Large health disparities in the performance of the SCORE2 cardiovascular risk prediction model in an ethnic and socioeconomic diverse population

Funding Acknowledgements Type of funding sources: Public Institution(s). Main funding source(s): Leiden University Medical Centre University Leiden. Background/introduction There is much evidence on cardiovascular health disparities in different ethnic and socioeconomic subgroups in (European) populations. Whether ethnicity and socioeconomic status should explicitly be taken into account as predictors in European risk prediction models, on top of traditional risk factors (blood pressure, cholesterol, age and gender), is still considered as a gap in evidence by the 2021 ESC Guideline on cardiovascular disease prevention. Purpose To assess the performance of the cardiovascular SCORE2 risk prediction model in ethnic and socioeconomic subgroups. Methods External validation of the SCORE2 risk model in a population-based study stratified by ethnicity (country of origin), and socioeconomic status (disposable household income). Results In total 6966 cardiovascular disease (CVD) events were observed versus 5495 events predicted by the SCORE2 model. Relative underprediction was the same in men and women (observed/mean predicted (OE-ratio) 1.3 and 1.2 in men and women, respectively). Underprediction was largest in the Surinamese subgroup (OE-ratio 1.9, in both men and women), particularly in the low socioeconomic Surinamese subgroups (OE-ratio 2,5 and 2.1 in men and women, respectively). In Dutch and the combined "other ethnicities", underprediction was observed in low socioeconomic subgroups too: OE-ratio 1.5 and 1.7 in men and women, respectively. Discrimination showed moderate performance in all subgroups, with C-statistics between 0.65 and 0.72, which is similar to discrimination of the SCORE2 model in the development study. Conclusions The SCORE 2 CVD risk model for low-risk countries (as the Netherlands are) was found to under predict CVD risk, particularly in low socioeconomic and Surinamese ethnic subgroups. Taking both socioeconomic status and ethnicity into account as predictors in CVD risk models is desirable for adequate CVD risk prediction and counselling.