Different Algorithm Models to Predict Postoperative Pulmonary Complications in Elderly Patients

Huazhong University of Science and Technology

Status: Not yet enrolling

Conditions: Postoperative Pulmonary Complications

Study type: Observational

Funder types: Other

Identifiers

NCT05671939

PPC20221201

Details and patient eligibility

About

Although a number of clinical predictive models were developed to predict postoperative pulmonary complications, few predictive models have been used in elderly patients. In this study, the researchers aim to compare different algorithms to predict postoperative pulmonary complications in elderly patients and to assess the risk of postoperative pulmonary complications in elderly patients.

Full description

Postoperative pulmonary complications occur frequently, which is an important cause of death and morbidity. Age has been an important predictor of postoperative pulmonary complications. As the world's population ages, more and more older people are undergoing surgery as indications for surgery expand. In order to better assess the risk of postoperative pulmonary complications in elderly patients, we plan to use database information and different algorithms such as logistic regression, random forest, and other algorithms to build models respectively and evaluate the effects of the models.

Enrollment: 10,000 estimated patients

Sex: All

Ages: 65+ years old

Volunteers: No Healthy Volunteers

Inclusion criteria

1. Age 65 years or older
2. receiving invasive ventilation during general anesthesia for surgery

Exclusion criteria

1. preoperative mechanical ventilation
2. procedures related to a previous surgical complication
3. a second operation after surgery
4. organ transplantation
5. discharged within 24 hours after surgery
6. cardiac surgery

Trial design

10,000 participants in 2 patient groups

Training set

Description:

The whole cohort is randomly assigned to a training cohort and validation cohort.

validation set

Description:

The whole cohort is randomly assigned to a training cohort and validation cohort.