Clipboard, Search History, and several other advanced features are temporarily unavailable. Skip to main page content
The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.
The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation Search: Search Advanced Clipboard User Guide Save Email Send to
Clipboard
My Bibliography
Collections
Citation manager
Display options Display options Format Abstract PubMed PMID
Email address has not been verified. Go to My NCBI account settings to confirm your email and then refresh this page. To: Subject: Body: Format: Summary Summary (text) Abstract Abstract (text) MeSH and other data Send email Cancel
Add to Collections
Create a new collection
Add to an existing collection
Name your collection: Name must be less than 100 characters Choose a collection: Unable to load your collection due to an error Please try again Add Cancel
Add to My Bibliography
My Bibliography
Unable to load your delegates due to an error Please try again Add Cancel
Your saved search
Name of saved search: Search terms: Test search terms Would you like email updates of new search results? Saved Search Alert Radio Buttons
Yes
No
Email: (change) Frequency: Monthly Weekly Daily Which day? The first Sunday The first Monday The first Tuesday The first Wednesday The first Thursday The first Friday The first Saturday The first day The first weekday Which day? Sunday Monday Tuesday Wednesday Thursday Friday Saturday Report format: Summary Summary (text) Abstract Abstract (text) PubMed Send at most: 1 item 5 items 10 items 20 items 50 items 100 items 200 items Send even when there aren't any new results Optional text in email: Save Cancel
Create a file for external citation management software
Create file Cancel
Your RSS Feed
Name of RSS Feed: Number of items displayed: 5 10 15 20 50 100 Create RSS Cancel RSS Link Copy
Full text links
Nature Publishing Group Free PMC article Full text links
Actions
CiteCollectionsAdd to Collections
Create a new collection
Add to an existing collection
Name your collection: Name must be less than 100 characters Choose a collection: Unable to load your collection due to an errorPlease try again Add Cancel PermalinkPermalinkCopyDisplay options Display options Format AbstractPubMedPMID
Page navigation
Title & authors
Abstract
Conflict of interest statement
Figures
References
Publication types
MeSH terms
LinkOut - more resources
Title & authors Abstract Conflict of interest statement Figures References Publication types MeSH terms LinkOut - more resources Full text links CiteDisplay options Display options Format AbstractPubMedPMID
Abstract
Soaring cases of coronavirus disease (COVID-19) are pummeling the global health system. Overwhelmed health facilities have endeavored to mitigate the pandemic, but mortality of COVID-19 continues to increase. Here, we present a mortality risk prediction model for COVID-19 (MRPMC) that uses patients' clinical data on admission to stratify patients by mortality risk, which enables prediction of physiological deterioration and death up to 20 days in advance. This ensemble model is built using four machine learning methods including Logistic Regression, Support Vector Machine, Gradient Boosted Decision Tree, and Neural Network. We validate MRPMC in an internal validation cohort and two external validation cohorts, where it achieves an AUC of 0.9621 (95% CI: 0.9464-0.9778), 0.9760 (0.9613-0.9906), and 0.9246 (0.8763-0.9729), respectively. This model enables expeditious and accurate mortality risk stratification of patients with COVID-19, and potentially facilitates more responsive health systems that are conducive to high risk COVID-19 patients.
PubMed Disclaimer
Conflict of interest statement
The authors declare no competing interests.
Figures
Fig. 1. Study design.
MRPMC mortality risk…
Fig. 1. Study design.
MRPMC mortality risk prediction model for COVID-19, SFT training cohort of…
Fig. 1. Study design. MRPMC mortality risk prediction model for COVID-19, SFT training cohort of Sino-French New City Campus of Tongji Hospital, SFV internal validation cohort of Sino-French New City Campus of Tongji Hospital, OV Optical Valley Campus of Tongji Hospital, CHWH The Central Hospital of Wuhan.
Fig. 2. Feature selection by LASSO.
a …
Fig. 2. Feature selection by LASSO.
a LASSO variable trace profiles of the 34 features… Fig. 2. Feature selection by LASSO. a LASSO variable trace profiles of the 34 features whose intracohort missing rates were less than 5%. The vertical dashed line shows the best lambda value 0.014 chosen by tenfold cross validation. b Feature coefficient of LASSO with best lambda value 0.014. High-risk (positive coefficient) and low-risk (negative coefficient) features are colored in red and blue, respectively. Gray features with coefficient 0 were considered redundant and removed, resulting in 14 features left for downstream prognosis modeling. LASSO least absolute shrinkage and selection operator, BUN blood urea nitrogen, RR respiratory rate, COPD chronic obstructive pulmonary disease, Hb hemoglobin, WB, white blood cell count, Cr creatinine, GGT gamma-glutamyl transferase, TB total bilirubin, AST aspartate aminotransferase, ALT alanine transaminase, MAP mean arterial pressure, ALB albumin, SpO2 oxygen saturation, CKD chronic kidney disease.
Fig. 3. Predictive performance of models across…
Fig. 3. Predictive performance of models across cohorts.
AUC to assess the performance of mortality…
Fig. 3. Predictive performance of models across cohorts. AUC to assess the performance of mortality risk prediction of models (LR, SVM, GBDT, NN, and MRPMC) in a SFV cohort, b OV cohort, and c CHWH cohort, respectively. Source data are provided as a Source Data file. Kaplan–Meier curves indicating overall survival of patients with high and low mortality risk in d SFV cohort, e OV cohort, and f CHWH cohort, respectively. The tick marks refer to censored patients. The dark red or blue line indicates the survival probability, and the light red or blue areas represent the 95% confidence interval of survival probability (p < 0.0001). AUC area under the receiver operating characteristics curve, SFV internal validation cohort of Sino-French New City Campus of Tongji Hospital, OV Optical Valley Campus of Tongji Hospital, CHWH The Central Hospital of Wuhan, LR logistic regression, SVM support vector machine, GBDT gradient boosted decision tree, NN neural network, MRPMC mortality risk prediction model for COVID-19.
Fig. 4. Statistical analysis of features included…
Fig. 4. Statistical analysis of features included in models.
a Heatmap representing the correlation between… Fig. 4. Statistical analysis of features included in models. a Heatmap representing the correlation between continuous features included in MRPMC using Spearman’s correlation coefficient. The colors in the plot represent the correlation coefficients. The redder the color, the stronger the positive monotonic relationship. The bluer the color, the stronger the negative monotonic relationship. The size of the circle represents the absolute value of the correlation coefficient, where a larger circle represents a stronger correlation. The numbers in the lower triangle represent the value of correlation coefficient. b Scaled importance rank of all features included in MRPMC for identifying high mortality risk COVID-19 patients included in the models. The size of circles represents the value of relative importance. The different color of circles represents the feature importance in different models. c Box and jitter plots showing distribution of continuous features included in MRPMC between deceased patients (n = 254) and discharged patients (n = 1906). The center line represents the median of the feature. Box limits represent upper and lower quartiles. Whiskers represent 1.5 times interquartile range. Gray points represent outliers. The median [IQR] of the features shown in Fig. 4c were listed in Supplementary Table 4. Wilcoxon test was used in the univariate comparison between groups and a two-tailed p < 0.05 was considered as statistically significant. ***p < 0.001. Source data are provided as a Source Data file. MRPMC mortality risk prediction model for COVID-19, ALB albumin, SpO2 oxygen saturation, BUN blood urea nitrogen, RR respiratory rate, LYM lymphocyte count, PLT platelet count, No. comorbidities number of comorbidities, CKD chronic kidney disease, IQR interquartile range. See this image and copyright information in PMC
References
Katz, J. N. et al. Disruptive modifications to cardiac critical care delivery during the Covid-19 pandemic: an international perspective. J Am Coll Cardiol.10.1016/j.jacc.2020.04.029 (2020). - PMC - PubMed
World Health Organization. Coronavirus 2019 (COVID-19) (World Health Organization, 2020). https://covid19.who.int/.
Phelan AL, Katz R, Gostin LO. The novel coronavirus originating in Wuhan, China: challenges for global health governance. J. Am. Med. Assoc. 2020;323:709–710. doi: 10.1001/jama.2020.1097. - DOI - PubMed
Guan WJ, et al. Clinical characteristics of coronavirus disease 2019 in China. N. Engl. J. Med. 2020;382:1708–1720. doi: 10.1056/NEJMoa2002032. - DOI - PMC - PubMed
Huang C, et al. Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China. Lancet. 2020;395:497–506. doi: 10.1016/S0140-6736(20)30183-5. - DOI - PMC - PubMed
Show all 39 references
Publication types
Research Support, Non-U.S. Gov't Actions
Search in PubMed
Search in MeSH
Add to Search
MeSH terms
Aged Actions
Search in PubMed
Search in MeSH
Add to Search
Betacoronavirus Actions
Search in PubMed
Search in MeSH
Add to Search
COVID-19 Actions
Search in PubMed
Search in MeSH
Add to Search
China / epidemiology Actions
Search in PubMed
Search in MeSH
Add to Search
Coronavirus Infections / mortality* Actions
Search in PubMed
Search in MeSH
Add to Search
Female Actions
Search in PubMed
Search in MeSH
Add to Search
Humans Actions
Search in PubMed
Search in MeSH
Add to Search
Logistic Models Actions
Search in PubMed
Search in MeSH
Add to Search
Machine Learning* Actions
Search in PubMed
Search in MeSH
Add to Search
Male Actions
Search in PubMed
Search in MeSH
Add to Search
Middle Aged Actions
Search in PubMed
Search in MeSH
Add to Search
Neural Networks, Computer Actions
Search in PubMed
Search in MeSH
Add to Search
Pandemics* Actions
Search in PubMed
Search in MeSH
Add to Search
Pneumonia, Viral / mortality* Actions
Search in PubMed
Search in MeSH
Add to Search
Risk Assessment Actions
Search in PubMed
Search in MeSH
Add to Search
SARS-CoV-2 Actions
Search in PubMed
Search in MeSH
Add to Search
Support Vector Machine Actions
Search in PubMed
Search in MeSH
Add to Search
LinkOut - more resources
Full Text Sources
Europe PubMed Central
Nature Publishing Group
PubMed Central
Other Literature Sources
The Lens - Patent Citations Database
Full text links [x] Nature Publishing Group Free PMC article [x] Cite Copy Download .nbib.nbib Format: AMA APA MLA NLM Send To
Clipboard
Email
Save
My Bibliography
Collections
Citation Manager
[x]
NCBI Literature Resources
MeSH PMC Bookshelf Disclaimer
The PubMed wordmark and PubMed logo are registered trademarks of the U.S. Department of Health and Human Services (HHS). Unauthorized use of these marks is strictly prohibited.
Search: Search Advanced Clipboard Save Email Send to
Nature Publishing Group Free PMC article Full text links 
