Machine Learning Model to Predict Outcome in Acute Hypoxemic Respiratory Failure (MEMORIAL)

Developing an Optimal Machine Learning Model to Predict ICU Outcome in Patients With Acute Hypoxemic Respiratory Failure

Acute hypoxemic respiratory failure (AHRF) is the most common cause of admission in the intensive care units (UCIs) worldwide. We will assess the value of machine learning (ML) techniques for early prediction of ICU death in 1,241 patients enrolled in the PANDORA (Prevalence AND Outcome of acute Respiratory fAilure) Study in Spain. The study was registered with ClinicalTrials.gov (NCT03145974). Our aim is to evaluate the minimum number of variables models using logistic regression and four supervised ML algorithms: Random Forest, Extreme Gradient Boosting, Support Vector Machine and Multilayer Perceptron.

Study Overview

• Status: Active, not recruiting
• Conditions: Acute Hypoxemic Respiratory Failure
• Intervention / Treatment: Other: machine learning analysis

Detailed Description

Acute hypoxemic respiratory failure (AHRF) is the most common cause of admission in the intensive care units (UCIs) worldwide. We will assess the value of machine learning (ML) techniques for early prediction of ICU death in AHRF patients on mechanical ventilation (MV). Few studies have investigated the prediction of mortality in patients with AHRF.

For model development, the investigators will extract data for the first 2 days after diagnosis of AHRF from patients included in the de-identified database of the PANDORA cohort. We had a database with 2,000,000 anonymized and dissociated demographics and clinical, data from 1,241 patients with AHRF enrolled in our PANDORA cohort (Prevalence AND Outcome of acute Respiratory fAilure) from 22 Spanish hospitals and coordinated by the principal investigator (JV). The investigators will follow the Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD) guidelines for model prediction. We will screen collected variables employing a genetic algorithm variable selection method to achieve parsimony. We evaluated the minimum number of variables models using logistic regression and 4 supervised ML algorithms: Random Forest, Extreme Gradient Boosting, Support Vector Machine and Multilayer Perceptron. We will use a 5-fold cross-validation in the dataset of 1,000 patients selected randomly in training data (80%) and testing data (20%). For external validation, we will use the remaining 241 patients.

• Study Type: Observational
• Enrollment (Estimated): 1241

Contacts and Locations

Study Locations

• Spain
  • Ciudad Real, Spain, 13005
    • Hospital General Universitario de Ciudad Real
  • Cuenca, Spain, 16002
    • Hospital Virgen de la Luz
  • Madrid, Spain, 28046
    • Hospital Universitario La Paz
  • Madrid, Spain, 28222
    • Hospital Universitario Puerta de Hierro
  • Murcia, Spain, 3012
    • Hospital Universitario Virgen de Arrixaca
  • Santa Cruz De Tenerife, Spain, 38010
    • Hospital Universitario NS de Candelaria
  • Valencia, Spain, 46010
    • Hospital Cinico de Valencia
  • Valladolid, Spain, 47012
    • Hospital Universitario Rio Hortega

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Adult; Older Adult
• Accepts Healthy Volunteers: No

• Sampling Method: Non-Probability Sample

Study Population

De-identified dataset inclusing 1,241 mechanically ventilated patients with acute hypoxemic respiratory failure admitted consecutively in a network of Spanish ICUs.

Description

Inclusion Criteria:

• endotracheal intubation plus mechanical ventilation (MV)
• PaO2/FiO2 ratio ≤300 mmHg under MV with positive end-expiratory pressure (PEEP) ≥5 cmH2O and FiO2 ≥0.3.

Exclusion Criteria:

• Post-operative patients ventilated <24 h
• Brain death patients.

Study Plan

How is the study designed?

Number of groups / cohorts

3

Cohorts and Interventions

Group / Cohort	Intervention / Treatment
Validation cohort; It will contain 200 patients randomly selected (20% of 1000 patients with AHRF	Other: machine learning analysis; We will use robust machine learning approaches, such as Random Forest, Extreme Gradient Boosting, Support Vector Machine and Multilayer Perceptron.; Other Names: Logistic regression, cross validation, and area under the ROC curves
Confirmatory cohort; It will contain the remaining 241 patients randomply selected (por external validation)	Other: machine learning analysis; We will use robust machine learning approaches, such as Random Forest, Extreme Gradient Boosting, Support Vector Machine and Multilayer Perceptron.; Other Names: Logistic regression, cross validation, and area under the ROC curves
Derivation cohort; It will contain 800 patients randomly selected (1,000 patients with AHRF)	Other: machine learning analysis; We will use robust machine learning approaches, such as Random Forest, Extreme Gradient Boosting, Support Vector Machine and Multilayer Perceptron.; Other Names: Logistic regression, cross validation, and area under the ROC curves

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
ICU mortality	death in the intensive care unit	up to 100 weeks (from inclusion to death or diascharge from intensive care unit

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
MV duration	duration of mechanical ventilation	up to 100 weeks (from inclusion to extubation)

Collaborators and Investigators

• Sponsor: Dr. Negrin University Hospital
• Investigators: Principal Investigator: Jesus Villar, MD, PhD, Fundación Canaria Instituto de Investigación Sanitaria de Canarias

Study record dates

Study Major Dates

• Study Start (Actual): March 19, 2024
• Primary Completion (Estimated): May 30, 2026
• Study Completion (Estimated): May 30, 2026

Study Registration Dates

• First Submitted: March 19, 2024
• First Submitted That Met QC Criteria: March 26, 2024
• First Posted (Actual): March 27, 2024

Study Record Updates

• Last Update Posted (Actual): March 25, 2025
• Last Update Submitted That Met QC Criteria: February 7, 2025
• Last Verified: February 1, 2025

More Information

Terms related to this study

• Keywords: mechanical ventilation; machine learning; outcome
• Additional Relevant MeSH Terms: Respiratory Tract Diseases; Respiration Disorders; Respiratory Insufficiency
• Other Study ID Numbers: PI24/00325

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No