Machine Learning Based Regional Epidemic Transmission Risks Precaution in Digital Society (preprint)

The contact and interaction of human is considered to be one of the important factors affecting the epidemic transmission, and it is critical to model the heterogeneity of individual activities in epidemiological risk assessment. In digital society, massive data makes it possible to implement this idea on large scale. Here, we use the mobile phone signaling to track the users’ trajectories and construct contact network to describe the topology of daily contact between individuals dynamically. We show the spatiotemporal contact features of about 7.5 million mobile phone users during the outbreak of COVID-19 in Shanghai, China. Furthermore, the individual feature matrix extracted from contact network enables us to carry out the extreme event learning and predict the regional transmission risk, which can be further decomposed into the risk due to the inflow of people from epidemic hot zones and the risk due to people close contacts within the observing area. This method is much more flexible and adaptive, and can be taken as one of the epidemic precautions with high efficiency and low cost.

Inferring Economic Condition Uncertainty from Electricity Big Data (preprint)

Inferring the uncertainty in economic conditions is significant for both decision makers as well as market players. In this paper, we propose a novel approach to measure the economic uncertainties by using the Hidden Markov Model (HMM). We construct a dimensionless index, Economic Condition Uncertainty (ECU) index, which ranges from zero to one and is comparable among sectors, regions and periods. We used the daily electricity consumption data of more than 18,000 firms in Shanghai from 2018 to 2020 to construct the ECU indexes. Results show that all ECU indexes, whether at sectoral or regional level, successfully captured the negative impacts of COVID-19 on Shanghai's economic conditions. Besides, the ECU indexes also presented the heterogeneities in different districts as well as in different sectors. This reflects the facts that changes in the uncertainty of economic conditions are mainly related to regional economic structures and targeted regulatory policies faced by sectors. The ECU index can also be readily extended to measure the uncertainty of economic conditions in various realms, which has great potentials in the future.

Engineered interferon alpha effectively improves clinical outcomes of COVID-19 patients (preprint)

Interferons are key to the antiviral host defense, yet the therapeutic value of interferon for coronavirus disease 2019 (COVID-19) is unknown. Recombinant super-compound interferon (rSIFN-co) is a new genetically engineered interferon, thus we conducted a multicenter, randomized controlled trial (ChiCTR2000029638) to evaluate the efficacy and safety of recombinant super-compound interferon versus traditional interferon alpha added to baseline antiviral agents (lopinavir–ritonavir or umifenovir) for the treatment of moderate-to-severe COVID-19. Participants received rSIFN-co (12 million international units [IU], twice daily) or interferon alpha (5 million IU, twice daily) nebulization added to baseline antiviral agents for no more than 28 days. The primary outcome was the time to clinical improvement. Secondary outcomes included the overall rate of clinical improvement assessed on day 28，the time to radiological improvement and virus nucleic acid negative conversion, and adverse events. 94 patients hospitalized with moderate-to-severe COVID-19 were included in the safety set (46 patients assigned to rSIFN-co group, 48 to interferon alpha group). Individuals in the rSIFN-co group showed shorter time to clinical improvement (11.5 days vs 14.0 days; P = 0.019) as compared to those in the interferon alpha group. The overall rate of clinical improvement on day 28 was much higher in the rSIFN-co group than that in the interferon alpha group (93.5% vs 77.1%; difference, 16.4%; 95% confidence interval 3% to 30%). The time to radiological improvement and the time to virus nucleic acid negative conversion were also much shorter in the rSIFN-co group (8.0 days vs 10.0 days, P = 0.002; 7.0 days vs 10.0 days, P = 0.018, respectively). Adverse events were reported in 13 (28.3%) patients in the rSIFN-co group and 18 (37.5%) patients in the interferon alpha group. No patients died during the study. Our study showed that rSIFN-co added to antiviral agents was safe and more efficient than interferon alpha plus antiviral agents in the treatment of moderate-to-severe COVID-19. Future clinical study of rSIFN-co therapy alone or combined with other antiviral therapy is warranted.