Implications of Cardiovascular Disease Risk Assessment Using the WHO/ISH Risk Prediction Charts in Rural India

Area of Science:

• Public Health and Epidemiology
• Cardiovascular Medicine
• Biostatistics and Machine Learning

Background:

• Cardiovascular disease (CVD) risk in India relies on World Health Organization/International Society for Hypertension (WHO/ISH) risk prediction charts due to the absence of population-specific models.
• The WHO/ISH charts offer two versions: a high-information (HI) model using total cholesterol (TC) and a low-information (LI) model without it, with limited guidance on their comparative use and performance.
• Discrepancies between the LI and HI models can lead to misclassification of patients requiring treatment for high CVD risk.

Purpose of the Study:

• To quantify the relative performance of the LI and HI WHO/ISH risk prediction models in rural India (WHO-South East Asian Region D).
• To develop and validate a simplified point-of-care (POC) test to identify patients who would benefit from a TC test, guiding the choice between LI and HI models.
• To improve CVD risk stratification strategies and resource allocation in large-scale screening programs.

Main Methods:

• Cross-sectional data from 1066 individuals in rural Andhra Pradesh with recorded blood cholesterol measurements were analyzed.
• CVD risk was calculated using both LI and HI WHO/ISH models, identifying individuals needing treatment (THR).
• Machine learning techniques (Support Vector Machine, Regularised Logistic Regression, Random Forests) were employed to develop the POC assessment for TC testing benefit, using age and systolic blood pressure as predictors.

Main Results:

• A 14.5% disagreement in CVD risk prediction between LI and HI models was observed, impacting 31% of identified high-risk patients.
• The developed POC assessment, utilizing age and systolic blood pressure, demonstrated high accuracy in predicting the benefit of TC testing (out-of-sample AUCs ranging from 0.82 to 0.85) and high sensitivity (98%).
• The POC test effectively identifies patients who would benefit from TC testing, guiding the selection of the appropriate risk prediction model (LI vs. HI).

Conclusions:

• The study highlights significant discrepancies between WHO/ISH LI and HI models in Indian rural populations, underscoring the need for refined risk assessment tools.
• The proposed POC assessment offers a cost-effective and efficient method for pre-screening patients for TC testing, optimizing CVD risk stratification.
• Implementing this POC test can enhance the planning of resource allocation and reduce costs in large-scale CVD screening programs in similar settings.