Comparison of the clinical features of human bocavirus and metapneumovirus lower respiratory tract infections in hospitalized children in Suzhou, China

Objective We compared the clinical data of hospitalized children with lower respiratory tract infections caused by human bocavirus (HBoV) and human metapneumovirus (hMPV). Methods In total, 8,430 children admitted to the Department of Respiration, Children's Hospital of Soochow University for lower respiratory tract infections from January 2017 to October 2021 were enrolled. Seven common respiratory viruses, including respiratory syncytial virus, influenza virus A, influenza virus B, parainfluenza virus (PIV) I, PIV II, PIV III, and adenovirus, were detected by direct immunofluorescence assay, whereas human rhinovirus and hMPV were detected by reverse transcription-polymerase chain reaction. Mycoplasma pneumoniae (MP) and HBoV were detected by real-time fluorescence quantitative polymerase chain reaction. Bacteria was detected in blood, nasopharyngeal secretion, bronchoalveolar lavage specimen or pleural fluid by culture. In parallel, MP was detected by enzyme-linked immunosorbent assay. In addition, we performed metagenomic testing of alveolar lavage fluid from some of the patients in our study. Results The detection rate of HBoV was 6.62% (558/8430), whereas that of hMPV was 2.24% (189/ 8430). The detection rate of HBoV was significantly higher in children aged 1 to <3 years than in other age groups, but there were no significant differences in positivity rates for hMPV by age. Before 2020, the incidence of HBoV infection peaked in summer and autumn, whereas that of hMPV peaked in spring. The epidemiology of both HBoV and hMPV has changed because of the impact of the novel coronavirus. Among the positive cases, the HBoV mixed infection rate was 51.6%, which was similar to that for hMPV mixed infection (44.4%). Comparing clinical characteristics between HBoV and hMPV single infection, the median age of children was 17 months in the HBoV group and 11 months in the hMPV group. In the HBoV single infection group, 31 patients (11.5%) had pulse oxygen saturation of less than 92% on admission, 47 (17.4%) had shortness of breath, and 26 (9.6%) presented with dyspnea. Meanwhile, four patients (3.8%) in the hMPV single infection group had pulse oxygen saturation of less than 92% on admission, eight (7.6%) displayed shortness of breath, and three (2.9%) had dyspnea. The proportion of patients requiring mechanical ventilation and the rate of PICU admission were higher in the HBoV group than in the hMPV group. Conclusion The prevalence of HBoV infection is higher than that of hMPV infection in children with lower respiratory tract infection in Suzhou, and HBoV is more likely to cause severe infection than hMPV. Public health interventions for COVID-19 outbreaks have affected the prevalence of HBoV and hMPV.

The relationship between nosocomial infections and coronavirus disease 2019 in the neurosurgery unit: Clinical characteristics and outcome (preprint)

Purpose The purpose of this study was to evaluate how the “COVID-19 prevention and control measures” changed nosocomial infections in neurosurgery.Methods To explore changes in nosocomial infections in neurosurgery during the COVID-19 pandemic, the clinical data of inpatients of neurosurgery from January 1, 2020, to April 30, 2020 (COVID-19 era) were first analyzed and then compared with the same period in 2019 (pre-COVID-19 era). We also analyzed data from May 1, 2020, and December 31, 2020 (post-COVID-19 era) at the same time in 2019 (second pre-COVID-19 era).Results The nosocomial infection rate was 7.85% (54/688) in the pre-COVID-19 era and 4.30% (26/605) in the COVID-19 era (P = 0.011). Between the pre-COVID-19 and COVID-19 eras, the respiratory system infection rate was 6.1% vs. 2.0% (P < 0.001) and the urinary system was 1.7% vs. 2.0% (P = 0.837). Between the pre-COVID-19 and COVID-19 eras, the proportion of respiratory system and urinary infections in total nosocomial infections was 77.78% (42/54) vs. 46.15% (12/26) and 22.22% (12/54) vs. 46.15% (12/26), respectively, (P = 0.006). Between the second pre-COVID-19( ) and post-COVID-19 eras, the proportion of respiratory system and urinary infections in total nosocomial infections was 53.7% (44/82) vs. 40.6% (39/96) and 24.4% (20/82) vs. 40.6% (39/96), respectively, (P = 0.022).Conclusions The COVID-19 pandemic reduced the incidence of nosocomial infection in neurosurgery, and the main reduction was in respiratory infection, while the proportion of urinary infections in total nosocomial infections increased significantly.

Establishing a Credit Risk Evaluation System for SMEs Using the Soft Voting Fusion Model

In China, SMEs are facing financing difficulties, and commercial banks and financial institutions are the main financing channels for SMEs. Thus, a reasonable and efficient credit risk assessment system is important for credit markets. Based on traditional statistical methods and AI technology, a soft voting fusion model, which incorporates logistic regression, support vector machine (SVM), random forest (RF), eXtreme Gradient Boosting (XGBoost), and Light Gradient Boosting Machine (LightGBM), is constructed to improve the predictive accuracy of SMEs’ credit risk. To verify the feasibility and effectiveness of the proposed model, we use data from 123 SMEs nationwide that worked with a Chinese bank from 2016 to 2020, including financial information and default records. The results show that the accuracy of the soft voting fusion model is higher than that of a single machine learning (ML) algorithm, which provides a theoretical basis for the government to control credit risk in the future and offers important references for banks to make credit decisions.

Understanding the Complex Adoption Behavior of Cloud Services by SMEs Based on Complexity Theory: A Fuzzy Sets Qualitative Comparative Analysis (fsQCA)

To survive in a competitive environment, small and medium enterprises (SMEs) have had to adapt to the digital environment in order to adjust to customer needs globally, particularly in the post-COVID-19 world. The advantages of cloud computing (e.g., flexibility, scalability, and low entry cost) provide opportunities for SMEs with a restricted budget and limited resources. To understand how SMEs adopt cloud computing in a complex digital environment, this study examines how antecedents combine with each other to explain the high adoption of cloud computing. From the perspectives of holism and set theory, we draw on complexity and configuration theories, present a conceptual model including seven antecedents based on the technology-organization-environment framework, and conduct an asymmetric fuzzy-set qualitative comparative analysis. Through an empirical study with 123 Chinese companies, we identify nine combinations (configurations) of determinant antecedents that lead to the high adoption of cloud computing. The results show that none of the factors are indispensable to explain a high adoption on their own;instead, they are insufficient but necessary parts of the causal combinations that explain a high adoption. This study contributes to the literature on cloud computing adoption by extending current knowledge on how antecedents combine to increase the adoption and identify specific patterns of SMEs for whom these factors are essential and greatly influence their adoption.

Integrative Analysis for COVID-19 Patient Outcome Prediction (preprint)

While image analysis of chest computed tomography (CT) for COVID-19 diagnosis has been intensively studied, little work has been performed for image-based patient outcome prediction. Management of high-risk patients with early intervention is a key to lower the fatality rate of COVID-19 pneumonia, as a majority of patients recover naturally. Therefore, an accurate prediction of disease progression with baseline imaging at the time of the initial presentation can help in patient management. In lieu of only size and volume information of pulmonary abnormalities and features through deep learning based image segmentation, here we combine radiomics of lung opacities and non-imaging features from demographic data, vital signs, and laboratory findings to predict need for intensive care unit (ICU) admission. To our knowledge, this is the first study that uses holistic information of a patient including both imaging and non-imaging data for outcome prediction. The proposed methods were thoroughly evaluated on datasets separately collected from three hospitals, one in the United States, one in Iran, and another in Italy, with a total 295 patients with reverse transcription polymerase chain reaction (RT-PCR) assay positive COVID-19 pneumonia. Our experimental results demonstrate that adding non-imaging features can significantly improve the performance of prediction to achieve AUC up to 0.884 and sensitivity as high as 96.1%, which can be valuable to provide clinical decision support in managing COVID-19 patients. Our methods may also be applied to other lung diseases including but not limited to community acquired pneumonia. The source code of our work is available at https://github.com/DIAL-RPI/COVID19-ICUPrediction.

Anti-cheating Online Exams by Minimizing the Cheating Gain (preprint)

Cheating prevention in online exams is often hard and costly to tackle with proctoring, and it even sometimes involves privacy issues, especially in social distancing due to the pandemic of COVID-19. Here we propose a low-cost and privacy-preserving anti-cheating scheme by programmatically minimizing the cheating gain. A novel anti-cheating scheme we developed theoretically ensures that the cheating gain of all students can be controlled below a desired level aided by the prior knowledge of students’ abilities and a proper assignment of question sequences. Furthermore, a heuristic greedy algorithm we developed can refine an assignment of questions from a cyclic pool of question sequences to efficiently reduce the cheating gain. Compared to the integer linear programming and min-max matching methods in a small-scale simulation, our heuristic algorithm provides results close to the optimal solutions offered by the two standard discrete optimization methods. Hence, our anti-cheating approach could potentially be a cost-effective solution to the well-known cheating problem even without proctoring.

Severe Acute Respiratory Syndrome Coronavirus 2 Infection in Patients Living with Human Immunodeficiency Virus: Case Reports and Review of the Literature (preprint)

Background: Novel coronavirus pneumonia (COVID-19) outbreak caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has rapidly spread over the globe. The knowledge about SARS-CoV-2 infection in immunocompromised patients was limited. Case presentation: We presented here two human immunodeficiency virus (HIV)-infected cases with laboratory confirmed COVID-19 and clinically confirmed COVID-19, respectively. The patients both presented with fever at illness onset and patchy shadows in radiological images of lungs. Laboratory findings revealed leukopenia, lymphopenia and positive anti-HIV antibody. The younger case had a moderate course and was discharged after a 28-day hospitalization. However, the elder case with multiple comorbidities developed dyspnea and died on the fourth day after admission. Conclusions: Combining our data with two case reports, we summarize that disease course varies in HIV-infected patients with COVID-19. More attention should be paid to the management of these patients. Whether there is any difference about clinical characteristics and prognosis of COVID-19 between HIV-infected and non-HIV infected patients, remains to be further investigated.

Scheme for Cheating Prevention in Online Exams during Social Distancing (preprint)

In the current pandemic of COVID-19, students and faculty are subject to social distancing and online learning. How to test students in this unprecedented environment is a new educational challenge with immediate and global impacts. The main contribution of this paper is to establish the feasibility that by a clever design we can control the average gain (which is referred to as the g-factor) from cheating behaviors to a degree as small as pre-specified so that accurate and reliable online exams can be administered. It is underlined that even after the pandemic the methods and systems in the spirit of our proposal are still valuable for cost-effective exams to promote open courses and internet-based education.