Predicting the Development of Surgery-Related Pressure Injury Using a Machine Learning Algorithm Model.

Area of Science:

• Medical Informatics
• Cardiovascular Surgery
• Patient Safety

Background:

• Surgery-related pressure injury (SRPI) poses a significant risk to patients undergoing cardiovascular surgery.
• Early identification of high-risk patients is crucial for effective SRPI prevention.
• Machine learning (ML) offers powerful tools for predictive analysis in healthcare settings.

Purpose of the Study:

• To develop and validate a machine learning (ML) model for predicting SRPI in cardiovascular surgery patients.
• To identify key risk factors associated with SRPI development in this patient population.

Main Methods:

• A secondary analysis of a prospective cohort study involving 149 cardiovascular surgery patients.
• Development of an ML model using the XGBoost algorithm to predict SRPI.
• Performance evaluation using receiver operating characteristic (ROC) curves and the C-index.

Main Results:

• The incidence of SRPI was 24.8% (37 out of 149 patients).
• Key predictors for SRPI included surgery duration, patient weight, cardiopulmonary bypass time, age, and disease category.
• The ML model achieved an area under the ROC curve of 0.806, indicating moderate predictive value.

Conclusions:

• ML-based prediction models show promise for assessing SRPI risk in cardiovascular surgery.
• Clinical application of this ML model could facilitate targeted interventions and improve patient safety.
• Further research should focus on deploying and validating the model in clinical practice.