Early Prediction of ICU Mortality in Patients with Acute Hypoxemic Respiratory Failure Using Machine Learning: The MEMORIAL Study

<b>Background</b>: Early prediction of ICU death in acute hypoxemic respiratory failure (AHRF) could inform clinicians for targeting therapies to reduce harm and increase survival. We sought to determine clinical modifiable and non-modifiable features during the first 24 h of AHRF associated with ICU death. <b>Methods</b>: This is a development, testing, and validation study using data from a prospective, multicenter, nation-based, observational cohort of 1241 patients with AHRF (defined as PaO<sub>2</sub>/FiO<sub>2</sub> ≤ 300 mmHg on mechanical ventilation [MV] with positive end-expiratory pressure [PEEP] ≥ 5 cmH<sub>2</sub>O and FiO<sub>2</sub> ≥ 0.3) from any etiology. Using relevant features captured at AHRF diagnosis and within 24 h, we developed a logistic regression model following variable selection by genetic algorithm and machine learning (ML) approaches. <b>Results</b>: We analyzed 1193 patients, after excluding 48 patients with no data at 24 h after AHRF diagnosis. Using repeated random sampling, we selected 75% (n = 900) for model development and testing, and 25% (n = 293) for final validation. Risk modeling identified six major predictors of ICU death, including patient's age, and values at 24 h of PEEP, FiO<sub>2</sub>, plateau pressure, tidal volume, and number of extrapulmonary organ failures. Performance with ML methods was similar to logistic regression and achieved a high area under the receiver operating characteristic curve (AUROC) of 0.88, 95%CI 0.86-0.90. Validation confirmed adequate model performance (AUROC 0.83, 95%CI 0.78-0.88). <b>Conclusions</b>: ML and traditional methods led to an encouraging model to predict ICU death in ventilated AHRF as early as 24 h after diagnosis. More research is needed to identify modifiable factors to prevent ICU deaths.

Authors

Institutions

Publication Details

Journal
Journal of Clinical Medicine
Published
2025-03-04
DOI
https://doi.org/10.3390/jcm14051711
Citations
1
Primary Topic
Respiratory Support and Mechanisms
Type
article
Field-Weighted Citation Impact
1.52
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Early Prediction of ICU Mortality in Patients with Acute Hypoxemic Respiratory Failure Using Machine Learning: The MEMORIAL Study

Tamás Szakmány, David Pestaña, Lorena Fernández de la Cruz, José M. Añón et al.
1 citations
Journal of Clinical Medicine
Respiratory Support and Mechanisms
1.52
article

Early Prediction of ICU Mortality in Patients with Acute Hypoxemic Respiratory Failure Using Machine Learning: The MEMORIAL Study

Tamás Szakmány, David Pestaña, Lorena Fernández de la Cruz, José M. Añón, Jesús M. González‐Martín, Cristina Fernández, Raquel Montiel, Juan M. Mora-Ordóñez, Carlos Ferrando, Isabel Murcia, Alfonso Ambrós, Ánxela Vidal, Ewout W. Steyerberg, Estrella Gómez-Bentolila, Jesús Villar, Miguel A. Romera, Domingo Martínez, Ana M. Domínguez-Berrot, Juan A. Soler, Fernando Mosteiro
article en
1 citations

Abstract

<b>Background</b>: Early prediction of ICU death in acute hypoxemic respiratory failure (AHRF) could inform clinicians for targeting therapies to reduce harm and increase survival. We sought to determine clinical modifiable and non-modifiable features during the first 24 h of AHRF associated with ICU death. <b>Methods</b>: This is a development, testing, and validation study using data from a prospective, multicenter, nation-based, observational cohort of 1241 patients with AHRF (defined as PaO<sub>2</sub>/FiO<sub>2</sub> ≤ 300 mmHg on mechanical ventilation [MV] with positive end-expiratory pressure [PEEP] ≥ 5 cmH<sub>2</sub>O and FiO<sub>2</sub> ≥ 0.3) from any etiology. Using relevant features captured at AHRF diagnosis and within 24 h, we developed a logistic regression model following variable selection by genetic algorithm and machine learning (ML) approaches. <b>Results</b>: We analyzed 1193 patients, after excluding 48 patients with no data at 24 h after AHRF diagnosis. Using repeated random sampling, we selected 75% (n = 900) for model development and testing, and 25% (n = 293) for final validation. Risk modeling identified six major predictors of ICU death, including patient's age, and values at 24 h of PEEP, FiO<sub>2</sub>, plateau pressure, tidal volume, and number of extrapulmonary organ failures. Performance with ML methods was similar to logistic regression and achieved a high area under the receiver operating characteristic curve (AUROC) of 0.88, 95%CI 0.86-0.90. Validation confirmed adequate model performance (AUROC 0.83, 95%CI 0.78-0.88). <b>Conclusions</b>: ML and traditional methods led to an encouraging model to predict ICU death in ventilated AHRF as early as 24 h after diagnosis. More research is needed to identify modifiable factors to prevent ICU deaths.

Journal of Clinical MedicineVol. 14(5)
St. Michael's Hospital (CA), Leiden University Medical Center (NL), Hospital Universitario La Paz (ES), Instituto de Salud Carlos III (ES), University Medical Center Utrecht (NL), Hospital Universitario Nuestra Señora de Candelaria (ES), Hospital Universitario de Gran Canaria Doctor Negrín (ES), Hospital Universitario Puerta de Hierro Majadahonda (ES), Hospital Regional Universitario de Málaga (ES), Hospital Clínic de Barcelona (ES), Hospital General Universitario de Ciudad Real (ES), Hospital Universitario Fundación Jiménez Díaz (ES), Complexo Hospitalario Universitario A Coruña (ES), Hospital Universitario Virgen de la Arrixaca (ES), Complejo Hospitalario Universitario de Albacete (ES), Hospital Universitario Río Hortega (ES), Consorci Institut D'Investigacions Biomediques August Pi I Sunyer (ES), Hospital Universitario Ramón y Cajal (ES), Complejo Asistencial Universitario de Palencia (ES), Cardiff University (GB)
Good health and well-being
Openalex Percentile: Top 21%
Respiratory Support and Mechanisms
1.52
DOIOpen Access PDFOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.