Constructing risk adjustment models for percutaneous coronary intervention: implications for quality assessment.

Area of Science:

• Cardiology
• Health Services Research
• Medical Informatics

Background:

• Assessing healthcare quality through patient outcomes requires risk adjustment due to variations in patient baseline clinical risk.
• Percutaneous coronary intervention (PCI) has seen increased volume and complexity, necessitating robust outcome evaluation.
• Risk adjustment models are crucial for reliable performance comparisons and quality improvement in interventional cardiology.

Purpose of the Study:

• To develop a risk adjustment model for in-hospital major adverse cardiac and cerebrovascular events (MACCE).
• To develop a risk adjustment model for in-hospital mortality following PCI.
• To utilize data from a national multicenter registry for model development.

Main Methods:

• A cohort study was conducted using data from the National Registry of Interventional Cardiology.
• The study included 10,399 procedures performed between June 30, 2003, and June 30, 2006.
• Multivariate prediction models were developed for in-hospital MACCE and mortality.

Main Results:

• Factors associated with in-hospital MACCE and mortality included advanced age, female gender, acute myocardial infarction, cardiogenic shock, renal failure, reduced ejection fraction, multi-vessel disease, and urgent PCI.
• The developed models demonstrated good discrimination (ROC AUC 0.83 for MACCE, 0.93 for mortality) and calibration.
• The models possess good discrimination and clinical value, indicating effective risk prediction.

Conclusions:

• A validated risk adjustment model for in-hospital MACCE and mortality after PCI was successfully developed.
• This model serves as a valuable tool for healthcare quality assessment.
• The model facilitates credible and reliable comparisons of outcomes among healthcare providers.