ABSTRACT
In this study, we aimed to identify the factors that were associated with mortality among continuing care residents in Alberta, during coronavirus disease 2019 (COVID-19) pandemic. Then, we examined pre-processing methods in terms of prediction performance. Finally, we developed several machine learning models and compared the results of these models in terms of performance. We conducted a retrospective cohort study of all continuing care residents in Alberta, Canada, from March 1, 2020, to March 31, 2021. We used a univariate and a multivariate logistic regression (LR) model to identify predictive factors of 60-day mortality by estimating odds ratios (ORs) with a 95% of a confidence interval. To determine the best sensitivity-specificity cut-off point, the Youden index was employed. We examined the pre-processing methods and then developed several machine learning models to acknowledge the best model regarding performance. In this cohort study, increased age, male sex, symptoms, previous admissions, and some specific comorbidities were associated with mortality. Machine learning and pre-processing approaches offer an assuring method for improving risk prediction for mortality, but more work is needed to show improvement beyond standard risk factors.
Subject(s)
COVID-19ABSTRACT
Background Decolonization is an important infection prevention and control strategy in the surgical context. Preoperative decolonization of hip and knee replacement patients reduces the incidence of surgical site infections (SSIs), but the implementation of decolonization protocols has been uneven. Understanding the clinical level barriers and facilitators that affect implementation may increase the effectiveness of preoperative decolonization strategies. Methods Leveraging ongoing quality improvement (QI) activity to reduce SSIs amongst hip and knee replacement patients in Alberta, Canada, qualitative methods were deployed. Semi-structured interviews (n=2) were conducted with surgeons, and focus groups (n=9) were conducted with seven nurses and two administrative staff to understand barriers and facilitators to the implementation of a provincial decolonization strategy. Interview questions were developed in conjunction with the Theoretical Domains Framework (TDF) and the research team. An inductive analysis derived from a Grounded Theory (GT) approach was conducted with the assistance of NVivo software. Results Knowledge and understanding of the decolonization strategy were central to implementation. When present, they acted as facilitators, but when absent or inconsistent, they were significant barriers to implementation. Specifically, clinics needed more knowledge and direction on how to deliver the decolonization strategy to patients receiving homecare; who had repeat surgeries; who required surgery during COVID-19 outbreaks. Conclusions Knowledge and understanding was a core category which summarized seven subcategories found within our GT analysis. A successful decolonization strategy will benefit from adopting further planning and development for specific patients and respiratory outbreaks such as COVID-19. Further aspects that may act as facilitators include having a champion in each clinic, regular reporting, and audit and feedback strategies. Findings from our study can provide information on the barriers and facilitators of a decolonization strategy and can be used in making the decolonization strategy successful.
Subject(s)
COVID-19ABSTRACT
Background: Pneumonia from SARS-CoV-2 is difficult to distinguish from other viral and bacterial etiologies. Broad-spectrum antimicrobials are frequently prescribed to patients hospitalized with COVID-19 which potentially acts as a catalyst for the development of antimicrobial resistance (AMR). ObjectivesWe conducted a systematic review and meta-analysis during the first 18 months of the pandemic to quantify the prevalence and types of resistant co-infecting organisms in patients with COVID-19 and explore differences across hospital and geographic settings.MethodsWe searched MEDLINE, Embase, Web of Science (BioSIS), and Scopus from November 1, 2019 to May 28, 2021 to identify relevant articles pertaining to resistant co-infections in patients with laboratory confirmed SARS-CoV-2. Patient- and study-level analyses were conducted. We calculated pooled prevalence estimates of co-infection with resistant bacterial or fungal organisms using random effects models. Stratified meta-analysis by hospital and geographic setting was also performed to elucidate any differences. ResultsOf 1331 articles identified, 38 met inclusion criteria. A total of 1959 unique isolates were identified with 29% (569) resistant organisms identified. Co-infection with resistant bacterial or fungal organisms ranged from 0.2 to 100% among included studies. Pooled prevalence of co-infection with resistant bacterial and fungal organisms was 24% (95% CI: 8-40%; n=25 studies: I 2 =99%) and 0.3% (95% CI: 0.1-0.6%; n=8 studies: I 2 =78%), respectively. Among multi-drug resistant organisms, methicillin-resistant Staphylococcus aureus, carbapenem-resistant Acinetobacter baumannii, Klebsiella pneumoniae, Pseudomonas aeruginosa and Candida auris were most commonly reported. Stratified analyses found higher proportions of AMR outside of Europe and in ICU settings, though these results were not statistically significant. Patient-level analysis demonstrated >50% (n=58) mortality, whereby all but 6 patients were infected with a resistant organism. ConclusionsDuring the first 18 months of the pandemic, AMR was moderately prevalent in COVID-19 patients and varied by hospital and geography although there was substantial heterogeneity. Given the variation in patient populations within these studies, clinical settings, practice patterns, and definitions of AMR, further research is warranted to quantify AMR in COVID-19 patients to improve surveillance programs, infection prevention and control practices and antimicrobial stewardship programs globally.
Subject(s)
COVID-19 , Klebsiella InfectionsABSTRACT
Background: Pneumonia from SARS-CoV-2 is difficult to distinguish from other viral and bacterial etiologies. Broad-spectrum antimicrobials are frequently prescribed to patients hospitalized with COVID-19 which potentially acts as a catalyst for the development of antimicrobial resistance (AMR).Objectives: We conducted a systematic review and meta-analysis during the first 18 months of the pandemic to quantify the prevalence and types of resistant co-infecting organisms in patients with COVID-19 and explore differences across hospital and geographic settings.Methods: We searched MEDLINE, Embase, Web of Science (BioSIS), and Scopus from November 1, 2019 to May 28, 2021 to identify relevant articles pertaining to resistant co-infections in patients with laboratory confirmed SARS-CoV-2. Patient- and study-level analyses were conducted. We calculated pooled prevalence estimates of co-infection with resistant bacterial or fungal organisms using random effects models. Stratified meta-analysis by hospital and geographic setting was also performed to elucidate any differences.Results: Of 1331 articles identified, 38 met inclusion criteria. A total of 1959 unique isolates were identified with 29% (569) resistant organisms identified. Co-infection with resistant bacterial or fungal organisms ranged from 0·2 to 100% among included studies. Pooled prevalence of co-infection with resistant bacterial and fungal organisms was 24% (95% CI: 8-40%; n=25 studies: I 2 =99%) and 0·3% (95% CI: 0·1-0·6%; n=8 studies: I 2 =78%), respectively. Among multi-drug resistant organisms, methicillin-resistant Staphylococcus aureus, carbapenem-resistant Acinetobacter baumannii, Klebsiella pneumoniae, Pseudomonas aeruginosa and Candida auris were most commonly reported. Stratified analyses found higher proportions of AMR outside of Europe and in ICU settings, though these results were not statistically significant. Patient-level analysis demonstrated >50% (n=58) mortality, whereby all but 6 patients were infected with a resistant organism.Conclusions: During the first 18 months of the pandemic, AMR was moderately prevalent in COVID-19 patients and varied by hospital and geography although there was substantial heterogeneity. Given the variation in patient populations within these studies, clinical settings, practice patterns, and definitions of AMR, further research is warranted to quantify AMR in COVID-19 patients to improve surveillance programs, infection prevention and control practices and antimicrobial stewardship programs globally.Funding: The Antimicrobial Resistance - One Health Consortium is funded through the Major Innovation Fund Program of the Ministry of Jobs, Economy, and Innovation (JEI), Government of Alberta, Canada.Declaration of Interest: We declare no competing interests.