Abstract
INTRODUCTION: Population aging represents a major demographic transformation in Italy, characterized by higher levels of education and socioeconomic resources compared to previous generations. Education is widely recognized as a key determinant of health, yet the mechanisms linking educational attainment to health in later life remain only partially understood. This study examines the association between education and self-perceived health among older adults in Italy, focusing on the mediating role of economic strain and lifestyle factors, and exploring differences by age, geographical area, and gender. METHODS: We use pooled cross-sectional data (2013-2019) from the ISTAT Aspetti della vita quotidiana survey regarding individuals aged 55 and older. Self-perceived health is analyzed using logistic regression models and mediation analysis is conducted using the KHB method to decompose the total effect of education into direct and indirect effects operating through economic strain and lifestyle variables (alcohol consumption, smoking, and physical activity). Subgroup analyses are performed by age group, geographical area, and gender. RESULTS: Higher educational attainment is significantly associated with a greater probability of reporting good health. The educational gradient remains robust after controlling for sociodemographic characteristics. Economic strain and physical activity emerge as the main mediators, jointly explaining approximately one-quarter of the total educational effect. Smoking and alcohol consumption play a limited mediating role. The mediating effect of economic strain is stronger among individuals aged 55-64, while the role of physical activity increases with age. DISCUSSION: Education plays a central role in shaping health inequalities in later life, both directly and through material and behavioral pathways. Policy interventions should address economic vulnerabilities and promote healthy lifestyles, adopting region-sensitive strategies to reduce health disparities in an aging society.