Cardiovascular risk estimation tailored to different clinical settings - the Tromsø study

Area of Science:

• Cardiovascular epidemiology
• Biostatistics
• Public health

Background:

• Cardiovascular diseases (CVD) remain a leading cause of mortality worldwide.
• Accurate risk prediction is crucial for effective primary prevention strategies.
• Existing risk models may not be optimal for all clinical settings or populations.

Purpose of the Study:

• To develop and validate a staged cardiovascular risk prediction model for myocardial infarction (MI) and stroke.
• To assess the incremental value of different data types (questions, examinations, blood samples) in risk prediction.
• To simulate applicability in diverse clinical settings.

Main Methods:

• Utilized data from 27,477 participants in the Norwegian Tromsø Study (1986-1995).
• Developed Cox regression models for MI and stroke, incorporating a competing risk framework.
• Gradually added explanatory variables and validated models using c-index, Brier score, IDI, and NRI.

Main Results:

• The developed models demonstrated excellent discrimination for both MI (c-index 0.833–0.946) and stroke (c-index 0.817–0.898).
• Significant improvements in discrimination were observed with the addition of variables (IDI).
• Models showed good calibration and fit across all stages and subgroups.

Conclusions:

• The staged cardiovascular risk model effectively predicts 10-year risk of MI and stroke.
• Even simplified models using only questionnaire data or blood samples provided valid cardiovascular risk estimates.
• The findings support the use of accessible data for robust risk stratification in clinical practice.