A prognostic model for classification of COVID-19 severity based on clinical and laboratory testing data

This study aimed to build an machine learning based model to predict the COVID-19 severity and reveal risk factors related to COVID-19 severity based on laboratory testing and clinical data for 420 participants, using tree-based models such as XGBoost, LightGBM, random forest. We calculated the Odds Ratios (OR) to investigate whether the top-ranked features were statistically significant for severity classification, turning out that high sensitivity C-reactive protein (hs-CRP) was the most important feature for determining of COVID-19 severity and XGBoost model showed the highest performance in classifying COVID-19 severity and healthy controls with F1score (0.84) and AUC (0.87). We expect that our results are of considerable significance for early screening for diagnosing COVID-19 severity, which, in turn, assist in further retrospective research for uncommon infectious diseases. © 2023 IEEE.

Towards Predicting Supplier Resilience: A Tree-Based Model Approach

With looming uncertainties and disruptions in today's global supply chains, such as lockdown measures to contain COVID-19, supply chain resilience has gained considerable attention recently. While decision-makers in procurement have emphasized the importance of traditional risk assessment, its shortcomings can be complemented by resilience. However, while most resilience studies are too qualitative in nature and to inform supplier decisions, many quantitative resilience studies frequently rely on complex and impractical operations research models fed with simulated supplier data. Thus there is the need for an integrative, intermediate way for the practical and automated prediction of resilience with real-world data. We therefore propose a random forest-based supervised learning method to predict supplier resilience, outperforming the current human benchmark evaluation by 139 percent. The model is trained on both internal ERP data and publicly available secondary data to help assess suppliers in a pre-screening step, before deciding which supplier to select for a specific product. The results of this study are to be integrated into a software tool developed for measuring and tracking the total cost of supply chain resilience from the perspective of purchasing decisions. © 2022 IEEE Computer Society. All rights reserved.

Socio-Demographic Factors Associated With COVID-19 Vaccine Hesitancy Among Middle-Aged Adults During the Quebec's Vaccination Campaign.

Introduction: The objective of this study was to characterize the combinations of demographic and socioeconomic characteristics associated to the unwillingness to receive the COVID-19 vaccines during the 2021 Quebec's vaccination campaign. Materials and Methods: In March-June 2021, we conducted an online survey of the participants of the CARTaGENE population-based cohort, composed of middle-aged and older adults. After comparing the vaccinated and unvaccinated participants, we investigated vaccine hesitancy among participants who were unvaccinated. For identifying homogeneous groups of individuals with respect to vaccine hesitancy, we used a machine learning approach based on a hybrid tree-based model. Results: Among the 6,105 participants of the vaccine cohort, 3,553 (58.2%) had at least one dose of COVID-19 vaccine. Among the 2,552 participants, 221 (8.7%) did not want to be vaccinated (91) or were uncertain (130). The median age for the unvaccinated participants was 59.3 years [IQR 54.7-63.9]. The optimal hybrid tree-based model identified seven groups. Individuals having a household income lower than $100,000 and being born outside of Canada had the highest rate of vaccine hesitancy (28% [95% CI 19.8-36.3]). For those born in Canada, the vaccine hesitancy rate among the individuals who have a household income below $50,000 before the pandemic or are Non-retired was of 12.1% [95% CI 8.7-15.5] and 10.6% [95% CI 7.6-13.7], respectively. For the participants with a high household income before the pandemic (more than $100,000) and a low level of education, those who experienced a loss of income during the pandemic had a high level of hesitancy (19.2% [8.5-29.9]) whereas others who did not experience a loss of income had a lower level of hesitancy (6.0% [2.8-9.2]). For the other groups, the level of hesitancy was low of around 3% (3.2% [95% CI 1.9-4.4] and 3.4% [95% CI 1.5-5.2]). Discussion: Public health initiatives to tackle vaccine hesitancy should take into account these socio-economic determinants and deliver personalized messages toward people having socio-economic difficulties and/or being part of socio-cultural minorities.

Impact of Community Mobility and Weather Variability on COVID-19 Case in the Provinces of Java Island

COVID-19 is easy to transmit from one infected person to a susceptible person through droplets. Human mobility and weather variable become the factors affecting COVID-19. However, the most influence variable needs to be investigated to effectively control COVID-19 spread. This paper studied the correlation between COVID-19, community mobility and weather variability in Java Island. We used the confirmed cases of COVID-19, community mobility data and weather data from the beginning of March 2020 until the end of February 2021 in each province of Java Island. Two decision tree-based models (Random Forest and XGBoost) in four experimental setups were implemented in this paper. We found that there is similarity trend between Random Forest and XGBoost method in prediction results. The performance of both has also no significant difference. The Capital City of Jakarta, Banten and the Special Region of Yogyakarta shows the best prediction result in the third experiment which used the community mobility variable as features. While, West Java shows the best result with a combination of all weather variables and mobility, Central Java and East Java with the combination of temperature and mobility. This shows that the community mobility gives an impact on COVID-19 cases in all provinces. The correlation analysis found that the community mobility percentage change in transit stations has a significant role in predicting COVID-19 cases. Based on the model performance, the prediction of COVID-19 cases in the Capital City of Jakarta has the best result. While the Special Region of Yogyakarta has the highest error. © 2021 ACM.

Epidemiological characteristics of the COVID-19 spring outbreak in Quebec, Canada: a population-based study.

BACKGROUND: By mid-July 2020, more than 108,000 COVID-19 cases had been diagnosed in Canada with more than half in the province of Quebec. In this context, we launched a study to analyze the epidemiological characteristics and the socio-economic impact of the spring outbreak in the population. METHOD: We conducted an online survey of the participants of the CARTaGENE population-based cohort, composed of middle-aged and older adults. We collected information on socio-demographic, lifestyle, health condition, COVID-19 related symptoms and COVID-19 testing. We studied the association between these factors and two outcomes: the status of having been tested for SARS-CoV-2 and the status of having received a positive test. These associations were measured with univariate and multivariate analyses using a hybrid tree-based regression model. RESULTS: Among the 8,129 respondents from the CARTaGENE cohort, 649 were tested for COVID-19 and 41 were positive. Medical workers and individuals having a contact with a COVID-19 patient had the highest probabilities of being tested (32% and 42.4%, respectively) and of being positive (17.2% and 13.0%, respectively) among those tested. Approximately 8% of the participants declared that they have experienced at least one of the four COVID-19 related symptoms chosen by the Public Health authorities (fever, cough, dyspnea, anosmia) but were not tested. Results from the tree-based model analyses adjusted on exposure factors showed that the combination of dyspnea, dry cough and fever was highly associated with being tested whereas anosmia, fever, and headache were the most discriminant factors for having a positive test among those tested. During the spring outbreak, more than one third of the participants have experienced a decrease in access to health services. There were gender and age differences in the socio-economic and emotional impacts of the pandemic. CONCLUSION: We have shown some discrepancies between the symptoms associated with being tested and being positive. In particular, the anosmia is a major discriminant symptom for positivity whereas ear-nose-throat symptoms seem not to be COVID-19 related. The results also emphasize the need of increasing the accessibility of testing for the general population.