tetano
Editor, Senior Moderator
J Med Virol
. 2020 Sep 25.
doi: 10.1002/jmv.26551. Online ahead of print.
Development and validation of a simplified nomogram predicting individual critical illness of risk in COVID-19: a retrospective study
Ranran Xu[SUP] 1 [/SUP], Junwei Cui[SUP] 2 [/SUP], Liu Hu[SUP] 3 [/SUP], Yiru Wang[SUP] 1 [/SUP], Tao Wang[SUP] 4 [/SUP], Dawei Ye[SUP] 5 [/SUP], Yongman Lv[SUP] 1 3 [/SUP], Qingquan Liu[SUP] 1 [/SUP]
Affiliations
Abstract
Background: This study aims to screen useful predictors of critical cases among COVID-19 patients and to develop a simple-to-use nomogram for clinical utility.
Methods: A retrospective study was conducted that consisted of a primary cohort with 315 COVID-19 patients and two validation cohorts with 69 and 123 patients, respectively. Logistic regression analyses were used to identify independent risks of progression to critical. An individualized prediction model was developed, and calibration, decision curve, and clinical impact curve were used to assess the performance of the model. External validations for the predictive nomogram were also provided.
Results: The variables of age, comorbid diseases, neutrophil-to-lymphocyte ratio, D-Dimer, C-reactive protein, and platelet count were estimated to be independent predictors of progression to critical, which were incorporated to establish a model of the nomogram. It demonstrated good discrimination (with a C-index of 0.923) and calibration. Good discrimination (C-index, 0.882 and 0.906) and calibration were also noted on applying the nomogram in two validation cohorts. The clinical relevance of the nomogram was justified by the decision curve and clinical impact curve analysis.
Conclusions: This study presents an individualized prediction nomogram incorporating six clinical characteristics, which can be conveniently applied to assess an individual's risk of progressing to critical COVID-19. This article is protected by copyright. All rights reserved.
Keywords: COVID-19; Coronavirus; Critical; Nomogram; Prediction model.
. 2020 Sep 25.
doi: 10.1002/jmv.26551. Online ahead of print.
Development and validation of a simplified nomogram predicting individual critical illness of risk in COVID-19: a retrospective study
Ranran Xu[SUP] 1 [/SUP], Junwei Cui[SUP] 2 [/SUP], Liu Hu[SUP] 3 [/SUP], Yiru Wang[SUP] 1 [/SUP], Tao Wang[SUP] 4 [/SUP], Dawei Ye[SUP] 5 [/SUP], Yongman Lv[SUP] 1 3 [/SUP], Qingquan Liu[SUP] 1 [/SUP]
Affiliations
- PMID: 32975825
- DOI: 10.1002/jmv.26551
Abstract
Background: This study aims to screen useful predictors of critical cases among COVID-19 patients and to develop a simple-to-use nomogram for clinical utility.
Methods: A retrospective study was conducted that consisted of a primary cohort with 315 COVID-19 patients and two validation cohorts with 69 and 123 patients, respectively. Logistic regression analyses were used to identify independent risks of progression to critical. An individualized prediction model was developed, and calibration, decision curve, and clinical impact curve were used to assess the performance of the model. External validations for the predictive nomogram were also provided.
Results: The variables of age, comorbid diseases, neutrophil-to-lymphocyte ratio, D-Dimer, C-reactive protein, and platelet count were estimated to be independent predictors of progression to critical, which were incorporated to establish a model of the nomogram. It demonstrated good discrimination (with a C-index of 0.923) and calibration. Good discrimination (C-index, 0.882 and 0.906) and calibration were also noted on applying the nomogram in two validation cohorts. The clinical relevance of the nomogram was justified by the decision curve and clinical impact curve analysis.
Conclusions: This study presents an individualized prediction nomogram incorporating six clinical characteristics, which can be conveniently applied to assess an individual's risk of progressing to critical COVID-19. This article is protected by copyright. All rights reserved.
Keywords: COVID-19; Coronavirus; Critical; Nomogram; Prediction model.