Blood Transfusion Management in Patients With Mitral Valve Replacement in China

Establishment of AI Prediction Model of Blood Transfusion Management in Patients With Mitral Valve Replacement

The evaluation uses ΔHb as an independent factor combined with artificial intelligence (AI) to predict its impact on the prognosis and blood transfusion of patients undergoing cardiac surgery, thereby guiding perioperative clinical blood use and improving patient prognosis.

1. Select 8 hospitals to form a multi-center team, and enter the keyword "mitral valve replacement" in the case system of 8 hospitals to collect information on 2000 heart surgery patients.
2. Mainly collect information on preoperative, intraoperative and postoperative test indexes (blood routine, liver and kidney function, coagulation function, blood gas), cardiac color Doppler, blood transfusion and prognosis of patients undergoing cardiac surgery, using statistical methods for analysis. Identify key observations.
3. Gradually incorporate and exclude data, and use statistical methods to conduct preliminary analysis on the collected data.

The AI prediction model was established by machine learning algorithm to predict intraoperative blood transfusion, verify the specificity and sensitivity of the blood transfusion prediction model, and scientifically guide clinical blood use.

Inclusion Criteria:

1. Patients undergoing "mitral valve replacement" surgery;
2. Age 18-75 years old;
3. 48h biochemical indicators before surgery, blood general indicators, coagulation indicators are complete.

Exclusion Criteria:

1. Lack of necessary tests and patient test information is not detailed, blood transfusion information is not detailed;
2. The blood transfusion information is not detailed in the same period as other cardiac operations (such as aortic valve surgery, coronary artery bypass grafting, etc.), but the tricuspid valvuloplasty is not ruled out at the same time;
3. Heart surgery or emergency surgery again.