Higher mortality in areas of lower socioeconomic position measured by a single index of deprivation in Japan

Public Health. 2007 Mar;121(3):163-73. doi: 10.1016/j.puhe.2006.10.015. Epub 2007 Jan 12.


Objectives: To formulate an index representing area deprivation and elucidate the relation between the index and mortality in Japan.

Study design: Ecological study for prefectures (N=47) and municipalities (N=3366) across Japan.

Methods: Based on socioeconomic indicators of seven domains of deprivation (i.e. unemployment, overcrowding, low social class and poverty, low education, no home ownership, low income and vulnerable group), an index was formulated using the z-scoring method. The relation between the index and mortality was examined by correlation analysis, hierarchical Poisson regression and comparison of standardized mortality ratio according to the index.

Results: The deprivation index ranged from -7.48 to 10.98 for prefectures and from -16.97 to 13.82 for municipalities. The index was significantly positively correlated with prefectural mortality, especially in the population aged under 74 years: r=0.65 for men and r=0.41 for women. At the municipal level, hierarchical Poisson regression showed a significant positive coefficient of the index to mortality for both men and women, and excess mortality in the most deprived fifth compared to the least deprived fifth was 26.4% in men and 11.8% in women.

Conclusions: We formulated a deprivation index, which was substantially related to mortality at the prefectural and municipal levels. This study highlights the higher risk of dying among populations in socially disadvantaged areas and encourages the use of indices representing area socioeconomic conditions for further studies of area effects on health.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Aged
  • Female
  • Humans
  • Japan / epidemiology
  • Male
  • Middle Aged
  • Mortality / trends*
  • Small-Area Analysis
  • Social Class*
  • Statistics as Topic
  • Vulnerable Populations / statistics & numerical data*