O-93 Cardiovascular disease risk prediction and cardiovascular mortality in Barbados

Methods: Combined Health of the Nation Barbados survey (2011–2013) data with national mortality records to December 2022, excluding non-CVD deaths. Logistic regression assessed associations between multiple baseline risk factors and CVD mortality after a >10-year follow-up. Calculated WHO 10-year CVD risks for ages 40–74 without prior CVD, categorising as low/medium/high; evaluated mortality associations by category comparison. Calibration was assessed by comparing predicted vs. observed cardiac death rates across risk strata.

Results: Among 1175 baseline participants, 69 died of CVD causes over 12 years (5.87% rate), comprising 40 women (58.0%) and 29 men (42.0%). Decedents were older and had higher baseline systolic blood pressure and HbA1c levels. Univariate analyses showed associations with hypertension [OR (95% CI): 5.02 (2.86, 8.79)], diabetes [5.42 (3.28–8.97)], age [65+ vs. 25–45: 136.33 (18.71, 993.49); 45-65 vs. 25–45: 13.68 (1.80, 103.97)], stroke [3.81 (1.52, 9.55)], and high cholesterol [3.94 (2.41, 6.46)]. Multivariable model: hypertension [2.08 (1.09, 4.00)], diabetes [2.11 (1.17, 3.80)], independent of age [1.12 (1.10, 1.15) per year] and male sex [1.66 (0.93, 2.96)]. Separate model with WHO categories: medium-risk [7.63 (0.81, 71.77)], highrisk [15.72 (1.15, 215.73)] vs. low-risk, independent of age [1.09 (1.02, 1.17) per year] and male sex [1.22 (0.52, 2.84)]. WHO predictions exhibited significant miscalibration [Hosmer-Lemeshow χ² = 29.53, df=5, p=0.00002]; mean absolute calibration error (MACE) 10.8%, root mean squared deviation (RMSD) 14.0%, indicating moderate deviation but reasonable informativeness across risk deciles.

Conclusion: Hypertension and diabetes independently drive long-term CVD mortality; WHO risk categories predict outcomes despite miscalibration. Early interventions targeting non-communicable diseases like hypertension and diabetes are vital to curb premature CVD morbidity and mortality.