Key result
FINRISK calculator accurately predicts 10-year CVD risk with strong discrimination and calibration in both sexes.
Why the study?
Online cardiovascular disease risk assessment tools are widely used, but only a fraction have undergone external validation. The FINRISK calculator needed external validation in a different Finnish general population sample than the one used for its development.
Does the FINRISK cardiovascular risk calculator accurately predict 10-year incident cardiovascular disease events in a general Finnish adult population?
Cohort (n=5,395)
Does the FINRISK cardiovascular risk calculator accurately predict 10-year incident cardiovascular disease events in a general Finnish adult population?
External validation in a large Finnish cohort confirms that the FINRISK calculator has good calibration and good-to-strong discriminative ability (AUC 0.777 in men, 0.824 in women) for predicting 10-year cardiovascular disease risk.
Answers are grounded in this paper and cite their sources.
Minor sex-specific calibration differences warrant caution in individual risk assessment; leaves open need for updated FINRISK validation in contemporary cohorts.
Background Online risk assessment tools such as ESC Score are widely used for cardiovascular disease (CVD) risk prediction, but only a fraction of them have gone through external validation. The FINRISK calculator was developed in 2007 to estimate the risk of acute coronary heart disease (CHD), stroke, or either of them (i.e., CVD) in a Finnish population. We validated the FINRISK calculator using different Finnish general population sample than originally used to develop the calculator. Methods The Health 2000 Survey was carried out in Finland in 2000–2001 in Finnish adult population aged 30-74 years. A total of 5,395 study participants with risk factor information required for the risk estimation by the FINRISK calculator and register-based 10-year follow-up on incident CVD events (2,421 men and 2,974 women), were included in this external validation study. Using risk formulas from the FINRISK calculator, predicted risks were calculated and compared with observed 10-year incidence rates of CHD, stroke, and CVD, separately for men and women. Area Under the Receiver Operating Characteristic Curve (AUC) values were used to assess discrimination for 10-year CVD risk. Results In men, the 10-year average predicted CVD-risk was 8.6% compared with the observed incidence rate of 9.5%, indicating a slight underestimation of the risk. In women, the risk was slightly overestimated, the predicted risk was 4.6% and the observed incidence rate 4.3%. The discrimination ability (AUC) was 0.777 (good) for men and 0.824 (strong) for women. The calibration plots demonstrated good agreement in both men and women. Conclusions The FINRISK calculator demonstrated good external validity. Calibration was generally good. Discriminative ability was good in men and strong in women. Overall, the FINRISK calculator demonstrated solid CVD risk predictive performance and appears well suited for cardiovascular risk assessment in the Finnish population. Key messages • In the external validation of the FINRISK calculator using the Health 2000 Survey data, overall calibration was good, with slight underestimation of CVD risk in men and overestimation in women. • The model`s good-to-strong discriminative performance for 10-year CVD risk supports the use of the FINRISK calculator for cardiovascular risk assessment in the Finnish population.
No takes yet. Share an insight, caveat, or question.
Elonheimo et al. (2026) conducted a cohort in Cardiovascular disease risk (acute coronary heart disease or stroke) (n=5,395). FINRISK cardiovascular risk calculator vs. Observed 10-year CVD incidence rates was evaluated on 10-year incident cardiovascular disease (acute coronary heart disease or stroke) (AUC 0.777 (men), 0.824 (women)). The FINRISK calculator showed good-to-strong 10-year CVD discrimination (AUC 0.777 in men, 0.824 in women) and good calibration (predicted vs observed: 8.6% vs 9.5% in men; 4.6% vs 4.3% in women).
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: