Bo Xu

Modeling the global dynamics of emerging infectious diseases (EIDs) like COVID-19 can provide important guidance in the preparation and mitigation of pandemic threats. While age-structured transmission models are widely used to simulate the evolution of EIDs, most of these studies focus on the analysis of specific countries and fail to characterize the spatial spread of EIDs across the world. Here, we developed a global pandemic simulator that integrates age-structured disease transmission models across 3,157 cities and explored its usage under several scenarios. We found that without mitigations, EIDs like COVID-19 are highly likely to cause profound global impacts. For pandemics seeded in most cities, the impacts are equally severe by the end of the first year. The result highlights the urgent need for strengthening global infectious disease monitoring capacity to provide early warnings of future outbreaks. Additionally, we found that the global mitigation efforts could be easily hampered if developed countries or countries near the seed origin take no control. The result indicates that successful pandemic mitigations require collective efforts across countries. The role of developed countries is vitally important as their passive responses may significantly impact other countries.

Mobility restrictions have been a heated topic during the global pandemic of coronavirus disease 2019 (COVID-19). However, multiple recent findings have verified its importance in blocking virus spread. Evidence on the association between mobility, cases imported from abroad and local medical resource supplies is limited. To reveal the association, this study quantified the importance of inter- and intra-country mobility in containing virus spread and avoiding hospitalizations during early stages of COVID-19 outbreaks in India, Japan, and China. We calculated the time-varying reproductive number (Rt) and duration from illness onset to diagnosis confirmation (Doc), to represent conditions of virus spread and hospital bed shortages, respectively. Results showed that inter-country mobility fluctuation could explain 80%, 35%, and 12% of the variance in imported cases and could prevent 20 million, 5 million, and 40 million imported cases in India, Japan and China, respectively. The critical time for screening and monitoring of imported cases is 2 weeks at minimum and 4 weeks at maximum, according to the time when the Pearson’s Rs between Rt and imported cases reaches a peak (>0.8). We also found that if local transmission is initiated, a 1% increase in intra-country mobility would result in 1430 (±501), 109 (±181), and 10 (±1) additional bed shortages, as estimated using the Doc in India, Japan, and China, respectively. Our findings provide vital reference for governments to tailor their pre-vaccination policies regarding mobility, especially during future epidemic waves of COVID-19 or similar severe epidemic outbreaks.

Significance: By linking seasonality of climate and changing human behavior, we demonstrate that collaboration on global efforts for prompt and intensive intervention is fundamental to coping with future pandemic waves of COVID-19. We propose that this collaboration can be started in locations with typically high population density and international travel, followed by other high-risk locations. We believe this tiered intervention strategy can greatly integrate global efforts and is effective and practical to improve the global emergency response to COVID-19 and many other infectious diseases.

The most effective interventions: By 23 January 2020, China had imposed a national emergency response to restrict travel and impose social distancing measures on its populace in an attempt to inhibit the transmission of severe acute respiratory syndrome–coronavirus 2 (SARS-CoV-2). However, which measures were most effective is uncertain. Tian et al. performed a quantitative analysis of the impact of control measures between 31 December 2019 and 19 February 2020, which encompasses the Lunar New Year period when millions of people traveled across China for family visits. Travel restrictions in and out of Wuhan were too late to prevent the spread of the virus to 262 cities within 28 days. However, the epidemic peaked in Hubei province on 4 February 2020, indicating that measures such as closing citywide public transport and entertainment venues and banning public gatherings combined to avert hundreds of thousands of cases of infection. It is unlikely that this decline happened because the supply of susceptible people was exhausted, so relaxing control measures could lead to a resurgence. Science , this issue p. 638

The 2009 pandemic influenza virus caused the majority of the influenza A virus infections in China in 2009. It arrived in several Chinese cities from imported cases and then spread as people travelled domestically by all means of transportation, among which road traffic was the most commonly used for daily commuting. Spatial variation in socioeconomic status not only accelerates migration across regions but also partly induces the differences in epidemic processes and in responses to epidemics across regions. However, the roles of both road travel and socioeconomic factors have not received the attention they deserve. Here, we constructed a national highway network for and between 333 cities in mainland China and extracted epidemiological variables and socioeconomic factors for each city. We calculated classic centrality measures for each city in the network and proposed two new measures (SumRatio and Multicenter Distance). We evaluated the correlation between the centrality measures and epidemiological features and conducted a spatial autoregression to quantify the impacts of road network and socioeconomic factors during the outbreak. The results showed that epidemics had more significant relationships with both our new measures than the classic ones. Higher population density, higher per person income, larger SumRatio and Multicenter Distance, more hospitals and college students, and lower per person GDP were associated with higher cumulative incidence. Higher population density and number of slaughtered pigs were found to advance epidemic arrival time. Higher population density, more colleges and slaughtered pigs, and lower Multicenter Distance were associated with longer epidemic duration. In conclusion, road transport and socioeconomic status had significant impacts and should be considered for the prevention and control of future pandemics.

Detecting the onset of influenza epidemic is important for epidemiological surveillance and for investigating the factors driving spatiotemporal transmission patterns. Most approaches define the epidemic onset based on thresholds, which use subjective criteria and are specific to individual surveillance systems. We applied the empirical threshold method (ETM), together with two non-thresholding methods, including the maximum curvature method (MCM) that we proposed and the segmented regression method (SRM), to determine onsets of influenza epidemics in each prefecture of Japan, using sentinel surveillance data of influenza-like illness (ILI) from 2012/2013 through 2017/2018. Performance of the MCM and SRM was evaluated, in terms of epidemic onset, end, and duration, with those derived from the ETM using the nationwide epidemic onset indicator of 1.0 ILI case per sentinel per week. The MCM and SRM yielded complete estimates for each of Japan’s 47 prefectures. In contrast, ETM estimates for Kagoshima during 2012/2013 and for Okinawa during all six influenza seasons, except 2013/2014, were invalid. The MCM showed better agreement in all estimates with the ETM than the SRM (R2 = 0.82, p < 0.001 vs. R2 = 0.34, p < 0.001 for epidemic onset; R2 = 0.18, p < 0.001 vs. R2 = 0.05, p < 0.001 for epidemic end; R2 = 0.28, p < 0.001 vs. R2 < 0.01, p = 0.35 for epidemic duration). Prefecture-specific thresholds for epidemic onset and end were established using the MCM. The Japanese national epidemic onset threshold is not applicable to all prefectures, particularly Okinawa. The MCM could be used to establish prefecture-specific epidemic thresholds that faithfully characterize influenza activity, serving as useful complements to the influenza surveillance system in Japan.

There is increasing concern about another influenza pandemic in China. However, the understanding of the roles of transport modes in the 2009 influenza A(H1N1) pandemic spread across mainland China is limited. Herein, we collected 127,797 laboratory-confirmed cases of influenza A(H1N1)pdm09 in mainland China from May 2009 to April 2010. Arrival days and peak days were calculated for all 340 prefectures to characterize the dissemination patterns of the pandemic. We first evaluated the effects of airports and railway stations on arrival days and peak days, and then we applied quantile regressions to quantify the relationships between arrival days and air, rail, and road travel. Our results showed that early arrival of the virus was not associated with an early incidence peak. Airports and railway stations in prefectures significantly advanced arrival days but had no significant impact on peak days. The pandemic spread across mainland China from the southeast to the northwest in two phases that were split at approximately 1 August 2009. Both air and road travel played a significant role in accelerating the spread during phases I and II, but rail travel was only significant during phase II. In conclusion, in addition to air and road travel, rail travel also played a significant role in accelerating influenza A(H1N1)pdm09 spread between prefectures. Establishing a multiscale mobility network that considers the competitive advantage of rail travel for mid to long distances is essential for understanding the influenza pandemic transmission in China.

After publication of the research paper

Striped field mice (Apodemus agrarius) are the main host for the Hantaan virus (HTNV), the cause of hemorrhagic fever with renal syndrome (HFRS) in central China. It has been shown that host population density is associated with pathogen dynamics and disease risk. Thus, a higher population density of A. agrarius in an area might indicate a higher risk for an HFRS outbreak. Here, we surveyed the A. agrarius population density between 2005 and 2012 on the Weihe Plain, Shaanxi Province, China, and used this monitoring data to examine the relationships between the dynamics of A. agrarius populations and environmental conditions of crop-land, represented by remote sensing based indicators. These included the normalized difference vegetation index, leaf area index, fraction of photosynthetically active radiation absorbed by vegetation, net photosynthesis (PsnNet), gross primary productivity, and land surface temperature. Structural equation modeling (SEM) was applied to detect the possible causal relationship between PsnNet, A. agrarius population density and HFRS risk. The results showed that A. agrarius was the most frequently captured species with a capture rate of 0.9 individuals per hundred trap-nights, during 96 months of trapping in the study area. The risk of HFRS was highly associated with the abundance of A. agrarius, with a 1–5-month lag. The breeding season of A. agrarius was also found to coincide with agricultural activity and seasons with high PsnNet. The SEM indicated that PsnNet had an indirect positive effect on HFRS incidence via rodents. In conclusion, the remote sensing-based environmental indicator, PsnNet, was highly correlated with HTNV reservoir population dynamics with a 3-month lag (r = 0.46, p < 0.01), and may serve as a predictor of potential HFRS outbreaks.