Variations in Individual and Population Health and Well-Being: a Multilevel Perspective
CitationKim, Rockli. 2017. Variations in Individual and Population Health and Well-Being: a Multilevel Perspective. Doctoral dissertation, Harvard T.H. Chan School of Public Health.
AbstractUnderstanding the complex heterogeneity in health outcomes and their determinants at multiple levels are important to prevent further increase in inequalities between- and within- populations. Population health research to date has been predominantly focused on the differences between populations or social groups, despite the evidence suggesting that dispersion within populations are mostly driving the overall inequalities. Multilevel modeling has become increasingly common in health literature given its flexibility to partition and explain residual variance at multiple levels of shared environment, but they have been mostly restricted to modeling complex variation at the contextual levels only and confined to two levels at the most. In this dissertation exercise, I attempted to explore the full richness of partitioning and explaining random variability at truly multiple (more than two) levels with examples of body mass index (BMI) measures from 57 low- and middle- income countries and poverty in India, and discuss their substantive significance. Chapter 1 evaluated the global BMI variability by population-, subgroup-, and individual- levels, and found that the differences between populations can be explained better with basic socioeconomic characteristics, but focusing on them alone will resolve only a small fraction of the global inequality. In light of the large within-population variation that was found in BMI, Chapter 2 further modeled individual heterogeneity and observed that there are systematic underlying differences by basic sociodemographic characteristics, which allowed identification of subgroups with particularly large BMI inequalities. Chapter 3 demonstrated that both micro and macro levels are equally important in shaping the geographic patterning of poverty in India and highlighted the importance of considering multiple meaningful units of analysis to correctly identify contextual level(s) at which interventions should be targeted. Findings from each of these chapters will not only contribute new scientific evidence but will also help inform policies to reduce inequality in health and well-being.
Citable link to this pagehttp://nrs.harvard.edu/urn-3:HUL.InstRepos:42066821