STATISTICAL MODELING OF LIFE EXPECTANCY USING SOCIOECONOMIC, MORTALITY, AND HEALTH INDICATORS ACROSS COUNTRIES

Abstract

Life expectancy is the result of country differences in socioeconomic conditions, mortality and population health features. Publicly available country-level records were used for a quantitative observational analysis between 2000 and 2015. The association between life expectancy at birth and schooling, income composition of resources, gross domestic product, health expenditure, adult mortality, infant deaths, under-five deaths, HIV/AIDS, body mass index, alcohol consumption, and indicators of immunization were explored. For descriptive analysis, group comparisons, Pearson correlation analysis and multiple linear regression, 1,649 complete country-year observations were analyzed. The mean life expectancy for developed countries' observations was significantly higher than the mean life expectancy for developing country observations. Life expectancy had the largest positive socioeconomic association for schooling and income composition of resources; strong negative associations were found for adult mortality and HIV/AIDS. Longevity was positively correlated with immunization indicators and BMI, and negatively correlated with infant and under-five mortality. The regression model accounted for about 75% of the variation in life expectancy, and found that income composition, schooling, adult mortality, alcohol consumption, and infant deaths were significant independent predictors. Results suggest that these factors, the social development, education, mortality reduction, and health conditions are interrelated and influence longevity. Policies designed to enhance population survival should thus include an educational, an economic, a preventive health-care, and a mortality-reduction component, taking into account that the relationships observed are statistical associations but not necessarily causal effects.