Advancing Cardiovascular Mortality Trend Analysis: A Machine Learning Approach to Predict Future Health Policy Needs

Area of Science:

• Gerontology
• Public Health
• Biostatistics

Background:

• Cardiovascular diseases remain a leading cause of mortality in elderly populations globally.
• Accurate forecasting of mortality trends is crucial for effective public health planning and resource allocation.
• Greece's aging demographic necessitates specific attention to cardiovascular mortality projections.

Purpose of the Study:

• To investigate and compare the efficacy of two distinct forecasting models in predicting cardiovascular mortality trends among the elderly in Greece.
• To evaluate the Autoregressive Integrated Moving Average (ARIMA) and Facebook's Prophet model for forecasting cardiovascular mortality.
• To identify which model performs best for different demographic subgroups within the elderly population.

Main Methods:

• Utilized longitudinal mortality data for Greece's elderly population from 2001 to 2020.
• Employed and compared the predictive performance of the Autoregressive Integrated Moving Average (ARIMA) model and Facebook's Prophet model.
• Assessed model accuracy for overall, male, and female populations aged 65-79 for the 2020-2030 decade.

Main Results:

• The ARIMA model demonstrated higher predictive accuracy for the general and male elderly population (65-79 years).
• Facebook's Prophet model exhibited superior forecasting performance for the female elderly population (65-79 years).
• Both models provided insights into future cardiovascular mortality trends, with varying subgroup performance.

Conclusions:

• Forecasting cardiovascular mortality requires models adaptable to demographic-specific characteristics.
• Advanced statistical methods like ARIMA and Prophet are valuable tools for informing health policy and planning in aging societies.
• Tailored forecasting approaches are essential for addressing the complexities of cardiovascular mortality in distinct population segments.