An office-based cardiovascular prediction model developed and validated in cohort studies of a middle-income country

Area of Science:

• Cardiovascular epidemiology
• Public health
• Health economics

Background:

• Existing cardiovascular disease (CVD) mortality prediction models often rely on laboratory measurements, limiting their applicability in resource-limited settings.
• There is a critical need for accessible and cost-effective tools to predict CVD mortality in low- and middle-income countries (LMICs).

Purpose of the Study:

• To develop and validate a non-laboratory-based model for predicting cardiovascular disease mortality.
• To assess the model's performance in a middle-income population, ensuring its suitability for resource-constrained environments.

Main Methods:

• Utilized data from three large cohort studies in Iran (Tehran Lipid and Glucose Study, Isfahan Cohort Study, Golestan Cohort Study) involving 40-80 year olds.
• Developed separate Cox proportional hazard models for men and women, employing cross-validation and bootstrapping for robust validation.
• Assessed model discrimination using the concordance statistic (C-index) and calibration via calibration plots, sensitivity, specificity, and net benefit fraction.

Main Results:

• The 10-year CVD mortality risk was estimated at 5.1% for men and 3.1% for women.
• The developed models demonstrated good discrimination, with optimism-corrected C-indices of 0.774 for men and 0.798 for women.
• Models showed favorable calibration (predicted-to-observed ratios of 1.07 for men, 1.09 for women) and acceptable sensitivity (0.76 men, 0.66 women).

Conclusions:

• A low-cost, non-laboratory prediction model effectively discriminates between individuals at low and high risk of CVD mortality.
• This model is suitable for implementation in screening programs within low- and middle-income countries, addressing a significant public health need.