ABSTRACT
Background. A naive Bayes classifier is a popular tool used in assigning variables an equal and independent contribution to a binary decision. With respect to COVID-19 severity, the naive Bayes classifier can consider different variables, such as age, gender, race/ethnicity, comorbidities, and initial laboratory values to determine the probability a patient may need to be admitted or transferred to an intensive care unit (ICU). The aim of this study was to develop a screening tool to detect COVID-19 patients that may require escalation to ICU status. Methods. Patients hospitalized with COVID-19 were gathered from the end of March 2020 to the end of May 2020 from four hospitals in our metropolitan area. We began searching for potential variables to include in the classification model using chi-square analysis or calculating the optimal cutpoint to separate ICU and non-ICU status. After identifying significant variables, we began using standard procedures to construct a classifier. The dataset was split 7:3 to create samples for training and testing. To appraise the model's performance, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), area under the curve (AUC), and the Matthew's correlation coefficient (MCC) were calculated. Results. A total of 574 COVID-19 patients were included in the study. There were 402 patients in the training sample and 172 patients in the testing sample. The naive Bayes classifier demonstrated an overall accuracy result of 75.6% (95% CI;68.5% - 81.8%) using the 14 variables listed in Table 1. The model was able to correctly classify 84.9% of ICU status patients (sensitivity), but only 54.7% of non-ICU status patients (specificity). The PPV and the NPV were 80.1% and 61.7%, respectively. The AUC was 0.717 (95% CI;0.629 - 0.805) and the MCC was 0.410. Conclusion. Our naive Bayes classifier operates by recognizing certain aspects of severe COVID-19 cases and looking for the probability of the variables in said patients. We present a classification model that potentially could be used alongside other tools to screen patients with COVID-19 early in their hospital course to identify those needing escalation to ICU level care.
ABSTRACT
TOPIC: Chest Infections TYPE: Original Investigations PURPOSE: The use of ascorbic acid or vitamin C in patients with COVID-19 pneumonia is a topic of significant investigation. Some argue it decreases the rate of mortality in severe cases of COVID-19 pneumonia, while others argue that it provides no additional benefits to a patient's outcome. This literature review looks at patients with COVID-19 pneumonia, and investigates the role of high dose vitamin C (HDVC) in reducing mortality, length of intensive care unit (ICU) stay, and length of hospital stay. METHODS: A thorough search of the literature was conducted to perform a meta-analysis of the available studies that compared the efficacy of HDVC to standard care or placebo in patients with severe COVID-19 pneumonia. This search was performed using PubMed, Embase, and Cochrane for data on the subject matter of study from inception to April 20, 2021. We considered randomized controlled trials, cohort studies, case-control studies, and case series. From each study, we collected the number of patients with severe COVID-19 pneumonia who underwent treatment with either HDVC or standard care/placebo. The primary outcome was the rate of mortality. Secondary outcomes were ICU length of stay and length of hospital stay. The random-effects model was used to calculate the risk ratios (RR), mean differences (MD), and confidence intervals (CIs). A p-value <0.05 was considered statistically significant. Heterogeneity was assessed using Higgins I2 index. RESULTS: Three randomized controlled trials and one retrospective cohort study involving 290 patients were including in the meta-analysis. The rate of mortality was not significantly lower in those receiving HDVC when compared to standard care/placebo (7.3% vs. 13.7%, RR 0.60, 95% CI 0.27-1.36, p= 0.22, I2= 15%). Two studies reported the length of hospital stay, which was not significantly lower in those receiving HDVC when compared to the standard care/placebo group (MD= 0.22 days, 95% CI -0.14-0.59, p= 0.23, I2= 0%). Two studies reported length of ICU stay, which was found to be significantly higher in patients receiving HDVC compared to those receiving standard care/placebo (MD= 0.38 days, 95% CI -0.01-0.75, p= 0.04, I2= 0%). CONCLUSIONS: Our meta-analysis demonstrates that the use of HDVC for patients with severe COVID-19 pneumonia does not significantly reduced the rate of mortality when compared to standard care/placebo. HDVC also does not did not significantly reduce the length of hospital stay and was in-fact found to increase the length of ICU stay. Further data should be collected to confirm our findings. CLINICAL IMPLICATIONS: The use of HDVC for patients with severe COVID-19 pneumonia is not effective in reducing mortality, and may in-fact come at the cost of longer ICU stays for patients who receive HDVC rather than standard care. DISCLOSURES: No relevant relationships by Hyder Ali, source=Web Response No relevant relationships by Modar Alom, source=Web Response No relevant relationships by Mohammed Awad, source=Web Response No relevant relationships by Azizullah Beran Beran, source=Web Response No relevant relationships by prabath herath mudiyanselage, source=Web Response No relevant relationships by Saffa Iftikhar, source=Web Response No relevant relationships by Nithin Kesireddy, source=Web Response No relevant relationships by Waleed Khokher, source=Web Response No relevant relationships by Rakin Rashid, source=Web Response
ABSTRACT
IntroductionCoronaviruses are one of the common viruses that can cause diseases for humans and animals. In December 2019, a new coronavirus (Novel Coronavirus) was detected as the cause of pneumonia and then respiratory failure. One of the modalities that have been used for treatment is prone positioning. The purpose of this meta-analysis was to investigate the efficacy of awake early prone positioning on COVID-19 patients with respiratory failure.Methods We performed a comprehensive search in the literature for studies that evaluated prone positioning in COVID-19. We searched the databases of PubMed/MEDLINE, Embase, World Health Organization COVID-19 Database, LitCOVID, and Web of Science Core Collection databases from January 1, 2020, until November 30, 2020. The search was not limited by language, study design, or country of origin. Two researchers (MA and SG) independently selected the studies;discrepancies were resolved by a third researcher (OS). We considered randomized controlled trials, cohort studies, case-control studies, and case series. We excluded animal studies, case reports, reviews, editorials, and letters to editors. We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The random-effects model was used to calculate the odds ratios (OR), and confidence intervals (CI). Our main outcomes were intubation rate and mortality rate for patients who were admitted to the hospital with hypoxic respiratory failure. Results We included 3 studies in our meta-analysis with a total of 290 patients. There was no significant difference between prone positioned awake patients compared to non-prone positioned awake patients regarding intubation rate OR 1.48 ( 95% CI: 0.751-2.927, P-value: 0.257 ) (Figure 1) and mortality rate OR 0.54 ( 95% CI: 0.219-1.326, Pvalue: 0.179) (Figure 2). Conclusion Early awake prone positioning does not have a significant impact on nonintubated COVID-19 patients with respiratory failure according to our meta-analysis. More randomized controlled trials should be done to evaluate the efficacy.