Validation and Application of the Globorisk-LAC Score in a Cohort of Patients from Northwestern Colombia

Abstract

Objectives: To internally validate the Globorisk-LAC model for predicting 10-year cardiovascular risk in a cohort from northwestern Colombia, adapting it to the epidemiological characteristics of Latin America and the Caribbean. Methodology: Data from prospective cohort studies in the region were used. Prediction models—one laboratory-based and one office-based—were developed using Cox proportional hazards regressions, employing age as the time scale. Both models were recalibrated by age and sex. Discrimination was assessed using Harrell’s C-statistic, and calibration was evaluated through linear regression analyses comparing predicted and observed risk. Results: Globorisk-LAC demonstrated good discrimination (C-statistic: 0.79; 95% CI: 0.69–0.89) and adequate calibration (calibration slope: 0.852). Sensitivity and specificity varied according to the selected risk threshold (10% or 20%). Both the laboratory and office-based models— the latter relying on easily obtainable predictors such as systolic blood pressure and body mass index—proved applicable in low-resource settings. Compared with recalibrated global models, Globorisk-LAC showed less risk underestimation, particularly among women. Additionally, relevant sex-specific differences were identified, and current epidemiological patterns of cardiovascular risk factors were incorporated. Conclusions: Globorisk-LAC represents a robust tool to optimize primary prevention, promote health equity, and support progress toward the Sustainable Development Goals. External validation in other Latin American populations is recommended to confirm its clinical applicability.