Recurrence risk prediction of acute coronary syndrome per patient as a personalized ACS recurrence risk: a...

Area of Science:

• Cardiology
• Medical Informatics
• Public Health

Background:

• Acute Coronary Syndrome (ACS) is a major global health concern.
• High recurrence rates in Coronary Heart Disease (CHD) necessitate effective secondary prevention strategies.
• Existing risk assessments for ACS recurrence are often binary, lacking individual patient precision.

Purpose of the Study:

• To develop and validate a machine learning model for predicting individual patient-level ACS recurrence risk.
• To provide personalized risk probabilities beyond binary outcomes for post-discharge care.
• To identify key predictors contributing to ACS recurrence.

Main Methods:

• Utilized logistic regression and machine learning on datasets from Korean health insurance and a university hospital.
• Included 6,535 patients diagnosed with ACS.
• Model predictors comprised age, gender, procedure codes, procedure reasons, prescription drug codes, and condition codes.

Main Results:

• The model achieved high performance metrics: accuracy (0.893), precision (0.894), recall (0.851), F1-score (0.869), and AUC (0.921).
• Identified specific procedure and condition codes related to acute transmural myocardial infarction as significant predictors.
• Reported high odds ratios for these predictors (97.908 for procedure reason, 58.215 for condition code).

Conclusions:

• The developed model offers personalized ACS recurrence risk assessment, potentially enhancing patient motivation for risk reduction.
• This tool can support more targeted secondary prevention efforts for CHD patients.
• Specific indicators of myocardial infarction severity significantly increase ACS recurrence risk.