Algorithmic fairness in cardiovascular disease risk prediction: overcoming inequalities

Area of Science:

• Cardiovascular Disease Research
• Health Equity Studies
• Biostatistics and Predictive Modeling

Background:

• Prognostic risk prediction models aim to identify at-risk individuals for early intervention.
• Existing cardiovascular risk models (e.g., SCORE2, SCORE2-OP) are the EU standard but often neglect socioeconomic determinants.
• This oversight can lead to disparities in risk prediction and healthcare resource allocation.

Purpose of the Study:

• To highlight the limitations of current cardiovascular risk prediction models in addressing socioeconomic factors.
• To advocate for the explicit inclusion of individual-level socioeconomic determinants in cardiovascular disease risk models.
• To explore the potential of fairness concepts in cardiovascular disease prediction to reduce health disparities.

Main Methods:

• Review of current prognostic cardiovascular risk prediction models and their methodologies.
• Analysis of socioeconomic determinants' impact on cardiovascular disease risk.
• Introduction and conceptualization of four distinct fairness concepts applied to cardiovascular disease prediction.

Main Results:

• Current gold-standard models (SCORE2, SCORE2-OP) generally overlook socioeconomic determinants.
• This omission contributes to disparities in cardiovascular risk prediction and resource allocation.
• The paper introduces fairness concepts as a potential tool to mitigate these disparities.

Conclusions:

• Explicitly incorporating socioeconomic determinants into cardiovascular risk models is a key recommendation.
• The potential for these enhanced models to promote health equity requires further investigation through research and implementation.
• Fairness concepts offer a framework for developing more equitable cardiovascular disease prediction tools.