Computer-Aided System for COVID-19 Using Semi-supervised-based Ensemble Learning and Reinforcement Learning

Coronavirus Disease 2019(COVID-19) has shocked the world with its rapid spread and enormous threat to life and has continued up to the present. In this paper, a computer-aided system is proposed to detect infections and predict the disease progression of COVID-19. A high-quality CT scan database labeled with time-stamps and clinicopathologic variables is constructed to provide data support. To our knowledge, it is the only database with time relevance in the community. An object detection model is then trained to annotate infected regions. Using those regions, we detect the infections using a model with semi-supervised-based ensemble learning and predict the disease progression depending on reinforcement learning. We achieve an mAP of 0.92 for object detection. The accuracy for detecting infections is 98.46%, with a sensitivity of 97.68%, a specificity of 99.24%, and an AUC of 0.987. Significantly, the accuracy of predicting disease progression is 90.32% according to the timeline. It is a state-of-the-art result and can be used for clinical usage. © 2022 IEEE.

Prediction and Analysis of Infectious Diseases Based on M-SIR Model

The epidemic of infectious diseases has become a major problem threatening the world public health, and the dynamic models of virus spreading are widely used for epidemic tracking and prediction. The existing dynamic models do not consider the synergistic effects of population migration factors and changes in transmission rates on diseases. Therefore, based on the SIR (Susceptible-Infectious-Recovered) model, the time-dependent M-SIR (Migration-Susceptible-Infectious-Recovered) model was proposed by introducing the population migration (Migration) factor. Meanwhile, introducing the machine learning LightGBM (Light Gradient Boosting Machine) method to track the infection rate and recovery rate, and explored the impact of cross-regional population movement and prevention and control measures on the development of the epidemic. Take the new crown epidemic as an example, firstly, the data of population migration and epidemic spread were statistically analyzed to monitor the relationship between population mobility and epidemic development. Then, the m-sir model is used to predict the infected cases and removed cases in Beijing and Shanghai. Through comparative analysis with the SIR model, the prediction accuracy of the model has been greatly improved. At the same time, the development trend of the epidemic situation in related cities before and after control is explored, which can provide some theoretical support for future epidemic prediction and control decisions. © 2022 IEEE.

Study on the Scavenging Effect of Steady-state Displacement Air Purification System on Indoor Bioaerosol

Many types of microorganisms, including SARS-CoV-2, can spread through aerosols. Indoor medical environments are abundant in bioaerosols, which can cause infections among medical staff members and patients in hospitals. Given the ongoing COVID-19 pandemic, using a steadystate displacement air purification system may reduce the spread of SARS-CoV-2 and other microorganisms. In this study, we analyzed the purification effect of the steady-state displacement air purification system on bioaerosols in the bronchoscopy room of the hospital. In particular, bioaerosols were collected from the bronchoscopy room at different periods from April to May 2021. Among them, the microorganisms contained in the bioaerosol were identified using nextgeneration sequencing (NGS) and culture and strain identification. During the experiment, we took 5 sampling points to collect the bioaerosols. The total purification efficiency was 88.0% (NGS) and 87.5% (microbial culture count and identification). The results were significantly different between the purified and unpurified groups. In an occupant environment in the bronchoscopy room, the steady-state displacement air purification system exerted a favorable removal effect on the bioaerosols. Such purification efficiency may help prevent the in-hospital spread of COVID-19 and various infectious diseases.

Sentiment Analysis of Weibo Posts on Public Health Emergency with Feature Fusion and Multi-Channel

[Objective] This paper proposes a multi-channel MCMF-A model for Weibo posts based on feature fusion and attention mechanism, aiming to further explore the semantic information of public health emergency. [Methods] Firstly, we generated word vectors with Word2vec and FastText at the feature vector embedding level, which were merged with the vectors of part-of-speech features and position features. Secondly, we constructed multi-channel layer based on CNN and BiLSTM to extract local and global features of Weibo posts. Thirdly, we utilized the attention mechanism to extract important features of the texts. Finally, we merged the multi-channel output results, and used the softmax function for sentiment classification. [Results] We examined MCMF-A model with 42 384 Weibo posts on COVID-19. The F1 value of the proposed model reached 90.21%, which was 9.71% and 9.14% higher than the benchmark CNN and BiLSTM models. [Limitations] More research is needed to expand the experiment data size to include more small and multi-modal information such as images and voices. [Conclusions] The proposed model could effectively conduct sentiment analysis with Weibo posts. © 2021, Chinese Academy of Sciences. All rights reserved.

Wastewater-Based Epidemiology for Managing the COVID-19 Pandemic

SARS-CoV-2 is shed by COVID-19 patients and can be detected in wastewater. Thus, testing wastewater for the virus provides a depiction of disease prevalence in a community. Virus concentration data can be utilized to monitor infection trends, identify hot spots, and inform decision makers regarding reopening efforts and directing resources. This perspective aims to shed light on the current situation relating to SARS-CoV-2 in the wastewater system and the opportunity to utilize wastewater to collect useful epidemiological data. First, the survivability of SARS-CoV-2 in different water matrices is examined through the lens of surrogate viruses. Second, the effect of wastewater treatment processes on SARS-CoV-2 is investigated. Current standards for sufficient reduction of the virus and the risk of exposure that arises at each stage in the wastewater treatment process are discussed. Third, the immense potential of wastewater-based epidemiology (WBE) for managing the ongoing COVID-19 pandemic is analyzed. Studies that have tested wastewater or sludge for SARS-CoV-2 are discussed, and results are tabulated. Lastly, the current limitations of WBE and opportunities of future research are explored. Using the wealth of knowledge that the scientific community now has about WBE, wastewater testing should be considered by regional governments and private institutions.

CLINICAL CHARACTERISTICS OF 537 CASES OF COVID-19 IN DIFFERENT ALTITUDE OF SICHUAN PROVINCE, CHINA: A RETROSPECTIVE STUDY

SESSION TITLE: Late-breaking Abstract Posters SESSION TYPE: Original Investigation Posters PRESENTED ON: October 18-21, 2020 PURPOSE: There are no published studies concerning the clinical characteristics of patients with coronavirus disease 2019 (COVID-19) resident in high altitude. We aim to analyze the unique cohort and investigated the relationship among the patients resident in different altitude areas of Sichuan province, China. METHODS: We retrospectively reviewed the medical data including clinical features, imaging findings, laboratory characteristics and received treatment of confirmed cases of COVID-19 with WHO interim guidance in Sichuan province from Jan 12th to March 12th, 2020. Outcomes were also compared between patients from high-altitude and non-high-altitude area. RESULTS: 537 patients were included in this study, 78 were from high-altitude area and 459 from non-high-altitude area. The high-altitude area group was less likely had a clear travel history of Wuhan (9.0% vs. 64.3%, p<0.001) or contact with people from Wuhan (10.3% vs. 25.5%, p=0.003) within recent two weeks, but more likely contact with people infected of COVID-19 never been to Wuhan (60.3% vs. 3.3%, p<0.001) or of unknown origin (20.5% vs.7.0%, p<0.001). The high-altitude area group was more likely asymptomatic before admission (50.6% vs. 8.7%, p<0.001). Less patients with cough, sputum production and fever in the cohort comparing with those in non-high-altitude area. Lower body temperature, slower heart rate, and lower blood pressure relatively on admission were also found in high-altitude patients as well. The period from onset of illness to visit was 3 days (range, 1-7 days) and to diagnosis was 4 days (range,1-7 days), which were both significantly shorter in the group of high-altitude area. The patients from high-altitude had significantly higher lymphocyte count, lower platelet count and lower level of erythrocyte sedimentation rate (p all＜0.05). Less patients with abnormally elevated C-reactive protein were also observed in high-altitude group (p＜0.05). The most common patterns seen in these population on chest CT were bilateral distribution, ground-glass opacity, and consolidation and nodules. Multifocal distribution and consolidation were noted and showed significant in high-altitude patients. The proportion of patients with mild illness was significantly higher in the high-altitude area group (16[20.5%] vs. 33[7.2%], p<0.001). More likely to receiving antiviral and glucocorticoid therapies and less likely to receiving antibiotic drugs were shown in the patients in high-altitude area. Multivariate logistic analysis indicated that age, gender, diabetes and recent travel history of Wuhan were predictors of severe pneumonia. CONCLUSIONS: The disease severity of COVID-19 in Sichuan province was milder than Wuhan. More patients in high-altitude area were asymptomatic. It could owe to the active prevention, early detection and timely treatment. CLINICAL IMPLICATIONS: Active prevention, early detection and timely treatment is critical in COVID-19. DISCLOSURES: No relevant relationships by Aamer Chughtai, source=Web Response No relevant relationships by Lu Guo, source=Admin input no disclosure on file for Caiyu Jiang;no disclosure on file for Xiayin Peng;no disclosure on file for Hong Pu;no disclosure on file for Weiwei Qiang;no disclosure on file for Yang Yang;