lancet-header

Preprints with The Lancet is part of SSRN´s First Look, a place where journals identify content of interest prior to publication. Authors have opted in at submission to The Lancet family of journals to post their preprints on Preprints with The Lancet. The usual SSRN checks and a Lancet-specific check for appropriateness and transparency have been applied. Preprints available here are not Lancet publications or necessarily under review with a Lancet journal. These preprints are early stage research papers that have not been peer-reviewed. The findings should not be used for clinical or public health decision making and should not be presented to a lay audience without highlighting that they are preliminary and have not been peer-reviewed. For more information on this collaboration, see the comments published in The Lancet about the trial period, and our decision to make this a permanent offering, or visit The Lancet´s FAQ page, and for any feedback please contact preprints@lancet.com.

Development and Evaluation of a Machine Learning-Based In-Hospital COVID-19 Disease Outcome Predictor (CODOP): A Multicontinental Retrospective Study

17 Pages Posted: 10 Sep 2021

See all articles by Riku Klén

Riku Klén

University of Turku - Turku PET Centre

Disha Purohit

Max Planck Institute of Experimental Medicine

Ricardo Gómez-Huelgas

University of Málaga (UMA) - Internal Medicine Department

José Manuel Casas-Rojo

Infanta Cristina Universitary Hospital - Internal Medicine Department

Juan Miguel Antón Santos

Infanta Cristina Universitary Hospital - Internal Medicine Department

Jesús Millán Núñez-Cortés

Hospital Universitario Gregorio Marañón - Internal Medicine Department

Carlos Lumbreras

Instituto de Investigación Sanitaria Hospital "12 de Octubre" - Department of Internal Medicine

José M. Ramos-Rincón

Universidad de Alicante - Department of Internal Medicine

Pablo Young

British Hospital - Clinical Medicine service

Juan Ignacio Ramírez

British Hospital - Clinical Medicine service

Estela Edith Titto Omonte

Caja Petrolera de Salud - Internal Medicine Service

Rosmery Gross Artega

Hospital San Juan de Dios - Epidemiology Unit

Magdy Teresa Canales Beltrán

Hospital Honduras Medical Centre - Instituto Hondureño of social security

Pascual Valdez

Hospital Vélez Sarsfield

Florencia Pugliese

Hospital Vélez Sarsfield

Rosa Castagna

Hospital Vélez Sarsfield

Nico Funke

Max Planck Institute of Experimental Medicine

Benjamin Leiding

Technical University Clausthal - Institute for Software and Systems Engineering

David Gomez Varela

Max Planck Institute of Experimental Medicine

More...

Abstract

Background: More contagious SARS-CoV-2 virus variants, breakthrough infections, waning immunity, and differential access to COVID-19 vaccines account for the worst yet numbers of hospitalization and deaths during the COVID-19 pandemic, particularly in resource-limited countries. There is an urgent need for clinically valuable, generalizable, and parsimonious triage tools assisting appropriate allocation of hospital resources during the pandemic. We aimed to develop and extensively validate a machine learning-based tool for accurately predicting the clinical outcome of hospitalized COVID-19 patients.

Methods: CODOP was built using modified stable iterative variable selection and linear regression with lasso regularisation. To avoid generalization problems, CODOP was trained and tested with three time-sliced and geographically distinct cohorts encompassing 40 511 blood-based analyses of COVID-19 patients from more than 110 hospitals in Spain and the USA during 2020-21. We assessed the discriminative ability of the model using the Area Under the Receiving Operative Curve (AUROC) as well as horizon and Kaplan-Meier risk stratification analyses. To reckon the fluctuating pressure levels in hospitals through the pandemic, we offer two online CODOP calculators suited for undertriage or overtriage scenarios. We challenged their generalizability and clinical utility throughout an evaluation with datasets gathered in five hospitals from three Latin American countries. 

Findings: CODOP uses 12 clinical parameters commonly measured at hospital admission and associated with the pathophysiology of COVID-19. CODOP reaches high discriminative ability up to nine days before clinical resolution (AUROC: 0·90-0·96, 95% CI 0·879-0·970), it is well calibrated, and it enables an effective dynamic risk stratification during hospitalization. The two CODOP online calculators predicted the clinical outcome of the majority of patients (73-100% sensitivity and 84-100% specificity) from the distinctive Latin American evaluation cohort.

Interpretation: The high predictive performance of CODOP in geographically disperse patient cohorts and the easiness-of-use, strongly suggest its clinical utility as a global triage tool, particularly in resource-limited countries.

Funding: The Max Planck Society.

Declaration of Interest: The authors declare no conflict of interest.

Ethical Approval: This study was approved by the Provincial Research Ethics Committee of Málaga (Spain) and the Institutional Research Ethics Committees of each participating hospital.

Keywords: COVID-19, in-hospital mortality prediction, machine learning

Suggested Citation

Klén, Riku and Purohit, Disha and Gómez-Huelgas, Ricardo and Casas-Rojo, José Manuel and Antón Santos, Juan Miguel and Núñez-Cortés, Jesús Millán and Lumbreras, Carlos and Ramos-Rincón, José M. and Young, Pablo and Ramírez, Juan Ignacio and Titto Omonte, Estela Edith and Gross Artega, Rosmery and Canales Beltrán, Magdy Teresa and Valdez, Pascual and Pugliese, Florencia and Castagna, Rosa and Funke, Nico and Leiding, Benjamin and Gomez Varela, David, Development and Evaluation of a Machine Learning-Based In-Hospital COVID-19 Disease Outcome Predictor (CODOP): A Multicontinental Retrospective Study. Available at SSRN: https://ssrn.com/abstract=3920914 or http://dx.doi.org/10.2139/ssrn.3920914

Riku Klén

University of Turku - Turku PET Centre ( email )

Turku
Finland

Disha Purohit

Max Planck Institute of Experimental Medicine ( email )

Germany

Ricardo Gómez-Huelgas

University of Málaga (UMA) - Internal Medicine Department ( email )

Málaga
Spain

José Manuel Casas-Rojo

Infanta Cristina Universitary Hospital - Internal Medicine Department ( email )

Spain

Juan Miguel Antón Santos

Infanta Cristina Universitary Hospital - Internal Medicine Department ( email )

Spain

Jesús Millán Núñez-Cortés

Hospital Universitario Gregorio Marañón - Internal Medicine Department ( email )

Madrid
Spain

Carlos Lumbreras

Instituto de Investigación Sanitaria Hospital "12 de Octubre" - Department of Internal Medicine

Madrid
Spain

José M. Ramos-Rincón

Universidad de Alicante - Department of Internal Medicine ( email )

Pintor Baeza, 11
03010 Alicante
Spain

Pablo Young

British Hospital - Clinical Medicine service ( email )

Buenos Aires
Argentina

Juan Ignacio Ramírez

British Hospital - Clinical Medicine service ( email )

Buenos Aires
Argentina

Estela Edith Titto Omonte

Caja Petrolera de Salud - Internal Medicine Service ( email )

Santa Cruz de la Sierra
Bolivia

Rosmery Gross Artega

Hospital San Juan de Dios - Epidemiology Unit ( email )

Santa Cruz
Bolivia

Magdy Teresa Canales Beltrán

Hospital Honduras Medical Centre - Instituto Hondureño of social security ( email )

Tegucigalpa
Honduras

Pascual Valdez

Hospital Vélez Sarsfield ( email )

Buenos Aires
Argentina

Florencia Pugliese

Hospital Vélez Sarsfield ( email )

Buenos Aires
Argentina

Rosa Castagna

Hospital Vélez Sarsfield ( email )

Buenos Aires
Argentina

Nico Funke

Max Planck Institute of Experimental Medicine ( email )

Germany

Benjamin Leiding

Technical University Clausthal - Institute for Software and Systems Engineering

Clausthal
Germany

David Gomez Varela (Contact Author)

Max Planck Institute of Experimental Medicine ( email )

Germany