Text Based Diagnosis of COVID-19 Using Data Mining Techniques: A Comparative Study

In the course of the recent pandemic, we have witnessed non-clinical approaches such as data mining and artificial intelligence techniques being exceedingly utilized to restrain and combat the increase of COVID-19 across the globe. The emergence of artificial intelligence in the medical field has helped in reducing the immense burden on medical systems by providing the best means for diagnosis and prognosis of COVID-19. This work attempts to analyze & evaluate superlative models on robust data resources on symptoms of COVID-19, consisting of age, gender, demographic information, pre-existing medical conditions, and symptoms experienced by patients. This study establishes paradigmatic pipeline of supervised learning algorithms coupled with feature extraction techniques and surpassing the current state-of-the-art results by achieving an accuracy of 93.360. The optimal score was found by performing feature extraction on the data using principal component analysis (PCA) followed by binary classification using the AdaBoost classifier. In addition, the present study also establishes the contribution of various symptoms in the diagnosis of the malady. © 2022 IEEE.

Biomarkers in COVID-19: Association with disease severity and outcomes

Background: In coronavirus disease 2019 (COVID-19) the need for intervention increases with disease severity and a risk prediction model that incorporates biomarkers would be beneficial for identifying patients for treatment escalation. Aim(s): To investigate biomarkers changes associated with disease severity and outcomes (mortality, thrombosis). Method(s): COVID-19 patients were sampled between April 15 and May 31 2020. Disease severity was assessed by World Health Organization (WHO) ordinal scale. 132 systemic biomarkers were investigated by routine and multiplex assays and statistical analysis performed to characterise the biomarker profile of COVID-19 patients associated with disease severity, duration, survival and thrombosis. Result(s): The study enrolled 150 COVID-19 positive adults and 16 healthy volunteers. The average age was 64 years, 59% were male, 85% had co-morbidities, 33% had a thrombotic event, and 13% died. A cross comparative analysis of biomarkers identified 13 biomarkers common to severity, mortality and thrombosis with significant correlation;including endothelial dysfunction (VWF, tPA, TFPI), hypercatabolism (low albumin, Hb, FXIII) and inflammatory response (IL-8, Osteopontin). Similarly, 14 biomarkers associated with severity and mortality included pro-inflammatory cytokines and their receptors (sTNFRII, STNFRI, sIL2a, IL6, MIP1a), neutrophils (elevated WBC, Neutrophils, TIMP1) and tissue remodelling (SCGF, EG3A). Nine biomarkers common across severity and thrombosis were angiogenesis (VEGF, LYVE1, Follistatin), acute phase response (SAP, AGP) and clot formation (Fibrinogen and PAPs). Conclusion(s): The biomarker profile associated with poorer outcomes indicates an inflammatory response, endothelial cell disruption, hypercoagulability and hypercatabolism. This study has identified several biomarkers that may be useful indicators of disease severity and progression. Further work is needed to determine how these may be used to direct clinical management. (Figure Presented).

Meditation Alleviates Psychological Distress and Sleep Quality in COVID-19 Patients during Home Isolation

Background: The COVID-19 pandemic have led to both physical and psychological outbreaks in lives of many people, especially among quarantined people. To handle this mood disorders of isolated COVID-19 patient's in home care, noninvasive OM chanting has been adopted to study its effect on stress, anxiety, depression, quality of sleep and life.Materials and Methods: 56 participants with the mean age of 41 years confirmed with COVID-19 infection under the Saveetha home care program were selected. The Experimental group practiced OM chanting for 20mins for a duration of 14 days. DASS-21 and Pittsburgh Sleep Quality Index was assessed before and after the intervention. Results: After 14 days' intervention, significant decrease in depression (13.26 +/- 4.52 to 7.84 +/- 3.96;p=0.01), anxiety (14.38 +/- 5.28 to 8.29 +/- 4.73;p=0.05) and stress (16.88 +/- 4.90 to 7.32 +/- 3.91;p=0.05) were noted. In addition to that over all sleep quality (11.24 +/- 3.89 to 6.70 +/- 3.51) and quality of life also improved among the patients after the 14 days OM chanting. Conclusion: The practice of OM chanting for two weeks showed notable reduction in the DASS-21 in comparison to the first base assessment. The increase in the Pittsburgh Sleep Quality Index was also recorded. In further, larger sample size and long-term intervention will be studied with a robust research design.

Pneumonia Classification Using Few-Shot Learning with Visual Explanations

Deep learning models have demonstrated state of the art performance in varied domains, however there is still room for improvement when it comes to learning new concepts from little data. Learning relevant features from a few training samples remains a challenge in machine learning applications. In this study, we propose an automated approach for the classification of Viral, Bacterial, and Fungal pneumonia using chest X-rays on a publicly available dataset. We employ distance learning based Siamese Networks with visual explanations for pneumonia detection. Our results demonstrate remarkable improvement in performance over conventional deep convolutional models with just a few training samples. We exhibit the powerful generalization capability of our model which once trained, effectively predicts new unseen data in the test set. Furthermore, we also illustrate the effectiveness of our model by classifying diseases from the same genus like COVID-19 and SARS. © 2022, Springer Nature Switzerland AG.

Automated detection of COVID-19 on a small dataset of chest CT images using metric learning

Coronavirus disease has caused unprecedented chaos across the globe causing potentially fatal pneumonia, since the beginning of 2020. Researchers from different communities are working in conjunction with front-line doctors and policy-makers to better understand the disease. The key to prevent the spread is a rapid diagnosis, prioritized isolation, and fastidious contact tracing. Recent studies have confirmed the presence of underlying patterns on chest CT for patients with COVID-19. We present a completely automated framework to detect COVID-19 using chest CT scans, only needing a small number of training samples. We present a few-shot learning technique based on the Triplet network in comparison to the conventional deep learning techniques which require a substantial amount of training examples. We used 140 chest CT images for training and the rest for testing from a total of 2482 images for both COVID-19 and non-COVID-19 cases from a publicly available dataset. The model trained with chest CT images achieves an AUC of 0.94, separates the two classes into distinct clusters;thereby giving correct prediction accuracy on the evaluation dataset.

Automated Diagnosis of COVID-19 from CT Scans Based on Concatenation of Mobilenetv2 and ResNet50 Features

Timely and precise identification of COVID19 is an arduous task due to the shortage and the inefficiency of the medical test kits. As a result of which medical professionals have turned their attention towards radiological images like Computed Tomography (CT) scans. There have been continued attempts on creating deep learning models to detect COVID-19 using CT scans. This has certainly reduced the manual intervention in disease detection but the reported detection accuracy is limited. Motivated by this, in the present work, an automatic system for COVID-19 diagnosis is proposed using a concatenation of the Mobilenetv2 and ResNet50 features. Typically, the features from the last convolution layer of the transfer learned Mobilenetv2, and the last average pooling layer of the learned ResNet50 are fused to improve the classification accuracy. The fused feature vector along with the corresponding labels is used to train an SVM classifier to give the output. The proposed technique is validated on the benchmark COVID CT dataset comprising of a total of 2482 images with 1252 positive and 1230 negative cases. The experimental results reveal that the proposed feature fusion strategy achieves a validation accuracy of 98.35%, F1-score of 98.39%, the precision of 99.19%, and a recall of 97.60% for detecting COVID-19 cases with 80% training and 20% validation scheme. The obtained results are better than the comparison models and the existing state of artworks reported in the literature. © 2021, Springer Nature Singapore Pte Ltd.

Empiric antibiotic therapy and community-onset bacterial co-infection in patients hospitalized with COVID-19: A multi-hospital cohort study

Background: Antibiotic therapy has no known benefit against COVID-19, but is often initiated out of concern for concomitant bacterial infection. We sought to determine how common early empiric antibiotic therapy and community-onset bacterial co-infections are in hospitalized patients with COVID-19. Methods: In this multi-center cohort study of hospitalized patients with COVID-19 discharged from 32 Michigan hospitals during the COVID-19 Michigan surge, we describe the use of early empiric antibiotic therapy (within the first two days) and prevalence of community-onset bacterial co-infection. Additionally, we assessed patient and hospital predictors of early empiric antibiotic using poison generalized estimating equation models. Results: Between 3/10/2020 and 5/10/2020, data were collected on 951 COVID-19 PCR positive patients. Patient characteristics are described in Table 1. Nearly two thirds (62.4%, 593/951) of COVID-19 positive patients were prescribed early empiric antibiotic therapy, most of which (66.2%, 393/593) was directed at community-acquired pathogens. Across hospitals, the proportion of COVID-19 patients prescribed early empiric antibiotics varied from 40% to 90% (Figure 1). On multivariable analysis, patients were more likely to receive early empiric antibiotic therapy if they were older (adjusted rate ratio [ARR]: 1.01 [1.00-1.01] per year), required respiratory support (e.g., low flow oxygen, ARR: 1.16 [1.04-1.29]), had signs of a bacterial infection (e.g., lobar infiltrate, ARR: 1.17 [1.02-1.34]), or were admitted to a for-profit hospital (ARR: 1.27 [1.11-1.45]);patients admitted later were less likely to receive empiric antibiotics (April vs. March, ARR: 0.72 [0.62-0.84], Table 2). Community-onset bacterial co-infections were identified in 4.5% (43/951) of COVID-19 positive patients (2.4% [23/951] positive blood culture;1.9% [18/951] positive respiratory culture). Conclusion: Despite low prevalence of community-onset bacterial co-infections, patients hospitalized with COVID-19 often received early empiric antibiotic therapy. Given the potential harms from unnecessary antibiotic use, including additional personal protective equipment to administer antibiotics, judicious antibiotic use is key in hospitalized patients with COVID-19. (Figure Presented).