Predicting Episodes of Intracranial Hypertension in Neuro-injured Patients: Development of a Decision Algorithm Using Artificial Intelligence (PREDICT-CE)

University Hospital, Brest500 enrolled

Overview

The investigators wish to build up a database of clinical data and physiological signals with a view to developing a predictive algorithm based on continuous analysis of the intracranial pressure waveform and other parameters commonly used in intensive care to predict the occurrence of an episode of intracranial hypertension (HTIC). This algorithm will be designed using supervised learning statistical methods based on innovative statistical analysis methods (artificial intelligence). These methods are classically used to exploit massive data such as sensor data.

Study Type

OBSERVATIONAL

Enrollment

500

Conditions

Creation of a Warehouse of Clinical Data and Physiological Signals at the Patient's Bedside

Eligibility

Sex: ALLMin age: 18 YearsMax age: 99 Years

Medical Language ↔ Plain English

Inclusion Criteria: * More than 18 years * admission to intensive care for less than 3 days for a neurological lesion * Sensor placement for intracranial pressure monitoring Exclusion Criteria: * patient under judicial protection * refusal to participate * patients under 18 years of age

Locations (1)

CHU de Brest

Brest, France

Outcomes

Primary Outcomes

Creation of a database of clinical data and physiological signals to develop a predictive algorithm based on continuous analysis of the intracranial pressure waveform to predict the occurrence of an episode of intracranial hypertension.

Occurrence of an episode of intracranial hypertension defined as ICP \>20mmHg for 30 minutes

Time frame: 30 minutes

Secondary Outcomes

Predict long-term functional outcome using intracranial pressure wave signal modeling and other parameters collected in real time

Occurrence of vasospasm diagnosed by cerebral perfusion angioscanner

Time frame: 28 days