Many prediction models estimating the risk of cardiovascular disease (CVD), such as SCORE and Globorisk, have been developed in cohorts from a general population. Although SCORE is applied in Dutch general practice, it has not been evaluated in this setting. Also, an external validation of the Globorisk models in a Dutch general practice setting is lacking. This study found that SCORE fatal and non- fatal (SCORE- FNF), Globorisk-laboratory, and Globorisk-office underestimate the 10-year risk of fatal plus non-fatal CVD in most of the patients who have their risk assessed by GPs. This study also showed that outcome definitions and eligibility criteria differed among the models, influencing their clinical applicability. |