New Epidemic Spread…Accurately Predicted Using ‘Mathematical Formulas’

– IBS Joint Research Team Proposes Methodology to Overcome Limitations of Existing Epidemic Spread Models,

Collaborative research team involved in this study. From left: Kim Jae-kyung CI (IBS/KAIST), Dr. Hong Hyuk-pyo (IBS/KAIST, currently at University of Wisconsin), Senior Researcher Choi Sun-hwa (National Institute for Mathematical Sciences), Dr. Eom Eun-jin (Korea University, currently at Disease Control Headquarters), Professor Choi Bo-seung (IBS/Korea University).[Provided by IBS],
,
, ‘[Herald Economy = Koo Bon-hyuk Reporter] Mathematics has provided scientific evidence for optimal defense systems in humanity’s war against epidemics. The research team led by Kim Jae-kyung CI (professor of mathematics at KAIST) of the Biomedical Mathematics Group at the Institute for Basic Science (IBS) collaborated with Choi Sun-hwa, Senior Researcher at the National Institute for Mathematical Sciences, Professor Choi Bo-seung at Korea University, and Professor Lee Hyo-jung at Kyungpook National University to propose a new predictive model for epidemic spread, drastically improving accuracy.’,
,
, ‘When an unknown virus emerges, scientists work to understand its structure and nature, while pharmaceutical companies develop vaccines and treatments against it. During the period of creating these weapons against the virus, prevention serves as a shield to protect the public and minimize damage. Mathematics is employed to accurately predict the damage, allocate medical personnel, and secure hospital beds, among other response measures.’,
,
, ‘The COVID-19 pandemic highlighted the importance of mathematic model-based epidemic spread models. Variables such as the estimated reproduction number (R-value), incubation period, and infectious period derived from these models play a crucial role in understanding the disease’s spread pattern and designing prevention policies.’,
,
, ‘However, existing models had limitations. Most traditional models assumed that all contacts of an infected person have the same probability of becoming infectious, regardless of the timing of contact. Predictions have been based on Markovian systems, where future states are determined solely by the current state and are unaffected by past events.’,
,
, ‘Yet, in real-world scenarios, both current and past states influence the future (non-Markovian systems). Due to the incubation period following exposure to an infected person, those who have been in contact earlier are more likely to develop infectiousness.’,
,
, ‘Professor Choi Bo-seung explained that “non-Markovian systems, which consider both current and past states, complicate mathematical estimations and modeling, making calculations difficult. Hence, traditional epidemic spread models have relied on Markovian assumptions, failing to accurately reflect the true spread pattern of infectious diseases.”’,
,
, ‘The joint research team led by IBS developed a new epidemic spread model that considers both current and past states. They overcame the limitations of existing models by introducing delay differential equations that explain future changes using both the current and past states instead of ordinary differential equations that only consider the current state.’,
,
,

New estimation method for epidemiological indices based on realistic assumptions.[Provided by IBS],
,
, ‘The research team evaluated the accuracy of the newly proposed model using the cumulative COVID-19 cases in Seoul from January 20 to November 25, 2020. During the initial virus outbreak period (2020.1.20.~3.3) when cases surged, the reproduction number was estimated as 4.9 by the existing model and 2.7 by the new model. The actual value obtained by tracing the transmission routes of the confirmed cases was 2.7. This demonstrates that the existing model could significantly overestimate the reproduction number, leading to an overestimation of COVID-19’s infectiousness.’,
,
, ‘Senior Researcher Choi Sun-hwa noted that “to address the issue of overestimation, existing models have used additional epidemiological information, such as infectious periods, to correct values. However, the new model can accurately estimate the reproduction number without the need for such additional information.”’,
,
, ‘Kim Jae-kyung CI stated, “Our research team developed a program called ‘IONISE’ based on the new model and made it freely available for other researchers in the field. We expect it to help public health experts better understand epidemic spread patterns and develop effective prevention strategies.”’,
,
, ‘The research findings were published in the international journal ‘Nature Communications’ on October 9.’,
,
,