强大流行防范。
