Geographic variation of life expectancy and all-cause mortality in Beijing

Abstract

BACKGROUND AND AIM: Life expectancy may vary substantially within Beijing, a globalized city with highly heterogenous physical and social environment. Small-area analysis helps reveal geographical variation. We aim to investigate geographic variation of life expectancy and all-cause mortality at the township level in Beijing. METHODS: All cause death counts in 2017 and 2018 by 5-year age group, sex and township were obtained from Beijing Center for Disease Control and Prevention. Population data for each township were from Beijing statistical yearbook, with weights for age and sex obtained from district-level population. We used a Bayesian spatial models to estimate age-specific mortality and life expectancy for 309 townships in Beijing, separately for men and women. RESULTS:The median of age-standardized all-cause mortality at township level was 445.3/10 0000 (interquartile range [IQR], 217.5/10 0000) for men and 329.5/10 0000 (IQR, 186.3/10 0000) for women in Beijing in 2017-2018. The relative geographic gap defined as 80th-to-20th percentile ratio of mortality was higher in women than in men (2.0 vs. 1.8). Townships with high mortality were primarily found in the northern and southwestern outskirts for both genders. The median of township life expectancies at birth in Beijing was 80.6 years (95% credible interval [Crl]: 79.1–82.0) for men and 85.3 years (Crl: 82.7–87.9) for women. The inequality across townships, measured as the difference between the 20th and 80th percentiles of township life expectancies, was 8.5 years for men and 8.8 years for women, respectively. The geographic pattern of life expectancy was similar in men and women and townships with high life expectancy were concentrated in central urban areas for both genders. CONCLUSIONS:Life expectancy and all-cause mortality varied substantially across townships in Beijing. Studies by small area may aid the development of local policies to reduce health inequalities. KEYWORDS: Life expectancy, Health inequalities, Bayesian spatial analysis, Small-area analysis