AiCR : Artificial Intelligence in Cardiac aRrest

NCT04462380 | Unknown

• 1 other identifier: 20reamed01
• Study Type: observational
• Target: 500
• Locations: 1 country
• Sites: 1

Brief Summary

The overall incidence of cardiorespiratory arrest in Europe is estimated at 350,000 to 700,000 cases per year. Survival rate is estimated at 10.7% for all rhythm disorders combined. Several examples of AI application in the medical field exist. Ting et al have developed a computer tool capable of diagnosing the presence of diabetic retinopathy with excellent power. In resuscitation, Celi et al proposed a tool capable of predicting the need for crystalloid vascular filling during a systemic inflammatory state. In Nature in 2018, Komorowski demonstrated the efficacy of AI in the hemodynamic management of sepsis. In a study of the renal response to fluid challenge, Zhang et al. demonstrate the effectiveness of the learning machine. Objectives: Determination of an algorithm capable of predicting the mortality of patients admitted to intensive care units (ICU) for ACR from hospitalization reports (CRH). Also use of the algorithm to predict the risk of recurrence of the arrest, the duration of mechanical ventilation, the appearance of sepsis, the development of organ failure, prediction of the CPC (Cerebral Performance Category), time to obtain catecholamine withdrawal, the appearance of acute renal failure with or without the need for extra-renal purification (EER) and duration under EER, the average length of stay. This project is part of a larger, nationwide project with greater power, and includes all the data generated during hospitalization in intensive care. Method: an estimated total number of patients included in this study to be between 300 and 500. The population will come from the intensive care units of Nice, Antibes, Cannes, Grasse. Inclusion will be retrospective, on CRH, CR of CT imaging (cerebral and thoraco-abdomino-pelvic), MRI, EEG, and daily follow-up words, from 2014 to the end of 2020. After anonymisation, application of semantisation using natural language processing (NLP) methods. The data to be extracted are entered in a document written by intensive care physicians. These data will then be stored in a database. In order to meet the main objective, we will develop a computer algorithm capable of predicting mortality in the study population. This algorithm, based on a large database, can be designed using machine learning or even deep learning techniques depending on the amount of data to be processed.

Conditions

Cardio Respiratory Arrest

Outcome Measures

Primary Outcomes (2)

• Prediction of mortality in the intensive care unit

  Definition of a semantic reporting tool, automated, transition from an anonymized report to an operational and relevant database.

  1day

• Prediction of mortality in the intensive care unit

  Use of the database thus created to create an intelligent mortality prediction algorithm. Use also on secondary judgment criteria in order to predict other parameters mentioned below.

  1day

Eligibility Criteria

• Age: 18 Years+
• Sex: all
• Healthy Volunteers: No
• Age Groups: Adult (18-64), Older Adult (65+)
• Sampling Method: Non-Probability Sample

Study Population

ACR for all causes, admitted in intensive care or intensive care, from the hospital centres of Nice, Antibes, Cannes, Grasse, etc.

You may qualify if:

• OCA recovered from: hypoxic, ischemic, pulmonary embolism, tamponade, rhythm or conduction disorder, shockable or not, intra or extra-hospital.
• CR computerized, typed in PDF format

Sponsors & Collaborators

• Centre Hospitalier Universitaire de Nice: lead
• Institut National de Recherche en Informatique et en Automatique: collaborator
• AIINTENSE: collaborator

Study Sites (1)

Nice Hospital

Nice, 06200, France

RECRUITING

MeSH Terms

Conditions

Heart Arrest

Condition Hierarchy (Ancestors)

Heart Diseases; Cardiovascular Diseases

Central Study Contacts

Jean DELLAMONICA

CONTACT

33 4 920 35 510; dellamonica.j@chu-nice.fr

romain LOMBARDI

CONTACT

lombardi.r@chu-nice.fr

Study Design

• Study Type: observational
• Observational Model: CASE ONLY
• Time Perspective: RETROSPECTIVE
• Sponsor Type: OTHER
• Responsible Party: SPONSOR

Study Record Dates

• First Submitted: July 3, 2020
• First Posted: July 8, 2020
• Study Start: February 1, 2020
• Primary Completion: December 31, 2020
• Study Completion: December 31, 2020
• Last Updated: July 8, 2020

Record last verified: 2020-07

Locations

FR (1)