Tracking post-ICU syndrome in critically ill patients using machine learning

Machine Learning-based Longitudinal Study of Post-ICU Syndrome Development Trajectory in Critically Ill Patients and Construction of Clinical Early Warning Models: a Research Protocol for Longitudinal Study

This study is trying to see how critically ill patients feel and function after leaving the ICU over six months, using machine learning to spot problems like memory issues and sleep disorders early on.

Quick facts

• Study type: Observational
• Enrollment: 840 (estimated)
• Ages: 18 Years to 100 Years
• Sex: All
• Sponsor: The Affiliated Hospital Of Guizhou Medical University (other)
• Locations: 1 site (Guiyang, Guizhou)
• Trial ID: NCT06427265 on ClinicalTrials.gov

What this trial studies

This project aims to monitor the development of post-ICU syndrome in critically ill patients over a period of 6 months after their transfer from the ICU. By employing a longitudinal study design, it will evaluate the occurrence of cognitive impairments, sleep disorders, and memory issues at multiple time points: 7 days, 1 month, 3 months, and 6 months post-ICU. The study will utilize latent category growth models to identify different trajectory patterns of post-ICU syndrome and apply machine learning techniques to create an early warning model for predicting these patterns.

Who should consider this trial

Good fit: Ideal candidates for this study are adults aged 18 and older who have spent at least 24 hours in the ICU and are able to communicate effectively with investigators.

Not a fit: Patients who have pre-existing cognitive impairments or serious communication barriers may not benefit from this study.

Why it matters

Potential benefit: If successful, this study could lead to early identification and intervention for patients at risk of developing post-ICU syndrome, improving their long-term recovery and quality of life.

How similar studies have performed: While there have been studies on post-ICU syndrome, the specific application of machine learning for trajectory modeling in this context is relatively novel.

Eligibility criteria

Inclusion Criteria patients :

* Length of stay in ICU ≥24h;
* Age ≥18 years old;
* Conscious when leaving ICU, communicating with investigators without barriers;
* Informed consent. Family members：
* One family member ≥18 years of age was selected for each patient;
* Assume the main role of caring for patients and medical decision-making;
* No history of mental illness or other serious organic diseases;
* Informed consent and voluntary participation in this study.

Exclusion Criteria patients :

* Have been in ICU for more than 24h within 3 months before this admission;
* Transferred to another ICU;
* Cognitive impairment existed before ICU admission (BDRS \> 4 points);
* Severe hearing impairment, dysarthria, etc., which cannot be followed up;
* Unable to complete the questionnaire survey due to serious illness.

Family members:

* Family members refuse to participate in the study due to their own reasons;
* Severe hearing and language impairment, unable to cooperate with researchers.

Where this trial is running

Guiyang, Guizhou

• Affiliated Hospital of Guizhou Medical University — Guiyang, Guizhou, China (RECRUITING)

Study contacts

• Study coordinator: Tingrui WANG

Conditions: Intensive Care Unit Syndrome, Prediction, Cognitive Impairment, Sleep Disorder, Memory Disorders, Post-icu syndrome, Prediction model, Development Trajectory